ii

iii

THE OCEAN, THE WINE, AND THE VALLEY

iv

v

The Ocean, The Wine, and The Valley: The Lives of Antoine Badan

Edited by

Edgar G. Pavia CICESE Ensenada, Mexico Julio Sheinbaum CICESE Ensenada, Mexico and Julio Candela CICESE Ensenada, Mexico

OCEANOGRAFÍA FÍSICA Centro de Investigación Científica y de Educación Superior de Ensenada (CICESE). Ensenada, México.

vi

ISBN 978-0-557-94026-4 Published by CICESE Carretera Ensenada Tijuana No. 3918 Zona Playitas Ensenada, B.C., 22860 MEXICO Translations and Technical Edition by María Isabel Pérez Montfort CNyN, UNAM Ensenada, Mexico All rights reserved © 2010 CICESE No part of this book may be reproduced in any form or by any means, with the exception of material supplied specifically for educational purposes by the purchaser of the book. Printed in the USA.

vii

TABLE OF CONTENTS Preface page xiii Chapter 1 Antoine, Antonio, Toño 9

1. Valley, Ocean and Wine 11 Hugo D’Acosta

2. A Remembrance of my Friend Toño: Scientist, Environmentalist and Wine- maker 19 Federico Graef

3. Toño 25 Juan Cristóbal Rubio Badan

4. Antoine, Friend and Collaborator 27 Alexis Lugo-Fernández

5. Recent Advances in Further Research on Irreproducible Results: A Generalization 31 Eric D. Barton Chapter 2 The Ocean 63

6. A Note on Eddy-Topography Interaction in the Northwestern Gulf of Mexico 65 Peter Hamilton

7. On the Acoustic Backscatter Strength and Vertical Motion Signals from ADCP Measurements in the Gulf of Mexico 85 José Ochoa, Julio Candela, Julio Sheinbaum, Helmut Maske

viii

8. Solutions of Continental Shelf Waves Based on the Shape of the Bottom Topography 103 Luis Zavala Sansón

9. The high Amplitude Internal Waves Generated at the San Esteban Sill of the Gulf of California 131 Anatoliy Filonov, Iryna Tereshchenko, César Monzón

10. Lagrangian Circulation in Todos Santos Bay and off Baja California During Spring 2007: Exploratory Experiments 173 David Rivas, Rocío Mancilla-Rojas, Ernesto García-Mendoza, Antonio Almazán-Becerril Chapter 3 The Wine 203

11. California Heat Waves with Impacts on Wine Grapes 205 Alexander Gershunov, Daniel R. Cayan, Bernard Retornaz

12. Chemistry, Materials and Equipment in Wineries: An Overview on the Guadalupe Valley 225 Michael Schorr, Benjamín Valdez -Salas, Mónica Carrillo, Blanca M. Arellano García, Alejandro Martínez-Ruiz

13. Voltammetric Studies of Baja California Red Wines 233 Alejandro Martínez-Ruiz, Gabriela Guzmán Navarro, Benjamín Valdez -Salas, Michael Schorr

ix

Chapter 4 The Valley 243

14. Preliminary Normalized Difference Vegetation Index for the Guadalupe Valley, Baja California, Mexico 245 Ignacio Galindo, Julián Barrón

15. Does wet Sand Evaporate more than Water in the Guadalupe Valley? 263 Edgar G. Pavia, Ismael Velázquez

16. Precipitation and Seasonal Variation of Surface Temperature-controlling Factors in the Sonoran Desert, Northwestern Mexico 279 Iryna Tereshchenko, Luis Brito-Castillo, Alexander Zolotokrilin, César Monzón, Tatiana Titkova Chapter 5 The Symposium 303 Index 329

x

xi

Acronyms CIBNOR: Centro de Investigaciones Biológicas del Noroeste. (Mexico) CICATA-IPN: Centro de Investigación en Ciencia Aplicada Tecnología Avanzada-Instituto Politécnico Nacional. (Mexico) CICESE: Centro de Investigación Científica y de Educación Superior de Ensenada. (Mexico) CICY: Centro de Investigación Científica de Yucatán. (Mexico) CNyN, UNAM: Centro de Nanociencias y Nanotecnología, Universidad Nacional Autónoma de México. (Mexico). CUICA U de C: Centro Universitario de Investigación en Ciencias del Ambiente, Universidad de Colima. (Mexico) DG UAB: Departamento de Geografía, Universidad Autónoma de Barcelona. (Spain) EG U de C: Escuela de Geociencias, Universidad de Colombia. (Colombia) FC-UABC: Facultad de Ciencias-Universidad Autónoma de Baja California. (Mexico) ICTA UAB: Instituto de Ciencias y Tecnología del Ambiente, Universidad Autónoma de Barcelona. (Spain) IG/RAS: Institute of Geography/Russian Academy of Sciences. (Russia) IIM (CSIC): Instituto de Investigaciones Marinas (Consejo Superior de Investigaciones Científicas). (Spain)

xii

II-UABC: Instituto de Ingeniería-Universidad Autónoma de Baja California. (Mexico) INEGI: Instituto Nacional de Estadística y Geografía. (Mexico) MMS: Mineral Management Service. (USA) SAIC: Science Applications Internacional Corporation. (USA) SIO: Scripps Institution of Oceanography. (USA) SMN: Servicio Meteorológico Nacional. (Mexico) UCSD: University of California in San Diego. (USA) U de G: Universidad de Guadalajara. (Mexico) USGS: United States Geological Survey. (USA) WHOI: Woods Hole Oceanographic Institution. (USA)

xiii

PREFACE

This volume and the symposium that preceded it have kept in mind the

unique personality of Antoine Badan. It is not an excuse: if you find in these

pages something that seems strange, wrong, disorganized or out-of-place, we

hope that you will consider it more like an invitation to ponder than as a

mistake. So we have divided this book in five chapters. In chapter one you

will find the more personal texts. For example in the first article Hugo

D’Acosta tried to summarize the interests of his “compadre” Antoine from

the point of view of a fellow wine-maker; a similar task was undertaken by

Federico Graef in the second article, but a little bit more from the point of

view of a coworker. The most intimate text is that of his nephew Juan

Cristobal Rubio Badan which is simply entitled “Toño”. The initial stages

of the very important activities of Antoine in the Gulf of Mexico are

recounted by Alexis Lugo-Fernández; while Eric Barton tells us in the last

article of this chapter a fascinating story involving the strong yearning of

Antoine for originality in research.

Chapter two is devoted to the articles about the ocean. It begins with a

description by Peter Hamilton of the interaction of an oceanic eddy with the

topography in the northwestern part of the Gulf of Mexico. Then José

Ochoa, Julio Candela, Julio Sheinbaum and Helmut Maske report on the

acoustic backscatter strength (including biological applications) and vertical

xiv

motion signals from ADCP measurements also in the Gulf of Mexico. In the

third work of this chapter Luis Zavala Sansón finds new solutions of

continental shelf waves based on the shape of bottom topography; the next

article is a study by Anatoliy Filonov, Iryna Tereshchenko and César

Monzón on internal waves generated at San Esteban sill in the Gulf of

California, a sea that before the Gulf of Mexico was the source of Antoine’s

research. The chapter ends with a work headed by Antoine’s PhD student

David Rivas (plus Rocío Mancilla-Rojas, Ernesto García-Mendoza and

Antonio Almazán-Becerril) about exploratory experiments to study the

lagrangian circulation of Todos Santos Bay, one of the first scientific

interests of Antoine Badan.

The third chapter is about wine. The impact of climate, in particular of

the ever-more-present heat waves, on wine grapes of California (and

possibly Baja California) is described by Alexander Gershunov, Daniel

Cayan and Bernard Retornaz. An overview of chemistry, materials and

equipment of Gudalupe Valley wineries is made on the second article of this

chapter by Michael Schorr, Benjamín Valdez-Salas, Mónica Carrillo, Blanca

Arellano García and Alejandro Martínez-Ruiz. The chapter ends with a

study using electrical methods of the red wines of Baja California by

Alejandro Martínez-Ruiz, Gabriela Guzmán Navarro, Benjamín Valdez-

Salas, and Michael Schorr.

Environmental factors of northwestern Mexico in general, and the

Gudalupe Valley in particular, are the subjects of chapter four. In the first

article Antoine’s long-time friend Ignacio Galindo and Julián Barrón

differentiate the satellite-derived vegetation index for the Guadalupe Valley.

The following work by Edgar Pavia and Ismael Velázquez, about the

difference between evaporation of plain water and wet sand from the drying

xv

riverbeds of the Guadalupe Valley, was originally suggested by Antoine

Badan with the purpose of providing scientific evidence about the negative

impacts of massive sand extraction. This chapter ends with a study of the

controlling factors of surface air temperature in the Sonora desert by Iryna

Tereshchenko, Luis Brito-Castillo, Alexander Zolotokrilin, César Monzón

and Tatiana Titkova.

Submission of all these sixteen articles (peer-reviewed and non peer-

reviewed) and all materials sent to us for publishing in this book, unless

explicitly stated otherwise, implies the transference of copyrights from the

authors to CICESE.

The final chapter five is a remembrance of the international symposium

to honor the life and work of Antoine Badan held at CICESE, November 16-

18, 2009. It contains a reproduction of the original announcement which

appeared several times in EOS and the original program, which of course

was afterwards modified, that reminds us that it all began with an icebreaker

(page 306) with local wines (mostly Julio Candela’s, but also a few bottles

of Luis Zavala’s and Edgar Pavia’s). Here we acknowledge also the early

interest of Professor Walter Munk in participating in this event;

unfortunately he had to cancel at the last minute due to personal reasons, but

Oscar Velasco Fuentes kindly and dutifully filled in. The first day of the

symposium ended with a wine tasting offered by the Cofradía del Vino de

Baja California (page 308). The second day we had to move the afternoon

sessions from CICESE to El Mogor (pages 310-312) in order to visit

Antoine’s winery and to have dinner and taste some of the last Mogor-Badan

wines made by Antoine himself. The symposium ended on its third day with

another full set of talks and yet another wine-tasting now offered by the

Asociación de Vitivinicultores de Baja California (page 314). Some of the

xvi

abstracts of submitted presentations received in time for the symposium

which for different reasons did not make it to the pages of this volume are

included at the end of this chapter (including an extended abstract by Art

Miller which somehow seems to explain why all this really makes sense).

We cannot include the long list of names of all the people who made this

possible. However we must acknowledge Dr. Federico Graef, General

Director of CICESE, and Dr. Alejandro Parés, Chairman of the Physical

Oceanography Department of CICESE, as their unreserved support allowed

us to truly honor the life and work of Antoine Badan.

Edgar G. Pavia Julio Sheinbaum Julio Candela María Isabel Pérez Montfort Ensenada, B. C., Mexico December, 2010.

LIST OF REVIEWERS

Ken Brink Julio Candela

Manuel Figueroa Alexander Gershunov

Federico Graef Edgar Pavia

Clinton Winant Luis Zavala-Sansón

2

3

Photographs provided by

Natalia Badan Julio Candela

Anatoliy Filonov Arthur Miller Edgar Pavia

François Puyplat

4

5

6

7

Miscellaneous assistance provided by

Miguel Lavín Alina Morales Lupita Pacheco

Ismael Velázquez Luis Zavala-Sansón

8

�

9

Chapter 1

Antoine, Antonio, Toño

�

10

�

11

Valley, Ocean and Wine*♦

Hugo D’Acosta ♠, Estación de oficios de El Porvenir

This title may sound as an excuse to evoke our admired and irreplaceable

friend Antonio Badan Dangon. It is, however, an exercise to join otherwise

“distant disciplines”, providing an inspiring opportunity to enrich the

evolution of the integral human being, while setting us apart from our

customary positions.

One of the great movements of humanity took place in the 17th century.

Although this cultural movement corresponds to an entire epoch, it is hard to

recognize a precise beginning, birth date, commemorative day, or concrete

period that draws its limits. The Renaissance was a time of human

restructuring; talents and skills were exalted, and a multidirectional path was

regained which, even if limited and limiting, drew closer as communities

accepted the value of individual understanding.

Present-day Western civilization, which rules over us and sets our limits,

is marked by imprecise but definitive phases that seek to define the

dominant tendency of man’s pursuits over his surroundings. To imagine the

future, and to model statistically plausible tendencies and events we avail

ourselves of a cultural background, sometimes daringly called knowledge,

which finds support in the exact sciences and in the currently inevitable

scientific rigor (the scientific method). These tools which are at our disposal,

accessible to individuals of the 21st century, only amount to an egocentric

* The Ocean, The Wine, and The Valley: The Lives of Antoine Badan, 11-18. © 2010 CICESE. ♦ This article has not been peer-reviewed. ♠ Contact: hda1958@hotmail.com

�

12

posture in this infinite dark space. The possibilities of granting a meaning, or

in more contemporary speech, a “value” to this “posture” can be found in the

context, a word which allows the elements of absolute value to be curbed,

thus permitting a momentary understanding of these concepts.

The end of the 20th century reminds me of man’s quest for specialization.

The beginning of the 21st century makes me wish to unchain myself from the

absolute, from the definitive. To search for a beginning that may not be

precise, which was acquired in the course of time, with a birth date of little

importance and a date of expiry that will be lost in human evolution.

Hyperrenacentism: a delicate remembrance of our previous experiences to

be confronted with current human actions, creeds, religions, specializations

and individual attitudes. The intent of allowing the other, that of the others to

enrich us, to even form and deform us, to disturb us, opening the possibility

of universalizing not only knowledge, but its consequences, accepting

ignorance and dissolving the boundaries between body and matter.

Confusing dreams and desires. Forgetting that we are dust and to dust we

shall return.

Today’s exercise is to bring together apparently different disciplines:

oceanology, agriculture, physics, art; to intertwine the exact sciences with

philosophical weavings and biological networks. All actions focused from

the viewpoint of any discipline may be interlaced to show a broader vision

of an activity. It is relatively obvious nowadays that climatic behavior of

wine-growing valleys largely depends on or responds to oceanic

temperatures. It may be more difficult to demonstrate the aesthetic value of a

valley as a measure of its sustainability or longevity, or the customs and

habits of consumption with respect to the cultural development of a society.

�

13

Even if we ramble on, the true hyperrenacentist exercise seeks to envisage

and be visualized within a specific but imaginary context.

What did we have? Which habits did we acquire? When is it fitting to

exacerbate biodiversity or to insist on the Mediterranean vocation of certain

grape cultivations? Which part of evolution is socially correct and a pillar of

forceful, innovative learning?

Atemporality, cultural views, diversity, hybrid value, multiculturalism,

specialty, context, precision, time, learning, dreams, knowledge, rigor,

flexibility, evolution, universe.

Modeling parameters

Mean annual rainfall: 125 mm

Infiltration capacity (IC): 4%

Volume collection (AVC): 1,250 m3/ha/year

Exploitation volume (AEV) (from current CNA permits): ~20,000,000 m3

Required surface to sustain the aquifer: ~400,000 ha resulting from

AEV/(AVC * IC)

�

14

Potential water infiltration from surface

1,250 m3 m3/ha ha (× 1000) potential m3

(× 1000)

4% 50 400 500

8% 100 200 230

12% 150 133 146

16% 200 100 105

10k m3 model

• Mean water requirements for a vineyard ha/year: 3,000 m3

• Family requirements (4 members per family): 2 m3/day

• Potential vineyard 10k m3: 3,300 has

• Families served: 13,698

Valley 66 Model 1,250 m3 vine/olive

(ha) Families (4 members)

diff. water (× 1000) m3

4% 1,100 4,520 79.2

8% 2,200 9,040 158.4

12% 3,300 13,560 237.6

16% 4,400 18,080 316.8

�

15

Human occupation model Ha m3 m3/year

(family) vine/olive (ha)

remnant (ha)

5 6,250 730 1.83 3.17

10 12,500 730 3.92 6.08

15 18,750 730 6.00 9.00

20 25,000 730 8.32 16.68

Further premises for modeling Landscape evaluation according to:

• Origin (degree of transformation)

• Use (introduced activities)

• Merely aesthetic

Definition of specific areas (at the time of modeling)

• Chaparral and woodland

• Riparian

• Housing

• Services

• Agricultural

• Stockbreeding

�

16

Transformation Chaparral and

woodland

A

Riparian

B

Agricultural

C

Urbane

D

Services

E

Nil Nil In accord with the ecosystem

In accord with A,B,C

In accord with D, while

respecting A,B,C

Own Aesthetics

Own Aesthetics

Integral Aesthetics

Integral Cultural

Aesthetics

Conditioned Aesthetics

• Definition and delimitation aesthetics

• No modification

• Respecting aesthetics and the ecological unity of each area

• Consolidated population centers

To be considered when modeling from the vineyard

• Cultural value of the activity

• Water production vs. consumption

• Seasonal, assisted or irrigation agriculture

• Highly adaptive varieties

�

17

Water use by grapes. Production in kg/ha vs. plant vigor Vigor Seasonal 1000 m3 2000 m3 3000 m3

High 1,000 2,300 4,800 8,200

Medium no/prod 1,300 2,200 5,000

Low no/prod no/prod 1,500 4,500

Association between vigor and management

• Seasonal or assisted irrigation production

• Lower water levels increase enological value

• Low or no incidence of fungal infection

• Minimum weed control

• Simple organic management

Waste disposal at the Valley

Recycling or preventing waste production?

• The only way to fight the problem is by making use or setting an

economic value to waste in rural areas.

�

18

Although we may be rambling, the real Hyperrenaissance-focused exercise

intends to visualize and be visualized from a very precise but imaginary

context.

• Which habits and customs did we have, which did we acquire, where

is it appropriate to exacerbate biodiversity or insist on the

Mediterranean calling of some crops?

Which part of evolution is socially correct and forms a pillar of innovative

and dynamic learning processes?

�

19

A Remembrance of my Friend Toño: Scientist,

Environmentalist and Wine-maker*♦

Federico Graef ♠, CICESE

To speak of Antonio Badan, my friend, “Señor Ministro”, as we used to call

each other, is to talk about a sensitive human being, extremely intelligent,

cultured, refined, gentlemanly and very talented. Few people have this rare

but enviable ability to negotiate, to get things done to his specifications,

without forcing anything, to handle things softly, in all aspects of life:

professional, academic and personal. I knew Antoine or Toño as a physical

oceanographer, a winemaker and an artist.

I had the great fortune of meeting Toño in 1985. While I was a doctoral

student at the University of Hawaii, I attended a course at UCSD and visited

CICESE in Ensenada. From the first time we met, there was good

chemistry. Then, shortly before I finished my doctoral studies in Honolulu, I

received an unexpected telephone call from Toño, telling me that if I was

really interested in moving to CICESE I should write a letter of intent. The

rest is history.

So, I largely owe Toño my arrival at CICESE. A few days after settling

down, in August 1990, Toño said “let's take a gastronomic walk so you

become acclimated to Ensenada”, and so we went to eat fish tacos in the

black market, clam cocktails at one of the carritos (street-carts), and finished

* The Ocean, The Wine, and The Valley: The Lives of Antoine Badan, 19-23.

© 2010 CICESE. ♦ This article has not been peer-reviewed. ♠ Contact: fgraef@cicese.mx

�

20

up with the pastelitos (French pastry) at the Rey Sol restaurant. He had

quickly realized my tendencies towards good food and good spirits. Shortly

afterwards he invited me to belong to the Cofradía del Vino de Baja

California (Baja California Wine Brotherhood). He recommended me and I

was enthroned in 1992; today, I am proud to be the third oldest of the current

members, Antoine being one of the founders.

On several occasions we traveled together, to Mexico City, to Ciudad del

Carmen, to San Francisco, to Napa Valley. We also traveled together to

Trieste, Italy, in 1994, to attend a fluid mechanics course, and enjoyed great

dining experiences and abundant intake of ethylic fluids. I remember the

restaurant "Pepi", in Trieste, suggested by Toño, which serves nothing but

Italian “carnitas” (that is, only pork—all parts of it). There were four of us

and we had two large orders plus several bottles of house wine. Each dish

could easily have fed three people. When we asked for the second order, we

saw the face of surprise of the owners of the place. After eating so

abundantly, we walked back to the ICTP, but on the way we stopped at a bar

where we drank several glasses of grappa, just to help digest the pig! On

that same trip, we went for one weekend to Salzburg in Austria, and I

insisted that they should try the "Salzburger Nockerl", a sensational dessert

prepared with egg whites, flour and sugar, which looks very large but slips

down easily. After searching diligently and persistently, we found a place

that served it and asked for the dessert after dinner. The surprise of my three

companions (Toño, Manuel López and Miguel Lavín) was tremendous when

they saw the size of the dessert. Toño often remembered that scene, as well

as another time when we traveled to San Diego and were planning to have

dinner at Ruth's Chris Steak House. I asked him if he would like a

milkshake at the UCSD library to put our hunger on hold, so we did and

�

21

Toño had it for the first time. He said that such “potion” effectively

removed not only hunger, but he named it “a time bomb”.

In the professional field at CICESE I had the opportunity to team-up with

Antoine: Toño as Head of the Department of Physical Oceanography and me

as Coordinator of the Graduate Program of Physical Oceanography; these

were the years of 1994-1997 or so. In 1997, while offshore in one of the first

cruises of the project Canek in the Caribbean Sea, Toño cleverly and

skillfully talked with Elena Enriquez, who at that time was the assistant of

Dr. Javier Mendieta, then general director of CICESE, to say that “Federico

is interested in the position of Director of Graduate Studies”. And so,

Mendieta appointed me.

More recently, in early 2006, we started together the negotiations with

PEMEX that culminated in September 2007 when we signed the project

“Metoceánico” funded by PEMEX, to study the physical oceanography of

the Gulf of Mexico. This is the largest and most important project in the

history of CICESE, a $52 million USD, 6 year project. We continued

complementing each other, in this case, Toño as Project Leader and me as

General Director. Several times he told me “I would not have thrown

myself into this huge project or adventure had there been another General

Director”; a phrase which I will take to the grave as I believe it was due to

our strong complementarities and friendship. I really think we made a great

team and I am very grateful and fortunate to have had these experiences.

Toño has left a big void at CICESE, impossible to replace. In large part

we owe him the academic strength and good vibes at the Department of

Physical Oceanography, since he ferociously opposed academic inbreeding,

or not to hire our own graduates. Toño formed and amalgamated the Canek

group, a team of researchers at CICESE with different skills and

�

22

personalities who have complemented each other extraordinarily: originally

the group had, besides Toño as the leader, a great field experimentalist, a

smart theoretician, and a skillful and knowledgeable expert in numerical

modeling. Before the PEMEX project, CICESE had already been hired

through the Canek group by U.S. institutions such as Deep Star-Texaco and

the Mineral Management Services to perform oceanographic research in the

Gulf of Mexico and the Caribbean Sea. I must not fail to mention that in the

1980s Antoine had obtained funds for the project “Pichicuco” (which, if I´m

not mistaken, was initially funded by the NSF), to study the ocean-

atmosphere interactions in the Gulf of California.

His great ability to negotiate really allowed CICESE to carry out the

Metoceanic project with PEMEX; just to give you an example, we paid

invoices worth several million dollars to RD instruments several months

after the equipment had been delivered, in fact it was already measuring

currents in the Gulf of Mexico.

In one of our more recent “lonchecines or refrigerios” (as he used to refer

to our not-so-light lunches), Toño talked about US president Obama; he told

me how impressed he was with his inaugural speech, being “so clear in his

ideas”, even to the point of getting emotional about it. I was a little

surprised about this but I would never know why… unfortunately, he left us

before I could ask him.

Toño leaves us with a legacy of ideas and concerns in CICESE and in our

brotherhood of the Cofradía. He was an advocate and supporter of the Baja

California wine, a creator, grower, winemaker and the founder of “Mogor

Badan”, a great defender of the Guadalupe Valley and especially of its

vocations. He was also fighting openly against the sand extraction of the

riverbed.

�

23

Several times he told me that the overhead of the PEMEX project ought

to leave something tangible to CICESE; he insisted on a new building for the

Physical Oceanography Department that should be specifically designed to

meet the requirements for a first-class construction that should be

environment-friendly and energetically efficient. At first I wasn’t sure about

it, but a few months later, he convinced me. Not only have I now thought

that we must build the new facility, but also that it should be named after

Antoine Badan!

Toño was certainly a character of our time, in all the areas he cultivated.

A great friend, a really great friend, believe me; I miss him a lot more than I

had imagined.

Salud, Señor Ministro!

�

24

�

25

Toño*♦

Juan Cristóbal Rubio Badan ♠, Mogor-Badan

Paying homage is simpler than what we would commonly think. On the

other hand, it is also more complicated. I had better elaborate a bit. It is not

a gesture of plain generosity, nor one that I know shall please and be thanked

for. No. It is only fair and I don’t have a say in it, for what I am, or rather,

whatever value there may be in what I am, I owe to someone. This homage

then, is dedicated to Antoine Badan, rightfully, inevitably. This is also a

family account of one of its members, and it shall be—there is no getting

around it—as unreliable as any; it is, nonetheless, authentic.

We regularly saw Antoine on the weekend; he was a man of rather untidy

ways, chaotic; the kind who went about his business in a hasty,

temperamental fashion. He would, however, dutifully examine his wine,

talk with the workers, and then retire to listen to classical music—that last

bit, always with a cat on his lap. I would go visit him and talk about

everything, anything. I have said this a tad loosely: he would seldom talk

nonsense; a conversation with my uncle should never portend a desire not to

have to think too much. And so we would set off, visiting the various ideas,

colors and faces of this world where, as it turned out, almost everything was

downright, fascinatingly wrong. Of course some conversations were

gloomier than others, but in general, I must admit this was the customary

mood. But then it came. You knew it would always come, the very startling

* The Ocean, The Wine, and The Valley: The Lives of Antoine Badan, 25-26 © 2010 CICESE. ♦ This article has not been peer-reviewed. ♠ Contact: juancrisbadan@hotmail.com

�

26

yet firm belief that it was all remediable; not just in the manner that all

impending catastrophes are naturally, forcefully tended to in the last hour,

but in the more important sense that we, you and I, could change the current

heartbreaking arrangement of things; anything, or at least almost anything

was possible. He never failed to think so, and he was, mind you, 56 years

old.

So, yes, Toño went into space, he obtained a pilot’s license, traveled

around the whole planet, decided that small vineyard proprietors could and

should make their own wine in Guadalupe; that Mexican scientists should be

made responsible of tackling the deep water studies for Pemex; he thought

so and did so. In this country, our country, this mind-boggling, unsettling

close-to-delusional “can do” discourse is, I think, a very valuable thing.

This has been but one impression, one account of a story very dear to me;

there’s nothing terminal or definitive about it, but anything else, any other

comment on my part would seem excessive, and is not for me to say. That, I

shall leave to you.

Tophe

�

27

Antoine, Friend and Collaborator*♦

Alexis Lugo-Fernández ♠, MMS

Antoine and I first crossed paths at the Transports and Linkages of the Intra-

Americas Sea workshop held in Cozumel, Mexico, in November 1997. This

was my first visit to Mexico and where I first heard of the Canek group and

CICESE. The talks by the Canek group (Antoine, Julio Sheinbaum, José

Ochoa, and Julio Candela) about their measurements across the Yucatan

Channel were captivating and exciting. My attendance at the workshop had

two objectives: first, to showcase my impending research plans for

deepwater, and second, to establish contacts within Mexico to collaborate in

research projects in Mexican territory. I accomplished both objectives.

Upon my return, at a meeting with an oil and gas consortium, I told them

about the Canek Group and their measurements in the Yucatan Channel.

Soon, the consortium contacted the Canek Group and signed a contract to

carry out further measurements in the Yucatan Channel. A meeting was

held in Houston, Texas to plan and initiate the measurements in Yucatan and

I finally met Antoine and his colleagues. Unfortunately, I was not able to

convince my Agency to participate in this program.

Later, I contacted Antoine for a contract to install a single full water

mooring at 25ºN and 90ºW in Mexican waters to complement my

exploratory studies in American waters, and Canekito came to be. From

* The Ocean, The Wine, and The Valley: The Lives of Antoine Badan, 27-29

© 2010 CICESE. ♦ This article has not been peer-reviewed. ♠ Contact: Alexis.LugoFernandez@mms.gov

�

28

then on, he and his CICESE colleagues became regulars at my meetings*.

Thereafter, we signed another contract to deploy moorings over the western

Mexican slope, and it was another success. Not long ago, he told me about

negotiations with PEMEX for a large contract. Concurrently, we began

discussions for another contract to go back to Yucatan for a large study of

the Loop Current. These were the last communications we had.

Over this time our friendship grew, and we talked about work, science,

and life. We shared many topics during dinners at meetings. He invited me

to visit Ensenada, Mexico and graciously showed me the town, his winery,

and the CICESE facilities. There I got to know Julio Candela, Julio

Sheinbaum, and José Ochoa more in depth and I have cherished these

friendships very much since then. We shared many scientific visions for

studies and meetings, and even collaborated in an article in the Journal of

Marine Research published in 2007. In 2005, when my family got

evacuated because of Hurricane Katrina, I received an unexpected phone call

in Austin, Texas; it was Antoine offering support. He did the same when he

learned of my son’s deployment to Iraq. I deeply enjoyed our conversations

and learned much of my Spanish oceanographic jargon from him and his

colleagues. However, our time was cut short with the promise of even

greater collaboration in the horizon. One thing I’m sure of, he was one of

those special people you rarely meet, people with a wide range of interests, a

diverse and full life with many friends and accomplishments, and with

* At a meeting, I’m talking to a good friend of mine in the hall of a hotel, when I saw

Antoine walking towards us with a large smile. To my surprise, they greet with a hug and start conversing in French as if they had known each other forever. Trying to disguise my surprised, I asked how they knew each other. That is how I found that Antoine, like my friend, was of European descent and that he spoke several languages. What a small world!

�

29

whom you enjoy talking and working because of their positive outlook.

Proof of this were the many friends, colleagues, and collaborators at the

symposium honoring him.

�

30

�

31

Recent Advances in Further Research on

Irreproducible Results: A Generalization**

Eric D. Barton ♠, IIM (CSIC)

Abstract

One of the key results arising from the JOINT experiments into coastal

upwelling during the decade of the 1970s was the unexpected discovery by

Badan-Dangon (1981) that events in different upwelling regions, in different

years, and in different oceans exhibited correlations at high levels of

significance. This pioneering work, published in the Journal of

Irreproducible Results, stimulated the present investigation into the meaning

of reproducibility, its relation to the peer review process, and its role in

recent scientific controversies. Rules of reproducibility are tentatively

suggested and then applied to several case studies. It seems that peer review

itself is not convincingly reproducible and that scientists are not always as

conscientious as they ought in observing accepted norms for reproducibility

of their results.

1. Introduction

While considering what might constitute a suitable topic with which to

contribute to this symposium in honor of the life of Antoine (Figure 1), it

was painfully clear that although we had known each other for most of our

* The Ocean, The Wine, and The Valley: The Lives of Antoine Badan, 31-62 © 2010 CICESE. ♠ Contact: e.d.barton@iim.csic.es

�

32

scientific careers, had enjoyed long conversations about oceanography,

perhaps over a glass of wine from the Valley of Guadalupe, and had

attended far flung conferences jointly, we had never actually worked

together. However, one of his early publications (Badan, 1981) provided

inspiration because not only does it reflect his senses of natural curiosity and

mischievous humor, but also it leads us into considerations of sometimes

uncomfortable issues in contemporary science. The paper, though in part a

whimsical contribution on irreproducible results, prompts us to contemplate

seriously the meaning of irreproducibility, its relation to our everyday

scientific endeavor, and its role in the less than perfect peer review process,

and also to review examples of both blatant and unwitting irreproducibility

in recent science. Along the way, the example of Badan’s paper is examined,

before moving on to consider irreproducibility in the peer review process

itself, then examining scientists’ views of peer review, and finally looking at

examples of its failings, and future possibilities.

1.a Irreproducible? Come again?

To start, what is meant by reproducibility in science? The dictionary

meaning is quite clear, for example, relevant definitions are: “Reproduce:

produce a copy or representation of, recreate in a different medium or

context”, http://www.askoxford.com/concise_oed, from the Compact Oxford

English Dictionary; “Reproduce: to cause to exist again or anew, to imitate

closely, to present again” http://www.merriam-webster.com/dictionary, from

Merriam-Webster. “Reproducibility is one of the main principles of the

scientific method and refers to the ability of a test or experiment to be

�

33

accurately reproduced or replicated by someone else working

independently” according to http://en.wikipedia.org/wiki.

Thus, in science a basic tenet is that it should be possible to reproduce

results of any experiment or study independently. Some guidelines or rules

that indicate ways in which results should be reproducible include:

1. Enough information should be provided in a report that the

conscientious reader can derive the result from the data or model

provided;

2. Other workers should be able to produce a convincingly similar

result using the approach and methodology described from new data

or independently derived model;

3. The same result should be attainable even using different

methodologies and approaches.

Although these might appear simple enough in principle, they can be

practically quite difficult to observe scrupulously in every case. For

example, where large data sets or complex analyses are involved, in practice

it is difficult to supply enough level of detail within the usual constraints of

journal publications. In such cases, however, good practice is to ensure all

data are banked in a secure but accessible repository, e.g., one of the national

data banks for oceanography, to publish data summaries and metadata in

formal data reports, and to describe detailed data, theory, computational and

laboratory analyses in technical reports. These can be cited in the article to

provide a complete audit path for the work. Where such information exists,

independent researchers may test reproducibility under Rules 1 and 2. In

any case, any result is testable under rule 3, whereby investigators need to

�

34

know only what the result is supposed to be in order to attempt its

reproduction by any means available.

2. Badan’s irreproducible breakthrough

For his doctoral thesis work at Oregon State University, Antoine was

working on data from the northwest African upwelling region, in particular,

on the relation between wind and currents over the continental shelf

observed during the JOINT I experiment. While he was engaged in data

analysis, a second experiment JOINT II was taking place off Peru. When

early results were coming in from JOINT II, a startling discovery was made

(Badan, 1981).

As may be seen in Figure 2, the variability of wind forcing in the

northwest African and Peruvian upwelling systems during the two

experiments was essentially the same despite their large physical and

temporal separation, i.e. they were linearly related at a level beyond that

expected from chance occurrence. To quote his results “Maximum lagged

correlation between the two linearly detrended series was calculated for a

common record length of 28 day beginning on March 6... and was found to

be 0.51 at 6 hours lag. The significance level was ... significant at the 98

percent level. It is inferred that atmospheric conditions off NW Africa lead

those off Peru by 3 years and 6 hours.” It may be noted that this correlation

value and significance level are at least as high as many results of

comparisons between climatic indices or of model results with observational

data.

Badan’s result may be tested against the rules of reproducibility indicated

in the Introduction. Clearly it complies with Rule 1, since the data are

�

35

provided graphically and the correlation calculation could be repeated with

the digital data. Moreover, the result probably complies with Rule 2 because

one could use different data sets from the same experiments, say winds

measured at coastal sites or on other nearby buoys, to find a similar result,

but it would be slightly surprising if we could use data one year later from

the same sites and closely reproduce the result. On the other hand, it is

evident that the result arises not from any physical causation or link between

the two observations but from the similar time scales of variability in the two

upwelling regimes. For that reason, any two samples of records of similar

lengths might conceivably produce levels of correlation as high as those

found by Badan. In the same way, looking for correlations between new sets

of other variables such as sea level or alongshore component of current

velocity might be successful, thus demonstrating compliance with Rule 3,

but this would be more unlikely. So Rule 1 is demonstrably obeyed, Rule 2

is possibly complied with, Rule 3 is probably breached, and it is tentatively

concluded that the article was published in the appropriate journal.

3. Irreproducibility in peer review – a review

3.a What is peer review for?

Peer review (PR) is one check on reproducibility, albeit in general an

indirect one. Peer Review is considered basic to modern science, and

findings that are not subject to PR are considered unscientific or at least

untrustworthy and less than reliable. The PR process is of relatively recent

broad implementation despite its apparent acceptance as a recognized critical

procedure in science by the Royal Society in the 17th century (Spier, 2002).

�

36

The basic procedure is to subject any piece of academic work to the

scrutiny of independent experts in the field of specialization before its

publication. In this way it is intended to:

i. improve the quality of reports,

ii. remove errors of procedure or fact,

iii. eradicate irrelevancies, unjustified inferences, unsupported

conclusions,

iv. prevent fraud, and

v. avoid personal views and conflict of interest.

On the other hand, it is possible to argue that it might allow the:

i. maintenance of elites,

ii. perpetuation of dogma,

iii. obstruction of new findings, and

iv. personal bias,

since reviewers tend to be principally “established” authorities in their field

who may not wish to upset the status quo.

Of course, not everyone is happy with peer review (Kennefick, 2005).

Einstein in 1935 wrote in response to the review of a paper on gravitational

waves, submitted to the journal Physical Review:

“Dear Sir,

We (Mr. Rosen and I) had sent you our manuscript for

publication and had not authorized you to show it to specialists

�

37

before it is printed. I see no reason to address the—in any case

erroneous—comments of your anonymous expert. On the basis of this

incident I prefer to publish the paper elsewhere.

Respectfully,“

As so often is the case, the referee appeared to have misunderstood the

thrust of the argument!

In fact the referee’s criticism was correct. Astutely, Einstein

subsequently recognized the error and the paper was published elsewhere

with a somewhat different conclusion.

Some journal editors have sardonic views of the PR process:

“There seems to be no study too fragmented, no hypothesis too trivial, no

literature too biased or too egotistical, no design too warped, no

methodology too bungled, no presentation of results too inaccurate, too

obscure, and too contradictory, no analysis too self-serving, no argument

too self-circular, no conclusions too trifling or too unjustified, and no

grammar or syntax too offensive for a paper to end up in print.”

(Rennie 1986)

One might conclude from this viewpoint that we may all aspire to have

even our most insignificant results published somewhere!

“We portray peer review to the public as a quasi-sacred process that helps

to make science our most objective truth teller. But we know that the system

of peer review is biased, unjust, unaccountable, incomplete, easily fixed,

�

38

often insulting, usually ignorant, occasionally foolish, and frequently

wrong.”

(Horton 2000).

It is unsurprising then that members of the general public often express

strong opinions on the subject in web-based discussion groups:

“I for one am rapidly coming to the conclusion that you and others who

mumble the mantras 'peer review' & 'respected journals' and so on, are

becoming as irrelevant to climate change as prayer wheels…”

K. Pierson in (http://climateaudit.org/2009/04/24/irreproducible-results-in-

pnas/ ).

On the other hand, no better method has been suggested to date, and it is

patently superior to tradition, received authority, revelation and intuition as a

means of approximating the “truth”. Over a century ago, seemingly magical

phenomena like telekinesis and electro-magnetism were being investigated,

but the former has yet to furnish useful, reproducible results, allegedly in

part because the act of controlled observation interferes with the process

(Humphrey, 1995), while the latter has progressed reproducibly and

methodically to provide amongst other benefits the “miracle” (if one may

use such an emotive term) of modern electronic technology.

3.b Scientific reviews of peer review

Given the importance accorded the PR process in scientific publication,

there are relatively few studies of its effectiveness. Medical and

�

39

pharmaceutical journals are amongst those that have published results of PR

studies. One such investigation, on the reproducibility of peer review

(Rothwell and Martyn, 2000), examined the agreement between two

independent referees as to whether the manuscripts submitted to two clinical

neuroscience journals should be accepted without revision, accepted after

revision or rejected. The results (Table 1) were somewhat disconcerting.

Table 1

JOURNAL A

Reviewer 1→ Reviewer 2↓

Accept Accept if revised

Reject Total

Accept 3 3 5 11

Accept if revised

15 50 30 95

Reject 6 36 31 73

Total 24 89 66 179

Agreement 47%

JOURNAL B

Accept 1 3 1 5

Accept if revised

8 45 17 70

Reject 4 12 25 41

Total 13 60 43 116

Agreement 61%

�

40

A substantial number of papers acceptable without revision by one

reviewer were rejected by the other. Referees also rejected up to two thirds

of papers accepted, subject to revision, by another.

The rather alarming conclusion was that the agreement between two

referees is statistically no different from what could be expected from

chance, for either Acceptance (as is, with revision, or reject) or Priority (low,

medium, high) . In the particular journals researched even if both reviewers

recommended acceptance, papers had no certainty of publication, because

the editor still rejected up to 15% of recommended papers. If this lack of

agreement between reviewers were general, it would be disturbing to the

community’s equanimity with regard to the PR process.

Personal editorial experience indicates the contrary: reviewers tend to

agree. The above study might be considered as unsubtle because of the

categorization solely in terms of accept, revise, or reject. Reviewers do not

simply sort manuscripts into three groups; they generally write extended

critiques of the paper, raising particular arguments about aspects of the work

and indicating areas for improvement. No account of the referees’ actual

comments was made in the study, but the substantial number of editorial

rejections of papers recommended suggests that editors were noting critical

remarks in the reviews. Many journals have a wider range of possible

categories, wherein reviewers choose between accept, minor revision, major

revision, reject with possible resubmission, and reject outright.

Again, subjective personal experience would indicate that reviewers

often specify a category contradicted by their comments. A demand for

“minor revision” might be accompanied by a litany of criticisms that

demand extensive reworking of the results, whereas a call for “major

revision” might be justified by insubstantial complaints about textual

�

41

expression or form of figures. Few reviewers are willing to “accept as is”

because there are always small errors in any manuscript. Similarly,

rejections outright are relatively few and tend to be related to papers that are

mis-targeted rather than incompetent.

Nevertheless, the fundamental result that PR does not function is a

disturbing one. There is clearly a need to investigate further the efficacy of

the PR process as a means of quality control in science generally. In

principle, it should be possible to devise relatively objective measures that

can be used to evaluate the reproducibility of reviewers’ opinions, based on

their actual comments rather than on the simple categorisations that they

signal.

3.c An editorial gripe about peer reviewers

One of the most difficult of editorial tasks is that of obtaining suitably

qualified, impartial reviewers who will return comments on a manuscript in

a prompt manner. A tentative scale for classifying invited reviewers in

terms of their response is summarised in Figure 3. It may be seen that editors

do not hold many types of reviewer in high regard. In reality, there are

multiple dimensions to the reviewer categorisation unaccounted for in the

figure. Further development of this topic might include the parameterization

of the dimensions of promptness, utility of comments, evident familiarity

with the field, personal bias and so on.

In general, one must be impressed by the diligence shown by many

reviewers and dismayed by the number of researchers who deign not to reply

to invitations to review. According to the review of Dye (2007), reviewers

under the age of 40 provide better reviews than their elders, 20% of reviews

�

42

are outstanding and 20% useless, and most reviewers detected only 25% of a

series of deliberate errors introduced into a set of test manuscripts. [Note

added in proof: In relation to the last point, it is noteworthy that neither

reviewers nor editors detected the amazingly good reproduction of the

results of Roy and Reason (2001) by Muni Krishna (2008). The latter paper

was however subsequently withdrawn (Jenkins et al. 2009)].

3.d So what do scientists opine about peer review?

The preliminary findings of the Peer Review Survey (2009) were

published late that year. The aims of the survey are to investigate the

opinions of the scientific community on questions such as the purpose of

PR, the detection of fraud and misconduct, the suppression or

encouragement of good ideas, and peer or author anonymity.

The Peer Review Survey was conducted by internet between 28 July and

11 August 2009. Some 40,000 researchers listed in the ISI author database

were randomly consulted, representing contributors to over 10,000 journals.

About 10% of the researchers completed the survey. The error margins on

the answers were estimated at ± 1.5% at the confidence level of 95%. A

subset of questions was aimed specifically at reviewers, who made up

something less than 10% of the total.

�

43

Table 2

Question Response Percentage

Why be a reviewer? To contribute to community

90

Could we do without PR? No 84

Is PR best system? Yes 32

Was your last paper improved by PR?

Yes 91

Lack of guidance on how to PR? Yes 54

Have you refused to review in last year?

Yes because of lack of expertise

61

Is the PR process too slow? Yes 43 Did you receive a decision on your paper in < 3 months?

Yes 65

Do you prefer to retain anonymity as a referee?

Yes 58

Would you prefer a double blind review process?

Yes 76

Do you think open peer review effective?

Yes 20

Does the public understand PR? Yes 30

Should PR detect plagiarism and fraud?

Yes 80

Does it detect plagiarism and fraud?

Yes 36

Would payment be an encouragement to participate in more PR?

Yes 41

If authors had to pay? Yes 3

�

44

The results (Table 2) show the majority of reviewers and scientists

undertake the PR task from a sense of community commitment, are

convinced that PR is necessary to control quality but believe that the PR

system could be improved. They consider that they could be better trained

to carry out PR (even though scientific journals generally provide quite

explicit instructions to reviewers). They consider the process too slow, even

though the majority received decisions on their submissions within 3

months. Interestingly, three quarters of respondents would prefer double

blind reviews, in which the identity of authors is not revealed to the

anonymous referees. Only one fifth of respondents considered that open

review, where comments are published along with the paper, would be

effective. A large majority believe that PR should prevent fraud, cheating

and plagiarism while one third consider it does not. A substantial proportion

said that being paid would encourage them to participate more fully in the

PR process, although few would be willing to pay for referees themselves.

At the moment, relatively few journals use either open or double blind

review. Some on line journals use the former system and invite comments,

anonymous or otherwise, from the community. First impressions are that

one sees few contributions other than exchanges between the official

reviewers and the authors; this, given the difficulty of obtaining reviewers, is

not greatly surprising. Journals that operate double blind reviewing accept a

higher proportion of papers from authors with clearly female names than

other journals in the same field (Budden et al., 2008). This indicates both a

perhaps disturbing sexism in a supposedly objective process and some

evidence that the method is effective. So despite the objection that authors

can probably be readily identified in some areas of research, double-blind

review does appear to make a difference. One point not questioned in the

�

45

survey was the possibility for authors to suggest names of reviewers.

Because finding appropriate reviewers is sometimes difficult, some journals

require authors to supply a list of up to five suitable experts, with the

possibility of naming potentially hostile reviewers to be avoided. Although

editors are not obliged to follow the suggestions, it has been shown in a

study of 788 reviews in 10 journals that authors who make use of this

facility have a roughly 10% higher acceptance rate than those who do not

(Grimm, 2005). The improved acceptance rates are not necessarily because

of cronyism; it could be that experienced authors know who the most

appropriate expert to judge their research is and who might be averse to their

work.

4. A little known correlation

One of the areas to which Antoine Badan contributed was that of public

well-being. Utilising similar correlation techniques as employed in Badan

(1981) crimino-oenological studies based on official USDA and USDOJ

statistics (see http://old.swivel.com/graphs/show/1013998) have

demonstrated that increased wine consumption correlates with reduced

crime rates in the US (Figure 4a). As wine consumption increases, so crime

rate decreases. In the light of this information, Badan’s general public

spirited desire to help reduce crime led to action in the form of his wine

making activities. Of course, it is difficult to be precise about exactly how

many crimes were prevented per case of Badan-Mogor produced and sold

(Figure 4b). While it has been argued that there is a genuine inverse relation

between alcohol consumption and homicide in the 20th century United States

�

46

(Miron, 1999), this may be more related to the enhanced opportunities for

criminal activity during Prohibition than to the soothing effect of wine.

5. Irreproducibility and creativity – or where peer review failed

Reich (2009) has documented in great detail the case of the German

physicist Jan Hendrik Schön (Figure 5), a highly creative individual whose

irreproducible scientific accomplishments illustrate clearly the fallible nature

of the PR process. His trajectory for the years 2000-2002, during his

employment at Bell Laboratories, is shown in Table 3.

Table 3

Year Achievement

2000 Eight first author papers in Science and Nature

2001 Eight first author papers in Science and Nature

2001 Otto-Klung-Weberbank Prize for Physics (annual German

award)

2001 Braunschweig Prize (awarded biennially by the city of

Braunschweig)

2001 Name appeared on a paper on average every 8 days.

2002 Outstanding Young Investigator Award of the Materials

Research Society

�

47

This would appear to be quite a remarkable record, especially viewed

from the field of oceanography, where it often takes years to work results

from extended field campaigns into publishable form. Schön was working in

nanotechnology and molecular electronics and claimed, amongst other

supposed advances, to have made a breakthrough using organic dye

molecules that behaved as a transistor, which would have permitted

tremendous reductions in scale for electronic circuits. Other laboratories and

scientist tried unsuccessfully to reproduce and follow up on his results with a

resultant waste of both time and money. Apparently none of his first author

papers were seriously questioned during PR, and perhaps even more

strikingly, none of his co-authors (who were benefiting from the prodigious

publication rate in prestigious journals) doubted the reliability of the results.

Several researchers in other institutes attempting to emulate his work

however discovered that he had fabricated and re-used the same data sets in

different papers on different topics. The subsequent enquiry in Bell

Laboratories adjudicated that the co-authors were free of responsibility in

what was clearly a major extended fraud by Schön lasting many years. None

of the supposed original experimental data were found because the computer

records and notebooks had been erased or destroyed. The scientific record

was corrected in the Schön case because individual scientists uncovered his

deceits in the course of their work, but not because peer reviewers for the

prestige journals detected any fraud. Science, Nature and Physical Review

Letters subsequently withdrew 21 of his publications.

The case could be thought a failure of the PR system, but one might

question whether PR can reasonably be expected to detect sophisticated,

deliberate fraud. The short-lived success of this particular web of deceit was

made possible by a combination of many factors. Personal ambition on the

�

48

part of Schön and his collaborators, wishful thinking by Bell Laboratories

managers hoping for quick results and the prestige journals’ desire for

papers in a newly breaking area all played their part. Above all perhaps, a

general atmosphere of credulity and absence of skepticism allowed the fraud

to prosper. It is hard to imagine a researcher capable of maintaining such an

extremely high rate of productivity, yet no-one closely involved - co-

authors, managers, editors, reviewers – appeared to express any suspicions.

On the other hand, it might be argued that the uncovering of the fraud is yet

another demonstration that the scientific method does work and in the end

will distinguish the genuine from the false result.

6. Irreproducible results in current science – conspiracy, cock up or

conundrum?

Much emotion has been exercised by the climate change discussion.

Some see the work of the Intergovernmental Panel on Climate Change as

disinterested counsel that may free the Earth from the grasp of a greedy

fossil fuel lobby that is destroying its environment. Others hold that the

IPCC is running a conspiracy designed to perpetuate the funding, careers

and hubris of a scientific community espousing a convenient orthodoxy.

Without entering the debate itself, it is possible to make interesting

observations on the way discussion is carried on by both sides.

Ohio State University is home to Lonnie Thompson - world leading

glaciologist and friend of Al Gore – and also to Huston McCulloch –

economist and climate change sceptic. The former, an ice core specialist,

published a 2000 year series of oxygen isotope ratios from four Himalayan

and three Andean glacial cores (Thompson et al., 2006). The ratio of

�

49

18O/16O decreases regularly with decreasing temperature because of differing

evaporation/condensation rates of the isotopes and is accepted as a

temperature proxy. The individual series were shown as z-score indices, i.e.

individually normalized, and used to derive a “Tropical composite series”.

The composite is compared favourably to the well-known Hockey Stick

curve and claimed as independent corroboration of the latter (Figure 6a).

Post-industrial revolution temperature increase appears evident during the

last two centuries (although no conversion to temperature units was actually

addressed).

The Tropical Composite record is derived from Andean and Himalayan

composites (Figure 6b) that track each other well through the strong rise in

19th and 20th centuries. The tropical curve appears to be the average of the

two regional composites. However, none of the individual series shows such

an increase (Figure 6c,d), and the paper does not explain how the composites

were calculated. According to the web site

(http://climateaudit.org/2009/04/24/irreproducible-results-in-pnas/)

McCulloch wrote repeatedly to the authors to query the results and request

access to the data and calculations, but received no reply. He then took the

data as archived with supplementary material on the journal web site and

attempted to deconstruct statistically the composites. He found that the

Tropical composite is the average of the Andean and Himalayan composites

(obvious from the figure) but found it impossible to reproduce the individual

composites from any linear combination of the original curves.

McCulloch therefore concluded that either the individual data series are

erroneous or that the composites are erroneous since the latter cannot be

obtained from the former. He submitted a comment to Proceedings of the

National Academy of Sciences, which had published Thompson’s article,

�

50

but it was rejected because comments are accepted only within 3 months of

publication of the original paper. The comment was subsequently published

elsewhere (McCulloch, 2009). As he developed the analysis McCulloch

maintained an up-to-date report of progress on the ClimateAudit web site.

He subsequently found that different data sets were used in different figures

in the Thompson et al. paper and rightly asks “which is the right data to use

to construct an ice-core proxy?” The controversy is the subject of a

continuing discussion on

http://www.econ.ohio-state.edu/jhm/AGW/Thompson/

and other web sites (e.g.

http://climateaudit.org/2009/12/10/calibrating-dr-thompsons-z-mometer/).

As of the time of writing (12 April 2010), there seems to have been no

response from Thompson et al., though his supporters in the discussion

groups have queried the credentials of Energy and Environment, which

although peer reviewed is not a Citations Index journal and has been a

vehicle for Climate Change sceptic articles.

Thompson et al. (2006) clearly fails Rule 1 of reproducibility, because

the result cannot be deduced from the materials supplied in the paper. Rule

2 is unlikely to be tested because of the logistic difficulty and expense of

mounting repeat expeditions to such inaccessible sites. Many groups are

working on a candidate for Rule 3.

Several points of interest arise from the controversy. First, it may be

thought that while PNAS comment policy seems overly restrictive, the

journal has not strictly enforced policy on the provision of supporting data.

Theoretically, the journal requires full submission of supporting data for any

article, but although it did not ensure this was done in this case, no

allowance was made in terms of an extension of the response time. Science

�

51

and Nature magazines have similar policies (six month cut off) to ensure

comments are timely, but such policies should be flexible and consider

factors such as unavailability of data. Second, peer review seems to have

been inadequate in this case since there clearly was an undetected problem

in the data analysis that should have been evident to discerning referees.

Third, Thompson and co-workers appear to be harming their reputation by

neither responding with a rebuttal to McCulloch’s criticisms nor

acknowledging errors in their analysis, though they may respond now the

complaint is published formally. Fourth, maligning the journal in which the

McCulloch comment was published is irrelevant (argumentum ad hominen);

regrettably, much of the public discussion of these scientific issues is

maintained at this or baser levels. Finally, it should be noted that a good

deal of the argument and development has taken place outside the

conventional framework of journals and their due process.

To some this controversy indicates conspiracy, to some it is a cock up,

but for the moment it remains a conundrum.

7. Closing remarks

Though peer review can be robustly criticized as inadequate, it is still

better than no system, and should be supported while improvements are

sought. PR is a check on the quality of published science, not an absolute

arbiter of truth and falsehood. In some cases it may reject valuable findings,

in others it may approve falsifications, but in general it will improve the

quality of published science. Whether it does so in the most efficient way

possible is unlikely. In general, scientists recognize their reviewing debt to

the community, and for every paper they publish, they review many more in

�

52

a diligent and timely manner. Journals, editors and scientists in general are

aware of the problems and difficulties and are actively seeking their

solutions. Although the major changes towards an all-electronic publishing

and open review system as foreseen by Judson (1994) have still not

materialized, it is clear that significant changes continue to take place as on

line journals experiment with new formats and review systems. Moderated,

journal-run sites for on-going review of publications and double blind

reviewing should be more widely explored.

Whatever developments take place, electronic publishing, blogs, and

special interest web sites have enabled a much wider and more rapid

scrutiny of scientific results than ever before. In this age of instant

communication, “unofficial” or uncontrolled review of controversial results

will inevitably occur outside the learned bodies and journals. Not all science

attracts the same degree of scrutiny by the wider public as Climate Change,

but it is clear that scientific work that impinges directly on everyday

concerns, economic factors, and vested interests will be subject to increasing

examination outside the traditional forums of academic research by the rise

of the so-called “peer-to-peer” review

(http://bigjournalism.com/pcourrielche/2010/01/08/peer-to-peer-review-

how-climategate-marks-the-maturing-of-a-new-science-movement-part-i/).

Since the oral presentation of this paper, the leaking of e-mail

correspondence between leading players in the Anthropogenic Climate

Warming debate from the Climate Research Unit of the University of East

Anglia has given rise to the international ‘Climategate’ affair. More than a

thousand illicitly released private messages were eagerly scanned by climate

change sceptics for indications of collusion, manipulation of the PR process,

and falsification of data. What emerged was evidence of a siege mentality

�

53

amongst a closely knit climate research community that felt itself threatened

by a relentless onslaught from mischievous, time-wasting amateurs, cranks

and special-interest groups pursuing an unscientific agenda. Even so, it

appears that scientists deliberately did not disclose data, have misplaced

important metadata, and have exchanged negative remarks about the validity

of papers under confidential review. The head of the CRU, Phil Jones, in

testimony to an investigating committee of the UK House of Commons

conceded, "I've obviously written some very awful e-mails". To date,

official enquiries have not revealed any significant wrong-doing, but the

Commons report (HC Science and Technology Committee, 2010) does say

“climate scientists need to take steps to make available all the data that

support their work and full methodological workings, including their

computer codes. Had both been available, many of the problems at CRU

could have been avoided", i.e., observe Rule 1 of reproducibility. However,

as remarked by one commentator “those in every profession should consider

how their reputation would survive if years of private correspondence were

filleted for dirt and handed over to critics” (Adam, 2010).

In science, honesty is not only the best policy, it is the policy. For that

reason, scientists should seek to expand the conscientious and open sharing

of data and methods to demonstrate that their results are reproducible.

Science stands or falls on its reproducibility, openness and honesty and so

the opening up of discussions to the widest possible audiences should be

accepted and welcomed. As Popper (1959) observed, it is not the

“possession of knowledge, of irrefutable truth, that makes the man of

science, but his persistent and recklessly critical quest for truth”. Pseudo-

science proffered by special interest lobbies is vulnerable because it has to

withstand the same recklessly critical scrutiny.

�

54

This contribution was inspired by and is dedicated to Antoine Badan

Dangon (Figure 8), whose friendship, vivacity, enthusiasm and bonhomie

are greatly missed, a truly irreproducible person.

What a piece of work is a man, how noble in reason, how infinite in

faculties, in form and moving, how express and admirable, in action

how like an angel, in apprehension how like a god! the beauty of the

world, the paragon of animals. (William Shakespeare)

Acknowledgements

I am grateful to the American Geophysical Union for contributing to

travel costs. Norman Sperling of Journal of Irreproducible Results kindly

provided permission to reproduce Figure 2. Julia Wilson of Peer Review

Survey allowed use of their preliminary findings in Table 2. Thanks are due

to Bill Hughes of Multi-Science Publishing for permission to reproduce

Figure 7 from McCulloch (2009).

�

55

Figure 1. Antoine Badan Dangon contemplating irreproducibility c. 1984.

�

56

Figure 2. Reproduction of Badan’s (1981) publication in Journal of Irreproducible Results.

�

57

Figure 3. Tentative scale for editors appreciation of reviewers.

Figure 4. Einstein was not the strongest supporter of peer review.

�

58

Figure 5. a) Wine consumption and serious crime relative rates of change in the United States. b) Badan’s contribution to combating crime.

Figure 6. Jan Hendrik Schön – the creative physicist responsible for supposedly major breakthroughs, later found to be irreproducible.

�

59

Figure 7. Composite d18O Z-score for 7 Tibetan and Andean ice cores, decadal averages, from Thompson et al., Data Set 3 (a) 2000 year record (b) record 1600 AD on, together with regional subindices; and individual d18O isotope ratio series, decadal averages constructed from 5-year averages in Thompson et al. Data Set 2 (c) Andean sites (d) Himalayan sites. Years indicate beginning of decade in each figure.

�

60

Figure 8. Antoine Badan Dangon in Gran Canaria during the ICES conference on the Canary Current in April 1978.

�

61

References

Adam, D., 2010: Climate wars damage the scientists but we all stand to lose

in the battle, The Guardian, Mar 1, 2010.

Badan-Dangon, A., 1981: The JOINT experiments rejoined. Journal of

Irreproducible Results, 27, 14-15.

Budden, A.E, T. Tregenza, L. W. Aarssen, J. Koricheva, R. Leimu, and C. J.

Lortie, 2008: Double-blind review favours increased representation of

female authors. Trends in Ecology and Evolution, 23, 1, 4-6.

Dye, L.R., 2007: A Review of the Review Process, Journal of Medical

Toxicology, 3, 4, 143-5.

Grimm, D., 2005: Suggesting or Excluding Reviewers Can Help Get Your

Paper Published. Science, 309, 1974.

Horton, R., 2000: Genetically modified food: Consternation, confusion and

crack up (Editorial), Medicine Journal of Australia, 172, 148-9.

House of Commons Science and Technology Committee, 2010: The

disclosure of climate data from the Climatic Research Unit at the University

of East Anglia, Eighth Report of Session 2009–10, House of Commons: The

Stationery Office Limited, HC 387-I, pp59.

Humphrey, N.K., 1995: Soul searching: human nature and supernatural

belief, Chatto and Windus. ISBN 0-7011-5963-4.

Jenkins, W. J., J. Johnson, and E. Delhez, 2009: Editorial Note, Ocean

Science Discussions, 6, 3055-3056.

Judson, H. F., 1994: Structural Transformations of the Sciences and the End

of Peer Review, JAMA, 272, 92-94.

Kennefick, D., 2005: Einstein versus the Physical Review, Physics Today,

43, 43-48.

�

62

McCulloch, J.H., 2009: Irreproducible results in Thompson et al., “Abrupt

tropical climate change: past and present (PNAS 2006)”, Energy &

Environment, 20, 367-373.

Miron, J.A., 1999: Violence and the U.S. Prohibitions of Drugs and Alcohol.

American Law and Economics Review, 1-2, 78-114.

Muni Krishna, K., 2008: Coastal upwelling along the southwest coast of

India – ENSO modulation, Ocean Science Discussion, 5, 123-134.

Peer Review Survey 2009: Preliminary Findings.

http://www.senseaboutscience.org.uk/

Popper, K.H., 1959: The logic of Scientific Discovery. Harper Torchbooks,

New York.

Reich, E.S., 2009: Plastic Fantastic: How the biggest fraud in Physics shook

the scientific world, Palgrave MacMillan, New York, 266 pp.

Rennie D., 1986: Guarding the guardians: a conference on editorial peer

review. JAMA, 256, 2391–2392.

Roy, C., and C. Reason, 2001: ENSO related modulation of coastal

upwelling in the Eastern Atlantic. Progress in Oceanography, 49, 245-255.

Rothwell, P.M., and C.N. Martyn, 2000: Reproducibility of peer review in

clinical neuroscience: Is agreement between reviewers any greater than

would be expected by chance alone? Brain, 123, 1964-1969.

Spier, R., 2002: The history of the peer-review process. Trends in

Biotechnology, 20, 8, 358-9.

Thompson, L. G., E. Mosley-Thompson, H. Brecher, M. Davis, B. León, D.

Les, P.-N. Lin, T. Mashiotta and K. Mountain, 2006: Abrupt tropical climate

change: past and present, Proceedings of the National Academy of Sciences,

103, 10536-10543.

63

Chapter 2

The Ocean

64

65

A Note on Eddy-Topography Interaction in the

Northwestern Gulf of Mexico*

Peter Hamilton ♠, SAIC

Abstract

A complex, evolving meso-scale eddy field characterizes the northwest

corner of the Gulf of Mexico, where the broad northern continental slope

turns into the steep western Mexican slope. Northwest slope eddies are

partially sustained by influxes of potential vorticity (PV) from the south,

where large Loop Current anticyclones interact with and dissipate against

the Mexican slope. Between April 2004 and July 2005, an array of 13 full-

depth moorings were deployed on the slope, north of 26°N and west of

94°W. Analysis of the velocity and vorticity fields in the upper 1000 m of

the water column showed a number of cases where a strong anticyclone

adjacent to the upper slope, induced offshore flows across the isobaths that

through stretching and PV conservation spun up a cyclone to the right of the

anticyclone when facing the slope. In at least one event, the interactions

between the upper slope, anticyclone and cyclone was associated with

velocity jets between 100 and 300 m below the surface. Sources of PV for

the slope eddy field appear to be where eddy flows are in contact with the

bottom between the shelf break and about 1000 m, consistent with numerical

modeling studies.

* The Ocean, The Wine, and The Valley: The Lives of Antoine Badan, 65-84.

© 2010 CICESE. ♠ Contact: PETER.HAMILTON@saic.com

66

1. Introduction

The northwestern Gulf of Mexico is a region of complex slope

topography where the broad, bathymetrically complex, northern continental

slope turns into the steep, east-west slope of the Mexican continental margin

(Figure 1). It is also oceanographically complex that displays a range of

eddy scales from major Loop Current anticyclones (Elliott, 1982; Brooks

and Legeckis, 1982) to small scale anticyclones and cyclones with radii of

order the baroclinic Rossby radius (Hamilton et al., 2002; Hamilton, 2007).

The impact of westward propagating Loop Current eddies with the steep

Mexican slope, south of ~ 25°N has been extensively studied from

observations (Elliott, 1982; Brooks, 1984; Lewis et al., 1989; Vukovich and

Waddell, 1991; Vidal et al., 1992) and models (Frolov et al., 2004; Nof,

1999; Shi and Nof, 1993; Smith, 1986). A principal result of Frolov et al.

(2004) was that offshore advection of potential vorticity (PV) over a

deepening slope by the anticyclonic flow of an eddy adjacent to a

continental shelf will spin up a cyclone on the right hand side of the

anticyclone when facing the slope. This has been observed a number of

times (e.g., Vulovich and Waddell, 1991). In a similar numerical study, Oey

and Zhang (2004) show that a cyclonic circulation can also be generated on

the left hand side of an anticyclone against a slope, and interaction of the

cyclone with the bottom boundary layer can generate a subsurface jet. Some

aspects of this model conform to some of the observations of subsurface jets

in the northwestern gulf discussed in Hamilton and Badan (2009). Hallberg

and Rhines (2000) show that a source of PV for upper layer interior flows of

the ocean is from where ocean currents are in contact with boundary sloping

67

topography. These ideas are explored using observations of cyclone spin-up

over the northwestern slope.

In 2004 and 2005 arrays of moorings were deployed in U.S. and Mexican

waters of the northwest gulf (Figure 1). The U.S. array, north of 26°N,

covered a roughly rectangular area from the 500 to 3000 m isobaths, and

was equipped with ADCPs, current meters and temperature and salinity

sensors to measure the water column with high resolution in the upper 450 to

500 m. The “W” array in Mexican waters was deployed about 6 months

later than the U.S. array (August versus April 2004) and overlapped by

about 11 months. In this study, use will be made of the W1 and W2

moorings to extend the current mappings southward. Unfortunately, the 75

kHz ADCP at 540 m on W3 failed to return any data. Details of the arrays

may be found in Donohue et al. (2008).

Hamilton and Badan (2009) used the same northwest gulf data to

investigate the occurrence of subsurface jets at depths between 100 and 400

m, with speeds exceeding 40 cm s-1. A total of 10 jets that were not caused

by vertically propagating inertial-internal waves, were identified in the upper

layer 75 kHz ADCP records. Though two of the strongest jets on the

northern slope occurred in water with bottom depths greater than 1000 m,

they both had precursors at locations nearer the upper slope. Maps of

relative vorticity implied that jets were associated with anomalous vorticity

fluxes of both signs that originated near the upper slope, and could be

associated with either an anticyclone or cyclone interacting with the slope

topography. In this paper, the spin-up of a cyclone on the western side of

the array and its subsequent eastward migration along the northern upper

slope in May and June 2004 strongly suggest that this eddy-topography

interaction was related to the jets observed at T3 and U4 on June 29 and July

68

10, 2004, respectively. Two jets were observed in 500 m water depth at W1,

and these also appear to be related to mesoscale eddies impinging on the

steep Mexican slope (Hamilton and Badan, 2009).

2. Database and Methods

A description of the mooring instrumentation is given in Hamilton and

Badan (2009) and Donohue et al. (2008). The locations of the full-depth

moorings on the northwest slope (T1 to V4) and Mexican waters (W1 to

W5) are given in Figure 1. All the NW slope moorings except T1, U1 and

V1 were equipped with 75 kHz ADCPs located at 450 m depths. Similarly

the W moorings were configured with 75 kHz ADCPs at nominal depths

between 435 and 550 m. The 540 m ADCP on W3 failed to return any

current measurements. All the NW slope moorings had temperature sensors

at 75, 150, 250, 350 and 450 m in the upper-layer, with moorings on the

500-m isobath, (T1, U1 and V1) having an additional temperature sensor at

90 m. The latter moorings were configured with a 300 kHz ADCP at 90 m

and conventional current meters (S4s) at 250 and 450 m. The W moorings

had no temperature measurements above the ADCPs. All the ADCPs

sampled at 60-minute intervals, and the independent upper-layer temperature

sensors at 30-minute intervals. After standard QA/QC and gap filling

procedures (see Donohue et al., 2008), the resulting time series with filtered

with a 3- and 40-hour low pass (HLP) Lanczos kernels, and decimated to 1-

and 6-hour intervals respectively. The latter were used to construct daily

averaged maps using Pedder’s (1993) method of successive corrections.

Where velocity vectors are mapped to a regular grid, the east and north

69

components are mapped separately, and then combined into vectors.

Therefore, the resulting field is not non-divergent.

Relative vorticity, ( / ) ( / )v x u yζ ∂ ∂ ∂ ∂= − , is calculated, following

Chereskin et al. (2000), for fixed depth levels from an array of moored

current measurements by fitting least square planes to the low-pass filtered

velocity data. Thus,

0

0

( , , , ) / /( , , , ) / /

u x y z t u x u x y u y HOTv x y z t v x v x y v y HOT

∂ ∂ ∂ ∂∂ ∂ ∂ ∂

= + + += + + +

(1)

where (x,y) are measured from the center position of the subset of the array

used for the estimates. The results are normalized by the local Coriolis

parameter (f = ~ 6.5 10-5 s-1), and typical standard deviations of the gradient

velocity terms are generally between 0.02f and 0.06f. The mooring locations

used for the relative vorticity estimates are given in Figure 1, where the

center position of each quadrilateral has been given a location number. For

example, calculations of ζ at location 30 use 40-HLP velocity records from

U1, T1, T2 and U2 at 3 depth levels (50, 250 and 450 m). The least-square

plane fits help to reduce noise in what is usually a noisy calculation if the

gradients are directly estimated from the measurements; however, it also

essentially averages ζ over the enclosing quadrilateral of stations.

Therefore, to be consistent, fluxes are estimated using (u0,v0) from (1), e.g.,

u0ζ,v0ζ( ), where ζ is not normalized.

70

3. Cyclone Spin-Up Event in May 2005

In May and June 2005, there was a good example of a cyclone being

spun up on the western side of the array from almost quiescent conditions.

The large change in the depth of the 11 °C isotherm surface is illustrated in

Figure 2. On May 15, the isotherms at depth are level with a weak warm

north to south ridge in the center part of the array (Figure 2a). A warm anti-

cyclonic intrusion moves in from the south, and begins to draw water from

upper western slope. The stretching of the water column produces cyclonic

vorticity, by PV = f + ς( ) h conservation, and through geostrophy a doming

up of the density surfaces (Figure 2b). This new western cyclone keeps its

integrity and under the influence of the anticyclonic swirl currents of the

intruding eddy begins to migrate northwards and eastwards, so that by June

15, the largest doming is on the northern T line (Figure 2c). The same

sequence, using objective maps of temperature at 75 and 450 m and currents

at 100 m (Figure 3), but at 5-day intervals, shows more clearly the strong

off-slope flows in the southern part of the array and the developing cyclonic

swirl currents between the anticyclone and the slope. The developing

cyclonic flow between May 20 and 30 enhances the anticyclonic on the

northern edge of the array by flows across shoaling isobaths. These

anticyclonic flows merge with the southern anticyclone around May 30, and

cause the western cyclone to migrate into the northwestern corner, and begin

to translate eastwards along the northern slope by June 20. Note that the

temperature maps in Figure 3 indicate that the spatial structures often change

with depth, particularly for the central anticyclone at the end of May. This

implies that vertical axis of rotation can become tilted as was as observed for

71

an anticyclone in July 2009, discussed in relation to subsurface jets in

Hamilton and Badan (2009); see their Figure 10.

Using (1) to calculate ζ and ζ fluxes at the two western locations, 30 and

31 (Figure 1), show visually coherent signals through the upper layer with

magnitudes decreasing with depth (Figure 4). The southwestern location,

31, has large northward fluxes of positive ζ at the end of May as the cyclone

forms roughly around location 30, where the fluxes are minimal but ζ

remains strongly positive through June. Therefore, the implication is that

the large northward flux of relative vorticity leads to an increase in positive

ζ in the northwest corner, assuming that the northern boundary is closed by

the shoaling topography and any eastward flux is blocked by the

anticyclone. The maps of ζ at 50 m (Figure 5) for May 15 and 30 show

convergence for the triangle 30, 31 and 33, which leads to larger values of ζ

on May 30 and June 15. Similarly, a divergence implies an increase in

anticyclonic vorticity, and this is apparent for the triangle 34, 37 and 41

between May 30 and June 15. The onshore flow between T2 and T3 that is

forced by the northwestern cyclone begins to develop around May 30

(Figure 5) and is associated with a weak divergence for the 32, 33 and 36

triangle has generated a second anticyclone in the northeastern corner of the

array by June 15. The two anticyclones and cyclone in array are clearly

interacting strongly over this June period.

4. Cyclone Intensification in May 2004

At the beginning of the U.S. array deployment in May 2004, there is a

nice example of a cyclone intensifying against the western slope. At the

72

start of the records (April 4, 2004), a cyclone – anticyclone pair were located

west to east across the slope. During April, the eddy pair fluctuated a little

in intensity and location, but remained configured similar to Figure 6a. Ten

days later on May 10, 2004, the western slope cyclone had intensified and

become more compact at approximately the same position with the lower

water column isotherms doming up by ~ 50 to 100 m (Figure 6b). The slope

currents on the western side of the cyclone, at U1, are almost barotropic, and

there is an indication of a null point on the southern part of the slope, around

V1, because of northwards anticyclonic slope flow south of 26°N that

produces off-slope flow on the southern boundary of the array (Figures 6a

and 6b). The cyclone intensification may be attributed to the generation of

cyclonic vorticity through off-slope stretching of the upper water column.

The strength of the western cyclone on May 10 is such that it distorts the

upper layer thermal structure above 200 m by interacting with the eastern

anticyclone to advect warmer water over the core of the cyclone. This

produces a characteristic dip of the upper layer isotherms in the U cross-

section in Figure 6b. This dip has been previously observed in hydrographic

surveys of other slope cyclones. For example, Hamilton et al. (2002)

showed similar cross-sectional thermal structures for lower slope cyclones

interacting with major Loop Current warm eddies. The implications of the

interaction of the cyclone – anticyclone pair in Figure 6b are that the flows

are becoming unstable with the center of the upper part of the cyclone,

represented by the 18 °C surface, being displaced to the southeast (compare

sections U and V in Figure 6b), and the eastern anticyclonic circulation

being split into two.

By the end of May (Figure 6c), the two parts of the eastern anticyclone

had moved further part, with the cyclone becoming weaker, with a more

73

regular vertical structure, and also moving off the slope and to the south as

the northern part of the anticyclone translates westward along the northern

slope. The cold dome is now more prominent in section V than section U

(Figure 6c). The westward movement of the warm eddy could be a result of

the anticlockwise flows of the northern part of the western cyclone. During

the first two weeks of June, the cyclone dissipates. It is noted that on May

28 (Figure 6c), that the southward nearly barotropic flows at U1 show

evidence of a subsurface jet developing with velocities at 250 m being larger

than at 75 m. Subsurface jets on the 500 m isobath were observed at W1

with high-resolution velocity profile data, as discussed in Hamilton and

Badan (2009), but here at U1, the vertical temperature profile indicates

stretching of the layers between 200 and 400 m relative to the interior

(Figure 6c). Similar layer PV anomalies were found for another strong

subsurface jet, at T3, associated with a slope cyclone in February 2005

(Hamilton and Badan, 2009).

Though the western cyclonic circulation begins to elongate and weaken

at the beginning of June, the northern part strengthens again into a cyclone

by the end of June as the northern half of the anticyclone moves westwards

to cause off-slope flow that contributes positive vorticity to the circulation.

The maps for the end of June and beginning of July are given in Figure 8 of

Hamilton and Badan (2009). At the end of June, a subsurface jet was

observed at T3 that probably arose from the interaction of the northern

cyclone with slope topography. This jet migrated off the slope with the

cyclone and was next observed at U4 on July 10-12, 2004. This subsurface

jet was also associated with an anomalous off-slope flux of negative ζ and

the strong flows between the anticyclone in the center of the array and the

cyclone on the western edge (see Figure 15b in Hamilton and Badan, 2009).

74

A case can be made that this southward flow off the northern slope

contributed to the increase in intensity of the western cyclone as it migrated

offshore through PV layer stretching.

5. Conclusions

The couple of examples of the eddy field over the northwestern slope

discussed herein, show that flow patterns are complex with the spin-up and

dissipation of both cyclones and anticyclones occurring as the flow fields

evolve and interact with the slope topography. In May 2005, a western slope

cyclone spun up from virtual rest over a period of 15 days and then

proceeded to migrate northwards and eastwards along the slope, interacting

with the anticyclone in the center and southern part of the array. In May

2004, an already established western slope cyclone strongly intensified over

a period of 5 to 10 days, causing the adjacent anticyclone to become

unstable and split into two. After which the cyclone elongated and

weakened, but later re-established itself, and similar to the May 2005

example, migrated northwards and eastwards around the anticyclone.

The spin-up of the cyclones can be partially explained through the

conservation of PV, where along slope flows in contact with the topography

are diverted to across slope by interior eddy circulations. The resulting

stretching generates positive relative vorticity that feeds the cyclonic flows.

There is some indication that the opposite can also occur with on-slope

flows compressing the water column and thus, generating anticyclonic

vorticity. The ~ 15 month observed slope eddy flows are not statistically

stationary (Donohue et al., 2008), and would need potentially many years to

resolve the lowest frequencies associated with the 6 to 15 month arrivals of

75

Loop Current warm eddies (Sturges and Leben, 2000) at the western slope.

These eddies and subsidiary cyclones formed from interactions with the

western slope (Vukovich and Waddell, 1991) interact and feed the northern

slope eddy fields (Hamilton, 2002; Donohue et al., 2008).

The observations are limited in spatial resolution, and only span a small

section of the slope, making it very difficult if not impossible to calculate

energy balances and transfers. One approach to a better understanding of

slope eddy fields and their instabilities would be by analyzing a high-

resolution, data assimilating circulation model to quantitatively evaluate the

eddy processes, followed by focused numerical experiments, such as those

by Frolov et al. (2004) and Oey and Zhang (2004) using simplified

circulations.

Acknowledgements

Minerals Management Service funded the northwest Gulf of Mexico study

through Contract 1435-01-03-CT-71562 to Science Applications

International Corporation (SAIC). The author wishes to thank Antoine

Badan and the CANEK group at CICESE for their contributions to an

excellent data set, and the program managers, Evans Waddell (SAIC) and

Alexis Lugo-Fernandez (MMS) for support during the preparation of this

paper.

76

Figure 1. Full-depth moorings (solid dots) deployed in the northwest Gulf of Mexico experiment. Inset of the Gulf shows the 200 and 2000 m isobaths. The numbered gray diamonds and the enclosing quadrilateral of dashed lines show the locations of relative vorticity calculations using velocities calculations using velocities from 4 surrounding moorings (see Section 2).

77

Figu

re 2

. C

onto

urs o

f dep

ths o

f the

11°

C (f

illed

), an

d 18

° C

(col

ored

thic

k lin

es) i

soth

erm

sur

face

s fo

r a) M

ay 1

5,

b) M

ay 3

0, a

nd c

) Jun

e 15

200

5. V

eloc

ity v

ecto

rs, i

n a

pseu

do-3

D v

iew

, are

at 7

5 m

(red

), 25

0 m

(gre

en),

430

m

(blu

e),

and

750

m (

purp

le).

Low

er p

anel

s: E

ast-w

est

verti

cal

tem

pera

ture

sec

tions

thr

ough

the

ind

icat

ed U

m

oorin

gs (a

and

b),

T m

oorin

gs (c

) (se

e gr

ay d

ashe

d lin

e on

map

s). A

ll qu

antit

ies

are

deriv

ed fr

om 1

-day

ave

rage

s of

40-

HLP

tim

e se

ries.

78

Figure 3. Sequence of maps at 5-day intervals of daily averaged 40-HPL temperature at 450 m (solid fill, horizontal upper scale) and 75 m (thick contour lines, vertical lower scale), and objectively mapped daily averaged currents at 100 m. Red arrows are the measured currents.

79

Figu

re 4

. R

elat

ive

vorti

city

(thi

ck li

nes,

frac

tiona

l f, r

ight

han

d sc

ale)

and

rela

tive

vorti

city

flux

(stic

k ve

ctor

s, up

is

nor

th, m

s-2 1

0-6, l

eft h

and

scal

e) a

t the

indi

cate

d de

pths

for

the

loca

tions

on

the

wes

t sid

e of

the

arra

y du

ring

Apr

il –

June

, 200

5.

80

Figu

re 5

. R

elat

ive

vorti

city

(fra

ctio

nal f

) con

tour

s (s

olid

fill)

at 5

0 m

, rel

ativ

e vo

rtici

ty fl

uxes

at 5

0 m

, 250

m a

nd

425

m (

red,

gre

en a

nd b

lue

arro

ws)

, and

opt

imal

ly m

appe

d ve

loci

ties

at 7

5 m

(gr

ay a

rrow

s), f

or d

aily

ave

rage

d va

lues

for t

he g

iven

dat

es a

t 15-

day

inte

rval

s dur

ing

May

and

June

200

5.

81

Figu

re 6

. To

p pa

nels

: Con

tour

s of

dep

ths

of th

e 11

° C

or 7

° C

(fil

led)

, and

18°

C (

colo

red

thic

k lin

es)

isot

herm

su

rfac

es fo

r a) A

pril

29, b

) May

10,

and

c) M

ay 2

8, 2

004.

Vel

ocity

vec

tors

, in

a ps

eudo

-3D

vie

w, a

re a

t 75

m (r

ed),

250

m (g

reen

), 43

0 m

(blu

e), a

nd 7

50 m

(pur

ple)

. Lo

wer

pan

els:

Eas

t-wes

t ver

tical

tem

pera

ture

sec

tions

thro

ugh

the

indi

cate

d U

and

V m

oorin

gs (s

ee g

ray

dash

ed li

nes

on m

aps)

. All

quan

titie

s ar

e de

rived

from

1-d

ay a

vera

ges

of

40-H

LP ti

me

serie

s.

82

References

Brooks, D. A., 1984: Current and hydrographic variability in the

northwestern Gulf of Mexico. J. Geophys. Res., 89, 8022-8032.

Brooks, D. A. and R. V. Legeckis, 1982: A ship and satellite view of

hydrographic features in the western Gulf of Mexico. J. Geophys. Res., 87,

4195-4206.

Chereskin, T. K., M. Y. Morris, P. P. Niiler, P. M. Kosro, R. L. Smith, S. R.

Ramp, C. A. Collins, and D. L. Musgrave, 2000: Spatial and temporal

characteristics of the mesoscale circulation of the California Current for

eddy-resolving moored and shipboard measurements. J. Geophys. Res., 105,

1245-1270.

Donohue, K., P. Hamilton, R. R. Leben, D. R. Watts, and E. Waddell, 2008:

Survey of deepwater currents in the northwestern Gulf of Mexico, Volume

II: Technical report, OCS Study MMS 2008-031, U.S. Dept. of the Interior,

Minerals Management Service, Gulf of Mexico OCS Region, New Orleans,

LA, 375 pp.

Elliott, B. A., 1982: Anticyclonic rings in the Gulf of Mexico. J. Phys.

Oceanogr. , 12, 1292-1309.

Frolov, S. A., G. G. Sutyrin, G. D. Rowe, and L. M. Rothstein, 2004: Loop

Current eddy interaction with the western boundary in the Gulf of Mexico. J.

Phys. Oceanogr., 34, 2223-2237.

Hallberg, R. and P. B. Rhines, 2000: Boundary sources of potential vorticity

in geophysical circulations. Developments in geophysical turbulence, R. M.

Kerr and Y. Kimura, Eds., Kulwer, 51-65.

83

Hamilton, P., 2007: Eddy statistics from Lagrangian drifters and

hydrography for the northern Gulf of Mexico slope. Journal of Geophysical

Research, 112, C09002, doi:10.1029/2006JC003988, 1-16.

Hamilton, P. and A. Badan, 2009: Subsurface jets in the northwestern Gulf

of Mexico. J. Phys. Oceanogr., 39, doi: 10.1175/2009JPO4158.1, 2975-

2891.

Hamilton, P., T. J. Berger, and W. Johnson, 2002: On the structure and

motions of cyclones in the northern Gulf of Mexico. J. Geophys. Res., 107,

3208.

Lewis, J. K., A. D. Kirwan, Jr., and G. Z. Forristall, 1989: Evolution of a

warm-core ring in the Gulf of Mexico: Lagrangian observations. J. Geophys.

Res., 94, 8163-8178.

Nof, D., 1999: Strange encounters of eddies with walls. J. Mar. Res., 57,

739-761.

Oey, L.-Y. and H.-C. Zhang, 2004: The generation of subsurface cyclones

and jets through eddy-slope interaction. Cont. Shelf Res., 24, 2109-2131.

Pedder, M. A., 1993: Interpolation and filtering of spatial observations using

successive corrections and Gaussian filters. Mon. Weather Rev., 121, 2889-

2902.

Shi, C. and D. Nof, 1993: The splitting of eddies along boundaries. J. Mar.

Res., 51, 771-795.

Smith, D. C., IV, 1986: A numerical study of Loop Current eddy interaction

with topography in the western Gulf of Mexico. J. Phys. Oceanogr., 16,

1260-1272.

Sturges, W. and R. Leben, 2000: Frequency of ring separations from the

Loop Current in the Gulf of Mexico: A revised estimate. J. Phys. Oceanogr.,

30, 1814-1819.

84

Vidal, V. M. V., F. V. Vidal, and J. M. Pérez-Molero, 1992: Collision of a

Loop Current anticyclonic ring against the continental slope of the western

Gulf of Mexico. J. Geophys. Res., 97, 2155-2172.

Vukovich, F. M. and E. Waddell, 1991: Interaction of a warm ring with the

western slope in the Gulf of Mexico. J. Phys. Oceanogr., 21, 1062-1074.

85

On the Acoustic Backscatter Strength and

Vertical Motion Signals from ADCP

Measurements in the Gulf of Mexico*

José Ochoa ♠, CICESE Julio Candela, CICESE Julio Sheinbaum, CICESE Helmut Maske, CICESE

Abstract

Zooplankton and micronekton daily migrations were detected in Acoustic

Doppler Current-meter Profilers (ADCPs) regarding both, their vertical

velocity and backscatter strength change. Measurements in 9 ADCPs

moored in the Gulf of Mexico were used in this study. The daily vertical

fluctuations showed maximum upward and downward velocities near sunset

and sunrise, as an odd function when referenced to midnight or noon. The

backscatter strength daily cycle appeared as an even function relative to

noon when, depending on the vertical level, either the minimum or

maximum occurs. The cycle amplitudes, defined as the maximum minus the

minimum within a cycle (i.e. within a day) in backscatter strength and

vertical velocity showed a highly correlated modulation throughout the year.

This high correlation may be due to: i) the simultaneous increase in

biological activity (i.e. swimming) and the concentration of specimens (i.e.

* The Ocean, The Wine, and The Valley: The Lives of Antoine Badan, 85-102.

© 2010 CICESE. ♠ Contact: jochoa@cicese.mx

86

biomass), and ii) the consequential increase in scattering perturbations

caused by faster swimming organisms.

1. Introduction

Before the advent of ADCPs most studies on the diel cyle inferred but

did not measure the vertical velocity. In the past 20 years the use of ADCPs

dramatically increased the available measurements related with diel cycles.

An early example recognizing the capacities of ADCPs in this respect is the

work by Plueddemann and Pinkel (1989). ADCPs are based on the acoustic

return caused to a large extent by ‘swimming’ organisms. The velocity

measured by ADCPs is a weighted mean of the motion of sound

perturbations, these being self-propelled or drifting within the sonified

volume. Measurements of ADCPs related with the biota include velocity

and backscatter strength. Multiple studies have dealt with the estimation of

zooplankton biomass via the backscatter strength of echo sounders such as

ADCPs (see for example Weeks et al. 1995; Flagg and Smith 1989; Jiang et

al 2007). In particular, Roe et al. (1996) showed a method to translate the

ADCP measurement of relative acoustic backscatter into mean volume

backscatter strength and, with the addition of simultaneous hydrographic

data, estimate biomass. In the absence of hydrographic data, the present

study employed the recorded ADCP backscatter strength as a biology-related

signal.

The purpose of this work is to show the high correlation between the

backscatter strength and vertical velocity signals, in the mean daily cycle

and in the amplitude from minimum to maximum every day throughout the

year. Some plausible explanations are offered. In particular, one

87

explanation is based on the observation that backscatter strength may depend

on sound speed perturbations due to stirred hydrographic fields (Ross and

Lueck, 2003). Not much attention has been paid to the acoustic properties of

the leftover in the wakes of swimmers as a source for backscatter recorded

by ADCPs.

The section that follows deals with a description of the data, in particular

of the time series and mean cycles of two layers, each 150 m thick, one

centered at 425 m and the other at 1075 m deep. Sheinbaum et al (2010)

offer an ample description of the data, and partial results of the ones

pertinent to this study. Mean diel cycles, in vertical velocity and backscatter

strength, are the average of individual cycles timed relative to sunrise and

sunset. The second section also shows time series of the daily amplitudes,

defined with the extreme values within 24 hours. These series show very

high correlations. The third and last section includes a discussion on the

possible cause of such high correlation.

2. Measurements

For a description of ADCPs, like the ones used in this study, the reader is

referred to RDI (1996). Basic to the ADCP measurement is the fact that the

sound-scattering features are perturbations in the sound speed distribution

with sizes near half wavelength of the emitted sound. This is the Bragg

scale, and for the near 75 KHz sound in use it implies scattering features of

the order of 1 cm. Another basic aspect of the measurement is that the

instrument uses a straightforward Doppler technique (narrowband) or its

generalization (broadband) to compute the velocity of the scattering features,

regardless of their movement in relation to the surrounding water.

88

The ADCPs were programmed for 10 m thick bins and recordings were

taken every half hour of the average velocities in 18 evenly distributed

‘pings’ (i.e. individual measurements). Figure 1 and Table I contain

information on location and length of available series.

2.a A shallow and a deep 150 m thick layers

In order to produce a high signal-to-noise ratio in the time series, we

averaged ADCP bins 5 to 20, which were neither too close nor too far from

the ADCP. Previous analyses have shown that these bins provide better

quality data (see for example Sheinbaum et al 2007). Figures 2 and 3 show

four available cycles averaged over 150 m thick layers as a function of time

corresponding to mooring WG-5 (Table 1 and Fig. 1 for location

information). They show the well-known timing of daylight change and

migration cycles. The daylight cycle is indicated in these figures, akin to

Figure 4 of Rippeth and Simpson (1998), by the step-wise function that is

high/low during daylight/night with jumps at sunrise (Sr) and sunset (Ss).

We use the conventional definition for sunrise and sunset: the times when

the upper edge of the disk of the sun is on the horizon. The sunrise and

sunset, which by definition must be close to 06:00 and 18:00 in standard

time, are for 96º 18.071’ W, the longitude of WG-5, close to 12:00 and

24:00 UT (see Figs. 2 and 3). The time series of backscatter strength shown

in Figure 3 are vertical averages (i.e. bins 5 to 20) as are the average

velocities in Figure 2.

To produce a meaningful mean diel cycle by averaging, the time scale

must be referenced to the sunrise and sunset times. The first half of each diel

cycle (the downward phase) was timed with differences relative to sunrise,

89

and each second half fraction of the cycle, the upward phase, was timed

relative to sunset. The original velocity time series was oversampled with

linear interpolations every 3 min interval, covering from 5 h before to 5 h

after sunrise and the mid-sample exactly at sunrise for each cycle or day of

the data set, and an analogous coverage was calculated for the sunset half

period. This allows the averaging of the daily cycles, all referenced to their

corresponding sunrise and sunset (see Fig. 4). Here, corrections for vertical

excursions of the ADCP were disregarded; the deep ADCP had a maximum

depth of 1264 m and a minimum of 1262 m, while the near surface reached

maximum and minimum of 557 m and 552 m. As illustrated in the examples

of individual cycles shown in Fig. 2 and of average cycles shown in Fig. 4

the deep and near surface cycles have peak velocities at different times. In

the downward phase, the near surface biota reaches the maximum speed

during twilight before the sunrise, but the deep biota has its maximum

downward motion well after sunrise. In contrast, the maximum upward

migration speed occurs earlier in the deeper layer. There is a longer nightly

shallower stay for deep than for near surface biota.

Regarding backscatter strength, to compare different levels it is

appropriate to define anomalies as in van Haren (2007) via:

( , ) ( , ) ( )I z t I z t I z∆ ≡ − (1)

where I is the recorded backscatter strength (in RDI relative counts), which

depends on depth (z) and time (t), the symbol denotes the averaging of

all the available samples in the time series of a given depth and ∆ defines

anomaly.

90

The mean cycle in backscatter strength anomaly is rather different at 425

m depth from that at 1075 m (Fig. 4).

2.b Time series of daily amplitudes in the two 150 m thick layers

For each diel cycle, in these 150 m-thick layers, we defined amplitudes

as the maximum minus the minimum within the cycle. The set of amplitude

pairs of vertical velocity and backscatter strength anomaly showed a high

correlation when lumped by depth level (Fig. 5). The linear regressions of

the velocity amplitudes in terms of the backscatter amplitudes were different

for each depth level but the correlations were similar. Taking all available

pairs of velocity and backscatter amplitudes, at the level centered at 425 m

the correlation was 0.83 (from 0.74 to 0.88 at 95% confidence level) and at

1075 m the correlation was 0.85 (from 0.77 to 0.90 at 95% confidence level)

(Fig. 5). Confidence intervals are computed by a Monte Carlo method. The

sample auto- and cross-correlation functions are

>+≡< )(')(')( τδδτ twtwCww , >+≡< )(')(')( τδδτ tItICii ,

and ( ) '( ) '( ) ,Cwi w t I tτ δ δ τ≡< + >

where ><−= wtwtw δδδ )()(' , ><−= ItItI δδδ )()(' and the assumption about

stationary allows the sole dependence of averages ( >< .. ) on the time lag (τ )

and their independence from the time reference ( t ) (i.e.

>+++>=<+< )()()()( ττ TtbTtatbta for any T). These functions are used to

design filters such that filtering Gaussian white noise produces synthetic

time series that follow the sample auto- and cross-correlation functions. The

91

synthetic series can be much longer than the available series of observations,

and an ample number of them can be built. Specifically, the series:

Cxgxhy ** 1111 += (2.1)

Cxgxhy ** 2222 += (2.2)

where 1x , 2x and Cx are independent Gaussian random variables of zero

mean and unitary variance, 0)()( >=+< τtxtx ij for i and j=1,2 and C, and

any ,...2, tt ∆±∆±=τ produce )()()( 11 ττ Cwwtyty >=+< , )()()( 22 ττ Ciityty >=+<

and )()()( 21 ττ Cwityty >=+< by choosing )||1((1 wwSCohh −= −1F ,

)|)|1((2 iiSCohh −= −1F , )(1 wwSCohg ⋅= −1F and )*)((2 iiSCohg ⋅= −1F where

is the Fourier Transform operator, and )( wwww CS F= , )( iiii CS F= ,

)( wiwi CS F= , iiwwwi SSSCoh /= are the spectra, co-spectra and coherence

functions. The use of synthetic series takes in accounts all time-lagged

correlations, yielding realistic confidence intervals.

The correlation is higher when filtering out high frequency components,

as shown in Figure 6, but the confidence diminishes (i.e. a wider interval),

because neighboring values are highly correlated. This high correlation of

neighboring values becomes the chief problem to compute confidence

intervals of any statistics, and was presently solved with the Monte Carlo

method just described.

92

3. Discussion

3.a Vertical variation in mean diel cycles of velocity and backscatter

strength

In the layer of 350 m to 500 m depth the maximum migration speed in

the daily cycle coincided with peaks in the backscatter strength: one peak

near sunrise and the other near sunset. The increase in backscatter is likely

due to the increase in the amount of individuals, whose concentration is low

at those depths near noon and midnight. The concentration is lower at

midnight when most of the nekton is at shallower levels. Some increase in

the backscatter intensity may be attributed to nekton orientation during the

swimming period. A speculation is that sound perturbations also increase

due to the wakes left by swimming individuals, i.e. to the stirring produced

during migration. This speculation is not easily discarded; the salt molecular

diffusivity is near scm /103.1~ 5−× , hence filaments or wakes of width size ~

1 cm have decaying time scales of ~ 20 h.

The speed diel cycle shows narrow peaks in shallow levels and broad

peaks in deeper levels. The timing between the maximum of the peaks is

closer to noon for the deeper than for the shallow level cycles. The resulting

mean cycle at the layer from 1000 m to 1150 m deep, as shown in Figure 4,

has the broad backscatter peaks lumped together at noon. In the deep

environment (at 1000 to 1150 m) the downward migration peaks ~2 h after

sunrise and slowly decays until it reverses, at noon, into the upward

migration which peaks ~1.5 h before sunset. In between the downward and

upward migration peaks the mean cycle shows an almost linear variation

with continuous swimming activity exhibiting no low backscatter strength

93

between both migrating phases. This produced the contrasting signal in the

mean backscatter strength cycle of the two layers shown in Figure 4.

3.b The correlation between vertical velocity and backscatter strength

daily amplitudes

The diel cycle amplitudes for vertical velocity and backscatter strength

are highly correlated (Fig. 5). Modulations throughout the year, in running

means of fifteen days, have correlations exceeding 0.8 at 95% confidence

level (Fig. 6).

The times when the minimum and maximum occur for vertical velocity

and backscatter strength are quite different. The velocity daily amplitude

comes from extremes that occur much regularly than backscatter strength

amplitudes. Even considering the delays for different depth levels, the

maximum upward migration is near sunset and the maximum downward

migration near sunrise. In contrast, for the layer from 350 to 500 m, the

backscatter strength amplitude is the difference from a high value near either

sunrise or sunset to a low value near noon. And, in the layer from 1000 to

1150 m the backscatter strength amplitude is the difference from the

maximum value near noon to the minimum near midnight.

As in the covariance in the mean diel cycles, the concentration of

individuals that occupy and abandon the layer is a cause for the modulation

of backscatter strength amplitudes. The question that arises is: Why does the

vertical velocity amplitude increase and decrease proportionally to the

amount of swimming biota? This high correlation is in agreement with the

speculation of a backscatter strength derived from the stirring of swimming

biota; there is no need for an increase/decrease in biota concentration. A

94

change in the swimming speeds, of the same distribution and amount of

individuals, changes the intensity of wakes and therefore of sound speed

perturbations. On the other hand, the rapid decay in the backscatter strength

for the shallow layer, between the migrating peaks, is evidence against the

production of wakes; the speed of sound perturbations associated with wakes

are very weak, or short lived, to significantly increase the backscatter

strength. The production of wakes with either insignificant or relevant

impact in sound speed perturbations merits further investigation.

Acknowledgements

This work was supported by grants from MMS (1435-01-02-85309) and

CONACYT (U50204-F). The participation of the crewmembers of the R/V

Justo Sierra and the technical staff of CICESE made it possible to collect the

data used in this study.

95

Date

yyyy/mm/dd

Name Latitude

°N

Longitude

°W

Water

Depth

Start End

Mean

Depth

WG-0 25.086 90.500 3640 2003/05/11 2004/08/27 368/1233

WG-1 25.436 96.314 447 2004/12/22 2005/11/09 436/NA

WG-2 25.389 95.439 2000 2004/08/26 2005/11/11 529/1240

WG-3 25.272 94.890 3524 “ “ Fail/1268

WG-4 24.651 96.078 1996 “ “ 501/Fail

WG-5 24.045 96.301 2003 “ “ 554/1263

Table 1. Positions, depths and time duration of moored ADCPs, all upward looking. All moorings had one ADCP close to 500 m below the surface and, when full depth was allowed, another ADCP near 1200 m. Depths are in m, the last column shows the instrument mean depth and nonexistent (NA) or failed (Fail) data collection.

96

Figure 1. Map with location of 6 moorings, refer to Table I for coordinates. The thin red line is the 2000 m isobath.

97

Figure 2. Four diel cycles in vertical velocity averaged over two 150 m thick layers, centered at 425 m and 1075 m in mooring WG-5. The time is in UT for 2004 and the thin black trace indicates the sunlight cycle (see text).

98

Figure 3. Four diel cycles in backscatter intensity (RDI measure in counts), in the same period and similar format as in Figure 2. Notice the backscatter strength (echo) intensification at sunrise and sunset during the downward and upward migrations in the near surface layer and, in contrast, the intensification during daylight in the deep layer signal.

99

Figure 4. Mean cycles, (A) for velocity, (B) for the backscatter strength anomaly (as defined in Eq. 1). The blue traces are for the layer centered at 425 m, the green for the layer at 1075 m.

100

Figure 5. Amplitudes (or difference from maximum to minimum) in each cycle of the vertical velocity ( wδ in cm/s) and relative backscatter intensity ( Iδ in RDI counts). Dots are for pairs at the 1075 m depth layer (moorings 2, 3 and 5), and crosses at the 425 m level (moorings 2, 4 and 5). Solid lines are the ordinary least square fits of the form kkk Iw εδβαδ +⋅+= such that

>< 2ε is a minimum.

101

Figure 6. Series of running means, of fifteen days length, of the amplitude in vertical velocity (solid trace with scale in the left vertical axis) and relative backscatter intensity (dashed trace with scale in right vertical axis). All frames are for averages over layers 150 m thick, the three top frames for layers centered at 425 m and the lower three at 1075 m. The numbers to the left in each frame show the 95% confidence limits for correlation.

102

References

Flagg, C.N., S.L. Smith, 1989: On the use of the acoustic Doppler current

profiler to measure zooplankton abundance. Deep-Sea Res., 36, 455-474.

Jiang, S., T.D. Dickey, D.K. Steinberg, and L.P. Madin, 2007: Temporal

variability of zooplankton biomass from ADCP backscatter time series data

at the Bermuda Testbed Mooring site. Deep-Sea Res., 54(4), 608-636.

Plueddemann, A.J., and R. Pinkel, 1989: Characterization of the pattern of

diel migration using Doppler Sonar. Deep-Sea Res., 36, 509-530.

Rippeth, T.P. and, J.H. Simpson, 1998: Diurnal signals in vertical motions

on the Hebridean Shelf. Limnol. Oceanogr., 43, 1690-1696.

Roe, H.S.J., G. Griffiths, M. Hartman, and N. Crisp, 1996: Variability in

biological distributions and hydrography from concurrent Acoustic Doppler

Current Profiler and SeaSoar surveys. J. of Mar. Sci., 53, 131-138.

Ross, T. and R. Lueck, 2003: Sound scattering from oceanic turbulence.

Geophys. Res. Lett., 30, 1344, doi:10:1029/2002GL016733.

Sheinbaum J., A. Badan, J. Ochoa, J. Candela, D. Rivas, and J.I. González,

2007: Full water column current observations in the central Gulf of Mexico.

U.S. Dept. of the Interior, Mineral Management Service, Gulf of Mexico

OCS Region, New Orleans, LA. OCS Study MMS 2007-022. xiv + 58pp.

van Haren, H, 2007: Monthly periodicity in acoustic reflections and vertical

motions in the deep ocean. Geophys. Res. Letters., 34, L12603,

doi:10.1029/2007GL029947.

Weeks, A.R., G. Griffiths, H. Roe, G. Moore, I. S. Robinson, A. Atkinson,

and R. Shreeved, 1995: The distribution of acoustic backscatter from

zooplankton compared with physical structure, phytoplankton and radiance

during the spring bloom in the Bellingshausen Sea. Deep-Sea Res., 42(4-5),

997-1019.

103

Solutions of Continental Shelf Waves Based on the

Shape of the Bottom Topography*

Luis Zavala Sansón ♠, CICESE

Abstract

Solutions of barotropic, rigid-lid topographic waves over an infinite family

of continental shelves characterized by a shape parameter are derived. The

bottom topographies are defined by a depth profile proportional to xs, where

x is the offshore coordinate and s is a real, positive number. As expected, the

resulting continental waves are characterized by subinertial frequencies and

by their propagation along the coast with shallow water to the right (left) in

the Northern (Southern) Hemisphere. The wave structure and the dispersion

relation are written as a function of the parameter s, which essentially

indicates that waves over steeper shelves have higher frequencies and phase

speeds. As an example, we determine the shape parameter of the continental

slope at four locations along the Eastern Pacific Ocean, and then we

calculate the corresponding wave properties. Another aim of the paper is to

underline the structure of the solutions in terms of the associated Laguerre

polynomials, which allow the introduction of the shape parameter. This

point is discussed to the light of previous studies that have reported solutions

in terms of Laguerre functions.

* The Ocean, The Wine, and The Valley: The Lives of Antoine Badan, 103-130.

© 2010 CICESE. ♠ Contact: lzavala@cicese.mx

104

1. Introduction

Coastal trapped waves in the oceans are oscillatory motions affected by

the Earth's rotation and are confined within a vicinity of coastal areas and

continental shelves. These waves have been thoroughly studied in the last 50

years from the theoretical, observational, numerical and experimental points

of view by a large number of researchers [important reviews were written by

Mysak (1980) and Brink (1991)]. This field of study is very extensive since

trapped waves can be associated with several factors, such as the shape of

the boundary (a straight or curved coast, an island), variable topography (the

continental shelf), ambient stratification, external forcing (storms, tides), and

combinations between them.

In order to develop theoretical models, past studies necessarily have done

strong simplifications of the physical mechanisms involved, as we shall do

here. A first consideration is the case of a homogeneous ocean, that is, in the

absence of stratification. Secondly, the geometry of the coastal boundary is

simplified as a straight coastline. In this barotropic system, the frequency of

topographic waves can be superinertial, inertial or subinertial (see, e.g.,

Huthnance 1975). Superinertial oscillations, also called edge waves, are

basically gravity waves affected by rotation and topography, which travel in

any direction along the coast. In cases where they are not trapped within a

vicinity of the boundary they are sometimes referred to as Poincaré "leaky"

waves (Mysak 1968). Subinertial motions are known as shelf waves, and

they are restricted to move along the coast with shallow water to the right

(left) when the Coriolis parameter is positive (negative). The mechanism of

these waves is associated with the conservation of potential vorticity.

105

In addition to their frequency and wave number, coastal oscillations are

also characterized by the shape of the depth profile. Indeed, wave

frequencies, and consequently phase speeds and group velocities, may vary

for a given wave number over different bottom topographies. Therefore,

there are neither universal solutions nor general dispersion relations for

coastal trapped waves (Huthnance 1975), since they depend directly on the

shape of the shelf. Thus, the bottom topography is often approximated with a

well-behaved analytical function.

Considering a semi-infinite domain in which [0, ]x∈ ∞ is the offshore

coordinate, several depth profiles h(x) have been proposed. One of the

earliest studies by Reid (1958) examined coastal waves over a linear profile

( )h x xα= where α is the bottom slope. That study shows the complete set of

solutions for superinertial and subinertial waves, as well as the asymptotic

limits for short and long waves. An additional complication arises when

considering a shelf of finite width L, as shown by Mysak (1968). In that

study, the wave solutions over the linear shelf were coupled with an external

solution outside, where the depth was considered constant. As a result, the

dispersion curves described by Reid (1958) are modified. Another well-

known topography is the exponential profile 20( ) xh x h e λ= with 1λ− the

length scale of the topography, which was used, among others, by Gill and

Schumann (1974) and by Gill (1982) in the context of shelf, barotropic

waves.

In this paper we examine a family of new solutions of coastal trapped

waves, whose behavior is determined by the shape of the continental slope.

The solutions are obtained for a bottom profile proportional to sx , where s is

a positive parameter that determines the monotonic shape of the shelf. A first

106

aim is to show the dependence of the wave frequency on the parameter s in

order to get a better understanding of the relevance of the topographic

steepness on the wave properties. The analysis is restricted to low-frequency

shelf waves, although in principle, it can be extended for edge waves. This

assumption means that temporal variations of the free-surface are ignored in

the continuity equation, allowing the introduction of a transport function

(rigid-lid approximation). In addition, the results apply for shelves with

widths small compared with the external radius of deformation (Gill and

Schumann 1974).

A second goal of this study is to call the attention to the offshore

structure of the solutions in terms of associated Laguerre polynomials. It is

shown that the analytical solutions of coastal trapped waves that include a

large family of bottom topographies are given in terms of these functions.

Furthermore, the solutions, characterized by a shape parameter of the shelf,

are obtained by a relatively simple procedure. This was also found in a

previous article for barotropic waves trapped around seamounts (Zavala

Sansón 2010), which suggests that similar solutions should be found in

problems with different topographic geometries.

The rest of the paper is organized as follows. The family of barotropic

wave solutions is derived in Section 2, and the structure and behavior of the

waves is given in detail. Section 3 includes some examples of continental

shelves at four locations characterized by a shape parameter; then, some

wave properties for these realistic cases are briefly discussed. In addition, we

comment on the structure of the waves in terms of the associated Laguerre

polynomials. The conclusions are presented in Section 4.

107

2. Wave solutions

2.a Derivation

Using Cartesian coordinates, the linear, shallow water equations for a

homogeneous fluid layer in a rotating system are

, t xu fv gη− = − (1)

, t yv fu gη+ = − (2)

( ) ( ) 0 , t x yhu hvη + + = (3)

where subindices denote partial derivatives, u and v are the velocity

components, η is the free-surface deformation, h is the fluid layer depth and

g is gravity. Hereafter we consider a semi-infinite domain (x,y) where

0 x≤ ≤ ∞ and y−∞ ≤ ≤ ∞ , bounded by a solid boundary at x=0 (the coast),

and in which the fluid depth is only a function of the off-shore coordinate x,

that is h(x). A second consideration is the rigid-lid approximation in the

continuity equation, by which we drop the time derivative of η. This is easily

understood by re-writing expression (3) in non-dimensional terms as

( ) ( ) 0 ,t x yhu hvδη + + = (4)

where 2 2/L Rδ = , L is a horizontal length scale and 1/2 2( ) /R gH f= is the

external radius of deformation, with H a depth scale (besides, we have made

108

use of η ~ UfL/g and t ~ 1/f). The rigid-lid approximation considers wave

motions with length scales much shorter than R, that is δ <<1. A strong

consequence of this approximation is that gravity waves are filtered out.

Thus, the velocity components can be defined in terms of a transport

function as

1 1 , . y xu vh hψ ψ= = − (5)

The vorticity equation is derived by subtracting the derivatives of the

momentum equations, which yields

( )( ) 0 .x y t x yv u f u v− + + = (6)

This well-known expression states that changes of relative vorticity are

associated with divergence or convergence of the flow as fluid columns

experience changes of depth. This is the basic mechanism of rigid-lid,

topographic waves (which can also be expressed as the conservation of

potential vorticity). Substituting the divergence from the continuity equation

and the corresponding expressions for the velocity components gives the

following equation for the transport function

0 .x xxxt yyt xt y

h hfh h

ψ ψ ψ ψ+ − + = (7)

Wave solutions are proposed of the form

109

( ) ( ) ( )1/2 ( ), , i ky tx y t h x x e ωψ φ += (8)

which yields an equation for ϕ:

2

21 1 0 .2 2

x x xxx

x

h h h fk kh h h

φ φω

⎡ ⎤⎛ ⎞ ⎛ ⎞+ − + − =⎢ ⎥⎜ ⎟ ⎜ ⎟⎝ ⎠ ⎝ ⎠⎢ ⎥⎣ ⎦

(9)

In the context of continental shelf waves, this expression is very suitable

when considering an exponential depth profile 2~ xh e λ , where 1λ− is the

length scale of the shelf, since the solutions are of the form ~ sin( )kxφ (see

Gill, 1982, p. 410). A different topography is considered here: the depth

profile is an arbitrary power of x with the following form:

( ) ( )0 ,s xh sh x h xh x

λ= ⇒ = (10)

where the parameter s>0 measures the shape of the shelf and h0 is a depth

scale. Evidently, larger s values mean steeper topographies for 1x λ−> .

Several studies on shelf waves use a linear profile s=1 (e.g. Reid, 1958).

Figure 1 shows some examples for different values of this parameter. The

advantage of this formulation is to find the wave properties as a function of

the topography shape. As a result, the following expression is obtained

2

22

1 0 . 2 4xxs s fks k

x xφ φ

ω⎡ ⎤⎛ ⎞

+ − + + − =⎢ ⎥⎜ ⎟⎝ ⎠⎣ ⎦

(11)

110

Figure 1. Depth profiles over continental shelves of the form

0( ) ( )sh x h xλ= for s = 0.5, 1, 2 and 3. Topographic parameters are ho = 1000 m and λ-1 = 50 km.

Applying the change of variable

( ) ( )2 ,kx xρ χ ρ φ= ⇒ = (12)

yields

2

2

1 1 0 . 2 4 2 4s s fs

ρρχ χρ ωρ

⎡ ⎤⎛ ⎞+ − + + − =⎢ ⎥⎜ ⎟

⎝ ⎠⎣ ⎦ (13)

111

The solution is obtained in terms of the associated Laguerre polynomials

with the following form (see Arfken, 1970, p. 620):

( ) ( )( 1)

2 2 ,j

jpe L

ρ

χ ρ ρ ρ+

−= (14)

where the indices j and p are defined by the following relationships

2 21 , 1 4 2 4

j s s j−= + > − ∈R (15)

2 1 , 0 2 2

p j sf pω

+ += ≥ ∈Z (16)

The solutions of the first equation are ( 1)j s= ± + . In general, j > -1 is a real

number, and therefore there can only be solutions for arbitrary s > 0 for the

positive root j = s+1. Index 0p ≥ , and the dispersion relation is derived

from expression (16):

2 2

sf p sω=

+ + (17)

Note that all waves are subinertial over a shelf with arbitrary s, and the

highest frequency corresponds to the wave with p=0. Also, there is no

explicit dependence of the wave frequency ω with the wave number k, in

112

contrast with waves over an exponential shelf (e.g. Gill, 1982). These and

other properties are further described in next subsections.

In order to find the complete solutions, we note first that

( ) ( ) ( )( 2)

122 2 ,s

kx spx Ae kx L kxφ

+− += (18)

where A is an arbitrary constant with appropriate units. When substituted in

(8), the full solution for the transport function is

( ) ( ) ( )2 1 1 ( )

0, , 2 2 ,2

s

skx s i ky tpx y t e kx L kx e

kωλψ ψ +− + +⎛ ⎞= ⎜ ⎟

⎝ ⎠ (19)

where 1/20 0Ahψ = is the arbitrary amplitude.

The horizontal velocity components are calculated by means of

expression (5) and taking the real parts:

( ) ( ) ( )10, , 2 sin ,kx s

pu x y t U kxe L kx ky tω− += − + (20)

( ) ( ) ( ) ( )10, , p s 1 2 2 kx s s

p pv x y t U e L kx kxL kx− +⎡ ⎤= − + + −⎣ ⎦

( ) cos ky tω× + (21)

where the (arbitrary) velocity amplitude is defined as:

113

( 1)2

00

( )0 2

(2 )( )[ ].

s

skU

hψ

λ

+

=

The x-derivative of the transport function was calculated by using the

following recurrence relation of the associated Laguerre polynomials

(Abramowitz and Stegun, 1972):

( ) ( ) ( ) ( )1 1 11 2 , s s s

p p pL pL p k L kxρ

ρ ρ ρ+ + +−⎡ ⎤ = − +⎣ ⎦ (22)

In order to write the polynomials with indices within the range of permited

values, an additional recurrence relation was also used:

( ) ( ) ( ) ( )1 11 1 2 , s s s

p p pL L p s L kxρ ρ+ +− = − + + (23)

2.b Spatial distribution

In order to describe the wave solutions, we first show the main structure

of the waves along the coast. Afterwards, we discuss the main

characteristics of the propagation of the waves in terms of the parameter s.

For these examples, the depth scale is 0 1000h = m and the length scale of

the topography is 1 50λ− = km. A positive Coriolis parameter 4 110f s− −= is

considered, which gives an inertial period of T ~ 0.72 days. Finally, recall

that the wave solutions have arbitrary amplitude.

Figure 2 shows the relative vorticity and velocity fields at time t = 0 for

three cases with different values of the alongshore wave number k, and using

114

s=2. The main structure of the waves is a set of positive and negative

relative vorticity patches arranged along the coast, traveling in negative y-

direction, i.e. with shallow water to the right. The patches have maxima and

minima at the coast and they rapidly decay offshore. The velocity field is

composed by the corresponding gyres near the boundary. Not surprisingly,

such a form is quite similar to other models of topographic waves. An

important point to notice is that the waves are trapped within a distance

determined by their own size along the boundary, i.e. 1k − , due to the factor kxe− .

How is the offshore structure of the waves? Figure 3 shows the relative

vorticity fields and offshore profiles for waves with p = 1, 2, 3 and setting k

= λ and s = 0.5. The vorticity profiles are taken from the positive maxima

located at (x = 0, y = 0) at time t = 0. These profiles show an oscillatory

behavior with strongly decreasing amplitude for large offshore distances.

Note that index p indicates the number of zero crossings of the vorticity. For

p = 1 there is one crossing: after reaching a minimum, the profile

asymptotically approaches zero for large x. For p = 2 there are two zero

crossings and, equivalently, for p = 3 there are three zero crossings. Thus,

index p is a natural measure of the offshore structure. Since the amplitude of

the profile rapidly decreases with x, however, such a structure is relatively

unimportant. Nevertheless, the parameter p is very important for the wave

propagation as shall be discussed in next subsection.

115

Figure 2. Alongshore structure of topographic waves over a shelf with 2, and traveling in the negative y-direction. Upper panels: Relative

vorticity contours for waves with alongshore wave number k λ= , 2λ and 3λ . In all cases p = 1. Thick (thin) contours indicate positive (negative) vorticity values of arbitrary magnitude. The domain is a 150 km × 300 km rectangular region. The horizontal scale 1λ− = 50 km is indicated by the vertical black line. Lower panels: Horizontal velocity vectors for the same waves. The magnitude of the vectors is arbitrary.

116

2.c Evolution of the waves

As the dispersion relation (17) indicates, the time evolution of the waves

depends directly of the shape parameter s and the offshore mode p. The

dispersion relation is plotted in Figure 4 for different cases. In the first panel,

the wave frequencies with different p values are shown for the case of a

linear sloping topography (s = 1). We show first this linear case as a

reference since it is considered in several previous studies (Reid 1958;

Mysak 1968). The plots are simple horizontal lines since there is no

dependence on k: all waves have the same frequency regarding their

alongshore size (given a fixed p). Dashed lines indicate the corresponding

curves derived by Reid (1958), who also included long waves in his analysis

(see his Figure 3). Of course, both models coincide for the short-wave

regime (compared with the deformation radius), where the frequency values

are given by

.2 3

fp

ω =+

(24)

117

Figure 3. Across structure of topographic waves over a shelf with s = 0.5. Upper panels: Relative vorticity surfaces for modes p = 1, 2 and 3. In all cases k λ= with 1λ− = 50 km. The domain is a 300 km × 600 km rectangular region. Contours as in previous figure. Lower panels: Offshore vorticity profiles of the same waves. The profiles begin at the origin and continue along the (eastward) horizontal direction. Vorticity values in the vertical axis are arbitrary.

In the work of Reid (1958) the modes are given by a positive integer n

( 1≥ ), which is related with index p ( 0≥ ) as n = p+1, i.e. the asymptotic

118

value found by Reid is / 2 1f nω = + (see his Table II showing the

asymptotic limits of the dispersion relation).

Wave frequencies as a function of the shape parameter for the first 5

modes (fixed p) are presented in the right panel of Figure 4. These curves

show one of the main points of the present study: low s values imply lower

frequencies, or waves with higher frequencies are developed over steep

slopes. Also, the gravest mode p = 0 implies higher frequencies. The

predicted values for linear slopes, s = 1, are shown with a circle

(corresponding with allowed frequencies shown in the first panel). On the

other hand, the star over the curve of the gravest mode p = 0 indicates the

frequency of a wave over a topography proportional to 1/2x . This value was

analytically calculated by Huthnance (1978) as 1/2( / ) [ 2 (9 5 ) ] / 5f skω = − + + , with S a stratification parameter; for the

barotropic case (S = 0), the present result is recovered, ( / ) 0.2fω = .

Trapped waves over this type of topography are dispersive with phase

speed

( )

. 2 2

fsck k p sω

= =+ +

(25)

Thus, larger waves (smaller k) travel faster. For a given wave (fixed k)

the phase speed as a function of the shape parameter has identical behavior

as the frequency curves in Figure 4 (with appropriate units). In other words,

waves over steep shelves travel faster than waves over weaker slopes. In

addition, over this type of topography, these waves do not transport energy

along the coast, since the group velocity dω/dk is null. This was also noticed

119

in the work of Reid (1958) for s = 1, who pointed out that energy

propagation is due to superinertial, edge waves (not present in this analysis).

Figure 4. Left: Dispersion relation for the first 5 modes of waves over a shelf with s = 1 (using f = 10-4 s-1). Dashed lines indicate the solutions of Reid (1958). Right: Wave frequency as a function of the shape parameter. Circles indicate the frequency for the linear shelf s = 1 and predicted by the model of Reid (1958). The star indicates the frequency of the gravest mode over a shelf with s = 1/2 calculated by Huthnance (1978).

3. Discussion

3.a Wave properties over realistic topographies

The main difference introduced in the present analysis in comparison

with previous studies, is the possibility of using different bottom

topographies according with the parameter s. It is therefore useful to

examine some wave properties over realistic bottom configurations obtained

from the ETOPO1 Global Relief Model (Amante and Eakins 2009).

120

In order to do this, we consider the continental topographies at the front

of four cities along the Eastern Pacific Ocean: two at the Northern

Hemisphere, Acapulco (Mexico) and Lincoln City (USA), and two at the

Southern Hemisphere, Lima (Peru) and Valparaiso (Chile). The continental

slope in each one of these cases is relatively uniform along the coast, i.e. it

has a more or less uniform profile at least for several tens of kilometers, as

shown in Figure 5. The bottom topography along the transect perpendicular

to the coast is plotted in Figure 6. Note that this transect has different

longitudes for each case. In addition, we plot the corresponding theoretical

topography 0( ) ( )sh x h xλ= (dashed lines), as defined in (10). The

topographic parameters are calculated as 0h h= , where the bar indicates the

average along the transect, and the horizontal scale 1λ− is the distance at

which the fluid depth is approximately the mean depth, i.e. 1( ) ~h hλ− . We

can reasonably assume the width of the shelf as 2 1λ− . The shape parameter

is indicated below the curves, and it is clearly different from unity (except in

Acapulco). Next, we can examine the wave properties according to these

values.

Table 1 shows the topographic parameters of the four topographies and

the corresponding frequencies and phase speeds of the first two modes (p =

0, 1), indicated with a subindex. Recall that the frequencies for the present

solutions do not depend on the wavenumber. The phase speeds are

calculated for a traveling wave with wavenumber k = λ/2, that is, of the same

order as the width of the topography. At the last column we have included

the ratio δ, which is verified to be much smaller than unity in order to justify

the rigid-lid approximation (when considering stratification effects this

approach might not be valid). How do the calculated values compare with

121

previous observations at these sites? In general, the comparison is pretty

satisfactory in some cases, and not so good in others, as shall be discussed

below. However, it is important to recall that the present results are not

intended to explain or reproduce observational results, but to provide a

theory that takes into account the shape of the topography.

Figure 5. Continental slopes in front of four cities (indicated with stars) along the Eastern Pacific Ocean. (a) Lincoln City (124° W, 45° N), (b) Acapulco (100° W, 17° N), (c) Lima (77° W, 12° S), and (d) Valparaiso (71.6° W, 36° S). In all cases the contour increment is 200 m. Maximum-minimum depth contours are: 1000-0 m, 4000-0 m, 4000-0 m and 3000-0 m, respectively. Bottom topographies have 1 min resolution (Amante and Eakis 2009). The bottom profiles along the straight lines ending at the locations are shown in Figure 6.

122

Figure 6. Bottom topography profiles along the transects indicated in Figure 5. Solid lines: data from ETOPO1 (Amante and Eakis 2009). Dashed lines: profile 0 ( )sh xλ . The topographic parameters are shown in Table 1.

For the case of the Oregon coast at Lincoln City, Table 1 shows

frequencies of 0.78 and 0.54 cpd for the first two modes. These values

overestimate the observations of Cutchin and Smith (1973) (0.2-0.3 cpd)

performed a few kilometers south of the same location, perhaps due to the

absence of bottom friction and/or stratification effects in the model. Near

Acapulco, Enfield and Allen (1983) reported winter phase speeds of about

1.6-2.7 m s-1, which are faster than the calculated for the first mode (1.0 m s-

1). The difference is expected to be larger in summer, since the waves along

123

the Mexican coast are mainly triggered by tropical storm forcing (see also

Martínez and Allen 2004). For the Lima transect, the predicted frequencies

for the first two modes are 0.21 and 0.14 cpd, which agree well with the

frequency band 0.1-0.2 cpd measured by Romea and Smith (1983) who

performed observations along the Peruvian coast. These authors estimate the

phase speed between 2-3 m s-1, while the present formulation indicates 2.6

and 1.7 m s-1 for the first two modes, respectively (Brink (1982) obtained

similar values with a more sophisticated numerical model). For the Chilean

coast there are reports of traveling disturbances at 3 m s-1 (Pizarro et al.

1994) which is of the same order than the calculated value for the zero

mode, 3.3 m s-1. Although the group velocity is usually more interesting,

recall that it is zero in the present model.

Units Lincoln C. Acapulco Lima Valparaiso

Φ 45° N 17° N 12° S 36° S

0h M 294 1995 1248 835

1λ− Km 50 35 86 48

S 2.5 1.0 2.0 1.3

Cpd 0.78 0.19 0.21 0.46

Cpd 0.54 0.12 0.14 0.29

m s-1 5.8 1.0 2.6 3.3

m s-1 4.0 0.6 1.7 2.0

∆ 0.0380 0.0005 0.0022 0.0083

Table 1. Wave properties at the four locations indicated in Figure 5.

124

It must be emphasized that the discrepancies observed between some of

the calculated properties and available observations are likely to be

explained by ambient stratification, external forcing or bottom friction

effects, among other relevant mechanisms. The main point of this

discussion, however, is to show that real continental slopes fit well with the

profile proportional to xs, and that s is between 1 and 2.5, at least for the sites

shown here. Furthermore, the wave properties (frequency and phase speed)

are easily estimated by using a very simple formula for the dispersion

relation, and in some cases realistic values are recovered.

3.b Some remarks on the use of associated Laguerre polynomial

The offshore structure of the waves studied here is given in terms of the

associated Laguerre polynomials. These functions are known since the 19th

century, when they were studied by the French mathematician Edmond

Laguerre (born in 1834). Several applications have been found in physical

problems since then. But these polynomials became particularly famous at

the beginning of the 20th century, when they were used for solving the radial

part of the Schrodinger wave equation applied to the hydrogen atom. The

integral character of the subindex (p in our case) gave rise to the quantization

of the energy in this atomic model, which constituted one of the greatest

achievements at the early stages of quantum mechanics (see, e.g., Arfken

1972).

In the context of topographic waves, previous models where the offshore

part of the wave solutions is given in terms of Laguerre polynomials is the

already cited work of Reid (1958), as well as the model reported by Mysak

(1968). In both cases the bottom topography is given by a shelf with a linear

125

profile, s = 1. The solutions of Reid are obtained over a semi-infinite plane

(as the one postulated here) and the offshore structure is given in terms of

Laguerre polynomials Ln with n an integer 0≥ . Mysak extended this theory

for a finite width shelf and the offshore solutions are Laguerre functions Lν

with ν a real number subject to some restrictions associated with the

topographic discontinuity. In the present case, we found associated Laguerre

polynomials (or generalized Laguerre polynomials) jpL , for waves over

bottom profiles proportional to xs. Thus, the point to underline here is that the

associated Laguerre polynomials appear as a consequence of considering this

infinite family of depth profiles, providing a more versatile, analytical model

of bottom topographies. Of course, the full solution of the present waves are

reduced to Reid's solutions when s = 1, as shown in Figure 4 (in the context

of rigid-lid waves).

The use of associated Laguerre polynomials was reported before in a

previous work by Zavala Sansón (2010) in the context of trapped waves

around axisymmetric seamounts. In that study, rigid-lid solutions were found

for seamounts with an exponential depth profile of the form exp( )s srλ ,

where r is the radial coordinate measured from the center of the seamount

and s is, as in this study, a shape parameter of the topography. Large s values

imply a flat-topped seamount, and small s means a sharp-peak mountain. The

radial part of the solutions was given in terms of the associated Laguerre

polynomials. Analogously to the present case, these functions give the wave

structure along the direction normal to the topography contours.

Another important point related with the polynomials is the form of the

wave solutions given by equation (8), which includes a factor h1/2. Some

studies proposing such a form are those of Gill (1982) for shelf waves and

126

Rhines (1969) for seamounts. This choice is essential to reduce the equation

for the offshore structure of the wave (radial, for the case of seamounts) into

a suitable form that can be solved with associated Laguerre polynomials.

However, the example shown by Gill uses an exponential shelf, while the

work of Rhines makes a further approximation to obtain solutions in terms of

Bessel functions. Other studies (like Reid's and Mysak's) do not include the

factor h1/2, and simply express the wave solutions of the form ( )( ) i kx tF x e ω+ . It

must be recalled that these latter studies, however, solve the problem

including gravity waves (no rigid-lid approximation). Given the present

results, we infer that the solution of the full problem (including superinertial

and subinertial oscillations) over arbitrary shelves proportional to xs might be

tractable with the present approach and the solutions should include

associated Laguerre polynomials. An analogous situation should apply for

seamounts, other topographic features or different basin geometries that can

be written in a general form determined by a shape parameter s.

4. Conclusions

We have derived solutions of barotropic, rigid-lid topographic waves that

travel over a continental shelf with depth profile proportional to xs, where s

is a real, positive number. The analysis is restricted to the case of shelf

widths much smaller than the external radius of deformation, for which

gravity waves are filtered out. The remaining set of waves are usually called

continental shelf waves (or sometimes quasigeostrophic waves), and they are

characterized by subinertial frequencies and by their propagation along the

coast with shallow water to the right (left) in the Northern (Southern)

Hemisphere. The waves are trapped in the sense that they rapidly decay in

127

the offshore direction and move in the alongshore direction. All these

properties are present in the family of solutions developed here. The novel

result is that the wave structure and characteristics are derived in terms of

the shape parameter s, so these solutions allow the description of the waves

over an infinite set of continental shelves defined as powers of the offshore

coordinate.

Wave frequencies as a function of parameter s are given by the dispersion

relation (17) and the corresponding curves were shown in Figure 4.

Basically, this plot describes that frequencies are increased over steeper

shelves (greater s values) for all modes of oscillation. For instance, the

frequency of the gravest mode (p = 0) over a parabolic depth profile (s = 2)

reaches a value of / 2fω = . In contrast, over a square-root shelf (s = 0.5)

the frequency is lower, / 6fω = .

We also showed some examples of real continental slopes for which a

profile proportional to fits quite well. These locations are situated at the

Northern (Lincoln City and Acapulco) and Southern (Lima and Valparaiso)

Hemispheres. The wave properties and shape parameter were presented in

Table 1. The shape parameter is within the range of 1 to 2.5. Repeating the

same procedure at other locations should be a rather simple exercise and it

might give some indications on the expected frequencies and phase speeds

of freely evolving barotropic waves.

Another aim of this paper is to call the attention to the offshore structure

of the solutions in terms of associated Laguerre polynomials. These general

functions allow one to solve the topographic wave problem in the presence

of depth profiles defined by a shape parameter, improving previous

analytical models. Besides the present case for continental shelves, this was

also shown by Zavala Sansón (2010) for barotropic waves trapped around

128

seamounts suitably written in terms of a shape parameter (see previous

section). Thus, we strongly suggest that there might be more geometries

allowing this type of structures in cases of oceanic interest, such as basins,

ridges or canyons. This is part of current research by the author and

collaborators.

Finally, an important remark must be made, namely, that waves over

different bottom topographies might show very different characteristics. For

instance, Gill (1982) considered a depth profile of the form 2~ xh e λ , where

now y is the offshore coordinate (just rotate 90° our present coordinate

system), and whose offshore solutions are proportional to ~ sin( )lyφ (l

represents the wave number in that direction). This is essentially a different

problem, since the depth field does not vanish at x = 0. In this case, the

dispersion relation is

2 2 2

2 kf k lω λ

λ=

+ + (26)

with k the alongshore wave number (a similar expression was reported by

Allen (1975) for coastal waves in a stratified ocean). These waves certainly

have a different behavior with respect to those presented here. For instance,

the group velocity is different from zero except at the highest frequency,

where the slope of the dispersion curve changes sign. Summarizing, the

behavior of topographic waves might strongly differ depending on the model

considered. Or, as pointed out by other authors (Huthnance 1975): there is

no universal dispersion relation for waves over arbitrary topography.

129

References

Abramowitz, M. and I.A. Stegun, 1972: Handbook of mathematical

functions. National Bureau of Standars, 1046 pp.

Allen, J. S., 1975: Coastal trapped waves in a stratified ocean. J. Phys.

Oceanogr., 5, 300–325.

Amante, C. and B.W. Eakins, 2009: Etopo1 1 Arc-Minute Global Relief

Model: Procedures, data sources and analysis. NOAA Tech. Mem. NESDIS

NGDC-24, 19 pp.

Arfken, G., 1970: Mathematical methods for physicists.

Academic Press, 815 pp.

Brink, K. H., 1982: A comparison of long coastal trapped wave theory with

observations off Peru. J. Phys. Oceanogr., 12, 897–913.

Brink, K. H., 1991: Coastal-trapped waves and wind-driven currents over

the continental shelf. Ann. Rev. Fluid Mech., 23, 389–412.

Cutchin, D. L. and R. L. Smith, 1973: Continental shelf waves: Low-

frequency variations in sea level and currents over the Oregon continental

shelf. J. Phys. Oceanogr., 3, 73–82.

Enfield, D. B. and J. S. Allen, 1983: The generation and propagation of sea

level variability along the Pacific coast of Mexico. J. Phys. Oceanogr., 13,

1012–1033.

Gill, A. E., 1982: Atmosphere-Ocean Dynamics. Academic Press, 662 pp.

Gill, A. E. and E. H. Schumann, 1974: The generation of long shelf waves

by the wind. J. Phys. Oceanogr., 4, 83–90.

Huthnance, J. M., 1975: On trapped waves over a continental shelf. J. Fluid

Mech., 69, 689–704.

Huthnance, J. M., 1978: On coastal trapped waves: Analysis and numerical

calculation by inverse iteration. J. Phys. Oceanogr., 8, 74–92.

130

Martinez, J. A. and J. S. Allen, 2004: A modeling study of coastal-trapped

wave propagation in the Gulf of California. Part I: Response to remote

forcing. J. Phys. Oceanogr., 34, 1313–1331.

Mysak, L. A., 1968: Edgewaves on a gently sloping contiental shelf of finite

width. J. Mar. Res., 26, 24–33.

Mysak, L. A., 1980: Topographically trapped waves. Ann. Rev. Fluid Mech.,

12, 45–76.

Pizarro, O., S. Hormazabal, A. Gonzalez, and E. Yanez, 1994: Variabilidad

del viento, nivel del mar y temperatura en la costa norte de Chile. Invest.

Mar., 22, 85–101.

Reid, R. O., 1958: Effects of Coriolis force on edge waves (i) investigation

of the normal modes. J. Mar. Res., 16, 109–144.

Rhines, P. B., 1969: Slow oscillations in an ocean of varying depth. Part 2

Islands and seamounts. J. Fluid Mech., 37, 191–205.

Romea, R. D. and R. L. Smith, 1983: Further evidence for coastal trapped

waves along the Peru coast. J. Phys. Oceanogr., 13, 1341–1356.

Zavala Sansón , L., 2010: Solutions of barotropic trapped waves around

seamounts. J. Fluid Mech. doi:10.1017/S0022112010003034.

131

The high Amplitude Internal Waves Generated at

the San Esteban Sill of the Gulf of California*

Anatoliy Filonov ♠, U de G Iryna Tereshchenko, U de G César Monzón, U de G

Abstract

Results are discussed from a short survey in the Gulf of California (GC) to

evaluate the parameters of internal tidal waves generated at the San Esteban

sill. Aboard a small oceanographic research vessel BIP XII, belonging to

the CIBNOR (Centro de Investigaciones Biológicas del Noroeste) at

Guaymas, different measurements of internal waves were performed using:

a) a CTD tow-yo SBE-19plus; b) arrays of HOBO thermographs (distributed

from the surface down to 150 m depths) on three moorings, deployed at the

north of a sill; and c) two temperature and pressure sensor arrays with an

ADCP, towed one behind the other at a distance of 475 m. Data showed that

the semidiurnal barotropic tide at San Esteban sill causes nonlinear internal

tides which propagate to the north during ebb, and to the south during flow.

These waves are dispersed at the surface layer forming short strongly

nonlinear waves of high amplitude and long internal waves; this generates

significant variations of temperature and salinity deeper than 150 m. The

waves at the leading edge of these groups have maximum amplitudes of 50-

* The Ocean, The Wine, and The Valley: The Lives of Antoine Badan, 131-171.

© 2010 CICESE. ♠ Contact: afilonov@prodigy.net.mx

132

80 m, wavelengths of 1200 m, and phase speed of about 1.2 m/s. Long

internal waves with semidiurnal periodicity and short nonlinear waves

disperse their energy generating turbulence, vertical and horizontal mixing.

Such processes improve the movement of nutrients to the surface and

contribute to maintain high biological productivity of the waters in this area

of the GC.

1. Introduction

In many regions of the world’s oceans, including the Mexican continental

shelf, researchers have determined the presence of internal waves (IW).

Alternating bands of sleeks and rips are clearly visible on satellite images of

the scarcely studied regions of the Mexican west coast (Apel and Gonzales,

1983; Fu and Holt, 1984; Howell and Brown, 1985; Filonov and Trasviña,

2000; Filonov, 2000), which demonstrate the existence of large amplitude

IW on the Mexican continental shelf.

As shown in a series of theoretical studies (Baines, 1982; Craig, 1987;

Holloway, 1987, 1996; Ostrovsky and Stepanyants, 1989; Miropol'sky,

2001), the generation of internal tides occurs mainly at the adjacent slope of

the continental shelf. Mechanisms of energy transformation from the

barotropic to the baroclinic tide also occur in the slopes of major ridges and

submarine elevations. This is because, in both cases, the slope of these

physiographic features does not differ much from that corresponding to the

continental slope. Such barriers, which rise near the ocean surface, strongly

intensify barotropic tidal currents. When they overflow this causes a major

collapse of the pycnocline. As it weakens the flow, the sinking thermocline

starts moving like an IW and, over time, it causes a nonlinear disintegration

133

and often the formation, in the front of the wave, of a group of internal

solitary waves of shorter period. In the presence of weak wind and wind

waves, these IW can manifest themselves at the sea surface in the form of

alternating flats of strips (divergence) and roughs (convergence), visible

both from aircraft and from space (Haury et al., 1979; Konyaev and Sabinin,

1992; Filonov and Sabinin, 1995; Filonov and Konyaev, 2003; 2006).

Large amplitude internal tides have been observed near the submarine

sills in: the Straits of Florida (Niller, 1968) and Gibraltar (Ziegenbein, 1969,

1970), the Gulf of Massachusetts (Chereskin, 1983; Hibiya, 1988) the Gulf

of California (Badan-Dangon et al., 1991; Fu and Holt, 1984), the Аndaman

(Osborn and Burch, 1980) and the Sulu Sea (Apel et al., 1985), the Indian

Ocean submarine ridges near the Macarena islands (Sabinin et al., 1992;

Filonov and Sabinin, 1995) and in many other parts of the world’s oceans.

Among the most interesting areas of the world’s oceans where internal

tidal waves of large amplitude are commonly generated are the sills, located

between the large islands that separate the northern and central parts of the

GC, Mexico (Filonov and Lavin, 2003). Tides in the GC are forced at the

mouth by the Pacific Ocean tides, and the length of the Gulf makes them

almost resonant to the semidiurnal tidal harmonics (Hendershott and

Speranza, 1971; Filloux, 1973). This causes large tidal ranges and strong

tidal currents in the shallow area at the northern end of the Gulf. The

presence of sills among the large islands in the mid-gulf (Figure 1) causes

even stronger currents there (up to 1.5 m/s). The strong tidal currents release

large amounts of turbulent kinetic energy, which has a tremendous impact

on the physics and biology of the GC (Filonov and Lavin, 2003).

The presence of sills and of high tidal energy, together with the year-

round strong stratification that characterize the GC, are appropriate

134

conditions for the generation of internal tides. Paden et al. (1991) found

evidence of the initial stages of the generation of internal motions in the GC

(distortion of the thermocline by tidal flow over the San Esteban sill). For

the first time in the case of the GC Fu and Holt (1984) analyzed SAR images

that show that groups of short-period IW are generated with semidiurnal

periodicity at the San Esteban sill during spring tides, and spread north-west

along the axis of the Tiburon basin. In that paper, the authors applied the

Korteweg-de Vries equation for analysis and determined the parameters of

soliton packets and individual waves within a packet. The first soliton of the

group had a length of about 1 km, a width of 50 m, and moved with a speed

of 1.2 m/s. Badan-Dangon et al. (1991) and Gaxiola-Castro (1994) reported

the passage of these wave packets from CTD casts.

Filonov and Lavin (2003) analyzed data from two moorings equipped

with current and temperature sensors (one for summer and the other for

winter) located between Angel de la Guarda Island and the mainland. They

applied the theory of IW and analyzed the modal structure of tidal waves in

detail, they calculated the energy balance between internal and barotropic

tide and presented an outline of the spread and transmission of energy from

these waves from the San Esteban sill to the position of the moorings. From

data recorded by the mooring deployed in summer they determined that: (a)

fluctuations in the currents were dominated by the semidiurnal frequency

band; (b) fluctuations of the horizontal baroclinic semidiurnal current were

aligned with the axis of the GC and had amplitudes of 10-15 cm/s; the

vertical displacement reached 4 m in this frequency band; (c) the

semidiurnal internal tide contained 45% of the barotropic tidal energy; (d)

the spectra of vertical wavenumber showed asymmetric peaks at high wave

numbers, indicating that the internal tidal energy flowing down from the

135

surface was larger than the tidal energy that flows up from the ocean floor.

This analysis also indicated that the energy of the barotropic tidal component

)0( =zk was approximately equal to that of the barotropic oscillations.

Though scientists from CICESE and SCRIPPS continuously obtain

numerous measurements from this area, measurements with tools specially

designed to study the processes of generation, distribution and dissipation of

internal tidal waves have not been carried out. In the present work we

discuss some results of a short experiment on measurement of propagation

and disintegration of internal tidal waves generated at the San Esteban sill in

the northern GC.

2. Measurements and data

For the investigation of IW in the area of the San Esteban sill we selected

only a narrow strip of the Gulf oriented perpendicularly to the sill and which

extended to the south for about 45 km and to the north for about 75 km

(Figure 1). SAR satellite images show the presence of groups of solitons in

this area (Fu and Holt, 1984). We obtained a detailed bathymetry of this

region based on the historical archive of measurements stored at CICESE

(thanks to Dr. Manuel Lopez). Maximum depth at the sill is less than 400 m

(Badan-Dangon et al., 1991; Lavin et al., 1997).

In this experiment we used three types of measurements for IW: with

moorings, using a towed CTD and antenna from two vertical chains of

thermographs and ADCP. All sea works were carried out during May 10-16,

2007, with well-formatted thermocline, and weather conditions favorable for

work from a small research vessel such as the CIBNOR’s BIP XII at

Guaymas. It was mostly calm and sea surface waves were absent.

136

To record time variability of the IW parameters beginning at the north of

a sill we dropped three moorings (Figure 1) with thermographs HOBO

(accuracy of 0.2 ºC, time rate of 3 min). The thermographs were set at 10 m

distance beginning from the surface to 150 m of depth. To measure the

inclinations in the vertical lines of the gauges every buoy had a depth of 45-

50 m (HOBO-LEVEL accuracy of 3 cm). On mooring N2, the thermographs

were set at different intervals from the top to the bottom of the ocean. Also,

this buoy was placed at a depth of 60 m where the ADCP RDI 600 kHz had

been suspended. Because of strong tidal currents in the study area we used a

bearing cable on the moorings which was a nylon cord with a break-up

resistance limit of 220 kg. On mooring N2 we used a metal corrosion-proof

cable with diameter of about 5 mm. Iron disks with weight of 100 kg were

used as anchors.

During the experiment we made five transects in which we measured

temperature and salinity vertical profiles to a depth of 220-300 m. All

transects were to the south and to the north of the San Esteban sill. A CTD

tow-yo SBE-19plus profiler was towed on the surface and behind the vessel

at full speed (about 10 knots) (Filonov et al., 1996). At every point of

measurement the vessel stopped and the profiler went freely down and then

up, measuring the temperature, conductivity and pressure with a 0.25-sec

sampling interval. The navigation data were logged by GPS.

For the recording of spatial parameters of IW at various distances from a

sill we used a towed antenna from two vertical chains. Each contained 10

thermographs (HOBO and SBE-39), which were evenly distributed from 5

m below the surface to a depth of 76 m. One of the chains was towed from

the vessel, and the second was moved 475 m behind it and fixed on a

specially equipped kayak on which we had also placed the ADCP RDI 300

137

kHz (sampling interval of 30 s, with 2 m vertical bins). For deep penetration

of the chains, special lattices were used. The depth of the lower end of each

chain was controlled by a pressure sensor, and the average was equal to

76±1 m. The speed of towage (about 4.5 knots) allowed the recording of

temperature every 150 m and of current structure every 75 m on a horizontal

direction.

In this cruise we had some trouble. At 3 am on May 14 on watch of the

captain the vessel towing the antenna of vertical chains was displaced at

great speed, and had passed at 5 m distance from the buoy N2, although the

original plan was to pass at a distance of 300 m. All thermographs had been

cut off from both chains and lost except for the uppermost and the bottom

devices. The buoy had also been torn off from its bearing cable

approximately at 40 m of depth, and all devices (ADCP also) located in the

deeper horizons were lost. As a result of the accident only two thermographs

HOBO (time rate of 1 min) survived on the horizons of 5 and 76 meters on

each chain.

The prognostic information of tidal level fluctuations for the Tiburon

Island provided by CICESE

(http://oceanografia.cicese.mx/predmar/predlin.php) was used for the

analysis. Despite the short time of the experiment, we obtained a large

amount of data. Therefore in this paper we only discuss data from one of the

CTD tow-yo SBE-19plus profiler transects across the San Esteban sill, and

data from three moorings, the chain thermographs and towed ADCP.

138

3. Analysis of the tow-yo data

The first transect was made on May 12, 2007. That day 25 vertical casts

up to 225 m every 5 km were obtained (Figure 1a). Its length was 130 km

and its duration was 10 h. On this transect the vessel crossed a sill precisely

when a hydraulic jump settled down from the southern part of a sill. Its

width was recorded on three consecutive vertical soundings near the San

Esteban sill and was equal to 10 km. Though the spatial step-type behavior

of measurements was large, despite the distance we saw in the centre of the

hydraulic jump the downward course of isotherms were at some tens of

meters. On the surface layers where the vertical temperature gradients were

relatively strong (equal to 0.09-0.10 ºC/m) this penetration was not so great.

For example, the isotherm of 17 ºC drops in the hydraulic jump only at 30

m. In contrast, at depths below 50 m vertical temperature gradients did not

exceed 0.015 ºC/m and the buoyancy force did not withstand the vertical

movements in the hydraulic jump. So, the isotherm of 15ºC dropped almost

at 100 m (from 90 to 190 m). The isotherms of 13 and 14 ºC went deeper to

about 300 m or more, maybe down to the bottom.

The vertical salinity distribution showed a similar behavior (Figure 1c,

d). It was evident that, at the moment of vertical casts near to a sill, salinity

coincided with low tidal barotropic level and absence of currents. However,

in the lapse of one hour, when the vessel moved 15 km to the north, strong

currents (Figure 1b) began to cause upwelling of warmer and saltier waters

to the north of a sill. These effects were significant in more than 60 km of

distance from a sill.

As shown previously (Filonov and Lavin, 2003) on each side of the San

Esteban sill two inclined internal tidal waves are generated: the tidal energy

139

propagates from the generation zone both downward (toward the bottom)

and upward (toward the ocean surface); when reaching the respective

boundaries, the rays reflect, interchanging direction of propagation (see

Figure 3 in Filonov and Lavin, 2003). The characteristic upper and lower

rays define the radial tube as they travel away from the site of generation. A

radial structure is observed for at least one or two cycles of ray reflections

from the bottom and the sea surface (Pingree and New, 1991; Holloway and

Merrifield, 1999). After several bottom-to-surface reflections of the rays, at

a distance of 40-60 km from the sill, a stationary oscillation in vertical

direction is established; that is, modes are formed.

However, not all the energy of the internal tides is directed from the sill

to the deep layers. Much of it is due to nonlinear dispersive transformation,

which forms a group of nonlinear waves with high amplitude that move in

the near-surface waveguide at the Brünt-Väisäla frequency rather than in the

deeper layers (internal waves can exist only if ( )inf N zω≤ ≤ (Miropol’sky,

2001). These groups of nonlinear waves (like solitons) propagating from the

sill on the north-east of the Tiburon basin have been described by Fu and

Holt (1984). Groups of short-period waves of large amplitude are always

located at the forefront of the internal tide and they cause strong vertical

mixing.

Figure 1c shows three regions of strong vertical mixing. The first was

located near the sill, and the second had its center at a distance of about 58

km. Its width was about 15 km and represents an area of the previous group

of short IW, which were formed earlier on a semi-diurnal period. However,

stratification in the upper layers after the passage of a short-wave was

quickly restored. This is well illustrated by the position of the isotherms at

20-50 km from the sill. We can also see traces of the weaker vertical mixing

140

caused by the package of short IW in the 30 km to the south of the sill,

which arose as a result of hydraulic jump on the north side of the sill.

Figures 2 a,c show all the profiles of temperature and salinity made at

transect. They also show average profiles (Figures 2 b,d) at the site of the

southern section (1) of the sill, north (13 casts) of it (2), and also profiles

registered at the center of the hydraulic jump. Mean temperature profiles on

both sides of the threshold similar to each other and the water layer from

surface to 200 m differed by no more than 0.5-0.6 ºC. Regarding salinity

these differences did not exceed 0.1-0.15 psu. Sounding in the centre of the

hydraulic jump showed the presence of a homogeneous subsurface layer (to

a depth of 26 m) of temperature and salinity. In it the temperature was 0.5-

1.0 ºC lower than its average profiles (1, 2), but it increased with depth and,

in the layer from 25 to 200 m, it was higher by about 1.0-1.5 degrees than

the “warmer” average profile to the north of the sill. The salinity profile at

the center of the hydraulic jump also had a homogeneous subsurface layer

and another uniform layer at depths from 50 to 200 m. Strong disturbances

in the vertical stratification caused by the hydraulic jump are apparently

always present near the San Esteban sill. However, stratification broken by

moving internal tidal waves to the north and south of the sill quickly were

restored only after 2-3 hours.

So, appearing with tidal periodicity, the hydraulic jumps caused at both

sides of a sill lowers local water layers some tens of meters, creating short-

term horizontal gradients of temperature of up to 0.5ºC/km and salinity of up

to 0.1 psu/km. Thus, indubitably, a strong vertical and horizontal turbulence

and mixing was formed.

141

4. Analysis of the moorings data

4.a The mooring N3

This mooring was the closest to the San Esteban sill, approximately 19

km, standing at this point almost 3.5 days (six semidiurnal tidal cycles) and

measurements showed what occurred at the 150 m layer of surface water

(Figure 1a). As the buoy was located fairly close to a sill, we can assume

that its devices show fixed processes of internal tide disintegration at their

initial stage, and formation of the groups of intensive nonlinear short waves.

As to temperature, we observed very strong vertical oscillations

sporadically accompanied by mixing, the formation of a layer of 40-60 m

thickness and a weak vertical temperature gradient. This layer is shown

between isotherms 16 and 17 ºC in Figure 3a.

Four (out of six) identified leading edges of internal tides are marked

very clearly in Figure 3a with Roman numerals I-IV. On these fronts the

aligned groups of 3-4 very short IW of large amplitude are visible. Before

each front, the isotherms were slightly pressed to the surface and increased

the stratification near the surface, but after the passage of short waves,

stratification was eroded. Although the barotropic tide during our

measurements displayed constant amplitude (Figure 3c), the intensity of the

internal tide varied in time, approximately with daily periodicity.

Apparently, short-period waves of large amplitude in one semidiurnal period

were intense and were then followed by a second period that was weak. It

resembled a "compression" and "stretching" of the isotherms in time. Figure

3b shows the temporal gradients of growth of the average temperature in the

layer 0-150 m at the forefront of semidiurnal tides. At the moments of

142

"compression", gradients reached 2.1-.3.1 ºC/hour, and during "stretching"

gradients declined to 0.9-1.1 ºC/hour. At the studied area the inertial period

was close to one-day’s length. Apparently inertial currents caused

compression and stretching of internal semidiurnal waves on counter and

passing directions. On the images published by SAR they are not visible and

(to the best of our knowledge) this modulation has not been described in the

literature.

Divergence and convergence of orbital currents of the internal tidal

waves were manifested in the temperature field as in-phase periodic raising

and lowering of the isotherms. Due to this, the surface layer of water (0-50

m) showed strong stratification with very short waves, which identified areas

of rectangles in Figure 1a, periodically weakened and shaped "spots" of cold

water, which only existed for 3-4 hours. This process is most clearly

manifested at the end of the measurements at this mooring. From 3 to 7 am,

May 16, 2010, the temperature at the surface of the bay did not exceed 16.5

ºC (6-8 hours before it had reached 22 ºC), at a depth of 120 m it was only

16 ºC, that is the vertical gradient of this layer at a specified time was

negligible.

4.b Estimates of the heights of intense short-period waves

We now consider in more detail the structure of waves in groups at the

forefront of internal tides (Figure 4). The first group of waves recorded at

the end of the day on May 12, consisted of three high soliton-like waves

with a time shift of 22 and 13 min between them. Their height was estimated

using a method commonly applied in the study of IW for temperature

measurements, which converts the temperature time series into vertical

143

displacements by means of (Konyaev and Sabinin, 1992):

1 2( ) ( ) ( ) ( ) / ( / )z z zt t t T t dT dzξ ξ ξ∆ = − = ∆ , where )(tzξ are the vertical

displacements of the water layers at level z at time t. These are reflected in

the difference in temperature fluctuations at two moments in time

1 0( ) ( ) ( )T t T t T t∆ = − , and /dT dt is the average vertical temperature

gradient, which was estimated from CTD data collected near the mooring.

Note that the accuracy of estimating wave height with this method depends

on the accuracy of the vertical temperature gradient. For the 16 ºC isotherm

we obtained ( )T t∆ = 2.4 ºC. The mean vertical gradient near the buoy at the

20-150 m layer was 0.02 ºC/m (Figure 5), hence 1.4/0.02 = 70 m.

Estimates of the height of the second and third waves were 65 m and 47.5

m. Note that the bottom of the leading wave moving ahead of its crest by

nearly 6 minutes led to the slope of the vertical axis of the back, and

apparently influenced the barotropic flows directed at this point of the

northern region of the bay with respect to the sill.

The second four-wave group was recorded at night (May 13-14). It was a

compact package of nonlinear waves not yet dispersed in space. They

showed a time shift of 10-15 minutes and height of about 40-50 m. The third

four-wave group was measured at night on May 15. The first two waves

were not completely separate; the remaining two, had a temporal shift of

about 45 and 50 minutes, and their height was 60 and 75 m, respectively.

Finally, the fourth group consisted of three waves. The leading wave had the

shape of a KdV soliton with height of about 90 m. The second wave in the

group was like an IW of second mode with zero vertical oscillations on the

horizon of 85 m (direction of the isotherm inclinations shown by arrows).

The third wave had a height of about 55 m.

144

4.c Hydraulic jump and internal tide generation

In accord with existing ideas, internal tides were generated at the buoy

N3 heading towards the south of the sill at ebb current, which form a

hydraulic jump (phase A). The hydraulic jump is formed from the wave

traveling toward the stream but with opposite direction, so that the jump

remains stationary on the bottom (Konyaev and Sabinin, 1992). Our

measurements allowed estimating the propagation time of the internal tide

from the area of its generation (from the position of the hydraulic jump) at

the San Esteban sill to the mooring N3.

Consider the circumstances for May 12, 2007, using the graphs in Figure

6. As shown, the ebb of the northern Gulf ended at 16:20. From then on

began the formation of the hydraulic jump; its center (see Figure 1c) was

located about 5 km south of the sill. Once the tide started to rise, the vertical

jump began to move with it towards the north, undergoing a nonlinear

transformation on its way to the buoy position. In order to divide the speed

of horizontal propagation of barotropic and baroclinic tides, we first

estimated the horizontal component of the orbital speed of barotropic waves.

The maximum water level in the vicinity of the sill was fixed at 22:00 h, i.e.

5 h 40 min (340 min) after low water, and the leading edge of the internal

tide was observed at the buoy at 22:32 h.

Assuming that the maximum speed of a barotropic tide in the vicinity of

the sill was approximately to maxV = 1.5 m/s (Badan et al., 1991), then its

average speed during this period can be taken as max / 2V V= = 0.75 m/s. The

lapse between the moment of high water and the time of arrival of internal

waves to the buoy N3 was at the initial stage of ebb. Assuming that the flow

145

increased to a maximum value of 1.5 m/s for about 3 h (180 min), we can

estimate its value at time 22:23 h, i.e. through 32 min (16º from the period of

growth rate of the wave), using the relation max (16 )V V Sin= ⋅ ° = 1.5 x 0.274

= 0.41 m/s. The average speed at this point in time may be half the value

obtained, i.e. 0.20V ≈ − m/s. It is negative, directed against the direction of

propagation of internal waves.

Thus, the water particles in the barotropic tide moved during 340 min

from the area of the hydraulic jump to the north with an average speed of

0.75 m/sec (they passed at a distance of 15.3 km). Then began the phase of

the ebb tide and water particles returned at 0.4 km in the opposite direction

(we found ourselves at 14.9 km from the starting point). However, the

leading edge of the tidal IW at this moment was already registered at

mooring N3. Total distance from the point of the hydraulic jump to the

mooring was 24 (19+5) km. The difference in spatial position of water

particles in barotropic and baroclinic tide was 9.1 (24-14.9) km. That is, the

barotropic flow did not change the direction and overcame the greater

distance. Since the average rate of the total flow (baroclinic + barotropic)

from the point of the hydraulic jump to the buoy was equal to (24 km) /

(340+32 min) = 1.25 m/s and above the average speed of barotropic flow

equal to 0.5 m/s, thus the average speed of the internal tide was propagated

to the north from the sill with a speed of 1.25-0.5 = 0.75 m/s.

On the basis of this work it is possible to assume the mechanism of

internal tide generation at the San Esteban sill, its subsequent disintegration

during its propagation at the Tiburon basin and the formation of a package of

short IW. Three phases of this process may be visualized (see Figure 7): (a)

stationary downward movement behind a sill during ebb tide (hydraulic

jump); (b) downward movement across a sill with tide inflow; (c) occurrence

146

of a group of short waves. This scheme requires detailed examination of

instrumental measurements directly on both sides of the San Esteban sill,

which we plan to do in the near future.

4.d The mooring N2

This buoy was located approximately 50 km away from the sill. As a

result of the ship’s accident only three thermographs of the mooring

remained at depths 10, 20 and 30 m. They worked from 17:30h on May 12,

till 03:00 h on May 13, a total of about 1.5 days (Figure 8a). Instruments

recorded the passage of five tidal IW. Of these, only the first and last waves

had the form of decaying waves. Some groups of nonlinear waves were

observed on the front of these decaying waves.

The first group of waves (Figure 8b) contained four soliton-like waves

with a time shift between 21, 18 and 22 min and heights of 35.5, 33, 17 and

9 m respectively (vertical gradient of the temperature in the upper 30 m layer

near the buoy at the time of passage of the wave packet was 0.076 °C/m).

This group of waves may be manifested on the surface of the Gulf, like the

waves on the satellite images presented in the study by Fu and Holt, 1984.

The second group of short waves (Figure 8c) contained the fifth tidal

wave and was less distinct. The first two waves had heights of 14 and 16 m

and a shift in distance between their troughs of 7 min. Behind them

stretched a long train of shorter waves with heights of up to 10 m. The

second, third and fourth long tidal IW (numbers II-IV on Figure 8) were at a

very early stage of disintegration and did not contain short waves of high

amplitude at their leading front.

147

It may be roughly assumed that the average horizontal velocity of the

tidal IW from the sill to mooring N2 is equal to 1.1 m/s (Fu and Holt, 1984).

Consequently, to arrive at the buoy, the waves took about 12.6 h; however,

some of them arrived quite disintegrated. In other words, some of the tidal

IW did not contain groups of nonlinear waves of great amplitude and,

therefore, should not manifest on the surface of the Gulf. Thus, processes of

internal tide disintegration at the Tiburon Basin are more complicated than

might be expected based on the analysis of satellite images, contained in

previously published works (Fu and Holt, 1984; Jackson and Apel, 2004).

Why this is happening remains to be studied.

4.e The mooring N1

This buoy was located 75 km away from the sill. Data received showed

(Figure 9a) that, as a result of strong disintegration, the tide IW were

"smeared" into space and lost their rhythm. Long IW appeared with irregular

fluctuations in temperature. They did not always cause large vertical

temperature gradients. For example on May 12 in the layer of 60-150 m, the

temperature varied by only 1°C (vertical gradient 0.01 °C/m). In contrast, in

the surface layers (0-60 m) the vertical temperature gradient reached 0.10

°C/m. Therefore, short IW were manifested in these water layers.

The water level fluctuations were measured by Hobo-Level guide; these

showed that at a depth of 40 m, depth variations did not exceed ±1.5 m

(Figure 9b). The total depth of moorings was above 400 m, so the inclination

of a mooring’s bearing cable during high tide and low tide (twice per tidal

cycle) did not exceed 2° from the vertical. This does not cause relevant

mistakes in the measured data, because when the gradient equals 0.1 °C/m

148

the error is not more than ±0.15 °C, i.e. within the accuracy of the

instrument.

Measurements were made at a time when the thermocline began to form

near the surface and internal tide disintegration activity occurred at the

surface layers of the Gulf, which resulted in a sporadic group of short

nonlinear waves. The most intense of them are marked in bold in Figure 9a

by rectangles with Roman numbers I and II. For the first rectangle, the

spectra of temperature fluctuations were calculated. The average spectrum

for horizons 10-60 m is shown in Figure 10. We can see variations from 5 to

38 min and heights from 10-15 m in the thermocline.

Figure 9 shows good conformity between the inclination of a spectral

exponent and frequency dependence ~ 1ω− in a range of 0.02 < ω < 0.2

cycle/min. This may be the result of strong nonlinear interaction of high-

frequency IW in a narrow near-surface waveguide in which they redistribute

their energy from larger waves to smaller waves, in an energy cascade

similar to that present in a vortex, which eventually leads to the universal

form of the slope of the spectrum. Filonov and Novotryasov (2005, 2007)

proposed an analytical model of such a spectrum on the continental shelf

with the ocean tides. Apparently the structure of groups of high-frequency

IW in a narrow waveguide is somewhat similar to the frequency structure of

the waves on the narrow shelf near the shore (at shallow depths), where the

destruction of the internal tide leads to the formation of a universal spectrum

with a slope ~ 1ω− .

149

5. Parameters of short nonlinear IW measured by towed sensor system

Four sections were made by towed thermographs and ADCP at the

studied area. We analyzed the most interesting one, for May 13, 2007

(Figure 11). It was carried from south to north at the time when the hydraulic

jump was located at the north of the sill. Despite the fact that the speed of

the vessel was constant throughout the towing (2.32 m/s), Figure 11b shows

that due to acceleration or deceleration of the ship by barotropic currents, its

movement in space relative to the Earth was uneven.

In a section, several groups of nonlinear IW were observed, but only two

of them were very distinct and had larger amplitudes. One of them moved to

the south and was located about 30 km from the sill; the other was 36 km

north of it. The vessel crossed the first group of waves in the opposite

direction, and the second group was crossed in the same direction. It is clear

that the first wave group was generated by half a tidal period earlier than the

second, which the ship “caught up” with after almost 7 h near buoy N2. We

did not have current meters on the moorings, so we could not take into

account the Doppler Effect in the measured flow field of IW in the section.

The presence of two towed vertical chains of thermographs in our

measurements allowed excluding the Doppler Effect in groups of short IW,

but only in the horizontal direction of the towed system (475 m). The second

group containing four waves propagated to the north of the sill was then

analyzed. For this we used the data of towed thermographs at a depth of

76(±1) m.

The speed of the vessel on the water during towing measurements was

4.5 knots, or 2.32 m/s; the coordinates of the vessel were monitored every

minute with the help of a GPS. Analysis of the ship’s location from 12:30 h

150

to 14:00 h on May 13, 2007 showed that it was moving at a speed of 1.44

m/s relative to the Earth. Figure 11b shows the maximum ebb against the

vessel and a group of propagating waves at the time of the intersection of the

IW’s towed sensor system. Barotropic flow rate can be calculated as the

difference between the speed of the vessel relative to the ground and its

velocity relative to the water: 1.44 – 2.32 = –0.88 m/s.

Figure 12 shows that the towed sensor system "catches up" at a speed of

2.23 m/s with a group of leading waves recorded at 13:31 h by the chain of

sensors attached to the vessel and, after 420 sec, by the second chain

hanging from the kayak. In fact, if the group of waves had not moved in the

same direction, the vessel would have overcome the distance between the

two sensors for the 475 m / 2.32 m/s = 205 s instead of 420 s. Consequently,

for 215 (402-205) sec a group of waves "moved" forward in the same

direction as the ship. Hence, the phase velocity of the leading wave in the

group was (2.23 x 215) / 420 = 1.19 m/s. Estimates of the velocity of the

remaining three waves in the group were identical with the leading wave,

that is, the whole group moved as a single formation.

Distances between the waves in the group were calculated with the

scheme described above. Their heights were found by equation (1), the

values of the vertical temperature oscillations were calculated, and the

average vertical temperature gradient within a layer of 76±5 m near the

wave packet was equal to 0.04 °C/m. Parameters of the waves in a group are

shown in Table 1.

The results of measurements of horizontal velocities in the groups of

short IW measured with the towed sensor system are beyond the scope of

this paper. To interpret these results (in addition to knowing of vessel

speed), the speed of the horizontal barotropic current along the tracks of the

151

towing devices must be at least roughly known. This could be done in the

future, using numerical simulation data of barotropic currents in the GC. For

this reason, we only analyzed the vertical velocity fluctuations measured by

the towed ADCP. This did not include the Doppler Effect caused by the

moving ship, and additions of the barotropic flow did not affect the nature of

the vertical IW velocity.

Temporal variability of the vertical component of the orbital velocity of

the internal wave ( )zw t at a fixed level is related to the vertical displacement

( )z tξ of a liquid particle by the equation: ( ) ( )z zt w t dtξ = ∫ , where z is the

depth at which fluctuations of the vertical component of currents were

measured (Konyaev and Sabinin, 1992; Plata and Filonov, 2007). This

equation allows for the recalculation of the data recorded by the ADCP as W

(z, t) in the shape of a matrix of vertical displacements ( , )z tξ in the total

layer of measurements.

By analogy, we calculated the vertical displacement of water particles in

the layer of 0-150 m for the time when the towed ADCP RDI 300 kHz

crossed a group of short IW, which was described above and shown in

Figure 12b. To reduce the high frequency noise which always accompanies

this type of measurements, the matrix was averaged over time for the three

adjacent values, as well as for the layers of 30-150 m. The result is shown in

Figure 13. As is evident, these results do not differ much from those shown

in Table 1. It seems that the height of waves is more accurate when obtained

from measurements of ADCP. Data of the towed thermograph gave slightly

overestimated wave heights. This was caused by the lack of precise

knowledge about the vertical temperature gradient.

152

Our results show that the waves in the group are somewhat different from

the wave parameters described in the article by Fu and Holt, (1984). This

difference could be due to the time of year at which our measurements were

made, the middle of May, when the near-surface thermocline is not

completely formed and short-period IW can propagate in a relatively narrow

waveguide located between the surface and the horizons of 40-60 m. As

shown above, this thermocline constantly eroded under the influence of long

tidal IW, which prevented the spread of short wave groups or led to their

collapse. Reproduced in the Fu and Holt, (1984) SAR image (Figure 2 in

their work), the analysis on which their bright work is based was made in

autumn when the thermocline has maximum gradients at a depth of 50 m.

This allowed short IW moving north from the San Esteban sill without

obstacles.

Wave number Wave height (m) ∆t, between troughs (numbers in parentheses)

Distance between troughs (m)

1 40 (1-2) 1260 1424

2 33 (2-3) 900 1017

3 28 (3-4) 390 441

4 9

Table 1. Parameters of nonlinear IW in the group shown in Figure 12.

6. Conclusions

The sill, which rises to the ocean surface, distorts the tidal currents.

Poured over the sill, the tidal flow creates a large-size penetration

153

pycnocline, like a hydraulic jump (Haury et al., 1979). When the tidal flow

weakens, this penetration starts to move as IW and begins its nonlinear,

dispersing transformation. Once released, these IW can go into the deep sea,

and move long distances, gradually transforming into a group of solitary

waves. These type of waves may observed in many parts of the ocean

(Ziegenbein, 1969, 1970; Osborn and Burch, 1980; Chereskin, 1983; Fu,

Holt, 1984; Liu et al., 1985; Apel et al., 1985).

In the GC, at a region of large islands and sills between them, we found a

similar process. Our measurements showed that the semidiurnal barotropic

tide at the San Esteban sill causes a hydraulic jump. Appearing with tidal

periodicity, hydraulic jumps cause on both sides of a sill downwards local

water layers at some tens of meters, creating short-term horizontal gradients

of temperature up to 0.5 ºC/km and salinity up to 0.1 psu/km. This

doubtlessly generates strong vertical and horizontal turbulence and mixing.

During the present study, we consistently recorded the hydraulic jump on

both sides of the sill in data received by a CTD tow-yo profiler and a towed

temperature and ADCP sensor system. On the basis of these measurements

we assume the mechanism of internal tide generation at the San Esteban sill

and its subsequent disintegration during its propagation in the Tiburon Basin

and formation of a package of short IW. It consists of three phases:

stationary downward movement behind a sill during ebb tide (hydraulic

jump), downward movement across a sill with tide inflow, and occurrence of

a group of short waves.

These waves propagate to the north, during ebb, and to the south, during

flow of the barotropic wave and generate significant variations of

temperature and salinity as far as 300 m of depth. Distinct groups of waves

are already formed at a distance of 20-25 km from the sill. At the leading

154

edge of the wave groups we regularly found the waves with maximum

amplitudes above 50-80 m. Wavelengths in one group of short waves at a

distance of about 36 km from the sill were found to be 400-1200 m. Their

height was 9-40 m and their phase velocity 1.2 m/s or less.

Measurements from the moorings showed that divergence and

convergence of orbital currents in the internal tidal waves were manifested

as periodic raising and lowering of the isotherms. Sometimes, as a result of

intensive mixing, on the surface layer of water (0-50 m) strong stratification

periodically weakened and near the surface appeared a "spot" of cold water,

which lasted for about 3-4 hours. Such processes usually improve the

movement of nutrients to the surface and contribute to maintain high

biological productivity of waters in this area of the GC.

It should also be added that the structure of the IW dynamics near the San

Esteban sill was very surprising. It appeared to be much more complicated

then expected, considering the results of the theoretical analysis of the SAR

images presented by Fu and Holt (1984). This stimulates us to continue with

similar investigations. More comprehensive and profound studies are

necessary of the dynamics of IW at different scales in the area of the sills of

the GC.

Acknowledgments

The work was supported by the Consejo Nacional de Ciencia y

Tecnologia (CONACYT, Mexico) and the University of Guadalajara

(project U 46674-F). The authors gratefully acknowledge the help and

technical assistance of M.Sc. students Diego Pantoja-Gonzalez, Edgar

Flores-Chavez, Hector Santiago-Hernandez and Dr. Luis Plata-Rosas from

155

the University of Guadalajara, as well as Ph.D. student Carlos Vargas-

Aguilera from CICESE. The authors also thank Elias and Michael Frenkel

for their help in translating and editing this manuscript.

156

Figure 1. a) Study area at the GC. Isobaths are drawn through the 100 m. The white dots mark the location of the CTD tow-yo transect profiler SBE-19plus. The asterisk shows the position near Tiburon Island for which CICESE forecasts tidal level variations. b) Corresponding sea level profile. c) Vertical section of temperature and salinity fields across the San Esteban Sill. d) Same as c) but for salinity.

157

Figure 2. a) Temperature and c) salinity profiles obtained from the section perpendicular to the San Esteban Sill on May 12, 2007 (points of vertical soundings are shown in Figure 1a). b) and d) The corresponding average profiles to the south of the sill (1), to the north of the sill (2) and just above the sill (3).

158

Figu

re 3

. a) V

ertic

al s

ectio

n of

tem

pera

ture

var

iatio

n fr

om m

oorin

g N

3. T

he re

ctan

gles

(with

Ara

bic

num

eral

s) a

re

mar

ked

by r

ecor

ding

with

hig

h-fr

eque

ncy

osci

llatio

ns i

n th

e up

per

40 m

lay

er o

f w

ater

. T

he l

eadi

ng e

dge

of

inte

rnal

tide

s w

ith g

roup

s of

sho

rt no

nlin

ear w

aves

of l

arge

am

plitu

de a

re m

arke

d w

ith R

oman

num

eral

s (th

ey a

re

show

n on

a l

arge

r sc

ale

in F

igur

e 4)

. Iso

late

d la

yer

with

tem

pera

ture

s be

twee

n 16

and

17

ºC (

shad

ed s

trip)

. b)

Ave

rage

tem

pera

ture

flu

ctua

tions

in

the

laye

r of

0-1

50 m

. A

vera

ge t

empo

ral

tem

pera

ture

gra

dien

ts (

abov

e).

c)

Leve

l flu

ctua

tions

nea

r the

isla

nd o

f Tib

uron

.

159

Figure 4. Vertical sections of temperature from mooring N3 at the moment of passage of groups of short IW of large amplitude, which are marked with Roman numerals on Figure 3a. Arabic numerals mark individual nonlinear waves in the group.

160

Figure 5. Averaged (4 casts) vertical profiles of temperature, salinity and the Brünt-Väisäla frequency near mooring N3. Dotted lines show the linear approximation of temperature and buoyancy frequency profiles.

161

Fi

gure

6.

Moo

ring

N3

tem

pera

ture

fluc

tuat

ions

on

the

horiz

on 5

0 m

(1).

The

dep

th o

f the

Hob

o-Le

vel g

auge

(2),

and

leve

l flu

ctua

tions

at

Tibu

ron

Isla

nd (

3). T

he d

otte

d lin

e (4

) sh

ows

a qu

alita

tive

asse

ssm

ent

of t

he p

hase

of

baro

tropi

c ve

loci

ty o

scill

atio

ns. T

he u

pper

left

corn

er sh

ows

a di

agra

m o

f the

inte

rnal

tide

mov

emen

t fro

m th

e ar

ea

of th

e hy

drau

lic ju

mp

to b

uoy

N3

(see

als

o ca

ptio

n to

Fig

ure

7).

162

Figure 7. Phases of the generation of internal tides on the San Esteban Sill. a) Stationary downward movement behind the sill during outflow (hydraulic jump). b) Downward movement across the sill with inflow. c) Occurrence of a group of short waves.

163

Figure 8. a) Vertical section of temperature fluctuations from mooring N2. Roman numerals mark the number of tidal IW. b) and c) The rectangles marked a group of nonlinear waves of large amplitude. Arabic numerals mark the number of waves in the groups.

164

Figu

re 9

. a) V

ertic

al se

ctio

n of

tem

pera

ture

fluc

tuat

ions

from

moo

ring

N1.

The

rect

angl

es m

ark

a gr

oup

of

nonl

inea

r wav

es o

f lar

ge a

mpl

itude

. b) T

he d

epth

of t

he H

obo-

Leve

l gau

ge

165

Figure 10. Average spectral density of temperature fluctuations in the layer of 10-60 m for the interval measurements shown in Figure (9) bold rectangle with Roman numeral I. The vertical line shows the 95% confidence interval. The spectrum was obtained by averaging 6 periodograms, smoothed to 5 frequencies.

166

Figure 11. a) The dashed line indicates the direction of tow chain temperature sensors and ADCP in the area of the San Esteban sill. The figure also shows the position of the short IW groups propagating to the south and north of the San Esteban sill. b). Sea level profile at the moment of towing. c). Temperature fluctuation on transect measured by chain of thermograph from the ship and (d) from the kayak. e) The bottom profile.

167

Figure 12 a) and b) The scheme for the measurement of nonlinear IW using a towed sensor system. Measurements made on May 13, 2007; 36 km north of the sill. c) Wave height in wave the group, measured towed Hobo thermographs at the 76 m horizon from the ship and kayak.

Figure 13. Average in the layer of 4-150 m vertical displacement ( )tξ in a group of short IW; measurements were made by the towed ADCP RDI 300 kHz (also refer to Figure 12). Sampling interval was 30 s, with 2 m vertical bins; towage speed was 2.32 m/s.

168

References

Apel, J. R., and F. I. González, 1983: Nonlinear Features of Internal

Waves off Baja California as Observed from the SEASAT Imaging

Radar. Journal of Geophysical Research, 88, 4459-4466.

Apel, J. R., Holbrook, J. R., Liu, A. K., and J. Tsai, 1985: The Sulu Sea

internal soliton experiment. Journal of Physical Oceanography, 15, 1625-

1651.

Badan-Dangon, A., M. Hendershott, M.F. Lavín, 1991: Underway

Doppler Current Profiles in the Gulf of California. EOS, 72 (19), 209, 217

and 218.

Bains, P.G., 1982: On internal tide generation models. Deep Sea Res., 29,

307-338.

Chereskin, T.K., 1983: Generation of internal waves in Massachusetts

Bay. Journal of Geophysical Research, 88, 2649-2661.

Craig, P.D., 1987: Solution for internal tide generation over coastal

topography. J. Mar. Res., 45, 83-10.

Filonov A.E. and K.D. Sabinin, 1995: Estimation of orbital velocities of

internal waves of short period and large amplitudes. Geofísica

Internacional, 34 (4), 395-398.

Filonov, A.E., C.O. Monzón and I.E. Tereshchenko, 1996: A technique

for fast conductivity-temperature-depth oceanographic surveys. Geofísica

Internacional, 35 (4), 415-420.

Filonov, A. E. and A.C. Trasviña, 2000: Internal waves on the continental

shelf of the Gulf of Tehuantepec, Mexico. Estuarine, Coastal and Shelf

Science, 50 (4), 531-548.

Filonov A.E. and K.V. Konyaev, 2003: Nonlinear Internal Waves near

the Mexican Pacific Coast.: En O. U. Velasco Fuentes, et al. Editors,

Nonlinear Processes in Geophysical Fluid Dynamics, Kluwer Academic

Publishers, Amsterdam, 257-274.

169

Filonov A.E and M.P. Lavín, 2003: Internal tides in the Northern Gulf of

California.. Journal of Geophysical Research, 108 (C5), 3151.

Filonov A. and K. Konyaev, 2006: Orbital motions and coherent

structures in the internal tide on a steep shelf off the Pacific coast of

Mexico. Geophys. Res. Lett., 33, l14612, doi:10.1029/2006gl026712,.

Filonov A. and V. Novotryasov, 2005: Features of the nonlinear internal

wave spectrum in the coastal zone, Geophys. Res. Letters, L15602,

doi:10.1029/2005GL023046,

Filonov A. and V. Novotryasov, 2007: On a spectrum of nonlinear

internal waves in the oceanic coastal zone, Nonlinear Processes

Geophysics, 14, 1-6.

Fu L.L., and B. Holt, 1984: Internal waves in the Gulf of California:

Observations from a spaceborn radar. Journal of Geophysical Research.,

89 (C2), 2053-2060,

Gaxiola-Castro, G., J. García Córdova, J.E. Valdéz Holguín and M.

Botello Ruvalcaba, 1995: Spatial distribution of chlorophyl and primary

productivity in relation with the winter physical structure of the Gulf of

California. Continental Shelf Research, 15 (9), 1043-1059.

Jackson C. R., 2004: An Atlas of Internal Solitary-Like Waves and their

Properties. Global Ocean Asociation.

Jackson, C. and Apel, J., 2004: An atlas of internal solitary-like waves

and their properties:

http://www.internalwaveatlas.com/Atlas2_index.html.

Konyaev, K.V. and K.D. Sabinin, 1992: Waves Inside the Ocean.

Hydrometeoizdat, Saint-Petersburg, 272 pp.

Liu, A. K., Holbrook, J. R., and J. R. Apel, 1985: Nonlinear internal wave

evolution in the Sulu Sea. Journal of Physical Oceanography, 15, 1613-

1624.

170

Miropol’sky, Y. Z., 2001: Dynamics of Internal Gravity Waves in the

Ocean. Kluwer Academic Publishers, Amsterdam, 406 pp.

Niller, P., 1968: On the internal tide motions in the Florida Straits. Deep

Sea Res., 15, 113-123.

Haury L.R., Briscoe M.G., and Orr M.H., 1979: Tidally generated internal

wave’s packets in Massachusetts Bay. Nature, 278, 312-317.

Hendershott, M.C., and A. Speranza, 1971: Co-oscillating tides in long

narrow bays; the Taylor problem revisited. Deep Sea Res., 18, 959-980.

Hibiya, T., 1988: The generation of internal waves by tidal flow over

Stellwagen Bank. Journal of Geophysical Research, 93, 533-542.

Holloway, P.E., 1987: Internal Hydraulic Jumps and Solitons at a Shelf

Break Region on the Australian North West Shelf. J. Geophysics. Res., 92

(C5), 5405-5416.

Holloway, P.E., 1996: A Numerical Model of Internal Tides with

Application to the Australian North West Shelf. Journal of Physical

Oceanography, 26, 21-37.

Howell, T.L. and W.S. Brown. 1985: Nonlinear internal waves on

California continental shelf. J. Geophys. Res., 90 (C4): 7256-7264.

Osborn, A.R. and T.L. Burch, 1980: Internal solitons in the Andaman

Sea. Science, 208 (4443), 451-460.

Ostrovsky, L.A. and Yu.A. Stepanyants, 1989: Do Internal Solitons exist

in the Ocean? Rev. of Geophysics, 27 (3), 293-310.

Paden, C.A., M.R. Abbot, and C.D. Winant, Tidal and atmospheric

forcing of the upper ocean in the Gulf of California. 1. Sea surface

temperature variability. J. Geophys. Res., 96, 18337-18359, 1991.

Plata L. and A. Filonov, 2007: Internal tide at the northwestern part of

Bahía de Banderas, Mexico. Ciencias Marinas, 32 (2), 197-215.

171

Sabinin K.D., Nazarov A.A. and Filonov A.E. 1992: An internal wave

train over the Mascaren Ridge. Izvestia, Atmosphere and Ocean Physics,

28 6, 625-633.

Ziegenbein J. 1969: Short internal waves in the Strait of Gibraltar. Deep

Sea Res., 16, 479–87.

Ziegenbein J. 1970: Spatial observations of short internal waves in the

Strait of Gibraltar. Deep Sea Res., 17, 867–75.

172

173

Lagrangian Circulation in Todos Santos Bay

and off Baja California During Spring 2007:

Exploratory Experiments*♦ David Rivas ♠, CICATA-IPN Rocío Mancilla-Rojas, CICATA-IPN Ernesto García-Mendoza, CICESE Antonio Almazán-Becerril, CICY

Abstract

Here we study numerically the circulation in Todos Santos Bay area

(31.88°N) and off Baja California during 2006 and 2007. This period was

selected after an intense toxic algal bloom occurred on April 2007 in this

area, which was most probably caused by environmental conditions

associated with the wind-driven upwelling in the region. We carried out

high-resolution, numerical model simulations to be used in a three-

dimensional Lagrangian analysis, which provides information about the

origin and distribution of the waters present in Todos Santos Bay by the

end of April 2007. The results show that such waters come mainly from

locations west of the Bay (even beyond the model's domain), approaching

to the continental shelf, embedded in a flow that bifurcates in a

equatorward current and a poleward current at 32-33°N, the so-called

* The Ocean, The Wine, and The Valley: The Lives of Antoine Badan, 173-201.

© 2010 CICESE. ♦ This article has not been peer-reviewed. ♠ Contact: drivasc@ipn.mx

174

Ensenada Front. Once the water parcels enter the Bay, mostly from the

west-northwest, two different regimes are identified. In the first one the

waters upwell and reach the surface (2-m depth), then leave the Bay

flowing to the south and affect only the ~200 km-long coastal portion

right south of TSB; they scatter south-southwestward, remaining at the

surface levels. The second regime consists of waters that do not reach the

surface within the Bay, but they remain at subsurface levels (38- and 75-

m depths) and eventually leave the Bay mostly to the northwest, reaching

deeper levels (below 200 m) and flowing poleward crossing the Southern

California Bight, most probably advected by the poleward undercurrent

over the slope. Estimations of residence times of the waters within the

Bay result in about one and a half months for the surface waters and one

month for the subsurface waters.

1. Introduction

The purpose of this paper is to study numerically the distribution of

the waters present in Todos Santos Bay during spring 2007. This period is

of particular interest because of the occurrence of an important toxic algal

bloom in this area (García-Mendoza et al., 2009). This bloom was the

southernmost report of the presence of domoic acid (DA) in the

California Current System and it is also the first report of the distribution

of toxic Pseudo-nitzschia species and DA on the Baja California west

coast. The accumulation of toxic cells during this event was most

probably caused by environmental conditions associated with the wind-

driven upwelling in the region (García-Mendoza et al., 2009). Therefore,

knowing the paths of the waters associated with this algal bloom is of

major importance for understanding this kind of phenomena.

175

Direct observations in Todos Santos Bay are scarce, hence numerical

models have been of paramount importance providing much of the

knowledge about its circulation (e.g., Mateos et al., 2009). In this paper

we present the results of some exploratory, process-oriented numerical

experiments, which provide useful information about the origin and

distribution of the water parcels that could be involved in the harmful

algal bloom mentioned above.

The rest of the paper is organized as follows: Section 2 describes the

model setups and their forcing and boundary data, as well as some

comparisons between the model outputs and observations and previous

simulations. In Section 3 the results from a Lagrangian analysis of the

upwelling waters are shown. In Section 4 the implications of this analysis

are discussed. Section 5 summarizes the main results.

2. Model setup and validation

2.a Setup

2.a.1 Regional model

The configuration used for the regional model is as in Rivas and

Samelson (2010), but adapted for the northern portion of the Baja

California Peninsula. The analysis is based on the Regional Ocean

Modeling System (ROMS) [e.g., Shchepetkin and McWilliams 2005]

version 3.0, configured on a spherical-coordinate domain extending from

26.8°N to 35.5°N and from 123.5°W to 113.7°W (see Figure 1), with a

horizontal resolution of ~1/30° (~3 km), resulting in a horizontal grid of

293x305 points. The vertical resolution consisted of 31 layers with

enhanced resolution near the lower and upper boundaries.

176

The model grid was prepared by using software described by Penven

et al. (2008), with topography from the ETOPO2 (Smith and Sandwell,

1997) gridded data set at 2’ resolution. No tides or riverine forcing were

used. The model was integrated for years 2006 and 2007 using wind

forcing and boundary data for this period. The daily varying wind-stress

forcing was obtained from a gridded QuikSCAT satellite scatterometer

product [Mean Wind Fields (MWF product) - User Manual - Volume 2:

QuikSCAT. C2-MUT-W-04-IF CERSAT - IFREMER]. These 1/2°-

resolution data were linearly interpolated to the model grid; for the model

grid points closest to the coastline, the values were extrapolated by

replacing them by the nearest interpolated values, a method that has

shown to work reasonably well (Rivas and Samelson, 2010). At the

model's boundaries, monthly mean fields of velocity, sea level,

temperature and salinity for the two-year period from the 1/8° version of

the Navy Coastal Ocean Model (NCOM; Barron et al., 2006) were used.

The model was initialized on 1 January 2006, using the corresponding

fields from the NCOM.

2.a.2 Todos Santos Bay model

A smaller-scale model for Todos Santos Bay (TSB), also based in the

ROMS, was implemented. Many of the details used in this model

configuration are the same as those described above, unless otherwise

noted. The model domain extends from 31.68°N to 31.99°N and from

116.86°W to 116.61°W (Figure 2). The horizontal resolution is ~300 m,

and the vertical resolution is given by 20 layers. A quadratic bottom

friction scheme was used. The model included tides, the first 10

components from Egbert and Erofeeva (2002).

177

As done for the regional model (Section 2.a.1), the TSB model was

integrated for the period 2006-2007. Indeed, the regional model provided

the initial condition and the boundary data for the TSB model, since this

latter model was initialized on 1 January 2006 by using the regional

initial fields, and daily averages of the regional outputs were used at the

boundaries of the TSB model. Nudging (relaxation) time scales are of 5

and 100 days were used for passive (outflow) and active (inflow) open

boundary conditions, respectively. In contrast with the regional model, in

the TSB model the wind stress and the heat fluxes at the surface were

calculated internally from the proper meteorological data (wind vector,

air temperature, atmospheric pressure, short-length radiation, etc.). Daily

averages of these atmospheric parameters were taken from the North

American Regional Reanalysis (NARR; see Mesinger et al., 2006), which

has a spatial resolution of 32 km. This resolution is not ideal, given the

size of the TSB model, but is apparently the best choice for such data.

2.b Validation

2.b.1 Observations

A limited number of observed data sets from period 2006-2007 are

used to assess the basic performance of the model, prior to the

Lagrangian analysis. Among the in-situ measurements available are the

sea surface temperature series from National Data Buoy Center (NDBC;

http://www.ndbc.noaa.gov/maps/Northwest.shtml) buoy 46086 (see

Figure 1). Additional in-situ observations during 2006-2007 include the

coastal sea level from two tide gauges at nearby Ensenada Port and El

Sauzal Port (see Figure 2), provided by CICESE data bank

(http://oceanografia.cicese.mx/). The most complete series was the one

from nearby Ensenada Port (about 96% of good data). Therefore, we took

178

this series and filled some short gaps by using data from nearby El Sauzal

Port and by harmonic fitting, including tidal frequencies. The resulting

sea level data were adjusted using local atmospheric sea-level pressure,

following Strub et al. (1987). The atmospheric pressure was taken from

the meteorological records at nearby El Sauzal Port.

The observed sea surface temperature (SST) at buoy 46086 shows

significant differences with respect to the model (Figure 3a). There is a

reasonable agreement between the model and the observations during

summers 2006 and 2007, but from fall 2006 to spring 2007 the model

tends to underestimate the temperatures. This is somewhat consistent with

the comparison between the spring-mean SST from the model and from

satellite imagery, which shows that, except for the Southern California

Bight where the model apparently overestimates the signal associated

with the wind-driven upwelling (colder coastal water), there is a

reasonably good agreement between the model and the observations (not

shown), but the model is generally colder and the horizontal gradients are

more intense. Note that point comparisons of modeled and observed SST

are complicated by the large spatial gradients, which can lead to large

SST differences from relatively small circulations errors; also, differences

in heat fluxes and mixed layer depth can contribute to such a mismatch.

The sea level diagnosed by the model shows a moderate agreement

with that observed in the tidal gauge (Figure 3b). The annual cycle is

reasonably well reproduced by the model, but the higher frequency shows

significant differences. The model generally underestimates the

magnitude of the peaks present in the observed series, and many of them

are not even reproduced. This problem can be associated with a poorly

resolved model-input wind close to the coastline, where orographic, air-

sea interaction, and other processes may modify the wind stress

substantially (e.g., Burk et al., 1999; Perlin et al., 2007). Another

179

possibility can be the effects of coastal trapped waves generated beyond

the model's southern boundary; hence their effects in our model's

dynamics are absent.

2.b.2. Previous simulations

We compare our TSB-model outputs with those obtained by Mateos et

al. (2009). The model used by these authors has a horizontal resolution of

half the one in our model and a vertical resolution similar to it; their

model domain extends well beyond the boundaries of our model, and

their runs include climatological monthly forcing. The near-surface July-

mean velocity shows several differences between the models (Figure 4a;

see also Figure 2a in Mateos et al., 2009). The most remarkable one is the

presence of a cyclonic circulation in Salsipuedes Bay that is not reported

in Mateos et al. (2009). This difference is probably associated with the

persistent upwelling in that area induced by the July-climatological wind

used by those authors (E. Mateos, pers. comm.). Another feature which is

not reproduced by our model is an anticyclonic circulation southwest of

El Sauzal. Nonetheless, as there is no data to compare the model outputs,

the veracity of these features remains to be clarified.

Vertical sections across the northern and western portions of TSB also

show significant differences between our model (Figures 4b-c) and that of

Mateos et al. (2009; their Figures 2b-c). The most remarkable feature in

the northern section in our model is an intense outflow in the narrow

coastal strip northwest of El Sauzal Port followed by a subsurface

maximum inflow at 5-15 m depth, in contrast with Mateos et al. model

which shows an outcoming flow attached to the bottom close to the 20-m

isobath, and a surface-intensified inflow at the outward side of the

section. The western section, on the other hand, shows some similarity

180

between both models in terms of outflow close to the coast and

subsurface inflow centered at 50-70 m at the outward side, but there is

also an important difference that is the intense inflow in the deepest

portion of the section, probably part of the poleward undercurrent of the

California Current System (e.g., Pierce et al., 2000; Gay and Chereskin,

2009), which is apparently absent within TSB in Mateos et al. model.

3. Lagrangian analysis

3.a Overview

The paths of the waters present in TSB during April 2007 are analyzed

by computing the paths of particles that are passively advected by the

model velocity field. The approach taken here is to initialize groups of

particles within TSB during the time of occurrence of the toxic algal

bloom reported by García-Mendoza et al. (2009), and then advect these

particles in two ways: backward and forward in time. The time-backward

advections provide information about where the water parcels come from.

To prevent confusion in the terminology, the paths are described here

from the standard, physical, forward-time perspective. Thus, the

backward-time integrations, by which the paths were computed,

originated at the particle final positions, and terminated at the particle

initial positions, while the physical, forward-time particle paths originate

at the initial positions and terminate at the final positions. The time-

forward advections, on the other hand, provide information about where

the water parcels go after they enter TSB. The combination of both kinds

of advections also provides information about the time residence of the

waters within TSB.

181

The initial conditions of the particles, which are the physical final

positions of the backward integrations, were chosen to be evenly

distributed horizontally within TSB (see Figure 2). Three different

advection experiments were done varying the initial depth of the

particles: 2, 38, and 75 m. A total of 830 particles were released in those

experiments starting at 2-m depth, whereas 331 and 160 particles were

released in those experiments starting at 38- and 75-m depths,

respectively. Such sets of particles were released on 24 April 2007, date

of the samplings reported on García-Mendoza et al. (2009). The

backward integrations were done through 24 April 2006 (i.e., one year-

long integration), whereas the forward integrations were done through the

end of year 2007 (i.e., 8 month-long integration).

The advections consist of two stages. The first one is carried out

within the TSB domain described in Section 2.a.2: the particles are

initialized as described above and are then advected by the flow in the

TSB interior. The second stage is carried out within the regional domain

described in Section 2.a.1: most of the particles leave the TSB domain

through its open boundaries during the advection, as soon as they have

left, they are advected by the flow in the regional domain and are not

allowed to re-enter the TSB domain. Thus, the advections in the TSB

domain provide the initial conditions for the advections in the regional

domain.

The model's three-dimensional velocities u = (u,v,w) wee used for the

advections, with standard bilinear interpolations to the instantaneous

particle positions. The particle-path time-integrations used an Euler

scheme and a time step of 10 minutes for the TSB domain, and of 3 hours

for the regional domain. No diffusion was used in such integrations.

182

3.b. Water-parcel paths

3.b.1 Surface waters

Here we discuss the paths of those particles starting the advection at 2-

m depth within TSB, a total of 830 particles per release. Despite the large

scatter of the advected-particle paths, there are some patterns

characterized by significantly higher particle concentrations (Figure 5a);

these patterns are the focus of this analysis. The particles come essentially

from locations far west of the continental shelf, many of them even

coming from locations beyond the model's western boundary (at

123.5°W). During their displacements, the particles depict meandering

trajectories, somewhat of the “M-shape” reported by Santamaría-del-

Ángel et al. (2002), affected by the mesoscale activity prevalent in the

region. Most of the particles are advected by a roughly zonal, inshore

flow at 32-33°N which bifurcates near the continental shelf in poleward

and equatorward flows, this latter one affecting directly the TSB

surroundings. This flow-bifurcation region is usually referred to as the

Ensenada Front (e.g., Haury et al., 1993; Venrick, 2000). Some particles

encounter the shelf and travel in the poleward flow just to be recirculated

within the Southern California Bight, immersed in the quasi-permanent

vorticity field in that region (e.g., Di Lorenzo, 2003), and ultimately enter

the equatorward flow and thus enters TSB. These particles are restricted

to depths of 40-100 m (Figure 5b). A smaller but significant amount of

particles show a different pattern. In this case the particles come from

locations west-southwest of TSB, within depths of 100-200 m (Figures

5a-b).

Most of the particles enter TSB through the northwest (northern

portion of the western boundary; west of Salsipuedes Bay), generally

within the first 70 m from the surface, and they distribute along the whole

183

Bay (Figures 6a-c). A much smaller amount of particles come from the

west and south, from deeper levels (100-200 m), but these paths are

probably meaningless, result from stochastic diffusive processes.

Nonetheless, as we discuss in the following section, these patterns are

indeed dominant for those particles starting the advection at deeper levels

within TSB.

After the particles upwell and reach the surface levels (2-m depth)

within TSB, they leave this domain mostly through the southern

boundary, remaining at the surface (Figures 6d-f). As before, there are

particles that follow different paths, leaving through the western

boundary at deeper levels (100-200 m), especially west of Salsipuedes

Bay.

As soon as the particles leave TSB, they scatter in the south-

southwestward direction, affecting only a ~200 km-long portion of Baja

California coast right south of TSB, following paths that separate from

the coast near Punta Baja (Figures 5c). The particles remain at the surface

layer during their paths (Figure 5d). Although in a much smaller amount,

some particles depict poleward paths roughly following the shelf, at

depths below 200 m, most probably advected by the undercurrent over

the slope (e.g., Pierce et al., 2000; Gay and Chereskin, 2009).

3.b.2 Subsurface waters

In this section we discuss the paths of those particles starting the

advection at 75-m depth within TSB, a total of 160 particles per release.

These results are essentially consistent with those obtained from the

advections starting at 38-m depth, hence these are not described in detail.

The paths followed by the particles to get into TSB are similar to those

described previously, also centered in depths between 100 and 200 m

184

(Figures 7a-b). As before, most of the particles enter TSB through the

northwest, at depths of 50-150 m (Figures 8a-c).

The most striking difference with respect to the paths described in the

previous section is that in this case most of the particles show a conversed

pattern. Most of the particles leave TSB through the northwest at depths

between 70 and 250 m (Figures 8d-f). Once these particles are out of

TSB, they travel poleward within a very well defined path adjacent to the

continental shelf, at depths below 200 m (Figures 7c-d). Interestingly,

south of the Channel Islands in the Southern California Bight, this

poleward flow bifurcates into a main flow following the continental slope

and an inshore current that flows poleward through the interior channel

(between the islands and the continent), surrounding the islands to

eventually re-join to the main flow; this inshore flow is usually referred to

as the Southern California undercurrent (e.g., Di Lorenzo, 2003).

3.b.3 Residence times

From the experiments described above, we can estimate residence

times for the waters present in TSB during April 2007. For such a

purpose, Figure 9 shows the percentage of the advected particles

remaining in TSB interior as a function of the time elapsed during both

the backward and forward advections; notice that zero corresponds to the

particle-release time, hence the percentage of particles at this time is

exactly 100%. We arbitrarily define the residence time as the period

between the time when 20% of the particles have entered the Bay during

the backward-time advections, to be accumulated and reach the 100%,

and the time when 80% of such maximum percentage have left the Bay

during the forward-time advections. Thus the resulting residence times

are 43 days for those particles starting the advections at the surface (2-m

185

depth), and 31 and 28 days for those particles starting at subsurface levels

(38 and 75-m depths), respectively. The longer time for the surface

waters is probably due to the time required to upwell within the Bay,

from depths of roughly 50 m to 2 m, against the local thermocline.

4. Discussion

Our analysis addresses, from numerical simulations, the origin and

distribution of the waters present in Todos Santos Bay (TSB) during

spring 2007. Circulation patterns for this period might help to explain the

presence of an intense toxic algal bloom occurred in this area, which was

the southernmost report of the presence of domoic acid (DA) in the

California Current System and it is also the first report of the distribution

of toxic Pseudo-nitzschia species and DA on the Baja California west

coast (García-Mendoza et al., 2009). The accumulation of toxic cells was

most probably caused by environmental conditions associated with the

wind-driven upwelling in the region.

Backward-time advections show that most of the water parcels come

from western locations, many of them even from locations well beyond

the model's western boundary (at 123.5°W). They travel eastward around

32-33°N, depicting a meandering paths, to eventually encounter the

continental shelf close to 32.5°N, the Ensenada Front (e.g., Haury et al.,

1993; Venrick, 2000), where the flow bifurcates and most of the waters

flow toward TSB and another significant amount flow into the Southern

California Bight (SCB). These latter parcels become immersed in a

vigorous vorticity field prevalent in that area (e.g., Di Lorenzo 2003),

staying there for a few weeks, but eventually re-enter the Ensenada Front

to flow toward TSB adjacent to the shelf. This suggests that the origin of

the toxic bloom mentioned above might have been occurred northward of

186

the study area. Therefore, the presence of toxic cells must have been

observed first in the SCB before its occurrence in TSB. Indeed,

accumulation of toxic Pseudo-nitzschia cells were detected near

Coronado Island at the beginning of April (E. García-Mendoza,

unpublished results) and high concentrations of DA were measured in

Central and Southern California in the same period

(http://www.cdph.ca.gov/HealthInfo/environhealth/water/Pages/Shellfishr

eports.aspx).

After the water parcels enter TSB, two different motion regimes occur.

Most of the waters upwell within the Bay to reach the surface and then

leave through the southern boundary of the Bay. As soon as they leave

the Bay, the waters scatter south-southward, many of them remaining

over the continental shelf of the ~200 km-long coastal portion south of

TSB, and then they move away from the coast south to Punta Baja. This

suggests that the waters associated to the toxic bloom would affect only

those regions right south of TSB, north of Punta Baja.

The second water-motion regime is a very striking result. Once they

have entered the TSB area, the waters upwell up to subsurface levels

(below 35 m), and eventually leave the Bay through the northern

boundary at deeper levels (below 100 m), probably favored by canyon-

induced vertical exchange (see Allen and Durrieu de Madron, 2009). As

soon as the waters leave the Bay area, they reach depths below 200 m and

flow poleward adjacent to the shelf north to SCB, most probably

embedded in the undercurrent over the slope (e.g., Pierce et al., 2000;

Gay and Chereskin, 2009), especially west of Salsipuedes Bay. This

result implies an export of subsurface waters from TSB to SCB.

The knowledge of the residence time of the water parcels within TSB

is of key importance to model the evolution of toxic blooms in the area.

Estimated residence times for surface water is of one and a half months,

187

whereas subsurface water remains in the Bay for one month only. These

estimated residence times corresponds to a “snapshot” of the TSB's

waters, since all the advection experiments start at the same time (24

April 2007), and they do not necessarily represent the expected period of

the presence of toxic cells. Unfortunately, there are not enough

Lagrangian observations to estimate residence times and dispersion of the

TSB's waters. Historical surface-drifter observations (Álvarez-Sánchez et

al., 1988) have provided useful information (for example,

convergence/divergence regions within the Bay), but given their very

limited spatial and temporal extension, the comparison with our results is

difficult and probably meaningless.

There is a historical “Lagrangian-like” study of the residence of TSB's

waters that is worthwhile mentioning it here. Edgar Pavia describes in

CICESE's internal gazette (De la Cerda 2008) that Antoine Badan carried

out one of his first experiments during 1970, in which thousands of

oranges with plastic markers in their interiors, were distributed and

released in several parts of the Bay. During the next days, local students

walked along the Bay's beaches looking for such oranges, only some of

them were recovered, the rest were lost. In spite of its anecdotal value,

this experiment showed evidence of the transitory character of the waters

in TSB.

The analysis presented in this paper offers a good framework to

understand the ecology processes behind the Pseudo-nitzschia blooms in

TSB and Baja California. Among other possible applications, our results

provide useful information for the definition of permanent monitoring

points which, together with real-time numerical models, would support

the ultimate goal of prediction of toxic algal blooms in the Mexican

eastern coasts.

188

5. Conclusions

The three-dimensional Lagrangian analysis based on a high-resolution

numerical model shows that the water parcels present in Todos Santos

Bay by the end of April 2007, which are most probably associated with

the intense toxic algal bloom occurred in spring 2007 reported in the

literature, come essentially from locations west of the Bay (even beyond

from the model's domain). They approach to the continental shelf,

embedded in a flow that bifurcates in an equatorward current and a

poleward current at 32-33°N, the so-called Ensenada Front. Once the

water parcels enter the Bay, mostly from the west-northwest, two

different regimes are identified. In the first one the waters upwell and

reach the surface (2-m depth), and then leave the Bay flowing to the

south, affecting only the ~200 km-long coastal portion right south of

TSB; they scatter south-southwestward, remaining at the surface layer.

The second regime consists of waters that do not reach the surface within

the Bay, but they remain at subsurface levels (38- and 75-m depths) for

eventually leave the Bay mostly to the northwest, reaching deeper levels,

and flowing poleward north to the Southern California Bight, at depths

below 200 m, most probably advected by the poleward undercurrent over

the slope. Estimations of residence times of the waters within the Bay

result in about one and a half months for the surface waters and one

month for the subsurface waters.

Acknowledgements

This work is dedicated to the memory of Antoine Badan, a great

oceanographer, mentor and friend. The numerical simulations were

carried out during DR's postdoctoral appointment at Oregon State

189

University; the hospitality of Prof. Roger M. Samelson and the financial

support by the National Science Foundation (NSF) Science and

Technology Center for Coastal Margin Observation and Prediction

(CMOP), through NFS award 0424602, are thankfully acknowledged.

The QuikSCAT product was available through the French Research

Institute for Exploitation of the Sea (IFREMER) website:

http://www.ifremer.fr/cersat/en/data/download/gridded/mwfqscat.htm.

The altimeter product was produced by Ssalto/Duacs and distributed by

Archiving, Validation, and Interpretation of Satellite Oceanographic data

(AVISO), with support from Cnes.

190

References

Allen, S. E., and X. Durrieu de Madron, 2009: A review of the role of

submarine canyons in deep-ocean exchange with the shelf. Ocean Sci., 5,

607-620.

Álvarez-Sánchez, L. G., R. Hernández-Walls, and R. Durazo-Arvizu,

1988: Drift patterns of lagrangian tracers in Todos Santos Bay. Ciencias

Marinas, 14, 135-162.

Barron, C. N., A. B. Kara, P. J. Martin, R. C. Rhodes, and L. F.

Smedstad, 2006: Formulation, implementation and examination of

vertical coordinate choices in the Global Navy Coastal Ocean Model

(NCOM). Ocean Modell., 11, 347-375.

Burk, S. D., T. Haack, and R. M. Samelson, 1999: Mesoscale simulation

of supercritical, subcritical, and transcritical flow along coastal

topography. J. Atmos. Sci., 56, 2780-2795.

De la Cerda, J., 2008: Antonio Badán: genio y amigo. TODoS@CICESE,

129, http://gaceta1.cicese.mx/.

Di Lorenzo, E., 2003: Seasonal dynamics of the surface circulation in the

Southern California Current System. Deep-Sea Res. II, 50, 2371-2388.

Egbert, G. D., and S. Y. Erofeeva, 2002: Efficient inverse modeling of

barotropic ocean tides. J. Atmos. Oceanic Technol., 19, 183-204.

García-Mendoza, E., D. Rivas, A. Olivos-Ortiz, A. Almazán-Becerril, C.

Castañeda-Vega, and J. L. Peña-Manjarrez, 2009: A toxic Pseudo-

nitzschia bloom in Todos Santos Bay, northwestern Baja California,

Mexico. Harmful Algae, 8: 493-503.

Gay, P. S., and T. K. Chereskin, 2009: Mean structure and seasonal

variability of the poleward undercurrent off southern California. J.

Geophys. Res., 114, C02007, doi:10.1029/2008JC004886.

Haury, L. R., E. L. Venrick, C. L. Fey, J. A. McGowan, and P. P. Niiler,

1993: The Ensenada Front: July 1985. CalCOFI Rep., 34, 69-88.

191

Mateos, E., S. G. Marinone, and A. Parés-Sierra, 2009: Towards the

numerical simulation of the summer circulation in Todos Santos Bay,

Ensenada, B.C., Mexico. Ocean Modell., 27, 107-112.

Mesinger, F., G. DiMego, E. Kalnay, K. Mitchell, P. C. Shafran, W.

Ebisuzaki, D. Jović, J. Woollen, E. Rogers, E. H. Berbery, M. B. Ek, Y.

Fan, R. Grumbine, W. Higgins, H. Li, Y. Lin, G. Manikin, D. Parrish, and

W. Shi, 2006: North American Regional Reanalysis. Bull. Amer. Meteor.

Soc., 87, 343-360.

Penven, P., P. Marchesiello, L. Debreu, and J. Lefèvre, 2008: Software

tools for pre- and post-processing of oceanic regional simulations.

Environ. Model. Softw., 22, 117-122.

Perlin, N., E. D. Skyllingstad, R. M. Samelson, P. L. Barbour, 2007:

Numerical simulations of air-sea coupling during coastal upwelling. J.

Phys. Oceanogr., 37, 2081-2093.

Pierce, S. D., R. L. Smith, P. M. Kosro, J. A. Barth, and C. D. Wilson,

2000: Continuity of the poleward undercurrent along the eastern

boundary of the mid-latitude north Pacific. Deep-Sea Res. II, 47, 811-

829.

Rivas, D., and R. M. Samelson, 2010: A numerical modeling study of the

upwelling source waters along the Oregon coast during 2005. J. Phys.

Oceanogr., in press.

Santamaría-del-Ángel, E., R. Millán-Núñez, A. González-Silvera, and F.

Müller-Karger, 2002: The color signature of the Ensenada Front and its

seasonal and interannual variability. CalCOFI Rep., 43, 155-161.

Shchepetkin, A. F., and J. C. McWilliams, 2005: The Regional Ocean

Modeling System: A split-explicit, free-surface, topography following

coordinates ocean model. Ocean Modell., 9, 347-404.

192

Smith,W. H. F., and D. T. Sandwell, 1997: Global sea floor topography

from satellite altimetry and ship depth soundings. Science, 277, 1956-

1962.

Strub, P. T., J. S. Allen, A. Huyer, R. L. Smith, and R. C. Beardsley,

1987: Seasonal cycles of currents, temperatures, winds, and sea level over

the northeast Pacific continental shelf: 35°N to 48°N. J. Geophys. Res.,

92, 1507-1526.

Venrick, E. L., 2000: Summer in the Ensenada Front: the distribution of

phytoplankton species, July 1985 and September 1988. J. Plankton Res.,

22, 813-841.

193

Figure 1. Regional model's domain and its bathymetry (in m). Location of Todos Santos Bay is shown. Black triangle indicates the position of the NDBC buoy (see text).

194

Figure 2. Todos Santos Bay model's domain and its bathymetry (in m). Dotted array indicates the initial positions for the advection experiments described in Section 3.

195

Figure 3. Temporal evolution of (a) Sea surface temperature (SST) and (b) adjusted sea level (ASL) at the position of NDBC buoy 46086 (see Figure 1) during 2006 and 2007. Red lines correspond to the observed variables, black lines correspond to the modeled ones (interpolated from the model's outputs). Tick marks in the time axis denote the start of the month.

196

Figure 4. Mean velocity (in cm s-1) at Todos Santos Bay during July 2007. (a) Values interpolated at 5-m depth are presented, as well as the velocity normal to the vertical cross-sections located at the (b) northern and (c) western portions of the Bay, indicated by the straight lines in panel (a).

197

Figure 5. Percentage of particles of a total of 830 (first column) and mean particle depth (second column, in m) per model's grid cell. All the trajectories starting at 2-m depth within TSB were used. Only those cells containing at least 10% of particles are plotted. Upper panels correspond to the backward-time advections, lower panels correspond to forward-time advections. The position of TSB domain is indicated.

198

Figure 6. Positions of the advected particles within TSB, starting at 2-m depth. The first column [panels (a) and (d)] shows their positions when they cross the open boundaries, divided into three groups for better visualization. The second column [panels (b) and (e)] shows their positions at the start of the advection, with the markers matching those in panel (a) to indicate where they cross to enter or leave the domain during the advection. The third column [panels (c) and (f)] indicates the depth (in m) when the particles were crossing the boundaries. The first row [panels (a)-(b)] corresponds to the time-backward advection, whereas the second row [panels (d)-(f)] corresponds to the time-forward advection.

199

Figure 7. Same as Figure 5, but for the advections starting at 75-m depth, with a total of 160 particles released.

200

Figure 8. Same as Figure 6, but for the advections starting at 75-m depth.

201

Figure 9. Percentage of particles remaining within the TSB domain during the advections described in Section 3. Notice that zero in the horizontal axis is the initial time for both the backward and the forward advections. Results from the experiments starting at three different levels are shown (see text). The horizontal dashed line indicates where 80% of the particles have left the TSB domain, which is arbitrarily chosen to estimate a residence time for the waters present within the Bay (see text). The estimated residence times are shown, with the font color matching the color of the lines.

202

203

Chapter 3

The Wine

204

205

California Heat Waves with Impacts on Wine

Grapes*

Alexander Gershunov ♠, SIO Daniel R. Cayan, SIO, USGS Bernard Retornaz, Maison Louis Latour

Abstract

Heat wave activity in the California region has undergone significant

changes over the last two decades. Regional heat waves in recent years were

more humid and their nighttime temperature expression has been greatly

accentuated. These recent changes appear to be an acceleration of a long-

term trend that is regionally consistent with global climate change. Humid

heat and elevated nighttime temperatures, especially in the form of stronger,

longer lasting and more spatially extensive heat waves can lay emphasis on

various social and environmental aspects stress society and environment.

Agriculture is susceptible to humid heat, particularly in a region

acclimatized to hot and dry summer days and cool nights, the ideal

conditions for wine grapes. Two decades of observations reveal potentially

detrimental effects of increased heat wave activity in California upon state-

wide wine production tendencies. In traditional “old world” wine-making

regions, wine harvest dates are known to reflect, in part, average climate

conditions over the growing season, in the sense that warmer seasonal

* The Ocean, The Wine, and The Valley: The Lives of Antoine Badan, 205-223.

© 2010 CICESE. ♠ Contact: sasha@ucsd.edu

206

temperatures tend to favor larger harvests. In California, however, we find a

non-significant tendency for warm summers on average to favor smaller

yields, especially for red wine grapes, while a much stronger relationship

exists with extreme temperatures rather than with mean summertime

temperatures—extreme hot events, especially those of the humid variety,

have been strongly associated with lower-yield harvests. This is a regional

state-wide relationship. A further indication of heat wave impacts is a

linkage between hot temperature extremes and delayed harvest dates of

Chardonnay grapes at a vineyard in the Carneros region of the Napa Valley.

Possible reasons for these findings are discussed along with their

implications and future research plans in climenology – the study of enology

and climate.

1. Introduction

Climatic vacillations and trends are keenly recorded by Bacchus since

time immemorial in grape harvests and in the quality of the wines they yield.

Grapes absorb heat and sunshine, the fruit optimally thrive when days are

warm and dry and nights are cool: conditions characteristic of

Mediterranean-type summer climates found in several regions around the

world. Grapes integrate climatic conditions during the warm growing season

and, therefore, the warm season’s climate signature is imprinted into the

vines and eventually the wine. Traditionally, cool and rainy summers result

in late harvests of immature fruit yielding weak sour wines. Hot and dry

summers typically result in early harvests of mature fruit and abundant

robust wines. The effects of average seasonal temperatures and related

insolation on harvest dates are well known and quantified, so much so, that

207

wine harvest dates are a prime example of historical proxy data used to

reconstruct past climate (spring and summer average temperature and

cloudiness/rainfall) in old-world traditional wine growing regions (e.g. Le

Roy Ladurie 1967). It is also understood that extreme short-duration weather

events, even if not clearly reflected in seasonal averages, can decisively

impact the grape if they occur at key times during the growing season. Such

is the case with heat waves, which tend to occur near the annual peak of

temperature in July and August, around the time of veraison or grape

ripening. Although dry heat is generally good for wine grapes, excessive

and/or humid heat can stress the grapes’ metabolism, delay ripening and

promote withering. In the less traditional “new world”, the relationships

between harvest times, yields and quality are complicated by non-traditional

wine-making practices. Without focusing on grapevine chemistry or

biology, we here provide empirical evidence of the effect of heat waves on

wine grapes in California.

Excessive heat waves in California were recently shown to be primarily

of two types – (I) dry daytime heat and strong nighttime cooling typical for

the region, and (II) humid heat waves expressed most strongly in nighttime

temperature anomalies (Gershunov et al. 2009, hereafter GCI09). Humid

heat waves display a clear and upward-accelerating trend culminating so far

with the July 2006 event – the strongest heat wave recorded in at least six

decades of station observations – twice as strong in minimum (nighttime)

temperature as the previous type II record, 2003, which was at the time of its

occurrence also unprecedented by a large margin. Even the daytime

(maximum) temperature expression of the 2006 primarily nighttime humid

event broke records set by type I events. The upward trend in California

208

humid nighttime heat waves is strongly reflected in temperature intensity,

duration, and spatial extent (GCI09).

Because nighttime heat waves have not traditionally been the typical

form of heat waves in this semi-arid Mediterranean climate, they may exert

greater stress on many living organisms acclimatized to dry daytime heat

and cool nights. The 2006 heat wave, therefore, impacted the health of

humans and animals, the ecosystem and energy supply (see references in

GCI09) and, as we shall see, the wine grapes. Powerful impetus for impact

quantification of humid heat waves on grapes, among other organisms, is

provided by the fact that a regional intensification of humid heat waves is

consistent with global warming via exceptional warming of the Pacific

Ocean just west of Baja California (GCI09) that is enhanced by a regional

cloud-feedback mechanism (Clement et al. 2009). This warming is

associated with increasingly humid air that is transported northward to

California by synoptic circulations responsible for large-scale regional heat

wave activity (GCI09).

Below, we briefly describe the weather and grape data as well as the

procedure for heat wave activity quantification locally and throughout the

region. We then show some relevant features of heat wave activity and their

observed changes. Red and white wine grape production is then described;

environmental influences are quantified and related to regional heat wave

activity. Finally, Chardonnay harvest dates at one Napa Valley vineyard are

examined together with local heat wave activity. In conclusion, results are

briefly discussed in the context of winemaking practices, future climate

challenges and topics for future research in climenology, our proposed term

for the study of climate and winemaking.

209

2. Data

Daily maximum and minimum temperature records were obtained, as in

GCI09, from the National Climatic Data Center (NCDC 2003) but updated

through summer 2008. Out of 353 weather stations in California, we chose

135 stations as representative of the State. These stations were quality

controlled and selected as in GCI09 but with somewhat denser coverage

practically throughout California – the station with the most complete

summertime records (1948-2008) in a 20 km radius was chosen. In hilly

terrain, typical of many vineyard locations, where elevation ranges at least

100 m in a 20 km radius, the highest elevation station with adequate data (no

more than 15% of data missing over the 1948-2008 period) was also

retained. This removed the urban bias present in the spatial distribution of

stations and resulted in a generally uniform state-wide coverage.

Every year, the US Department of Agriculture publishes a publicly

available report on all wine grapes officially crushed in the state of

California (CDFA 2009). Information on tonnage and price is available for

every year since 1988. The data are classified by red and white wine types in

addition to raisin and table types. Grape-crush in thousands of tons as well

as price per ton for various types of grapes is utilized here for red and white

wine types.

Harvest dates for Chardonnay grapes grown at the Brown Ranch

vineyard in northeast Carneros, Napa Valley, were obtained from Saintsbury

Vineyards (http://www.saintsbury.com/) courtesy of David Graves.

Unfortunately, not enough recent data are available for Baja California,

particularly for the Guadalupe Valley, Mexico’s primary wine region.

210

However, its proximity to the State of California makes the state-wide

results presented here relevant to impacts on wine making in Mexico.

3. Results

Clearly, there has been a general warming across California, which is

particularly apparent in minimum temperatures (Figure 1). The summer

(June, July and August, or JJA) of 2006 stands out as the hottest summer in

average nighttime temperatures (GCI09). Some of this warmth had to do

with the late July 2006 heat wave that was accentuated in nighttime

temperatures because the elevated greenhouse effect associated with

extraordinary levels of humidity impeded radiational cooling (GCI09). This

event powerfully marked most of California including, as we shall see, the

Napa Valley (Figure 2). Because stronger warming occurred at nighttime or

minimum temperature (Tmin) compared to daytime maximum temperature

(Tmax), the diurnal temperature range (Tmax – Tmin) decreased. This

pattern has been largely typical of observed global warming (e.g. Easterling

et al. 1997).

211

a) Average summertime Tmax

ooo

o

o

oo

oo

o

o

oo

o

o

o

o

o

o

o

oo

o

o

ooo

oo

o

oo

o

o

o

o

o

o

o

o

o

oo

o

o

o

o

o

o

oo

o

oooo

oo

o

oo

1950 1960 1970 1980 1990 2000 2010

29.0

30.0

31.0

32.0

b) Average summertime Tmin

o

oooo

oooo

o

ooo

o

oo

o

o

o

o

o

ooo

o

oo

oo

o

o

oo

o

o

o

oo

o

o

o

o

o

o

o

o

o

o

o

oo

o

ooo

o

o

o

o

oo

1950 1960 1970 1980 1990 2000 2010

12.5

13.0

13.5

14.0

14.5

15.0

Figure 1. Maximum and minimum temperatures (Tmax and Tmin) averaged over summer (JJA) in 135 California stations.

3a Heat waves

We define heat waves from daily observed maximum and minimum

temperatures, as in GCI09. Local Tmax and Tmin climatologies are

constructed at each station using daily data from the 1950-1999 base period

summers. For each summer, temperature exceeding over a high threshold

(99th or 95th percentiles of the local baseline climatology) are summed up as

degree days at each station for each summer giving the local summertime

heat wave activity index (HWAI). Figure 2 shows the data and thresholds at

the Napa State Hospital station. Heat waves that exceed local thresholds can

last from several days to two weeks. Not every summer registers local heat-

wave activity at the 99th percentile threshold. The late July 2006 event was

unprecedented in many California stations. In the Napa station, it was not

extraordinary in terms of its peak one-night intensity (June 14, 2000, was

record high at 25ºC), but due to its overall magnitude (intensity and

212

duration), local HWAI peaked in 2006, relative to this part of the record, and

this is reflected in Figure 7 below.

date from June 10 20 40 60 80

1020

3040

xx

xxx

x

x

x

xx

xx

xx

xxx

xxx

xx

xx

x

xx

xx

xx

xxxx

xxxxxx

xxx

xx

xx

xxxxx

x

xxx

xxxx

x

x

xx

x

xx

xx

xx

xxxxx

x

xx

xx

xx

xxxxx

xx

x

x

x

xxxxxx

x

xx

xx

x

xx

xx

x

xxx

x

xx

x

xxxxxx

xxx

x

xxx

xxx

x

x

xx

x

xx

xxxxxx

xx

x

x

xx

xx

xxxxx

xx

xx

xxxx

xx

x

xx

xxxx

x

xxx

xx

x

x

xx

xxx

x

x

xx

xx

xxx

x

x

xx

x

x

x

x

x

xxx

x

x

xx

x

xx

xxx

xxx

xx

xxxx

x

x

x

xx

x

xx

xx

x

xxxx

x

x

xx

x

x

x

x

x

xx

xxxxx

xx

xxx

xx

xx

xxx

xxx

x

x

xxx

xx

x

xx

xx

xx

x

xxx

xx

xx

xx

xx

x

xxx

x

xxxx

xxxx

xx

xxx

xxxxxxxx

x

xx

x

xxxxx

xxx

xxxx

x

x

x

xxxx

x

xx

xxx

xx

xxxxxxxxx

xx

x

x

xx

x

xxxxxx

x

xx

x

xxx

xxxx

x

x

xx

xx

xx

xxx

x

xx

xx

x

xx

x

xx

x

xx

x

xxxx

xxxx

xx

x

x

x

xxx

x

x

x

x

xxxx

x

xxxx

xx

x

xxxx

x

xxxx

x

xx

xx

xxxxx

xxx

xxxxxxxxxxx

xx

x

x

xx

xx

x

x

x

xxxx

xx

xx

x

x

xxxxxx

xxxxx

xxx

xxx

xx

xx

xxxx

x

xx

xxxxx

xx

xx

x

xx

x

xxx

xxxxxx

xxx

xx

xx

x

x

x

x

xx

xx

x

xxxx

x

x

xxx

xx

xxxxxx

xxx

xxxxx

xx

xx

xxx

x

x

xx

xx

x

xxx

x

x

xx

xxx

xxxx

xx

xxx

xx

xxxxx

x

xxx

xxx

x

x

x

xxx

x

x

xx

x

xx

x

xxx

xxxx

xxx

xxx

xx

x

xxx

xx

x

xx

xx

xx

xxxx

x

xx

xx

xx

xx

x

x

xxxxx

xxx

xxxx

xx

xx

xx

xxx

xx

xx

x

xx

xx

xx

xxx

x

x

x

xx

x

xx

x

xx

xx

x

xxx

xxx

xx

xx

xxxxxx

x

xx

x

xxx

xx

x

xx

x

xx

xxxx

x

xx

xxx

xxxx

xx

xxxx

xx

xx

x

xxx

x

xx

xxxx

xx

xx

x

xx

x

xxx

xx

xx

xx

xxxxxxx

xx

xx

xxxxxx

xxx

x

x

xxx

x

xxxx

xx

xx

x

xxxxx

xx

x

xx

xxxxx

xx

xxxx

xxxxxx

xxx

xxxxx

xxx

xx

x

xxx

x

x

xxx

x

x

xxx

xx

xx

xx

x

xxx

x

xxx

xx

xxxxxxx

x

xxxx

xxx

xx

xxx

x

xxx

xx

x

x

xxxx

xxxx

xx

xx

x

xxxxxx

xxx

x

xx

x

x

xx

xx

xx

xx

x

xxxxx

x

x

x

xx

xxxxxx

x

xxx

xxx

x

xx

x

x

x

x

xxx

xx

xxxxx

xx

x

xxxxx

x

xx

x

x

xxxxxx

xx

xxx

xx

xxxx

xxx

xx

x

x

xxx

x

xx

xx

x

xx

xx

xxxx

xx

x

x

x

xxx

xx

xx

xx

xx

xx

xxxxx

x

x

xx

x

x

x

xx

x

xxx

x

x

xx

xxxxx

x

x

xx

x

x

x

xx

x

xxx

x

xx

xx

x

x

x

xx

x

xx

xxx

xxx

x

x

xx

x

xxxx

x

xx

x

xx

xxxx

xxxx

x

x

xx

xxx

x

xxx

xx

x

xxx

x

xxx

x

xxxx

xxx

xxx

x

xx

xx

x

xxx

xxx

xx

xx

xx

xxx

xxx

xx

xxxxx

xxx

xx

x

xx

xxx

x

xxxxxx

x

xxxx

xxx

x

x

xx

x

Figure 2. Minimum (Tmin, blue circles) and maximum (Tmax, red x’s) temperatures as functions of summer date (from June 1 to August 31) observed at the Napa State Hospital meteorological station (38.27ºN, 122.25ºW) along the 14 years 1994-2007, the period of overlap with phenological observations at close-by Brown Ranch vineyard, Carneros (Figure 7). 2006 Tmin and Tmax are shown as blue and red curves, respectively. July 23, the peak date of the 2006 event is marked with the vertical line. Horizontal lines delineate the 95th and 99th percentiles of the 1950-1999 Tmin and Tmax climatologies.

Local HWAI computed at every station and averaged throughout all

stations in California constitutes the regional California HWAI (CAHWAI,

Figure 3). Regional nighttime-accentuated (Type II) heat waves are clearly

213

on the rise, as discussed in detail in GCI09. The crucial causal difference

between daytime and nighttime-accentuated heat waves is not atmospheric

circulation, but rather the availability of humidity that is advected by the

heat-wave circulations. Heat-wave circulations typically involve a high

surface pressure dropping into the southern Great Plains and a low pressure

developing concurrently along California’s central coast. This synoptic

configuration of surface pressure centers favors northward advection of air

into California notably from a marine region west of Baja California that has

been experiencing a pronounced warming (1.5°C since 1948) and a related

warming and moistening of the air above it. Advection of this gradually

warmer and especially moister air by episodic synoptic heat wave

circulations is largely responsible for the rising trend in nighttime humid

Type II heat-wave events. Please refer to GCI09 for more detail.

a) CAHWAI95: 95th percentile

threshold

1950 1960 1970 1980 1990 2000 2010

510

1520

25

TmaxTmin

b) CAHWAI99: 99th percentile

threshold

1950 1960 1970 1980 1990 2000 2010

02

46

8

TmaxTmin

Figure 3. CAHWAI computed using the 95th (a) and 99th (b) percentile exceedance thresholds.

214

The more extreme CAHWAI99 index shows a steeper and apparently

accelerating rate of increase compared to the milder CAHWAI95.

3b California grape crush

Figure 4 shows total tons of red and white wine grapes crushed in the

entire state along with their prices averaged over all vineyards and for all

varieties. In the following discussion, we use grape crush and production as

interchangeable terms.

a) Red and white grapes crushed

(K tons)

R

RR R

RR

R

R R

R

RR

R

R

R

R R

R

R R

R

K to

ns

1990 1995 2000 2005

1000

1500

2000

WW

WW

W

WW

WW

W

W W

W

W WW

W

W

W

W W

R

RR R

RR

R

R R

R

RR

R

R

R

R R

R

R R

R

b) Red and white grape prices

(USD/tonn)

R R

RR

R R

R

R

R

RR R

R

R

R RR R R R

R

$/to

n

1990 1995 2000 2005

300

400

500

600

W

WW

W

W

W W

W

W

WW W

W W

W W

WW W

W

W

R R

RR

R R

R

R

R

RR R

R

R

R RR R R R

R

Figure 4. California Grape Crush Report data. Red and white grapes crushed. a) Red and black curves, respectively, and b) Wine grape prices.

It is clear from Figure 4b that most of the variability in grape crush

(Figure 4a) is due to wine style and other resource management decisions.

Originally, more white wine grapes were grown in California. As red wine

grapes have always sold for a higher price (Figure 4b), the production of red

wine grapes increased, while that of white wine grapes remained relatively

215

constant since 1988. Probably more vineyards were planted to grow red

wine grapes, perhaps at the expense of other agricultural commodities, but

apparently not too much at the expense of white wine grapes whose

production, although decreasing somewhat in the mid-nineties, subsequently

recovered. Prices for both types of wine grapes increased the most in the

early and mid-nineties. Possibly in response to the price of red wine grapes

remaining higher than that of white wine grapes, production of red wine

grapes increased most quickly in the second half of the 1990s, by about 80%

in five years. The production increase lagged the price increase by about

three years required for vines to take root and mature. Prices remained

relatively stable since 1997 and the red wine production trend subsided since

2000 and stabilized at a level above that of white wine grape production.

Although the red wine grape crush trend subsided, the year-to-year

variability or production volatility increased since 1997 with an all-time

peak in production registered in 2005, for both grape types. To remove

trends due to economic decisions and accentuate the possible effects of

environmental variability including that of climate on crush volatility, we

produced a series of innovations or gains, i.e. changes in grape production

from one year to the next (Figure 5).

216

R

R

R RR

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

K to

ns

1990 1995 2000 2005

-200

020

040

060

0

W

W

WW

W

W W W

W

W

W

W

W

WW

W

W

W

W

W

R

R

R RR

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

Figure 5. Red and white grape crush gains, i.e. production in year (n) – year (n-1).

The series of innovations in red and white wine grape crush (Figure 5)

more clearly show a relative stability of production until a big gain in 1997,

and strong volatility (i.e. large year-to-year changes) in production since

then. Big jumps in innovations of red and white grape production are clearly

in phase with each other, which is to be expected if large-scale

environmental factors are responsible for these jumps1. The largest year-to-

year drop in production of both types of wine grapes occurred in 2006, while

the second largest drop in red wine grape production occurred in 2003, the

1 It is also possible that environmentally-driven overproduction affects prices and production decisions the following year. However, this idea is not pursued here. Instead, we show how changes are related to heat.

217

second most intense humid heat-wave year. We next compare these gains

with the daytime and nighttime statewide heat-wave indices, the CAHWAI

(Figure 6).

1990 1995 2000 2005

-2-1

01

23

4

CAHWAI (Tmax99)CAHWAI (Tmin99)Red Grape Crush Loss (-Gain)White Grape Crush Loss (-Gain)

Figure 6. This figure combines Figures 5 and 3b over their overlapping period. Grape crush gains are inverted for ease of visual interpretation. Units are scaled to standard deviations. Original magnitudes can be seen in Figures 3b and 5. The correlation coefficient (r) of CAHWAI (Tmin99, blue curve) with red wine grape crush gain (solid black curve) is -0.60, p = 0.0027, and that with white wine grape crush is -0.38, p = 0.047; correlations with Tmax99 as well as with Tmin95 and Tmax 95 are progressively lower, but all significant at the 95% confidence level, while correlations with average summertime Tmin and Tmax are not significant (-0.32 with red grape crush gain and only half that magnitude with white grape crush gain, for either Tmin or Tmax summertime average).

Large gains in crushed grapes were registered in years with mild

summers (1997, 2000, 2005) while big drops in 1997 and 2003 and 2006

clearly coincided with active heat wave summers. We know from GCI09

218

(their Figures 12d and 13d) that the summer of 2001 saw strong humidity

and nighttime-accentuated heat wave activity, although daytime

temperatures were depressed due to convective cloudiness and even

relatively widespread rainfall in early July. This mugginess together with

reduced sunshine could have affected the grapes adversely in 2001.

We know that, traditionally, dry daytime heat is good for grapes. In

California, where the bulk of wine grapes is nowadays grown in the very hot

Central Valley, we see a weak state-wide tendency for cooler summers on

average to yield bigger harvests. Too much heat appears to be harmful, as

the much stronger relationship between grape crush and the more extreme

CAHWAI99 clearly suggests. Moreover, it is specifically humid heat that

appears to be particularly detrimental to wine grapes and red wine grapes are

especially affected.

3c Napa valley harvest dates

Given results of Le Roy Laduire (1967) and the fact that heat builds up

sugar in the grapes, we expected to see early harvest dates in particularly hot

summers. However, results presented in Figure 7 suggest the opposite, at

least for the Chardonnay grapes at Brown Ranch vineyard of Saintsbury in

northeastern Napa Valley.

219

1994 1996 1998 2000 2002 2004 2006

-2-1

01

23

Napa HWAI (Tmax95)Napa HWAI (Tmin95)Harvest date

Figure 7. Local Napa Valley HWAI and Chardonnay harvest dates at nearby Brown Ranch vineyard, northeast Carneros, Saintsbury, Napa Valley. Indices are scaled to standard deviations. The HWAI data are scaled from original values displayed in Figure 3b; harvest dates ranged from August 27 in 1997 and October 17 in 2006. Correlation coefficients of harvest date with nighttime and daytime heat wave activity are, respectively, 0.43 and 0.42 (p = 0.065).

Both nighttime and daytime HWAIs evaluated at Napa State Hospital

weather station are correlated with Brown Ranch Chardonnay harvest dates

at above 0.42 (p = 0.065). Fourteen years is a small sample and these

correlations are only significant at the 90% confidence level. Obviously,

many other environmental factors besides heat waves influence harvest

dates, including those influencing earlier phenological milestones (e.g. bud

break and veraison), and factors such as summertime cloudiness and

220

springtime climate, etc. Other grapes may possibly react differently to the

same environmental pressures and other growers may make different

harvesting decisions in the face of the same environmental pressures. Heat

wave impacts differ from one event to another. We recognize that besides

the types of heat waves (humid vs. dry), timing of events can also be very

important vis a vis the grape’s phenology or stages of maturity. We do not

consider this important topic here. With the above caveats, heat waves do

appear to make a difference, but why towards later harvests? Moreover and

interestingly, average summertime Tmax and Tmin are uncorrelated with

harvest dates of Brown Ranch Chardonnay.

4. Discussion and conclusions

California statewide harvest records indicate that heat waves often

associate with lower-yield harvests. Furthermore, a record of Chardonnay

harvest timing from the Carneros region of Napa Valley reveals that harvest

dates associate in a seemingly counterintuitive manner—heat waves may

actually provoke later harvests.

Although excessive heat builds up the grape’s sugar level necessitating

earlier harvests in traditional winemaking societies and making climate

reconstruction from harvest date records possible (e.g. Le Roy Ladurie

1967), new-world practices are less strict allowing irrigation and other

practices considered anathema in traditional wine growing and making.

Increased irrigation, of course, mitigates growing stress due to excessive

heat. Among other new-world practices is the addition of water at grape

crush to dilute the grapes’ sugar content. The only characteristic that cannot

be easily manipulated is character. Character is a function of the grape’s

221

maturity, and maturation slows down with slower metabolism. Metabolism

is quick under optimal environmental conditions that include warm dry days

and cool nights. Humidity acts to slow down metabolic processes in grapes

at least (without considering direct biochemical influences) by

disproportionately increasing nighttime temperature when daytime

temperature is already at detrimentally high levels. Under such extreme

conditions, new-world wine growers and makers, preferring to enhance the

grapes’ maturity and the character of the vintage, apparently choose to

harvest later even at the risk of loosing a significant portion of the crop to

heat stress. We do not know how prevalent this practice is, but certainly it is

more frequent in California than in France, Spain, Italy, and other

winemaking societies where the process is more strictly controlled by laws

steeped in tradition. It is possible, therefore, that the effect of heat waves on

the quantity of the harvest (Figure 6) is greater in new-world winegrowing

regions, while their effect on the quality is less.

This fine point is somewhat speculative and warrants further research,

perhaps as much with the palate as with less subjective statistical tools.

Statistics should also be done in a way that is more sensitive to the complex

geography of California. Heat waves can have different effects on grapes in

the normally hot Central Valley where most wine grapes are grown,

compared to the cooler coastal valleys of northern and central California –

regions growing the higher quality grapes. Effects on different grape

varieties should be considered separately. Furthermore, we do not

adequately know how marine influences, including low-level stratus clouds,

react to inland heat waves and modulate coastal temperatures. A more

detailed study is certainly warranted focusing on California and northern

Baja California including the Guadalupe Valley, where the impacts of humid

222

heat waves are expected to intensify in the future. It would also be extremely

useful to compare similar results obtained in California and other

winemaking regions, especially those immersed in “old world” winemaking

traditions. What’s more, because humid heat waves in California appear to

be associated with ocean warming, particularly in the California current

outflow region west of Baja California (GCI09), future enjoyment of

respectable quantities of quality California wines necessitates further

oceanographic exploration off the cost of Baja California.

Acknowledgments

Following a presentation given by Gershunov at CICESE in April 2008

on heat waves and their oceanic influences, Antoine Badan articulated

concerns about impacts of humid heat waves on wine grapes and later

expounded upon these in a fascinating and gracious chat over a bottle of his

exquisite Chasselas. That chat provided inspiration for this work. We thank

David Graves for making the Chardonnay phenological data from

Saintsbury’s Brown Ranch available to us, Kimberly Nicholas Cahill for

learned discussion, and Mary Tyree for data handling. We are also grateful

to Hugo D’Acosta and several other winemakers in the Guadalupe Valley of

Baja California for fruitful and tasteful discussions.

223

References

CDFA, 2009: Grape crush report, preliminary 2008 crop. California

Department of Food and Agriculture, Sacramento, CA, 9pp, [Available

online at

http://www.nass.usda.gov/Statistics_by_State/California/Publications/Grape

_Crush/Prelim/2008/200802gcbnarr.pdf].

Clement, A.C., R. Burgman, and J.R. Norris, 2009: Observational and

model evidence for positive low-level cloud feedback. Science, 325, 460-

464.

Easterling, D. R., and Coauthors, 1997: Maximum and minimum

temperature trends for the globe. Science, 277, 364–367.

Gershunov, A., D. Cayan and S. Iacobellis, 2009: The great 2006 heat

wave over California and Nevada: Signal of an increasing trend. Journal of

Climate, 22, 6181–6203.

Le Roy Ladurie, E., 1967: Histoire du climat depuis l’an mil. Flammarion,

287 pp.

NCDC, 2003: Data documentation for data set 3200 (DSI-3200): Surface

land daily cooperative summary of the day. National Climatic Data Center,

Asheville, NC, 36 pp. [Available online at

http://www.ncdc.noaa.gov/pub/data/documentlibrary/tddoc/td3200.pdf].

USDA, 2008: California grape crush report.

http://www.nass.usda.gov/Statistics_by_State/California/Publications/Grape

_Crush/Prelim/2008/200802gcbnarr.pdf.

224

225

Chemistry, Materials and Equipment in Wineries:

An Overview on the Guadalupe Valley*♦ Michael Schorr ♠, II-UABC Benjamín Valdez Salas, II-UABC Mónica Carrillo, II-UABC Blanca M. Arellano García, FC-UABC Alejandro Martínez-Ruiz, FC-UABC

1. Introduction

Human life and health depend upon the balanced nutrition of the body and

the soul. The two oldest nutritive beverages known to mankind are milk and

wine, both symbolic elements in many religious rituals. Milk for

nourishment, wine for the joy of the heart: “ יין ישמח לבב אנוש ”, “wine to

gladden man's heart” ( Psalms 104-15 ).

Wine is as old as human history, being referred to in the earliest literature

and religious writings. In the Old and New Testaments of the Bible, wine is

frequently mentioned as a drink, a blessing and a medicine.

* The Ocean, The Wine, and The Valley: The Lives of Antoine Badan, 225-232.

© 2010 CICESE. ♦ This article has not been peer-reviewed. ♠ Contact: mschorr2000@yahoo.com

226

2. Vines and grapes

The first human being to plant a vineyard was Noah, upon descending

from his ark after the deluge and, in the laconic biblical style, the writer lets

us know that he drank his wine and became drunk. In one sentence:

" נח איש האדמה ויטע כרם וישת מן היין וישכר" (“Noah began the planting of

vineyards. He drank some wine and became drunk”) (Genesis, 9:20,21 ).

The stems of the vine (Vitis vinifera) called sarments bear the leaves and

flowers and, later on, the sweet grapes. The sarments require external

support for maintaining a healthy, productive vineyard. The vines are

pruned to increase yield and improve quality.

Propagation is achieved mainly by means of stem cuttings. This method

has been practiced in Europe for centuries and was brought to California and

to the Guadalupe Valley by missionaries and colonists. The appliances

needed for the vintage are shears, trays and baskets. These should be clean,

in order to avert diseases caused by fungi, insects and bacteria.

In modern times viticulture has spread from Europe to temperate areas of

South America, Western USA, Australia, New Zealand and Mexico.

3. Wine production

Wine is produced by fermentation. In this process, sugar in ripe grapes

is converted into ethanol (ethyl alcohol), a bio-catalytic process promoted by

enzymes from yeast:

C6 H12 O6 → CH3 CH2 OH + CO2

(Sugar) (Anaerobic fermentation) (Ethanol) (Carbon dioxide)

227

In a subsequent stage, wine may be converted into vinegar (a dilute

solution of acetic acid) by another bio-catalytic process, this one

accomplished by the aerobic bacterium Acetobacter aceti. During this

process, the previously formed ethanol is oxidized by oxygen from the air:

CH3 CH2 OH + O2 → CH3 COOH + H2 O

(Ethanol) (Oxygen) (Aerobic fermentation) (Acetic acid) (Water)

It is called wine vinegar; it adds an acidic taste to green salad.

Wine production is one of the oldest industries established by mankind.

In the ancient cultures of the Middle East, such as Egypt, Phoenicia, Israel,

Greece and Rome, vines, vineyards and wine were part of everyday life.

Moreover, wine became associated with art, literature, religion, and even

magic.

4. Wine quality

Physical and chemical analyses are applied to ensure the quality of wine,

as well as the conservation of its organoleptic characteristics. Human

sensory examination by wine experts and testers detect the subtle differences

in wines, their origin, age, sugar content and so on. The vintage year, its

weather and the grape’s harvest are dominant factors in wine quality and

taste. In many countries, “new” wines are pleasant, easy to drink and are

meant for early consumption at modest prices. Modern technologies have

nowadays been introduced into wine production. These are used, for

example, for safe wine transportation, while microchips embedded in the

crates monitor temperature and track location. Wineries, winemakers and

228

wine inventories utilize software for management, production and marketing

to assure high efficiency, cost control and profit on a global scale.

An instrumental method to differentiate and classify wines, and investigate

their organoleptic characteristics has been recently developed. It is based on

voltammetry and spectroscopic techniques, and predicts the relation between

pigments and antioxidant species in wine. A similar method is applied to

determine the oxidative-reductive properties of water (ASTM D 1498.)

6. Corrosion control

As part of modern science and technology, an extended range of

engineering materials have been developed and are presently used in the

wine industry improving the equipment and facilities. They encompass

metals, alloys, plastics, glasses and composites. Among their features are

increased corrosion resistance and improved mechanical strength. Wines and

distilled alcoholic beverages are aqueous solutions of neutral pH, containing

carbohydrates and small amounts of minerals. Wine production requires

large quantities of water, which are used for cleaning, bottling and hygiene-

securing procedures. Most of the equipment used in these plants, such as

tanks, fermentation vats, steamers, pasteurizers, refrigerators, blenders,

filters and packaging/bottling machines, is made of authentic stainless steel

UNS S 31600 to avoid corrosion. On the other hand, classical caskets, casks

and barrels of hooped staves are made of special woods, and these impart to

the contained wine its particular body, taste, aroma and color.

Scaling deposited in production vessels is removed with mild acids, such

as phosphoric acid. The use of stainless steel guarantees that no corrosion

products are formed, while following strict sanitation regulations avoids the

229

formation of contaminants. In so doing, the taste and organoleptic

characteristics of the wine are not affected.

6. Overview on the Guadalupe Valley

Before the arrival of the Spaniards to Mexico, the indigenous peoples of

the Baja California Peninsula prepared beverages from local wild grapes

species, such as: Vitis rupestris, Vitis labrusca, and Vitis barlandieri

sweetened by adding honey. These varieties were not the most appropriate

for making wine; nevertheless, these drinks were the best means to calm the

thirst of the ancient inhabitants.

The discovery and the colonization of the peninsula by Spain introduced

winemaking from the variety Vitis vinifera. These grapes and wines were

utilized as food, and also for their religious rituals in church.

Viticulture was started in Baja California by Jesuits, Franciscan and

Dominican monks, who built the misiones (missions) along the peninsula in

the XVII and XVIII centuries with the help of the local native tribes. This

civilizing enterprise transformed parts of the semiarid desert region with a

harsh climate, into productive green gardens, by way of planting fruit trees,

vines and olives. The main growth and industry during those centuries were

due to the vineyards and the wineries for production of all types of wines.

The grapes and vines that the misiones builders brought to the regions of

“New Spain” and South America was the variety called “Misión” derived

from the Spanish species called “Mónica”. It is a vine resistant to plant

plagues and hard climates; it is cultivated up to this day, with vines 100

years old, in the Santo Tomas region, to the south of Ensenada City.

230

The world wine belts are localized between 30° and 50° degrees of

latitude, in each of Earth’s hemispheres, where the best climate conditions

for the vine cultivation are to be found. The State of B.C. is located in the

northern hemisphere belt; thus, it enjoys a warm climate during the summer

and is cold during the winter. The vineyards are located in a region between

the Pacific Ocean and the California Desert; this particular situation

generates geographic and climate conditions favorable for the production of

high quality wines.

Since 1948, the National Association of Wineries have reinforced the

presence and operation of the 29 wineries displaying advanced and modern

technologies for production of different types of wines with a wide range of

alcohol content. Owing to this progress, Mexico was nominated head office

for the Annual Convention of the International Office of Wine promoting the

Mexican wine industry.

The Ensenada area, the land of wine, operates several regions namely: the

valleys of Santo Tomas, Guadalupe, San Antonio de las Minas, San Vicente,

Tecate and Ojos Negros, due to the high influence of the local

Mediterranean climate and large production of a vast range of wines:

Tempranillo, Cabernet, Merlot, Grenache, Carignan, Petit Syrah, Nebbiolo

and Sangiovese, Malbec, Pinot Noir among others.

During frequent visits by local and foreign tourists, they are surprised by

the devotion and passion of many scientists, technicians, growers, producers,

and tasters involved in this science, technology and industry who aim to

produce the best wines for the joy of humankind. Numerous enologists,

some of them originally from other countries, have successfully added their

experience to improve and diversify the wines made with the grapes of our

Baja California grounds. Such are the cases of the Argentinian Sebastián

231

Suárez of Domeq industry, the Chilean José Luis Durand of Wines and

Terruños and Sinergi-VT, the Swiss Christoph Gaertner of Vinisterra, the

Italian Camilo Magoni of LA Cetto, who joined an already excellent group

of local enologists, among them, Hugo D’Acosta of Casa de Piedra, Hans

Backoff of Monte Xanic, Fernando Martain of Cavas Valmar, Víctor Torres

Alegre of Chateau Camou, Baron Balché and La Llave, Laura Zamora of

Santo Tomás and last but not least Antonio Badán Dangón of The Mogor.

There is no doubt that this noble human, agricultural and industrial

activity, one of the oldest in the world, will continue to prosper for the

benefit of migrants, producers, workers, entrepreneurs, and all others

devoted to this spiritual-material enterprise of making wine from grapes.

Acknowledgements

The authors are grateful to the wineries of the Guadalupe Valley for

providing technical information on wine production.

232

References

Australian stainless steel development, 2009: En: www.assda.au.

ASTM Standard D1498-08. Test method for oxidation-reduction potential of

water, 2009: En: www.astm.org.

Escuela de Enología y Gastronomía, UABC, 2009: XVII Concurso

Internacional “Ensenada, Tierra del Vino”.

Etchairen, E. R., Editor, 2008: Guía de Viñedos de México, Editorial Sabor.

Magoni, C. 2009: Historia de la Vid y el Vino, Vinicola L. A. Cetto.

Martínez, A., G. Guzmán, B. Valdez and M. Schoor, 2010: Voltammetric

studies of Baja California red wines. In this book.

Stainless steel for food and beverages, 2009: En:

www.hygienics.stainless.org.

Valdez, B. and M. Schorr, M., 2004: Stainless steel for corrosion control in

the food industry, Stainless Steel World.

233

Voltammetric Studies of Baja California Red

Wines*♦

Alejandro Martínez-Ruiz ♠, FC-UABC Gabriela Guzmán Navarro, FC-UABC Benjamín Valdez-Salas, II-UABC Michael Schorr, II-UABC Abstract

The present work examines the feasibility of developing a qualitative model

based on cheap and simple instrumentation to differentiate and classify red

wines. Voltammetry and spectroscopy were used to predict simultaneously

the relation between pigments and antioxidant species in red wine. Four

commercial red wines from the L. A. Cetto Baja California winery were

used for this study. A correlation between maximum wavelength λmax and

potential oxidation facilitated wine analysis. It was concluded that

spectroscopy and linear voltammetry may be used as a rapid alternative

method to predict the concentration of active antioxidant compounds in red

wines.

* The Ocean, The Wine, and The Valley: The Lives of Antoine Badan, 233-241.

© 2010 CICESE. ♦ This article has not been peer-reviewed. ♠ Contact: alejandro@uabc.mx

234

1. Introduction

It is well known that red wine is a complex medium containing a huge

quantity of substances able to undergo a large number of changes along the

period of oxidative ageing in oak wood. Antioxidants found in wine are

beneficial for the human body, because they neutralize the free radicals by

providing an electron. Antioxidants are desirable molecules because the free

radicals hurt our cells by stripping them of electrons and transforming

damaged cells into cancerous cells. In this work we attempted to distinguish

the red wines of Baja California by measuring these substances.

The physicochemical processes that occur during the ageing in barrels are

directly responsible for most of the final organoleptic characteristics of

wines, such as, for example, the color. Most studies examining the influence

of oak wood on the chemical and sensory characteristics of wines are related

to the analysis of particular families of compounds such as biogenic amines,

mono-saccharides, phenolic derivatives, and other molecules which confer

specific effects to the organoleptic and physicochemical characteristics of

red wines. Despite the usefulness of these analyses, they do not consider the

global nature, harmony and chemical macro-equilibrium of wine.

In recent years, scientists have been developing an arrangement of

electrochemical sensors for the characterization of complex liquids. In these

studies, potentiometric and voltammetric sensors are used to discriminate

liquids like mineral water, juice or wines. Electronic tongues and noses

usually consist of a few tens of sensors. A method for reducing the number

of sensors in these arrays has not been provided so far.

The aim of this work is to describe a procedure to reduce the number of

sensors. We propose that red wines may be discriminated by spectroscopy

235

data and voltammetry curves that contain information about the redox

processes occurring in the solution when a voltage is applied.

2. Experiments

Electrochemical measurements were performed in a conventional three-

separate compartment electrochemical cell using graphite for the working

electrodes. Platinum wire was used as a counter electrode and served as

reference electrode (Ag/AgCl). All potential values reported in this paper are

compared vs. the (Ag/AgCl) electrode. Base solutions were 50 ml wine and

50 ml 0.1 M KCl.

3. Results

The ions interfere in the way of the particles of our interest, which are of

the order of armstrongs, to interact with the electrode. These interactions are

registered in Figure 1, which shows potential vs. current density as “peaks”.

According to the literature, these peaks, depending on the interval of

potential, indicate the type of compound it can be (tocopherols or

polyphenols), the amount, and the density.

The potential peaks observed in Fig. 1 also depend on the diffusion

velocity of ions that are moving through the mix and arrive at the electrodes.

This can be seen by changing the sweep velocity and comparing the graphs

of a single wine. It is easier to detect if we draw a graph of density current

vs. v-1/2 (the velocities were 100 mV/s, 75 mV/s, 50 mV/s, 25 mV/s, 15 mV/s

and 10 mV/s). It shows an exponential behavior.

236

In Figure 3, the positions of peaks vary according to the sweep velocity

because if the velocity is high in a wine with low particle concentration,

these particles become ionized for a short period of time and, due to the

Brownian movement, they will not be able to reach the electrode and interact

with other electrons. Thus, a slower velocity would provide more

probabilities for the electrons of the substance and the electrode to interact.

According with these results, the detection of the particles depends on the

scan rate of the instrument, the “concentration/adsorption” and the diffusion

of active particles.

We also expected to see differences between the wines by comparing the

respective oxidation potentials shown in Figure 5.

The spectroscopy study showed that these wines absorb in a λ of 450 to

550 nm, as shown in Figure 6. Maximum of absorbance is also shown.

4. Conclusions

Four commercial red wines from the L.A. Cetto Baja California winery

were utilized in this study. A correlation between λmax and potential

oxidation facilitated wine analysis. It was concluded that spectroscopy and

linear voltammetry may be used as a rapid alternative method to predict the

concentration of active antioxidant compounds in red wines.

Acknowledgments

L. A. Cetto winery is gratefully acknowledged for providing samples of

red wines.

237

Figure 1. Linear voltammetry of red wines.

0.0 0.2 0.4 0.6 0.8 1.0 1.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

Cur

rent

Den

sity

(J)

Potential (V) vs Ag/AgCl

BARBERA

Figure 2. Linear voltammetry of Barbera red wine.

0.0 0.2 0.4 0.6 0.8 1.0 1.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Cabernet Nebbiolo Barbera Petit Verdot

Cur

rent

Den

sity

(J)

Potential (V) vs Ag/AgCl

23

8

Figu

re 3

. Eac

h w

ine

pres

ente

d no

t onl

y di

ffus

ion

cont

rol d

ue to

the

“con

cent

ratio

n/ad

sorp

tion”

of t

he a

ctiv

e

parti

cles

in th

e w

ine.

We

can

asso

ciat

e th

is b

ehav

ior t

o th

e bo

dy o

f the

win

e (c

ollo

idal

inte

ract

ions

).

0.10

0.15

0.20

0.25

0.30

0.35

-1.2

-1.0

-0.8

-0.6

-0.4

-0.2

Current Density (J)

υ-1

/2

Bar

bera

0.10

0.15

0.20

0.25

0.30

0.35

-1.2

-1.0

-0.8

-0.6

-0.4

-0.2

Current Density (J)

υ-1

/2

Neb

bio

lo

0.1

20

.16

0.2

00

.24

0.2

80

.32

-1.1

-1.0

-0.9

-0.8

-0.7

-0.6

-0.5

-0.4

-0.3

Current Density (J)υ

-1/2

Pe

tit V

erd

ot

0.12

0.16

0.20

0.24

0.28

0.32

-0.7

-0.6

-0.5

-0.4

-0.3

-0.2

Current Density (J)

υ-1

/2

Ca

ber

net

239

0.0 0.2 0.4 0.6 0.8 1.0 1.2

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

-J [m

A/c

m2 ]

V vs Ag/AgCl

Sweep velocities (mV/s)

100 75 50 25 15 10 15

Barbera

Figure 4. Cyclic voltammetry for several scan rates.

Neb Cab Pv Bar

0.46

0.48

0.50

0.52

0.54

0.56

0.58

0.60

Pot

entia

l (V

) vs

Ag/

AgC

l

Wines

Figure 5. Maximum oxidation potential for polyphenols.

240

900 800 700 600 500 400 3000.0

0.4

0.8

1.2

1.6

2.0

2.4

2.8

3.2

Petit Verdot Barbera Cabernet Nebbiolo

Abs

orba

nce

λ(nm)

Figure 6. Visible spectroscopy study of red wines.

241

Figure 7. Maximum wavelength λmax according to the

maximum absorbance of each wine.

242

243

Chapter 4

The Valley

244

245

Preliminary Normalized Difference Vegetation

Index for the Guadalupe Valley, Baja California,

Mexico*♦

Ignacio Galindo ♠, CUICA U de C Julián Barrón, CUICA U de C

Abstract

A first attempt is presented to establish the characteristics and variations of

the main photosynthetic indicators of the vineyard cover along the year at

the Valle the Guadalupe. Landsat 5 TM satellite images were used to

determine the Normalized Difference Vegetation Index (NDVI). These

results were then related with global solar radiation, precipitation and

temperature. Obtained data show that the Valle de Guadalupe has

Mediterranean-type climate, i.e., the highest NDVI values are found during

the rainy winter season. Although solar radiation incidence is minimal in

January, the NDVI values are high because precipitation occurs during 6.5

days/year with an average of 56 mm. The month of March with the second

highest values of solar radiation of the year and the highest precipitation

levels, offers the best conditions for vineyard photosynthesis, as it presents

the highest NDVI values. August presents the maximum incidence of solar

radiation but the amount of precipitation falls to around 10% compared with

* The Ocean, The Wine, and The Valley: The Lives of Antoine Badan, 245-262 © 2010 CICESE. ♦ This article has not been peer-reviewed. ♠ Contact: igalindo@ucol.mx

246

March. This reduction is clearly reflected in the NDVI values. October

shows the minimum NDVI values although incoming solar radiation holds

the third place along the year, and precipitation is low although higher than

in August.

1. Introduction

The Guadalupe Valley is located in the state of Baja California at the

extreme northwest of the Mexican Republic, and is part of the municipality

of Ensenada. Its coordinates are 31° 57’ 34’’ to 32° 08’ 57’’ N and 116° 28’

05’’ to 116° 42’ 14’’ W. The importance of this small valley with surface of

11,840 ha, is that it produces the best wines in Mexico. See Figure 1.

To determine the photosynthetic activity of the vegetation cover at any

location, the temporal variations of the main climatic elements need to be

initially investigated. Fortunately, we found two meteorological stations of

the National Mexican Meteorological Service (SMN) with 30-yr records.

These stations are Agua Caliente and Olivares Mexicanos, located very close

to the Guadalupe Valley. Figure 2 and 3 show the 30-yr average of the mean

monthly air temperature and precipitation, respectively. The “dry” summer

and “rainy” winter are immediately apparent. Both meteorological stations

indicate this seasonal variation: the period from November to April marks

the wet winter while the months from May to October correspond to the dry

summer. This is a clear indication of a Mediterranean climate, which has

historically been known in Europe to be very favorable for wine production.

Thus, we considered an analysis of the vegetation cover of the Guadalupe

Valley a helpful tool.

247

Figure 1. Localization of the Guadalupe Valley in Ensenada, Baja California, Mexico, ( INEGI).

248

Figure 2. Mean monthly 30 years average temperature and precipitation (SMN).

Figure 3. Mean monthly 30 years average temperature and precipitation (SMN).

249

We studied some vegetation cover characteristics of the Guadalupe

Valley during the dry and wet seasons in relation with vineyard cover. The

method uses vegetation indices derived from the analysis of Landsat 5 TM

data. The study area is found within the physiographic province of Baja

California Norte inside the sub province of Sierras de Baja California Norte.

Around of the Guadalupe Valley there is A variety of low height sierras and

several hills surround the Guadalupe Valley, which provide water to the

aquifer. Their main water sources are the Sierra Juárez and the Guadalupe

River (T. Kretzschmar, personal communication)

2. Satellite Imagery

Remote satellite-sensed data were recorded in raster format. The U.S.

Landsat satellites have produced the most widely used imagery worldwide

since 1972. Landsat 1, 2, and 3 obtained images by the Multispectral

Scanner (MSS) with a spatial resolution of 79 meters. When launched in

1982, Landsat carried a new sensor, the Thematic Mapper (TM) scanner.

TM was a significant improvement over MSS, providing a spatial resolution

of 30 meters. A second TM was launched aboard Landsat 5 in 1984, and is

presently still in operation. Its orbital altitude is 705 km. Each 16 days it

generates an image of the same point. On board the satellite, the Thematic

Mapper (TM) sensor detects the Earth’s reflected radiation in seven spectral

channels ranging from visible to mid-infrared and thermal infrared (Lowell

and Jaton, 1999). The spectral wavelength interval of each band and its

spatial resolution are shown in Table 1.

250

Table 1. Spectral channels of Landsat TM. From Soria, et al. (1998)

Channel Spectral Interval (λ) Spatial Resolution (m)

Band 1 Visible (0.45 – 0.52 µm) 30

Band 2 Visible (0.52 – 0.60 µm) 30

Band 3 Visible (0.63 – 0.69 µm) 30

Band 4 Near-Infrared (0.76 – 0.90

µm)

30

Band 5 Near-Infrared (1.55 – 1.75

µm)

30

Band 6 Thermal (10.40 – 12.50 µm) 120

Band 7 Mid-Infrared (2.08 – 2.35

µm)

30

The Landsat series were designed to study land areas and the Landsat TM

has a thermal infrared channel with very fine spatial resolution. Therefore,

the idea of studying vegetation indices was originally developed using the

Landsat satellite series. It has been extensively used to study vegetation

changes and the land culture development as pointed out by Márquez

(1998). This author gives examples of how this satellite is able to detect and

monitor the different types of land resources, e.g., band 2 detects green

reflectance from healthy vegetation, while band 3 detects chlorophyll

absorption and band 4 is useful for detection of vegetation health and water-

land interfaces.

In the last few years a large number of studies have been done to detect

different properties of the vegetation cover, which have used vegetation

251

indices generated mainly by a combination of the bands 3, 4, and, 5 from the

TM sensor (Lillesand & Kiefer, 1999, Márquez, 1998, Soria., et al. 1998).

The combination of bands 4 and 5 define the so-called Normalized

Differential Index of Vegetation (NDVI) (see Equation 1) which monitors

photosynthetic activity. It ranges from -1.0 ≤ NDVI ≤ 1.0, i.e., it increases

with rising chlorophyll-production of the vegetation. Low index values

therefore indicate reduced production. Another interesting property of the

NDVI is that it discriminates between soil and vegetation: as the index

increases it indicates a stronger vegetation cover; low contrast values

indicate illness or old vegetation. Thus, we calculated NDVI for the

Guadalupe Valley in this work as follows:NDVI was computed from the

following equation:

Bnir BredNDVIBnir Bred

−=

+ (1)

3. Methodology

3.a NDVI analysis

To analyze the NDVI for the Guadalupe Valley region, four Landsat 5

Row Paths 039038 were acquired with six multispectral channels of 30 m,

plus one thermal infrared image of 60 m spatial resolution. See Table 2.

252

Table 2. Processed images used to calculate NDVI

Row Path Date

039038 January 12, 2009

039038 March 1st, 2009

039038 August 8, 2009

039038 October 24, 2008

The images are provided with radiometric calibration and are also geo-

referenced.

Image preparation consisted in joining each of the channels in order to

generate a unique multispectral archive (layer stack). Since the images show

a large area (170 x 185 km2) with respect to the study region it was

necessary to concentrate in an area of 58 x 48 km2 within the coordinates 31°

51’ 08.7’’-32° 17’ 06.2’’ N and 116° 18’ 58’’- 116° 56’ 06.6’’ W. Once the

area was defined, the NDVI was calculated.

The NDVI was processed using the specialized digital treatment of

satellite images from Erdas Imagine Version 9.2. The images described in

Table 2 were used for a preliminary analysis of the vegetation cover in

different months, which in turn informs about the photosynthetic activity of

the Guadalupe Valley vineyard under different environmental conditions

along the year.

253

4. Results

4.a Combination of satellite images

Figure 4 shows an image for each season obtained by combining the 3-2-

1 visible bands from the LandSat 5 TM sensor to illustrate the different

humidity regimes of the year.

Figures 4A and 4B are representative of the wet season. Vineyards appear in

colors ranging from dark green to clear green, while areas without

vegetation show gray colors.

Figures 4C and 4D are representative of the dry season at Guadalupe

Valley. Vineyards show light to dark brown colors, zones without vegetation

appear in clear gray tones while the characteristic thick bramble-bushes of

the area, mainly composed of xenophile briers, show dark brown and dark

gray colors.

254

4.b NDVI results

Figure 5 shows NDVI images for each of the selected satellite passes in

false colors where clearer colors represent values nearer to 1.0; i.e., greater

vegetation activity, where darker colors represent higher vegetation activity

values, nearer to 1.0, i.e. greater photosynthetic activity.

The NDVI values in Figure 5A for January 12, 2009shows large green

areas and high values of photosynthetic activity. Table 3 compares vineyard

NDVI with irrigated agriculture NDVI on the same date. These measures

reflect the high photosynthesis level of vineyards during this time of the year

due to the characteristic humidity regime of this season. Note the low NDVI

values for temporal agriculture.

Table 3. NDVI values for the image dated January 12, 2009.

Vegetation cover NDVI

Vineyards 0.35 – 1.00

Irrigated agriculture land 0.35 – 1.00

Temporal agriculture land 0.10 – 0.35

Scarce vegetation cover or bare

soils

-1.00 – 0.06

255

Figure 4. Combination of the bands 3-2-1 from the TM sensor for Guadalupe Valley .

256

Figure 5. Guadalupe Valley, NDVI images obtained from the combination of bands 3 y 4 from the Landsat 5 TM.

Figure 5B shows the NDVI values for March 1st, 2009. NDVI values are

higher than in the previous image; there is also a clear increase in the NDVI

values of areas with temporal agriculture, i.e., cultivated land shows high

photosynthetic activity, similar to irrigated agriculture and vineyards making

257

it difficult to discriminate between these areas, as shown in Table 4.

However, some temporal agriculture areas show a fall in photosynthetic

activity, probably because crops are reaching the end of their physiological

or cultivation cycle.

Table 4. NDVI values for the image dated March 1st, 2009.

Vegetation cover NDVI

Vineyards 0.40 – 1.00

Irrigated agriculture land 0.40 – 1.00

Temporal agriculture land with high photosynthetic activity

0.40 – 1.00

Temporal agriculture with low photosynthetic activity

0.10 – 0.40

Scarce vegetation cover or bare soils

-1.00 – 0.10

Figure 5C shows the NDVI values for the dry season with high incoming

solar radiation. Small areas are visible at the center and to the northeast of

the Guadalupe Valley with high photosynthetic activity characteristic of

irrigated agriculture land. With lighter red colors are the extensive vineyards

while areas with lower photosynthetic activity can be recognized by the red

colors of less intensity, and, finally, the lighter red to dark pink colors

represent zones without vegetation.

Table 5 shows the NDVI values for the dry summer season. Two types of

agriculture are displayed, the minimum value for irrigated agriculture

during this season is 0.3, while for vineyards, the NDVI oscillates between

0.12 and 0.3 and for bare soils or soil without scarce vegetation it falls below

0.12.

258

Table 5. NDVI values for the image dated August 8, 2009.

Vegetation cover NDVI

Vineyards 0.12 – 0.30

Irrigated agriculture land 0.30 – 1.00

Scarce vegetation cover or bare

soils

-1.00 – 0.12

Figure 5D shows the NDVI values determined for October 24, 2008.

Although this image is similar to Figure 5C, it may nevertheless be

appreciated that some areas show an increase in photosynthetic activity

mainly in the north and northeast regions of Guadalupe Valley where the

highest NDVI values were obtained.

Table 6 shows the NDVI values registered for October 24, 2008. For the

vineyards, NDVI oscillated between 0.06 and 0.25, the irrigated agriculture

land showed values ranging from 0.25 to 1.00, and uncovered areas varied

between -1.00 and 0.06.

Table 6. NDVI values for the image dated October 24, 2008.

Vegetation cover NDVI

Vineyards 0.06 – 0.25

Irrigated agriculture land 0.25 – 1.00

Scarce vegetation cover or bare

soils

-1.00 – 0.06

259

Finally, a summary of the main photosynthetic indicators of the

Guadalupe Valley vineyards along the year is shown in Table 7. The highest

NDVI values occur during the rainy winter season. Although overall solar

radiation incidence is minimal during January, the second maximum of

annual precipitation is registered during this month with more than 6.0 days

per month. March registers the second maximum of solar radiation during

the year and highest precipitation values and presents the best conditions for

vineyard photosynthesis. During August, the maximum annual incidence of

solar radiation occurs, however, precipitation is reduced to around 10%

compared with March. The reduction in photosynthetic activity is clearly

reflected in the NDVI values. October shows the lowest NDVI values

despite incoming solar radiation, which ranks third place; precipitation is

still low, although higher than in August.

26

0

Tabl

e 7.

Gua

dalu

pe V

alle

y vi

neya

rd p

hoto

synt

hetic

act

ivity

indi

cato

rs d

urin

g th

e ye

ar.

ND

VI

Ove

rall

Sola

r

Rad

iatio

n*

Agu

acal

ient

e O

livar

es M

exic

anos

Mon

th

-1.0

≤ N

DV

I ≤1.

0(M

J/m

2 ) T

(° C

) P

(mm

) D

ays w

ith

prec

ipita

tion

T (°

C)

P (mm

)

Day

s with

prec

ipita

tion

Janu

ary

0.35

- 1.

00

12

12.4

55

.6

6.5

1.18

56

.6

6.0

Mar

ch

0.40

- 1.

00

20

13.7

59

.8

6.6

12.7

64

.6

6.1

Aug

ust

0.12

-0.3

0 21

25

.2

6.8

0.9

22.7

3.

7 0.

4

Oct

ober

0.

06 -0

.25

17

19.7

12

.7

1.8

18.5

19

.4

3.3

*Gal

indo

I.,

y V

aldé

s, M

., 19

92:

Méx

ico-

Atla

s de

Rad

iaci

ón S

olar

, Pr

ogra

ma

Uni

vers

itario

de

Ener

gía,

C

oord

inac

ión

de la

Inve

stig

ació

n C

ient

ífica

, Uni

vers

idad

Nac

iona

l Aut

ónom

a de

Méx

ico.

(Sol

ar R

adia

tion

Atla

s, U

nive

rsity

En

ergy

Pro

gram

, Sci

entif

ic R

esea

rch

Coo

rdin

atio

n, N

atio

nal A

uton

omou

s Uni

vers

ity o

f Mex

ico.

)

261

5. Conclusions and Recommendations

The present study confirms that the Guadalupe Valley has Mediterranean

climate with rainy winters and dry summers. The highest NDVI values

reached for vineyards together with solar radiation incidence and high

precipitation are present in March while the minimum NDVI values occur in

October during the dry season when precipitation is almost nill, in spite of

the high incidence values of solar radiation. Further studies extending to the

whole year are advisable.

Acknowledgments

The authors are in debt with Miss Cristina Rivera-Godinez for her

invaluable help during manuscript preparation.

262

References

Galindo I., y Valdés, M., 1992: México- Atlas de Radiación Solar.

Universidad Nacional Autónoma de México. Mexico City, Mexico.

Lillesand, Th. M. and Kiefer, R.W., 1999: Remote Sensing and Image

Interpretation. John Wiley & Sons, Inc., New Jersey, USA.

Lowell, K. and Jaton A., 1999: Spatial Accuracy Assessment. Ann Arbor

Press. Chelsea, Michigan, USA.

Márquez L. M. A., 1998: Cartografía de Asociaciones Arbóreas Mediante

Imágenes de Satélite LandSat TM y Modelos de Relación Asociación-

Topografía. Masters thesis. Universidad Autónoma Chapingo. Chapingo,

Mexico.

Ponce C. J. E., 1998: Clasificación de Uso de Suelo y Vegetación en un Área

del Estado de Tabasco Mediante Imágenes de Satélite. Masters thesis.

Colegio de Postgraduados. Chapingo, Mexico.

Soria R. J., et al., 1998: Sensores Remotos, Principios y Aplicaciones en la

Evaluación de los Recursos Naturales, Experiencias en México. First

edition. Sociedad Mexicana de la Ciencia del Suelo. Chapingo, Mexico.

263

Does wet Sand Evaporate more than Water in the

Guadalupe Valley?*

Edgar G. Pavia ♠, CICESE Ismael Velázquez, CICESE

Abstract

We compare evaporation rates from wet sand and water estimated

experimentally under the same varying ambient conditions throughout the

day using 10 kg-microlysimeters. Our daily-total estimates in turn compare

well with independent estimates at nearby climatological stations. Our

experiments were carried out during 2009 and 2010 at 31° 52' 09" N, 116°

39' 52" W. Wet sand with more than 5% of moisture content on a wet-mass

basis evaporated at a greater rate than open water most of the time during

clear daytimes. At night evaporation was null and occasionally even a small

condensation was produced. Drying sand (~5% of moisture content on a

wet-mass basis) during bright days exhibited almost typical evaporation

rates in the morning hours before decreasing to nearly zero values, while

simultaneously open water evaporation rates continued to increase. That is:

a complementary relationship (Bouchet 1963) behavior, despite our

relatively small temporal and spatial scales, which precludes this result from

invoking the alleged physics: surplus evaporative energy going from the

drying sand tray to the water tray. Daily mean evaporation is almost

exclusively a function of daily mean solar radiation; but other

* The Ocean, The Wine, and The Valley: The Lives of Antoine Badan, 263-278

© 2010 CICESE. ♠ Contact: epavia@cicese.mx

264

meteorological variables are important in particular cases. Hopefully these

findings would help to evaluate the impact of sand extraction from the dry

riverbeds of the Guadalupe Valley.

1. Introduction

In this research we estimate simultaneously evaporation rates from thin

layers of water (eW, EW) and wet sand (eS, ES) by the differential-weight

method. The main experiment consists of two bird-guarded shallow trays,

one with a ~35 mm-thick layer of water and one with a ~21 mm-thick layer

of wet sand, each resting on a continuously-recording electronic balance

during several days; where weight-differences are taken as amounts of water

evaporated, and the weights on the balances are sampled at regular intervals

(see Velázquez-Mendoza 2010, for a complete description of all

experiments). A small random noise on the weight data generated by the

wind-effect on the trays is filtered out digitally before evaporation rates are

calculated. The ~0.22 m2 (0.58 m by 0.38 m) empty trays weigh 1.52 Kg

each; initially one contains a sun-protected temperature sensor weighing

0.72 Kg plus ~7.76 Kg of water, and the other 4.93 Kg of air-dry sand that

was spread out uniformly across the tray (with a 0.05 Kg temperature sensor

underneath) plus ~3.5 Kg of water. Thus initially and after refill both

evaporating trays weigh about 10 Kg. Water temperature (Tw), wet-sand

temperature (Ts), air temperature (Ta), solar radiation (R), relative humidity

(H), wind speed (S), precipitation (P) are also recorded at the site during the

experiment. The evaporating trays are re-filled with tap water most days at

13:00 h, local time (20:00 h UTC), to ~10 Kg total weight; but sometimes (if

ambient conditions are expected to be favorable for evaporation) the wet

265

sand is allowed to depart from moist conditions by avoiding refill during a

few days. There was no measurable rain throughout the experiment; i.e. P =

0. Our goal is to compare eW and eS in their evolution throughout the day

every 5 minutes (as sand moisture decreases), and in their total daily values

EW and ES (for comparison to climatological pan evaporation).

The goal of this paper is to contribute to evaluate the effect of sand

extraction from the dry riverbeds of the Guadalupe Valley.

2. Results

We start by sampling the electronic balances output at δt = 5 minute

intervals. This gives two time series of evaporating-tray weights (wi,j) which

are smoothed with a zero-phase, one-hour (12-value) time filter. We then

compute weight-differences (δwi,j = wi,j − wi+1,j) where i is a time index, for

example: i = 1, 2,…, 287 for one day, and j is a tray index, i.e. j = 1 (water),

and 2 (wet sand). Evaporation rates: eW and eS (in mm per δt) are estimated

by dividing the corresponding δwi,j (in Kg per δt) by tray area (in square

meters), assuming that δwi,j represent differences in mass (i.e. constant

gravity), and that one kilogram of tap water (i.e. one liter) is equivalent to

one millimeter (per square meter). Correspondingly robust daily totals (in

mm per day) may be calculated by adding up evaporation rates, for example:

287

,11

W ii

E e=

= ∑ , and 287

,21

S ii

E e=

= ∑ ,

or from daily weight-differences: ∆wi,j = wi,j − wi+287,j, i = 1, 288, 2 × 288, 3

× 288, etc. (per day), divided by tray area. Figure 1 presents the time series

266

of EW and ES; and the time series of eW and eS, showing highest values during

the daytime in the early afternoon, and lowest values during the nighttime.

We can see that most of the days EW < ES but clearly not on 27 April (d27A),

and 4 and 25 May 2009 (d04M and d25M).

Figure 1. Time series of evaporation rates: ES (big red dot) is total evaporation from wet sand in 24 h (scale on the right side); eS (small red dot) is the evaporation from wet sand in 5 minutes (scale on the left side); EW (big blue dot) is total evaporation from water in 24 h (scale on the right side); eW (small blue dot) is the evaporation from water in 5 minutes (scale on the left side).

For these three days eW and eS are considered atypical with respect to

their mean diurnal variation, that is: the mean values at a particular time

during the day throughout the 38-day experiment (see Fig. 2):

267

38

1(( ) ) ,W i W i k

ke e

=

< > = ∑ and 38

1(( ) ) ;S i S i k

ke e

=

< > = ∑ i = 1, 2, 3,..., 287, (1)

Figure 2. Mean diurnal variation of eS (red) and eW (blue).

Figure 3 shows 24-h details of the time series for these three days

centered on 00:00 h, and we can see that most of the time during these

periods eW > eS, contrasting with the full Figure 1 where mostly eW < eS. To

see if during these three days evaporation rates crossed a critical point in

sand dryness we look at the same information, but using now a linear sand

dryness index Dws instead of time.

268

Figure 3. Details of time series corresponding to d27A, d04M and d25M of eW (blue) and eS (red).

With 0 ≤ Dws ≤ 1; i.e. Dws = 0 when the evaporating wet sand tray weighs

10 Kg at refill (with 3.5 Kg of water, equivalent to 71% [100 × (3.5 / 4.93)

~ 71] on a dry-mass basis, or 42% [100 × (3.5 / (4.93+3.5)) ~ 42] on a wet-

mass basis); and Dws = 1 when the evaporating wet sand tray weighs 6.5 Kg

(once its water has completely evaporated). Figure 4 clearly shows that for

the three cases considered the critical point is Dws = 0.92 ± 0.01

corresponding to sand with (5.7 ± 0.7)% water on a dry-mass basis or about

5% on a wet-mass basis.

269

Figure 4. Evaporation rates: eW (blue) and eS (red) as a function of Dws corresponding to d27A, d04M and d25M; arrows point toward the suspected sand dryness threshold.

This is the threshold where wet sand can no longer evaporate as expected

for the prevailing ambient conditions and as evidenced by the anticipated

evaporation of water; i.e. this critical point depends not only on the dryness

of wet sand but also on the ambient conditions. At this time-scale both eW

and eS are highly correlated with R, Ts and Tw, but not so highly correlated

with Ta, D (D = 100 – H), and S (see Table 1).

270

Table 1. Correlation coefficient matrix of main variables involved in this

study.

eS eW Ts Tw R Ta D S

eS 1.00 0.91 0.86 0.86 0.94 0.66 0.52 0.54

eW 0.91 1.00 0.87 0.81 0.93 0.64 0.58 0.60

Ts 0.86 0.87 1.00 0.95 0.93 0.81 0.41 0.44

Tw 0.86 0.81 0.95 1.00 0.86 0.90 0.28 0.40

R 0.94 0.93 0.93 0.86 1.00 0.68 0.54 0.56

Ta 0.66 0.64 0.81 0.90 0.68 1.00 0.15 0.31

D 0.52 0.58 0.41 0.28 0.54 0.15 1.00 0.46

S 0.54 0.60 0.44 0.40 0.56 0.31 0.46 1.00

Moreover these local evaporation rates can be modeled as a sole function

of R, because Ts and Tw are functions of R; and Ta, D, and S do not contribute

significantly to the variance explained by the model. Hence by a systematic

statistical method (Pavia 2008) we can obtain (eW)model ~ f1(R) and (eS)model ~

f2(R), where f1 and f2 are linear functions; and thus R may be seen as a proxy

for model evaporation rates, or as representative of ambient conditions

favorable for evaporation. Figure 5 compares R for the three selected days

with the diurnal variation < R > as defined in equations (1); showing that

d04M and d25M were indeed above-average favorable for evaporation (i.e.

brighter than average), and d25M even also on the preceding evening; while

d27A was favorable only on the preceding evening (see Figure 5).

271

Figure 5. Solar radiation (*) corresponding to d27A, d04M and d25M, and the mean diurnal variation < R > (solid line).

Finally to see how our evaporation data compare to other local estimates

of evaporation we calculate our mean total daily value:

38

1( ) 3.75W W k

kE E

=

= =∑ mm/day,

at our site (31° 52' 09" N, 116° 39' 52" W). This is somewhat lower than the

1971-2000 pan evaporation average (April and May) climatology at the two

closest climatological stations (4.22 mm/day at Ensenada station at 31° 53'

45" N, 115° 35' 50" W; and 3.93 mm/day at El Ciprés station at 31° 47' 25"

N, 116° 35' 17" W). These latter values are calculated by adding the

reported pan evaporation climatologies for April (122.1 at Ensenada, and

112.6 mm at El Ciprés) and May (135.6 at Ensenada, and 127.2 mm at El

Ciprés) and dividing the sum by 61 (see also Velázquez-Mendoza 2010).

272

Neither pan evaporation nor climatological records of ES corresponding to

our experiment period are available for these other sites. Our mean total

daily value for the latter is

38

1( ) 5.26S S k

kE E

=

= =∑ mm/day.

3. Discussion

Our results suggest that although time series of eW and eS (Figure 1) are

highly correlated (Table 1), wet sand evaporation is always different than

pan evaporation (indeed the only thing that seems to evaporate like water is

water itself). The high correlation between eW and eS in this study may be

explained because most of the time the evaporating wet sand was evaporable

(i.e. far from the critical point on the wet side); but even in this case there

are noticeable differences. The mean diurnal variations (Figure 2) show that

on average only during the first 8 hours of the day < eW > and < eS > are

nearly equal to each other and very small; however during the daytime (from

08:00 until 18:00 h, local time) evaporable wet sand evaporates at a greater

rate than plain water, and this situation is reversed for the remaining 6 hours

of the day with a much smaller difference.

Drying wet sand in our experiment (i.e. on the dry side from the critical

point) seemed to evaporate somewhat as Bouchet’s (1963) complementary

relationship stipulates (Figure 4), although the scales of our experiment are

smaller than Bouchet’s (1963) and Kahler and Brutsaert (2006). This

suggests that at least part of the complementary-relationship behavior may

be achieved if conditions are otherwise favorable for evaporation (i.e. bright

273

days), even if the physics is different because of the difference in scales (i.e.

there is no “communication” between evaporating trays). For example our

water-refill procedure, involving only the evaporating trays, is a very gross

surrogate of ambient precipitation, while in previous works soil moisture is

considered to be due to natural generalized conditions [although preliminary

experiments (Velázquez-Mendoza 2010) under rainy days did not yield

different results]. That is: the water-only evaporating tray seems to be

affected very little if at all by the moisture amount of the evaporating wet-

sand tray. This means that in this case there is no energy-surplus from the

deficit of wet sand evaporation that could be used to evaporate plain water.

Nevertheless this demonstrates that under favorable conditions (i.e. bright

days as in d04M and d25M) drying soils quickly loose their evaporative

capacity, eventually evaporating less than water until evaporation becomes

negligible, even if the soil retain a small amount (< 5%) of moisture (Figure

4). This result may be important, because of the difference in physics (here

dominated solely by solar radiation) and scales (here much smaller), in the

sense that Bouchet’s (1963) hypothesis was invoked by Brutsaert and

Parlange (1998) when offering a first explanation to the so called “pan

evaporation paradox”. For example if “complementary behavior” can be

achieved without “complementary physics” one could ask: What are the

consequences for the above explanation, or for other questions raised about

the complementary relationship (e.g. Lhome and Guilioni 2006).

The time-series correlations indicate that R is the most important variable

for both evaporations, compared with Ta, D, and S. Although when

comparing < eW > and < eS > with the mean diurnal variation of these

meteorological variables Ta, D, and S, as defined in equations (1), their

relative importance improve; i.e. these latter correlations are somewhat

274

higher than the former correlations obtained with their time-series. This

means that in average or for time-scales of days (or larger) other

meteorological variables (especially Ta) besides R, may be also important to

model evaporation: e.g. EW ~ f3(R, Ta) or ES ~ f4(R, Ta). Nevertheless the

relevance of R for local evaporation rates: eW ~ f1(R), eS ~ f2(R), EW ~ f3(R,

Ta), and ES ~ f4(R, Ta), in terms of variance explained, suggests that changes

in R may affect not only pan evaporation (Roderick and Farquhar 2002) but

soil evaporation on the wet side from the critical point as well. Soil

evaporation on the dry side from the critical point is not modeled by R (see

Figs. 3−5).

Thus this study seems relevant when considering the opposing trends in

pan and soil evaporation (Khaler and Brutsaert 2006), or the opposing trends

in evaporation and precipitation (e.g. Brutsaert 2006), especially in regards

to the global hydrological cycle (Wentz et al. 2007). However we believe

that more definitive answers to the global questions raised by these

considerations only could be achieved once the uncertainties in oceanic

evaporation (Yu 2007) are reduced, simply because oceanic values are one

order of magnitude greater than continental values.

4. Summary and conclusions

This paper compares the evolution of water and wet sand evaporation

throughout the day and in their daily totals (Figure 1). It shows that they are

highly correlated at r = 0.91 (Table 1), but that their diurnal variations are

somewhat dissimilar (Figure 2), for example in average wet sand

evaporation is greater than water evaporation during the daytime. However

evaporable drying sand behaves differently (Figure 3), at a critical point its

275

evaporative capacity decreases, while open-water evaporation increases.

This critical point, or threshold, is reached at around Dws = 0.92 (Figure 4),

which corresponds to about 5% of moisture content. This is not the

minimum point where sand evaporation ceases which happens at night (Figs.

3 and 4), when even negative values (condensation) may be reached.

Therefore evaporation from open water and evaporation of drying sand

(around 5% of water on a wet-mass basis) during bright days (Figure 5)

behave very similarly to the complementary relationship (Bouchet 1963),

even at our small scales and without the physics that has been suggested

before (e.g. Morton 1983, Szilagyi 2001, etc.). Therefore despite the fact

that physical- and time- scales are smaller than in previous studies (e.g.

Bouchet 1963; Kahler and Brutsaert 2006), the three plots in Figure 4, to the

right of the critical point, somewhat resemble a mirror image (in the

abscissas axis) of Figs. 4 and 5 of Kahler and Brutsaert (2006), and indeed

Bouchet’s (1963) well-known complementary relationship plot. The mirror

image is because we plot eW and eS as a function of a dryness index while

Kahler and Brutsaert (2006) and Bouchet (1963) use a moisture index. We

use a dryness index because we believe it better reflects the time-direction of

the evaporation process (see Figure 3). In these three days: d27A, d04M and

d25M, wet sand seems to have lost gradually its evaporative capacity while

experiencing ambient conditions most favorable for evaporation (as shown

by the increasing eW), even though it retained a small amount of moisture.

Wet sand and water evaporation, far from the critical point on the wet

side, may be modeled mainly by solar radiation because of their high

correlations (Table 1), which may lead to further simplifications (Priestley

and Taylor 1972) to Penman’s (1948) equation; e.g. EW ~ f5(R), ES ~ f6(R).

A more thorough discussion may be found in Velázquez-Mendoza (2010).

276

Acknowledgments. This work, as well as the master’s thesis of one of us

(I.V.), resulted from the original question asked by Antoine Badan: Does

wet sand evaporate more quickly than plain water? We hope that we have

finally answered his query. Our research is supported by the Mexican

CONACYT system. We thank Santiago Higareda for technical support, and

Federico Graef, Thomas Kretzschmar and Luis Zavala-Sansón for helpful

suggestions.

277

References

Bouchet, R. J., 1963: Evapotranspiration réelle, évapotranspiration

potentielle, et production agricole. Ann. Agron., 14, 743-824.

Brutsaert, W., 2006: Indications of increasing land surface evaporation

during the second half of the 20th century. Geophys. Res. Lett., 33, L20403,

doi:10.1029/2006GL027532.

Brutsaert, W., and M. B. Parlange, 1998: Hydrologic cycle explains the

evaporation paradox. Nature, 396, 30.

Kahler, D. M., and W. Brutsaert, 2006: Complementary relationship

between daily evaporation in the environment and pan evaporation. Water

Resources Research, 42, W05413, doi:10.1029/2005WR004541.

Lhome, J. P., and L. Guilioni, 2006: Comments on some articles about the

complementary relationship. J. Hydrol., 323, 1-3.

Morton, F. I., 1983: Operational estimates of areal evapotranspiration and

their significance to the science of hydrology. J. Hydrol., 66, 1-76.

Pavia, E. G., 2008: Evaporation from a thin layer of wet sand. Geophys. Res.

Lett., 35, L08401, doi:10.1029/2008GL033465.

Penman, H. L., 1948: Natural evaporation from open water, bare soil and

grass. Proc. R. Soc. London, Ser. A, 193, 120-145.

Priestley, C. H. B., and R. J. Taylor, 1972: On the assessment of surface heat

flux and evaporation using large-scale parameters. Mon. Wea. Rev., 100, 81-

92.

Roderick, M. L., and G. D. Farquhar, 2002: The cause of decreased pan

evaporation over the past 50 years. Science, 298, 1410-1411.

Szilagyi, J., 2001: On Bouchet’s complementary hypothesis. J. Hydrol.,

246, 155-158.

278

Velázquez-Mendoza, J. I., 2010: Cuantificación y análisis temporal de la

evaporación natural en arenas húmedas (Quantification and temporal

analysis of natural evaporation from wet sand). Master in Science thesis,

CICESE. Ensenada, B. C., México. 55 pp.

Wentz, F. J., L. Ricciardulli, K. Hilburn, and C. Mears, 2007: How much

more rain will global warming bring? Science, 317, 233-235.

Yu, L., 2007: Global variations in oceanic evaporation (1958-2005): The

role of changing wind speed. J. Climate, 20, 5376-5390.

279

Precipitation and Seasonal Variation of Surface

Temperature-controlling Factors in the Sonoran

Desert, Northwestern Mexico* Iryna Tereshchenko ♠, U de G Luis Brito-Castillo, CIBNOR Alexander Zolotokrilin, IG/RAS César Monzón, U de G Tatiana Titkova, IG/RAS

Abstract

According to the Fourth Assessment Report of the Intergovernmental Panel

on Climate Change working group I, an enhancement of desert conditions is

expected to occur in the next 50 years in Southwestern USA and

Northwestern Mexico (where the Sonoran desert is located) with

implications regarding water availability in the region. The present study is

focused on the seasonal cycle of precipitation and on the parameters that

control surface temperature in the Sonoran desert. Understanding these

processes is important in desertification monitoring, because, in any given

year, the period during which the radiation factor is predominant is crucial in

the desertification process. The most significant feature of the summer

precipitation in the Sonoran desert is the abrupt change in moisture

conditions, with a maximum in August. The main features of the ratio

* The Ocean, The Wine, and The Valley: The Lives of Antoine Badan, 279-302. © 2010 CICESE. ♠ Contact: irina.tereshchenko@cucei.udg.mx

280

between albedo and surface temperature are discussed analyzing the monthly

means (albedo, temperature, NDVI) in the state of Sonora; in particular

within the box 30-31° N, 112-113° W. The analysis of synchronous time

series of albedo, surface temperature and NDVI has shown that the dominant

temperature-controlling factors can switch within the year in the study area.

The radiation factor is dominant in dry months (April – May) and the

surface temperature is negatively correlated with albedo. This can cause the

generation of a positive albedo-precipitation feedback, which in turn

contributes to the desertification process.

1. Introduction

The main competing factors underlying surface temperature control

constitute an inherent feature of dry lands (Becker et al., 1988); they are

listed below:

First – Radiation. As albedo increases the radiation absorbed by the

surface decreases; this also causes the surface temperature to decrease, and

viceversa.

Second – Evapotranspiration (i.e. transpiration and evaporation from

the soil surface). Prolonged lacks of precipitation as well as

anthropogenic effects often cause reduction of vegetation. This, in turn,

results in significant evapotranspiration decrease and a slight increase in

albedo leading to an overall increase of the surface temperature, and vice

versa.

Third – Aerodynamic control. As the density of short vegetation (i.e.

grass and bushes) decreases, the surface becomes smoother (i.e. the

281

roughness parameter decreases). This causes a decrease in the vertical fluxes

of sensible and latent heat and an increase in diurnal surface temperature.

Temperature control by evapotranspiration and aerodynamic controls are

closely linked via a roughness parameter. But in this case it constitutes a

purely informational statement, because in regard to Sonora’s vegetation

there may be different scenarios of aerodynamic control; however,

unfortunately, any field measurements that would allow accounting for this

factor are presently unavailable.

Correlation between albedo and dry land surface temperature can serve as

an indicator of processes that control temperature. The term dry land is used

in reference to arid, semi-arid and dry sub-humid regions, whose

humidification aridity index, (according to Thornthwaite, 1948), ranges

between 0.05 and 0.65 as determined by the United Nations Convention to

Combat Desertification (UNCCD, 1994).

Interest in different ratios between the temperature-controlling factors

was raised after a number of researchers (Jackson and Idso, 1975; Ripley,

1976 a, b; Idso, 1977, 1981) interpreted numerical experiments by Charney,

1975, who studied the impact of albedo of dry land on regional climate.

According to Charney, the radiation factor is predominant in circumstances

where anthropogenic effects cause a reduction of vegetation, which leads to

an increase in albedo and a decrease in temperature (negative correlation).

As a result convection, cloudiness and precipitation decrease, which in turn

causes a further increase in albedo. By this means an albedo-precipitation

positive feedback is formed (i.e. albedo-feedback mechanism of

desertification, according to Charney), which contributes to anthropogenic

desertification.

282

The in-situ observations of albedo and surface temperature were first

published in the 1980s, followed by the publications of remote sensing

observations. These publications not only indicate the drop in temperature

with the increase in albedo in deserts, but also some increase in surface

temperature of land with highly developed vegetation during its

anthropogenic or natural reduction (Wendler and Eaton, 1983; Goward et al.,

1985; Otterman and Tucker, 1985; Zolotokrylin, 1986; Vukovich et al.,

1987; Seguin et al, 1987; Menenti et al, 1988). An in-depth analysis of in-

situ, helicopter and satellite data was published by Zolotokrylin (2003)

examining how the albedo-to-temperature ratio varies in deserts and semi-

deserts of different dry regions of Asia and Africa.

The theoretical explanation of positive albedo-temperature feedback

based on the thermal balance equation was provided by Idso (1981). Later, a

linear statistical model of positive feedback generation was suggested by

Avissar and Pielke (1989).

This study is focused on the seasonal cycle of parameters which control

surface temperature, together with the intra-annual precipitation cycle in the

Sonoran desert (northwestern Mexico). Understanding this process is

important to monitor desertification. This is because in any given year, the

time span of the period during which the radiation factor is predominant is

an important issue in the desertification process. One indirect characteristic

of prevalence of the radiation factor is the Normalized Difference

Vegetation Index (NDVI), which is an indicator of green phytomass. More

specifically, this is a threshold value that allows the switch of the surface

temperature-controlling factors within a seasonal cycle (Zolotokrylin, 2003.)

NDVI is a satellite-based remotely sensed measure of the “greenness” of the

vegetation cover. Its use and applications will simplify the monitoring of

283

desertification and restrict it to the NDVI data from AVHRR and MODIS

[(Watts et al (2007), Vivoni et al. (2008) and Mendez-Barroso et al. (2009)].

2. Study area

The Sonoran desert is the wettest and warmest of four deserts in North

America (Fig. 1). It covers most of the Mexican states of Sonora and Baja

California extending to the U.S. states of Arizona and California (Shreve and

Wiggins, 1964).

Due to its altitude and geographic location, the Sonoran desert displays

many different climates along the year as well as a fairly varied flora. The

desert receives 100 to 300 mm of yearly precipitation, mostly as rain.

Rainfall occurs primarily during the summer so it has a strong seasonal

signature, and its interannual variability is significant. Daytime temperatures

can reach or exceed 40° C during the summer months of May to September.

There are two pronounced rain periods: winter – December to March, and

summer - with showers from July to September. The winter rainfall is

associated with the passage of mid-latitude systems and their frontal parts

over the North American continent (Jáuregui, 1995). The humid airflow

transports the showers from the tropical Pacific and the Gulfs of Mexico and

California to Mexican territory mainly due to the monsoon system (Douglas

et al., 1993; Higgins et al., 2003, Vivoni et al. (2008).). Droughts occur in

early summer (May-June) and early fall.

The study area includes the Sonoran desert and adjacent areas within

23-35oN, 109-118ºW. The main focus is on the Mexican part of the territory,

in the state of Sonora (large box 29-32oN, 111-116.5oW), in particular,

within the box of 30-31ºN, 112-113ºW which extends to the Gulf of

284

California (Fig 1.) Typical scrub vegetation dominates this region of the

Sonoran desert (Fig.2).

3. Data and methods

The present study made use of monthly mean sums of precipitation

observed at 23 stations within the region of 28-33°N, 110-118°W over the

period of 1985-1994 (ERIC, 1996). These data were used both to determine

the temporal variations of precipitation in the study area and for verification

at individual stations.

Time series around the Sonoran desert display the long-term average of

monthly rainfall totals, in three gage stations: one located in the small box

(Pitiquito), and two in the large box (Sonoyta and Atil) on the mainland of

Mexico (Fig. 1).

To compute monthly means of precipitation at regular grid nodes, 23

stations within this region were selected. For each month, the sum of

precipitation was calculated and averaged over the period of 1985-1994. The

on-station data was interpolated onto the regular 0.5° grid by means of the

WINSURF program.

Within the smaller region of 29-32° N, 111-116.5° W, the precipitation

values that correspond to the water and elevated parts of the land were

rejected, thus, the precipitation values show good agreement with the values

of NDVI, albedo and surface temperature in the study area.

Next, the mean sums of precipitation were derived for two regions: 29-

32°N, 111-116.5°W and 30-31°N, 112-113°W (within the box). The time

series of the intra-annual precipitation cycles in these areas were then plotted

(Fig. 1) for the 1985-1994 period.

285

Mean NDVI, calibrated albedos from Advanced Very High Resolution

Radiometer (AVHRR), channels 1, 2; calibrated temperatures from

AVHRR, channels 4, 5 with a 0.15º spatial resolution collected from

NOAA-9 and NOAA-11 satellites for the period of April, 1985 to

September, 1994, were used in this study (Gutman et al, 1995). Yet,

discontinuities and residual trends can be traced in time series of NDVI and

temperature from Channels 4. Discontinuities result from the switch from

NOAA-9 to NOAA-11 in 1988, and the Mount Pinatubo eruption in 1991.

Trends are a combined effect of satellite orbit drift and a possible persistent

error in post-launch calibration of channels. The orbit drift affects the solar

and thermal IR channels through systematic variation of illumination

geometry and diurnal heating/cooling of the surface and atmosphere,

respectively. The corrections are incorporated in the present data.

We also used monthly means of albedo and surface temperature data

(Global AVHRR-Derived Land Climatology) prepared by the NOAA/

NESDIS (National Environmental Satellite, Data, and Information Service)

and the NGDC (National Geophysical Data Center). This is one of the few

climatological datasets that have been brought up to the point of being

practically usable and considered as a reference, because of their meticulous

preparation. Ground-truthing of these albedo and temperature data sets as

well as inter-comparisons with helicopter data were performed in the deserts

of Central Asia (Zolotokrylin, 2003).

4. Results

The most significant feature of the summer season in the Sonoran desert

is the abrupt change in moisture conditions, from hot, dry conditions in

286

May-June to relatively cool and rainy conditions in July. The period of July

to September contributes 40 to 60% of the total annual precipitation (see

time series in stations: Pitiquito, Sonoyta and Atil, Fig. 1), with a maximum

in August on the mainland, and August-September near the coast of the

Peninsula (Vivoni et al. (2008)). The precipitation maximum observed in

August near the coast on the continental side has also been documented as a

regular feature of the monsoon circulation in western Mexico (Brito-Castillo

et al., 2010).

As shown in Fig. 1 (lower-right side) average monthly precipitation

over the entire study area presents one maximum in August. This maximum

is higher within the small box and lower within the large box. In other

words: the broader the territory from which we calculate the average

maximum precipitation, the lower the value of that maximum precipitation.

3.a The ratio between monthly means of albedo and surface

temperature during a year

The main features of the ratio between albedo and surface temperature

are discussed in terms of analysis of monthly means (albedo, temperature,

NDVI) in the state of Sonora (29-32°N, 111-115°W) in particular, within the

box of 30-31°N, 112-113°W.

Figure 3 shows the cluster of points of temperature monthly means

observed for the corresponding monthly albedos and values of NDVI. As

shown, the correlation between albedo and surface temperature was not

pronounced along the year. The points were noticeably differentiated by

months. Clearly visible is the intra-annual variability of surface temperature

at a given value of albedo. For maximum albedo, the amplitude of surface

287

temperature peaked at almost 22ºС with NDVI in the 0.09 – 0.11 range.

However, no clear surface temperature dependence of albedo was observed.

Within the territory the green phytomass underwent significant changes

during the year: NDVI ranged from 0.19 in August to 0.07 in May-June,

clearly reflecting the progress of intra-monthly rainfall. The highest

temperatures at lowest NDVI (values ≤0.08) were encountered in the study

area mostly in June, the end month of the dry season. The seasonal cycle of

NDVI manifested itself in albedo variations from 0.15 to 0.23. At the same

time the range of seasonal cycle of temperature was about 25ºС (23-48ºС).

As mentioned, no comparisons with in situ measurements of albedo and

temperature were made for the Sonoran desert, but the NDVI values of 0.07

- 0.19 differed little from those in deserts and semi-deserts of Central Asia,

where field work was performed (Zolotokrylin, 2003), and the semi-desert

part of Sahel (Hydrologic Atmospheric Pilot Experiment, HAPEX-Sahel).

3.b The ratio between monthly means of albedo and monthly surface

temperature

Month-by-month inspection of data helps to illuminate the relationship

between albedo and surface temperature. As shown in Fig.4 a, b in dry

months (May, June) the radiation factor controlling the surface temperature

was predominant, there was a negative correlation (-0.75) between these

parameters and the values of NDVI declined as albedo increased and the

temperature dropped.

The situation changed to the opposite in the humid months July and

August (Fig.4 c, d). Vegetation responded completely (in a week or so); in

July the NDVI increased on most of the territory peaking at 0.19 in August

288

(the month of a maximum precipitation) when the green transpiring

phytomass grew in quantity. The portion of radiation balance consumed by

evapotranspiration became larger, while the portion of energy which goes

into turbulent heating of the surface boundary layer decreased. In the end,

evapotranspirational control of the surface temperature dominated in most of

the study area and the correlation between albedo and surface temperature

was positive. This is shown in the left part of the approximation curve of

Fig.4d. In September, when NDVI started to decline, the albedo-temperature

correlation weakened in August when the correlation coefficient was 0.81,

and in September it dropped to 0.50, but the positive sign of the correlation

remained. At the same time, some zones with minimal green phytomass still

remained in the study area (NDVI did not exceed 0.09) on the rightmost

approximation point of the same curve (Fig.4d). Most likely, this was a

manifestation of both evapotranspiration and radiation factors.

3.c Average threshold NDVI: when the values are less than this limit,

the predominance of the radiation factor is possible

For monitoring purposes it is important to know the threshold value of

monthly mean NDVI, which is a characteristic of the predominance of the

radiation factor in the Sonoran desert. This value can vary depending on the

spatial scale. Fig.5 presents histograms of NDVI in May (a), June (b) and

July (c) for the area 30-31° N, 112-113° W (local scale in Fig.1). It follows

from the comparative analysis of the histograms that average NDVI

decreased from 0.103±0.013 in May to 0.086±0.01 in June and 0.08±0.01 in

July. This situation is possible if the effective precipitation (i.e. the

precipitation sufficient for grass layer growth) falls mostly in mid- and late

289

July. The growth of the grass layer takes time; greening will be observable

from the satellite no sooner than August. (According to Zolotokrylin (2003)

it takes a month for the deserts of Central Asia.) As a result, the threshold

value of 0.08±0.01 can be accepted for this territory. At lower values, the

temperature-controlling radiation factor becomes predominant.

Figure 6 shows the change of threshold NDVI values with the extension

of the study area. The histogram is almost bimodal. This points to the fact

that the radiation factor was predominant in areas where NDVI values were

less than 0.06-0.07. Such areas made up two thirds of the study area. It

should be noted that in the areas where NDVI exceeded 0.17-0.19 (i.e.

higher phytomass content) the evapotranspiration factor played a more

important role.

Therefore, as the scale increases, the threshold NDVI value is somewhat

reduced and more stable with a value less than or equal to 0.07. (A similar

result was obtained by Zolotokrylin (2003) for dry lands of Central Asia and

Sahel).

Significantly, the part of the Sonoran desert where NDVI was lower than

the threshold value of 0.07 reached its maximum in July. In the most humid

month (August) this area was reduced to its minimum size.

3.d Periods with predominant controls on surface temperature

Three periods were identified in which either the radiation,

evapotranspiration or radiation-evapotranspiration type prevailed. The

radiation type was predominant prior to the rainfall season in April-June.

During this period the amount of phytomass dropped down to a minimum,

and the surface temperature correlated with the albedo. This creates

290

favorable conditions for generation and preservation of the desert climate,

until the external factor (precipitation) creates a new situation.

The evapotranspirational control dominated during a very short period

(August-September), when soil moisture content and water storage capacity

of plants markedly increased thanks to the rise in precipitation rates. Vivoni

et al. (2008) described the seasonal progression of the link between

evapotranspiration and soil moisture for northwestern Mexico. The increase

in the amount of green phytomass adds to the evapotranspirational moisture

loss. This reduces the portion of energy spent on turbulent heating of the

bottom atmosphere and the heating of soil. The radiation is also reduced by

the presence of the plant canopy (leaves) itself and by the evapotranspiration

factor, but it is still significant. Thus, the climate of the area takes on semi-

desert features. However, in some rocky parts where vegetation is reduced,

certain local desert-type climates can be preserved.

During the rest of the year (October-March) neither the radiation nor the

evapotranspiration factors prevailed. The latter is mainly characterized by

evaporation from soil. The reduction of solar radiation diminishes the effect

of the radiation factor, but the evapotranspiration factor is boosted due to the

increase in winter precipitation. In this way, the state of equilibrium between

the factors that control temperature is maintained, while the correlation

between albedo and temperature drops.

4. Discussion

Inferences about the seasonal variability of temperature-controlling

factors in dry lands may become a scientific background to understand the

dynamics of desertification and its monitoring. Desertification is defined as

291

a reduction of vegetation due to climate change and human impacts. We

strongly suggest it should be monitored. Changes in the areas of arid regions

with NDVI ≤0.08 can be considered as a characteristic of evolution of

desertification of the territory. A long-standing increase of the area is an

indicator of reinforcement of desertification, while the long-standing

decrease of the area is suggestive of its weakening. Two aspects are of

importance here. The first one is related to the investigation of switching

mechanism between the temperature-controlling factors, based on the energy

balance. Our study states that the switch is possible at a certain value of

NDVI. To understand the switching mechanism better, more specific

experiments need to be conducted, based on energy balance studies and

calculation of radiation and vertical turbulent fluxes of heat and moisture in

the atmosphere boundary layer and soil.

The second aspect of this study is related to the switch indicator. Strictly

speaking, the green phytomass content may serve as this indicator, i.e. the

amount of the assimilating vegetation per square unit. The work on

formulation and definition of this indicator will take many years of multiple

experiments in different vegetation types. This indicator can be viewed as a

base (reference), which will subsequently be transformed into remote NDVI

indicator, because green phytomass content and NDVI are highly correlated

(Tucker et al, 1985). The reference indicator is also essential for applying

corrections to the NDVI data obtained by different radiometers (AVHRR

and MODIS). One of the first and foremost tasks is to obtain uniform NDVI

time series. This issue has gathered importance because the MODIS

radiometer replaced the AVHRR radiometer on a new satellite series.

292

5. Conclusions

The analysis of synchronous time series of albedo, surface temperature

and NDVI showed that the dominating temperature-controlling factors could

switch within the year in the study area.

The radiation factor was dominant in dry months (April – May), when

the threshold value of 0.08±0.01 exceeded NDVI values in most of the study

area. In these circumstances the surface temperature is negatively correlated

with albedo. This can cause generation of positive albedo-precipitation

feedback, which in turn contributes to the desertification process.

The evapotranspiration temperature-controlling factor prevailed in the

most humid month, August. Positive correlation between albedo and

temperature, which occurred during this month, created conditions for the

generation of negative albedo-precipitation feedback, which impedes

desertification.

In the autumn and winter months equilibrium was achieved between the

radiation and evapotranspiration types of temperature control. During these

periods the correlation between albedo and surface temperature was weak.

To monitor desertification it is important to estimate the areas with NDVI

values lower than the threshold value both in dry and humid months and to

outline the trend of the changes. This will be the subject of a separate study.

293

References

Avissar R., Pielke R.A., 1989: A parameterization of heterogeneous land

surface for atmospheric numerical models and its impact on regional

meteorology. Month. Weather Rev., 117, 2113-2136.

Becker F., Bolle H.J., Rowntree P.R., 1988: The International satellite land-

surface climatology project. Berlin Free Univ. Press, 100 pp.

Brito-Castillo L., E.R. Vivoni, D.J. Gochis, A. Filonov, I. Tereshchenko, C.

Monzon, 2010: An anomaly in the occurrence of the month of maximum

precipitation distribution in northwest Mexico. J. Arid Environ., 74 (5), 531-

539.

Charney J.G., 1975: Dynamics of deserts and drought in Sahel. Quart. J.

Royal. Meteorol. Soc.,.101, 193-202.

Douglas M.W., Maddox R.A., Howard K., Reyes S., 1993: The Mexican

Monsoon. Journal of Climate, 6, 1665-1677.

ERIC, 1999: Extractor Rápido de Información climatológica. IMTA,

Instituto Mexicano de Tecnología del Agua: Morelos, México. CD-ROM, 4th

version. Also in:

http://www.imta.gob.mx/productos/software/meteorologia.html (2008).

Goward S.N., Cruickshanks G.D., Hope A.S., 1985: Observed relation

between thermal thermal emission and reflected spectral radiance of a

complex vegetated landscape. Remote Sens. Environ., 18, 137-146.

Gutman G., Tarpley D., Ignatov A., Olson S., 1995: Global land products

from AVHRR: The enhanced NOAA GVI dataset. Bull. Amer. Met. Soc., 76,

1141-1156.

Higgins R.W., Douglas A., Hahmann A., Berbey E.H., Gutzler D.,

Shuttleworth J., Stensrud D., Amador J., Carbone R., Cortez M., Douglas

M., Lobato R., Meitin J., Ropelewski Ch., Schemm J., Schubert S., Zhang

294

Ch., 2003: Progress in Pan American CLIVAR Research: The North

American Monsoon System. Atmosfera, 16, 29-65.

Idso S.B., 1977: A note on some recently proposed mechanisms of genesis

of deserts. Quart. J. Royal. Meteorol. Soc., 103, 369-370.

Idso S.B., 1981: Surface energy balance and genesis of deserts. Arch.

Meteorol. Geophys. Bioklimatol., 30, 253-260.

Jackson R.D., Idso S.B., 1975: Surface albedo and desertification. Science,

189, 369-370.

Jáuregui, E., 1995: Rainfall fluctuations and tropical storm activity in

Mexico. Erkunde, 49, 39-48.

Mendez-Barroso L.A., E.R. Vivoni, Ch.J. Watts, J.C. Rodriguez, 2009:

Seasonal and interannual relations between precipitation, surface soil

moisture and vegetation dynamics in the North American monsoon region.

J. Hydrology., 377, 59-70.

Menenti M.W., Bastiaanssen D. van Eick and M. A. Abd Karim, 1989.

Linear relationships between surface reflectance and temperature and their

application to map actual evaporation of groundwater. Adv. Spase Res., 9,

165-176.

Otterman J., Tucker C.J., 1985: Satellite measurements of surface albedo

and temperatures in semi-desert . J. Climate and Appl. Meteor., 24, 228-

235.

Ripley E.A., 1976: Comment on the Paper by J.G. Charney ‘Dynamics of

deserts and drought in Sahel. Quart. J. Royal. Meteorol. Soc., 102, 466-467.

Ripley E.A., 1976: Drought in the Sahara: Insufficient biogeophisical

feedback. Science, 191, 100.

295

Seguin B., Assad E., Freteaud J.P. et al., 1987: Suivi du bilan hydrique á

l’aide de la télédetection par satellite: Application au Senegal. Rep. to the

EEC-DG8. Bruxelles.

Shreve F., I.L. Wiggins., 1964: Vegetation and Flora of the Sonoran Desert.

Stanford University Press, Stanford, 1627 pp.

Tucker C.J., Vanpraet C.L., Sharman M.J., Van Ittersum G., 1985: Satellite

remote sensing of total herbaceous biomass production in the Senegalese

Sahel: 1980-1984. Remote Sens. Environ., 17, 233-249.

Thornthwaite, C. W., 1948: An approach toward a rational classification of

climate. The Geographical Review, 38, 55-94.

UNCCD, 1994: United Nations Convention to combat desertification in

those countries experiencing serious drought and desertification,

particularly in Africa. WMO, Geneva.

Vivoni E.R., Moreno H.A., Mascaro G., Rodriguez JC., Watss Ch.J.,

Garatuza-Payan J., and Scott R.L., 2008: Observed relation between

evapotranspiration and soil moisture in the North American monsoon region.

Geophysical Research Letters, 35, L22403, doi:10.1029/2008GL036001.

Vukovich F.M., Toll D.L., Murphy R.E., 1987: Surface temperature and

albedo relationship in Senegal derived from NOAA-7 satellite data. Remote

Sens. Environ., 22, 413-421.

Watts, Ch. J., Scott R.L., Garatuza-Payan J., Rodriguez J.C., Pruegger J.H.,

Kustas W. P., Douglas M., 2007: Changes in Vegetation Condition and

Surface Fluxes during NAME 2004. J. of Climate, 20, 1810-1820.

Wendler G., Eaton F., 1983: On the desertification of the Sahel zone: Part 1.

Ground observations. Climate Change, 4, 365-380.

296

Zolotokrylin A.N., 1986: The relationships between surface albedo and

temperature in arid regions. Data of meteorological studies, Russian

Academy of Sciences. Geophysical Committee. 10, 38-45 (in Russian).

Zolotokrylin A.N., 2003: Climatic desertification. A. N. Krenke, Editor.

Nauka. 246 pp. (in Russian)

297

120 118 116 114 112 110 108 106Longitude (W)

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

Latit

ude

(N)

México, D.F.

Guadalajara

Hermosillo

Arizona

California

Baja California

Baja California Sur

Sonora

1 2 3 4 5 6 7 8 9 10 11 12Month

020406080

100

Pre

cipi

tatio

n, m

m 26096 Sonoyta

The Sonoran desert

1 2 3 4 5 6 7 8 9 10 11 12Month

020406080

100

Pre

cipi

tatio

n, m

m 26038 Atil, Atil

1 2 3 4 5 6 7 8 9 10 11 12Month

020406080

100

Pre

cipi

tatio

n, m

m 26093 Pitiquito

1 2 3 4 5 6 7 8 9 10 11 12Month

020406080

100

Pre

cipi

tatio

n, m

m 29 - 32°N 118 - 111°W30 - 31°N 113 - 112°W

N

23 climatic stationsStations in the Sonoran Desert

Figure 1. The study area of the Sonoran desert. The large box includes part of the states of Sonora and Baja California (29-32oN, 111-116.5oW). The small box includes 30-31oN, 112-113oW. Time series around the Sonoran desert display the long-term average of monthly rainfall totals in three gage stations. The intra-annual cycle of average monthly precipitation sums from the large box (solid line) and the small box (dashed line) are displayed at the lower right side of the figure.

29

8

Figu

re 2

. The

typi

cal s

crub

veg

etat

ion

dom

inat

es in

the

Sono

ran

Des

ert.

(P

hoto

grap

h by

Edg

ar A

lcán

tara

Raz

o, C

IBN

OR

).

29

9

Fi

gure

3. R

elat

ions

hip

betw

een

albe

do a

nd s

urfa

ce te

mpe

ratu

re in

the

Sono

ran

dese

rt du

ring

the

year

. The

dat

a po

ints

are

lab

elle

d w

ith N

DV

I va

lues

. Th

e da

ta a

re a

vera

ged

over

the

per

iod

1985

-199

4. O

ne c

ell

of d

egre

es

cont

ains

39

pixe

ls 1

6x16

km

.

30

0

Fi

gure

4. R

elat

ions

hip

betw

een

albe

do a

nd s

urfa

ce te

mpe

ratu

re in

the

Sono

ran

dese

rt. T

he d

ata

poin

ts a

re

labe

lled

with

ND

VI v

alue

s. Th

e da

ta a

re a

vera

ged

over

the

perio

d 19

85-1

994.

a )

May

, b) J

une,

c) J

uly,

and

d)

Aug

ust.

301

Figure. 5. The histogram of NDVI in the Sonoran desert. a) May, b) June, and c) July.

302

Figure 6. The histogram of NDVI in the Sonoran desert. A) June, and b) July.

303

Chapter 5

The Symposium

304

30

5

306

307

The Ocean, the Wine and the Valley November 16-18, 2009

(Sunday 15, November: Icebreaker at Coral & Marina)

9:00-9:05

9:05-9:15

Introduction: E.G. Pavia

Welcome: F. Graef The Gulf of Mexico I Chair: J. Sheinbaum

9:15-10:45

J. Candela: Circulation in the western Gulf of Mexico. A review of recent measurements by the Canek group. E.P. Chassignet: Gulf of Mexico circulation: What we can learn from numerical models? P. Hamilton: Eddy spin-up through eddy-topography interaction over the northwest Gulf of Mexico continental slope.

10:45-11:15 Coffe break The Gulf of Mexico II

Chair: J. Candela

11:15-12:45 J. Zavala: Seasonality of the Yucatan upwelling. A.Lugo-Fernandez: The Florida Current effects on the Loop Current: A Bernoulli gate.

A.M. Moore: Yucatan transport inferred from ROMS 4D-Var.

12:45-14:15 Lunch The Ocean and the Wine

Chair: T. Cavazos

14:15-15:45

W. Munk: The ocean and the wine. (CANCELLED) O. Velasco Fuentes (FILLED IN) A. Gershunov: Ocean warming, heat waves and grapes. A.J. Miller: The confluence of wine and ocean currents: Oenological forcing of climate science interactions.

Monday, 16

15:45-16:15 Coffe break

308

The Ocean, the Wine and the Valley Chair: E.G. Pavia

16:15-17:15

17:15-18:00 Posters

H. D’Acosta: Valley, Ocean, Wine. M. Schorr: Chemistry, materials and equipment in wineries: An overview on the Guadalupe Valley. POSTER J. Färber: Winter trophic conditions off the southwest coast of Mexico. POSTER A Martínez: Voltammperometric studies for Baja California red wines. POSTER

Monday, 16

18:00

Wine tasting offered by

Cofradía del Vino de Baja California.

309

The Ocean, the Wine and the Valley

The Pacific Ocean Chair: J.L. Ochoa

9:00-11:00

P. Niiler: New concepts of upper ocean circulation of the eastern Pacific. J.C. Herguera: Droughts and sea surface temperature variability in the southern California Current: High resolution records for the past 1.2 Ka. J. Gómez: The deepening of the wind-mixed layer in the southern part of the California Current. D. Rivas: Lagrangian circulation in Todos Santos Bay during spring 2007.

11:00-11:30 Coffe break “plus” The Ocean

Chair: M. López

11:30-13:00 K.H. Brink: Vanishing drag and increasing flow inhibition. L. Zavala: Linear and nonlinear barotropic motions around seamounts. E.D. Barton: Recent advances in further research on irreproducible results: a generalization.

13:00-13:30 Transportation to Cavas del Mogor Winery Antoine Badan

Chair: L.G. Alvarez

13:30-15:00

F. Martain: Antoine Badan, Fundador y Gran Maestre, Cofradia del Vino de Baja California. M. Hendershott: Remembrances of Antoine Badan. F. Graef: A remembrance of my friend Toño, scientist, environmentalist and wine-maker.

Tuesday, 17

15:00-

Dinner at “El Mogor”

310

311

312

313

The Ocean, the Wine and the Valley

The Gulf of California Chair: G. Marinone

9:00-11:00

C. Winant: Tidal circulation in the Sea of Cortez. M. López: Flow at the sills bounding Delfin Basin, northern Gulf of California. A. Filonov: Internal tide at San Esteban sill in the Gulf of California. M. Figueroa: Microstructure observations in Salsipuedes Channel, Gulf of California.

11:00-11:30 Coffe break The Gulf of Mexico III

Chair: J. Zavala

11:30-13:00

P. Pérez-Brunius: Surface circulation in the southern Gulf of Mexico. J. Sheinbaum: Recent numerical modeling efforts in the Gulf of Mexico. J. Ochoa: Deep Migrating Zooplankton in the Gulf of Mexico.

13:00-14:30 Lunch The Guadalupe Valley I

Chair: J. Gómez

14:30-16:00

T. Kretzschmar: Un plan de manejo del agua sustentable para Valle de Guadalupe, Baja California. L. Lizarraga: Sand mining from Guadalupe valley riverbeds. E.G. Pavia: Evaporation from wet sand.

16:00-16:30 Coffe break The Guadalupe Valley II

Chair: A. Gershunov

16:30-17:30 I. Galindo: Satellite climatology of Valle de Guadalupe. T. Cavazos: Regional climate change scenarios for Baja Calfornia.

17:30-18:00 J. Sheinbaum: Closing remarks

Wednesday, 18

18:00 Wine tasting offered by

Asociación de Vitivinicultores de Baja California.

314

315

Abstracts*

The Ocean and the Wine

Walter Munk ♣, SIO UCSD

I will cover two of the three assigned topics by recalling an event early in

World War II: how an amphibious landing in Northwest Africa came close

to being aborted by the effect of red wine on ocean waves.

Droughts and sea surface temperature variability in the southern

California Current: High resolution records for the past 1.2 Ka

J. C. Herguera ♦, CICESE

Graham Mortyn, ICTA UAB

Miquel Àngel Martínez-Botí, DG UAB

Gladys Bernal Franco, EG U de C

An outstanding issue in our understanding of coastal ocean dynamics is how

the increasing anthropogenic CO2 injection into the atmosphere will change

upwelling patterns in eastern boundary current systems and their implication

for droughts in the Baja California arid region. In this study we develop a

* These works were either withdrawn from the symposium or not submitted to these proceedings for different reasons. ♣ Contact: wmunk@ucsd.edu ♦ Contact: herguera@cicese.mx

316

~1100-year high resolution proxy record of summer sea surface

temperatures (SSTs) based on the Mg/Ca derived summer SSTs on the

plankton foraminifer Globigerinoides ruber, shown to calcify in the upper

15 m of the water column during the summer season in a well stratified

shallow upper surface layer in this region and on the organic C content in the

sediments. The SST and organic C time series reconstruction from sediments

of San Lázaro Basin (SLB) (25˚N 112˚W) for the past century are shown to

capture the variability between the instrumental summer SSTs on decadal

timescales fairly well for the last century. Here we will show a millennium

long time series of the reconstructed summer surface temperatures and the

organic C record and their links with the Northern Hemisphere (NH)

temperature variability. One of the implications of these observations is that

the processes that control sea surface temperatures, droughts,

biogeochemical cycling of nutrients, ocean-atmosphere carbon fluxes and

biological productivity of the CCS may intensify as a consequence of global

warming and increasing atmospheric CO2 concentrations, with still largely

unknown biogeochemical and ecological implications.

New concepts of upper ocean circulation of the eastern Pacific

Peter Niiler, SIO UCSD

New concepts in the circulation of the eastern Pacific Ocean have principally

come from direct measurements of the ocean velocity and altimeter derived

sea level and Ocean General Circulation Model (OGCM) solutions at high

spatial and vertical resolution. Direct measurements show that the

317

ageostrophic velocity that is directly driven by wind is generally larger than

the geostrophic velocity derived from hydrographic and sea level data. These

data define vortical regions of surface convergence of mass both at 30N and

30S that not only collect drifting buoys but also large islands of flotsam

discarded by ships and injected by flow off land. OGCM solutions at 5km

resolution show that the flow patterns and eddies generated on the western

continental shelves have effects thousands of kilometers to the west. Strong

SST anomalies formed in the California and Baja current systems propagate

rapidly to the south, even reaching the equator at times when Los Niños are

formed. The best recipe for digesting these new ideas is to eat Baja

California Red Fish baked with salt lemons and accompanied of Mogor

Badan wines, as served for many years at Badan’s ranch in the Guadalupe

Valley.

Vanishing Drag and Increasing Flow Inhibition

K.H. Brink ♠, WHOI

S.L. Lentz, WHOI

Strange effects are known to arise near the bottom when an interior along-

isobath flow occurs in a stratified ocean. Specifically, either up- or down-

slope flow in a bottom boundary layer modifies the density field in such a

way that geostrophic shear leads to the bottom velocity (hence bottom stress)

vanishing. We report new results on the contrasting ways that the boundary

♠ Contact: kbrink@whoi.edu

318

layer evolves for the two flow directions. Then, we explore how the

consequent boundary layer properties -layer thickness and absence of stress-

affect cross-isobath exchanges in the steady limit. This work harks back to

some of the earliest oceanographic efforts by Antoine Badan.

The Confluence of Wine and Ocean Currents: Oenological Forcing of

Climate Science Interactions

Arthur J. Miller ♣, SIO UCSD

Scientific productivity in oceanography and climatology is often instigated

by collaborations forged by mutual affinities for vinicultural product

assessment and enjoyment. Antoine Badan had a deep understanding of this

effect as evidenced by his outstanding wine-making skills and his extensive

coterie of productive scientific collaborators with enthusiastic oenological

interests.

Previous studies have addressed the important issue of climate forcing of

the quality of wines. For example, I once noted in a seminar that a simple

ENSO index correlated well with highly rated wines in the California

regions (Fig. 1). This is an intuitively appealing and plausible result.

♣ Contact: ajmiller@ucsd.edu

319

Figure 1. Wine quality in California Cabernet Sauvignons (blue line) and the Southern Oscillation Index (red line), revealing the significant and direct impact of climate on wine. (Wine quality data taken from a Lett’s pocket calendar guide to wines.)

It is not so obvious, however, that the reverse is also plausibly true. That

is, oenological forcing exerts a significant influence on scientific

productivity in ocean-atmosphere research. This is the hypothesis that we

shall explore in this discussion. I have been fortunate in experiencing some

first-hand effects of this dynamic process so I provide the representative data

here.

The forcing datasets include (1) Italian wines, c. 1992-1994, (2) French

versus California wines, c. 2000-present, (3) German wines, c. 2006, and (4)

Ensembles of wines, 2005-present. The response datasets include (1)

Forecasting Iceland-Faroe Frontal variability, (2) ROMS adjoint model and

320

data assimilation platforms, (3) Climate-fisheries linkages, and (4) Regional

coupled ocean-atmosphere dynamics.

My first data point was established during my years living and working

in Lerici, Italy, where Italian wines imbibed at a charming wine bar (Fig. 2),

as well as in countless other restaurants throughout Italy, instigated

numerous close collaborations with my local (Alex Warn-Varnas, Pierre

Poulain of SACLANTCEN) and remote (Allan Robinson, Hernan Arango,

Pierre Lermusiaux, Wayne Leslie of Harvard) collaborators. This directly

led to four peer-reviewed publications in well-respected journals.

Figure 2. Examples of the Italian vintage that forced a productive collaboration on Iceland-Faroe frontal variability.

321

My second data point derives from a friendly challenge motivated by

Hernan Arango (Rutgers) who claimed that an expensive St Julien Bordeaux

was superior to a modestly priced California Zinfandel [favored by the

author]. A blind wine taste testing challenge ensued [in which the Zin

prevailed] that consequently resulted in a series of blind wine taste testing

competitions (Fig. 3) for which vigorous and detailed tasting notes were

painstakingly assembled and validated. The highly productive collaborations

(which also included Andy Moore, Manu Di Lorenzo, Bruce Cornuelle and

Julio Scheinbaum, among many others) precipitated (at least) four

benchmark papers in ROMS data assimilations and its applications.

Figure 3. Examples of wine forcing at various locales during the highly productive time period of the development of ROMS adjoint model and data assimilation platforms.

322

My third data point comes from a workshop hosted in Berlin by Jurgen

Alheit, whose vinicultural expertise and voluminous reserves of German

Rieslings (Fig. 4) drove an intensive, albeit short-term, interaction between

physical and biological oceanographers (especially Andy Bakun). The result

was a splendid refereed synthesis article on climate forcing of marine

ecosystems and fish populations that was recently published.

Figure 4. Examples of the forcing functions derived from German Rieslings during a scientifically productive workshop on climate forcing of marine ecosystems in Berlin.

323

My fourth and final data point is associated with an ensemble of wine

forcings from bottles obtained from around the world. It elicited a lively and

continuing collaboration on coupled ocean-atmosphere feedbacks between

my former Ph.D. student Hyodae Seo (now at Wood Hole), Ragu

Murtugudde (University of Maryland), Markus Jochum (NCAR) and yours

truly. The scientific production currently totals six refereed publications in

distinguished journals, and continues today in an effervescent mode.

As a test of our hypothesis, I have quantified the magnitude of the

oenological forcing and established a productivity index for the scientific

output that ensued from each of the forcings considered here. A simple

linear model relating oenological intensity to scientific creativity can thereby

be developed. The result (Fig. 6) is manifestly evident and our hypothesis is

verified!

Antoine Badan recognized this fundamental relationship and applied it

throughout his scientific career. His consequent influence on the field of

oceanography and on the collegial culture of wine-making in Baja California

will always be remembered.

324

Figure 5. Examples of the wines contributing to the ensemble forcing of our research collaborations on regional coupled ocean-atmosphere interactions.

Figure 6. Wine Forcing Intensity Index versus Research Productivity Output Index for cases of both strong and weak forcing. The highly significant correlation validates the hypothesis.

325

Flow at the Sills Bounding the Delfin Basin in the Northern Gulf of

California

Manuel López ♦, CICESE

Julio Candela, CICESE

In order to conceptualize the currents at the two sills, bounding the Delfin

Basin, we first provide an overview of the circulation and exchange in the

northern Gulf of California. At first sight the currents at both of these sills

appear markedly different. The eastern sill is characterized by having the

largest mean velocities near the bottom in an overflow that discharges water

into the Delfin Basin, whereas the western sill has the largest velocities close

to the surface and directed towards the head of the gulf. The energy of the

subinertial current fluctuations is also quite different, with most of the

energy in the eastern sill concentrated in the lowest frequencies (periods

around 60 days), whereas the spectra at the western sill are not red and much

of the energy is concentrated at periods of 15 days and less, albeit the energy

is concentrated in the lowest frequencies near the surface. Near bottom

currents at the overflow of the eastern sill are well correlated with surface

currents at the western overflow, especially in the lowest frequency bands.

This is interpreted as part of the circulation in which the near bottom flow

into the deepest basins of Delfin and Ballenas is compensated by near

surface flow out of the same basins. Deeper, higher frequency subinertial

current fluctuations (periods of 8 to 20 days) at both basins appear coherent

and have some characteristics of topographic Rossby waves with currents at

the eastern sill leading those at the western sill. Tidal currents at the eastern ♦ Contact: malope@cicese.mx

326

sill are quite strong (M2 amplitudes of about 50 cm/s) and surface

intensified. The fortnightly signal in the western sill appears to be related to

the MSF fortnightly tide, although some of the energy may also be

associated with current fluctuations propagating from the eastern sill.

Tidal Circulation in the Sea of Cortez

Clinton Winant ♠, SIO UCSD

The three-dimensional tidal circulation in the Sea of Cortez is described with

a linear, constant density model on the f-plane. Rotation fundamentally alters

the lateral flow, introducing a lateral recirculation comparable in magnitude

to the axial flow, as long as friction is not too large. This circulation is due

to the imbalance between the cross channel sea level gradient, which is in

near geostrophic balance with the Coriolis acceleration associated with the

vertically averaged axial flow, and the Coriolis acceleration associated with

the vertically sheared axial flow. During flood condition for example, the

lateral Coriolis acceleration near the surface exceeds the pressure gradient,

tending to accelerate the lateral flow, while the converse is true near the

bottom. As a result, with rotation, fluid parcels tend to corkscrew into and

out of the basin in a tidal period. The axial flow is only weakly modified by

rotation. When friction is small, the axial velocity is uniform in each

section, except in a narrow bottom boundary layer where it decreases to

zero. The boundary layer thickness increases with friction, so that with

♠ Contact: cwinant@ucsd.edu

327

moderate or large friction axial velocities are sheared from bottom to

surface. When friction is large, the local and the Coriolis acceleration are

both small and the dynamics are governed by a balance between friction and

the pressure gradient. The three dimensional residual circulation is described

with a small amplitude, constant density model on the f-plane. The inclusion

of rotation fundamentally alters the residual flow. With rotation, fluid is

drawn into the basin on the right side of an observer looking

toward the closed end (in the northern hemisphere), and the return flow is on

the opposite side. A lateral circulation is superposed on the axial flow, with

upwelling over the deeper part of each section and downwelling near the

sides. The residual flow is driven by a combination of advective terms,

including the lateral advection of axial momentum associated with the

Coriolis acceleration, and Stokes forcing.

328

329

Index

A

Acetic acid, 225

ADCP, 68, 85, 131

aerodynamic, 280

Albedo, 281

Anaerobic

(fermentation), 224,

Andean, 49

Anthropogenic, 315

axisymmetric, 124

AVHRR, 284

B

Baccus, 204,

Backscatter, 85

Barotropic, 103, 131, 134

141

Bell labs., 48

Biota, 93

C

Canek, 27, 75

Carbon

(dioxide), 224

Chemistry, 223

Climategate, 52

Climenology, 204

Coriolis, 69, 104

Corrosion, 226

CTD, 131

D

Desert

(Sonoran), 279

diel, 88

Drag

(vanishing), 317

Droughts, 315

E

Eddy, 65

Einstein, 36

Ethanol, 224

ETOPO2, 174

Evaporation, 263

Evapotranspiration, 280

G

Gaussian, 90

Gore (A.), 48

GPS, 136

Grapes, 203

H

Heat waves, 203

330

Himalayan, 49

HOBO, 136

Hyperrenacentism , 12

hyperrenacentist, 13

I

ICTP, 20, 48

ions, 233

J

Jump

(hydraulic), 144

K

KdV, 143

L

Lagrangian, 171

Laguerre, 103, 123

Landsat, 245

Loop (crnt.), 65

M

Model

(Gbl. Rel.), 118

MODIS, 291

N

NDBC, 175

NDVI, 245, 284

NOAA, 285

O

Ocean

(upper), 316

OGCM, 316

Organoleptic, 232,

Oxygen, 225

P

PEMEX, 21, 26, 28

Phenolic, 232

Photosynthetic, 258

Physiographic, 249

Polyphenols, 233

Popper, 53

Precipitation, 279

Q

QuickSCAT, 174

R

Radiation

(solar), 260, 264, 280

reproducibility, 31

review (peer), 35, 42

ROMS, 173

S

Saccharides

(mono), 232

Sand

(wet), 263

SAR, 135

SBE, 136

331

Schön (J.H.), 46

Shelf, 132

Sills, 319

SST, 176

Sugar, 224

T

Temperature

(surface), 279, 315

Tidal

(circulation), 320

tide, 134

tocopherols, 233

topography, 65, 103

transistor, 47

V

Vegetation

(index), 245

Veraison, 205

Vitis vinifera, 224

Vitis (sp), 227

Vinegar, 225

Vorticity, 71, 104, 180

Volatility, 213

Voltammetric, 231

W

Waves

(Cntl. Shelf), 103

(gravity), 104

(internal), 131

War

(World II), 315

Winsurf, 284