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: [email protected]
�
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: [email protected]
�
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: [email protected]
�
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: [email protected]
�
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: [email protected]
�
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: [email protected]
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: [email protected]
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: [email protected]
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: [email protected]
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: [email protected]
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: [email protected]
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: [email protected]
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: [email protected]
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: [email protected]
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: [email protected]
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: [email protected]
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: [email protected] ♦ Contact: [email protected]
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: [email protected]
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: [email protected]
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: [email protected]
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: [email protected]
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