IntroductionThe 1997/1998 El Niño event wasprobably one of the most importantever observed. The whole planet hasrecorded abnormal events that might berelated to El Niño effects. Some of themare the direct consequence of El Niño:droughts in Indonesia and northernBrazil, floods in Peru, fires in Indonesia[Buongiorno et al. 1997], destruction ofa large part of the Brazilian Amazonianforest (equalling the size of The Nether-lands), all provoking serious environ-mental damage. The economic impacthas been major, ranging from pertur-bation in fishing and farming, to greatnatural disasters causing severe casualties.

El Niño is characterised by a sea-sur-face temperature increase and anabnormal sea-surface rise in height inthe tropical area of the eastern Pacific.

ERS, thanks to its particular payloadconfiguration, was able to provide use-ful information for monitoring the El Niñoevent. Indeed, the Radar Altimeter (RA)provides very high accuracy of the sea-surface height while the Along-TrackScanning Radiometer (ATSR) gives veryprecise measurements of the sea-sur-face temperature. The repetitive cycle ofERS allows a frequent global coverageof the Earth. ERS-2 was able to

observe the premise of El Niño and tomonitor the phenomena throughout theevent. A description of the data usedand the processing performed isdetailed below.

Description of El NiñoThe whole oceanographic and atmos-pheric system of ocean currents andwinds are constantly redistributing thesun energy absorbed by the water andthe atmosphere. In the equatorial zoneof the Pacific, trade winds blow fairlyconstantly from east to west. Thesewinds transport water and cause thesea level to slope upwards from east towest. In the west, water piles up and inthe east, upwelling of water from deepcold layers occurs at the surface. Thetrade winds are maintained by the high-pressure region above South America,where cool and dry air converges, andby the Indonesian low-pressure region,where warm moist air rises, producingcumulo-nimbus clouds and heavy rain-fall.

During the Northern Hemisphere winter(Southern Hemisphere summer), thepressure above South America dropsand the pressure above Indonesia rises.As a result, trade winds become weaker and upwelling diminishes, suchthat less cold, fresh and nutritious waterrises to the surface in the eastern part.

The sea level becomes slightly higherand the ocean temperature increases.Due to this temperature increase, fishpatterns alter. Fish start to migrate tocooler zones and fishing yields becomesignificantly smaller. This annual appear-ance of warmer ocean surface wateralong the South American coast aroundChristmas is what Peruvian fishermenused to call El Niño (Christmas Child).This period of two to three months wasused to repair boats, nets, etc.

However, sometimes this situation getstriggered into an extreme state. Tradewinds are extremely weak and mayeven turn westerly in the westernPacific under the influence of thechanging pressure distribution. Thepotential energy of water (which waspiled up by these winds) in the west isreleased, exciting an eastward propa-gating Kelvin wave (equatorially trapped)and causing a drop in sea level on thewestern side of the ocean. At the east-ern boundary, the equatorial Kelvinwaves split into northward and south-ward travelling coastal Kelvin waves aswell as being partially reflected asRossby waves. This enhances the sealevel and temperature increase on theeastern side of the Pacific, creating aneven lower pressure region and thusdecreasing the trade winds even further.

el niño27

1997/98 El Niño observed by ERSK. Cardon1, P. Goryl1, R. Scharroo2, J. Benveniste1

1ESA/ESRIN, Directorate of Application, Remote Sensing Exploitation DepartmentCP 64, Via Galileo Galilei, 00044 Frascati, Italy

2 Delft University of Technology, Delft Institute for Earth-Oriented Space Research (DEOS)Kluverweg 1, 2629 HS Delft , The Netherlands

The recent El Niño Southern Oscillation (ENSO) of 1997/1998 is one of the strongest of the century. This event is characterisedby abnormally high sea level and sea surface temperatures along the west coast of South America, near the Equator.

With the Radar Altimeter (RA) and the Along-Track Scanning Radiometer (ATSR) on-board the second European RemoteSensing satellite (ERS-2), this event was monitored in great detail. Animations clearly show the development of the latest ElNiño. Particularly, the sudden Kelvin wave at its onset is clearly recognised. It can also be seen that this El Niño actually hit theSouth American coast twice, once in July 1997 and then again in October; both are of great magnitude and initiated unusual-ly early in the year.

This phenomenon, nowadays called El Niño, can persist for a period of several months taking up to an entireyear to return to ‘normal’ conditions.

The set of climatic anomalies referred toas ENSO is the most energetic andbest-defined pattern of inter-annual vari-ability. El Niño is the oceanographiccomponent of the climatic fluctuation.The most obvious sign of an El Niñoevent is the appearance of a massivepool of unusually warm water off thecoast of Peru and Ecuador, whichbuilds up within a period of a fewmonths. The warm water pool can becompared to an iceberg: most of it issubmerged, but a part of it protrudesabove the sea surface as the wedgefloats in the surrounding ocean.

Importance of Earth-observationdataThe most severe El Niño episode inrecent history occurred during 1982-83.The event came as a total surprise andcaused enormous devastation: $ 2.4billion due to flooding, 280 million dueto hurricanes and 5.4 billion due to firesand droughts, totalling over 8 billion(NOAA, El Niño theme page).

Following the 1982-83 disasters, scien-tists and governments set up extensiveobservation projects. The observationsof the atmosphere and the ocean permit the monitoring of such an event,and the data can be assimilated into climatological models. The models arebased on physical laws, but the input ofmeasurements can avoid models fromdeviating too much from the actualstate of the ocean.

El Niño does not appear regularly, butsome researchers claim that the fre-quency might be increasing due to theCO2 gasses in the atmosphere. Thisyear’s El Niño was different for two rea-sons:– scientists forecasted it as one of the

worst; – for the first time, the event was pre-

dicted six months in advance.

Although some damage has been limited to certain areas thanks to theearly-warning predictions (some govern-

ments stocked up on cereals or rice,others advised farmers on their planta-tions and crops, efficient evacuationmethods were put into place, and roadsand bridges were reinforced, etc.), thisyear’s El Niño left scars.

In order to monitor an El Niño event,four parameters are very important:wind, ocean temperature, sea-levelheight and atmospheric pressure at sealevel. The Radar Altimeter (RA) onboardERS-1 and ERS-2 provides sea-surfaceheight (SSH) and the ATSR providessea-surface temperature (SST).

RA instrument characteristicsThe ERS-2 RA is a Ku-band (13.8 GHz)nadir-pointing active microwave sensor

designed to measure the time-returnechoes from ocean, ice and land sur-faces [ESA Bulletin 65, 1991]. Thesemeasurements are repeated 1000 timesper second and are averaged for noise-reduction purposes to provide onemeasurement every 7 km along thetrack.

Satellite altimeters consist of a transmit-ter that sends out short chirped pulses,a receiver to record the pulses afterthey are reflected by the surface below(e.g. the ocean surface) and an ultra-stable and accurate clock to measurethe time interval between emission andreception. The height of the sea surfacecan be determined by differentiating theorbital altitude of the satellite and the

ers 28

ATSR spatial averaged sea-surface temperature coverage for 1 day.

ATSR spatial averaged sea-surface temperature coverage for 4 consecutivedays.①b

①a

el niño29

Onset of El Niño in the beginningof May 1997, start of the elevated sealevel of about 7 cm along the equatorand 17 cm near the west coast ofSouth America.

Developed El Niño which reachesthe South American coast in July 1997with sea-level elevation maxima of 30cm along the equator.

Rebounced El Niño in September1997. There are still considerableanomalies in the central Pacific alongthe equator (max. 30 cm), but the sealevel dropped back to 7 cm off theSouth American coast.

The second hit of the severe ElNiño episode reaches the SouthAmerican coast in its most developedstate on 7 Dec. 1997, with anomalieshigher than 35 cm. Meanwhile the sealevel at the Australian side hasdecreased.

Fading of the intense and highanomalies at the South Americancoast, the signal is partially reflected ascoastal Kelvin waves propagate north-ward and southward along the coast.

②e

②c

②a

②b

②d

radar range measurement. Furthermore,the slope of the echo leading edge isrelated to the height distribution of thereflecting facets and thus to the oceanwave height. Finally, the power level ofthe echoed signal depends on small-scale surface roughness and thus onwind speed.

ATSR instrument characteristicsThe Along-Track Scanning Radiometerdeveloped by a consortium of labora-tories led by the Rutherford AppletonLaboratory has been flying onboard theERS satellites since 1991 [ESA 1992].ATSR-2 has three visible and near-infrared channels centred at 0.55, 0.65,0.86 µm and 4 infrared channels cen-tred at 1.6, 3.7, 10.8 and 12 µm[Stricker et al. 1995].

The ATSR is an instrument using conical scan system producing a

double view of the same surface (55°forward and nadir) with a 512-km swathallowing a full repetitive coverage in 4days with a resolution of 1 km (Fig. 1).

The instrument is equipped with a veryprecise onboard calibration system,which provides an excellent accuracy ofthe measurements. Its high-radiometricsensitivity (signal-to-noise better than0.05 in all MIR and IR channels for tem-peratures > 270 K) [Mason 1991]together with 12-bit digitalisation enablethe detection of fine sea structures andsea temperature variation.

Both of these instrument types, butgreatly improved, will continue to fly onEnvisat-1 to be launched end 1999[ESA Bulletin 76, 1993]. This will ensurethe continuity of the measurements.

RA data flow, products and processingFor the monitoring of the sea-levelheight, the ERS-1 and ERS-2 RadarAltimeter OPR (Ocean Product) [ESA1993] data are used and enhanced withprecise orbits form the Delft Universityof Technology (from April 1992 untilNovember 1997). The sequence hasbeen kept up-to-date (until end March1998) with ERS-2 OPR data takenbefore its final stage. This has allowedthe monitoring of the El Niño event witha delay of only a few weeks. This timelag is necessary in order to:– obtain the data transmitted through

the Prince Albert ground receiving station, which covers the zone around the equator near Peru and Ecuador;

– calculate the wet tropospheric cor-rection, which has a magnitude similar to the El Niño signal.

ers 30

The warm sea-surface signalintensifying and persisting eastwardlyat the end of August 1997. The anom-alies are bigger than 6 K.

Development of warm surfacelayer during the first week of July 1997.The mean anomaly reaches about 5 Kin the eastern equatorial Pacific.

Start of the warming-up of thesea surface in the eastern Pacificaround the equator, measured byATSR during the last week of May1997.

③a

③c

③b

The correction is obtained from measurements of the water-vapour content by the Microwave Radiometer (MWR) onboard ERS-2. This correction is absolutely neces-sary for the observation of El Niño because warm waters stimulate cloud production and water evapo-ration.

These OPR data have been correctedfor bias jumps, oscillator drift, dry tropo-spheric, wet tropospheric and ionos-pheric path delays, solid earth and pole

tides, ocean tides and loading, sea-state bias, inverse barometer. OhioState University MSS95 was taken as areference mean sea-surface height. Theorbit correction from the OPR was notapplied as such, but data refinementwas achieved by performing a minimi-sation of the sea level differences atcrossovers of ascending and des-cending tracks with a maximum intervalof 17.5 days.

The relative sea heights were mappedto regular grids with a resolution of 1° x

1°. The grids are calculated every 7days, but the data span for each solu-tion is 16 days (being a sub-cycle ofERS-2’s repeat cycle) centred aroundthe epoch. The grid mapping uses aGaussian function with a spatial scale of1.2° to weight the along-track measure-ments to the grid points.

ATSR data flow, products and processingThe sea-surface temperature was moni-tored using the SADIST-2 productsaveraged sea-surface temperature

el niño31

Time/longitude diagram of sea level anomalies along theequator observed by ERS-1 and ERS-2 radar altimeters,spanning from April 1992 until March 1998. Kelvin wavesassociated with El Niño events can be seen propagatingeastwardly at a velocity of about 3.6 m/s.

Time/longitude diagram of sea-level anomalies along the 5°Nas observed by ERS-1 and ERS-2 radar altimeters. Rossbywaves migrate towards the west at regular intervals and witha velocity of about 70 cm/s.

④a ④b

(ASST) [Bailey et al. 1995]. RutherfordAppleton Laboratory (RAL) producedthe ASST, which contain spatially-aver-aged sea-surface temperatures at 10-arcmin and half-degree resolution, usingnadir-only and dual-view retrieval algo-rithms. In order to respond to ClimateUser Community requirements(TOGA/WRCP), ATSR has beendesigned to measure sea-surface tem-perature with an accuracy better than0.3 K, which is a significant improve-ment when compared to the previousinfrared sensor [see Zavody et al. 1995for more details].

A composite was made using 10-dayperiods. During overlaps, priority wasgiven to the highest temperature inorder to reduce the remaining atmos-pheric effect.

Because the time range in which theERS products has been available is

relatively short (since July 1991), noaccurate long-term mean has yet beenestablished. Instead, differences of theASST were created with respect toASST data concurrently acquired in1995. This implies that no annual cycleis present in sea-surface temperatureanomalies, as shown in the images.1995 was considered to be a relativelynormal year, without any importantirregularities.

In order to fill the missing data due topersistent cloud coverage, an interpola-tion in time and space has beenapplied.

El Niño observationThis year El Niño was also observed atESRIN using ERS RA and ATSR data,and data visualisation tools. From May1997, a slight rise in the sea level alongthe equator and at the South Americancoast was noticed, probably a Kelvin

wave propagating from west to east,transporting warm water and anincrease in the sea level. By July 1997,the wave reached the South Americancoast and seemed to partially rebouncein August and September 1997. InOctober 1997, the wave reached theSouth American coast for the secondtime, but with more strength than inJuly. It heavily and rapidly intensifieduntil December 1997, from which timethe signal partially rebounced, but alsosplit into northern and southern coastalKelvin waves (Fig. 2). The warm water(higher sea level) was visible all alongEcuador, Peru and part of Chile towardsthe south, and along Colombia, CostaRica, Nicaragua, Honduras, Guatemala,San Salvador and Mexico towards thenorth (Fig. 3).

The propagation of waves can best beseen in time/longitude diagrams (Fig. 4).These plots show the sea-level

ers 32

ATSR image over Chile (South of Atacama desert), SoilAdjusted Vegetation Index. The sea has been masked. Thedata was acquired on 9 Jan. 1997, before the start of ElNiño. The area in the north of the image is very dry (low SAVIvalue). The snow coverage in the mountain area is not veryextended.

Same area as 4a. Data acquired on 25 Dec. 1997 during theEl Niño event. The vegetation as measured by the SAVI hasincreased 25% in area 1 and 50% in area 2. The snow coverage (area 3) is heavier.

⑤a ⑤b

anomalies for particular latitude bandsas a function of longitude (horizontalaxis) and time (vertical axis). Along theequator (Fig. 4a), the Kelvin wavesassociated with El Niño can be seen at irregular intervals and strengths, but allwith a typical velocity of almost 360cm/s. Note that the 1997-98 El Niñoactually consisted of two Kelvin wavesand that the sea-level anomaly wasabout three times stronger than duringthe 1993 and 1994 events.

Just north of the Equator (at 5°N, Fig.4b) Rossby waves can be seen propa-gating westward. Again, these phenom-ena progressed in very sharp trainsacross the Pacific. This took about ayear during which their signal remainedevident. Their propagation velocity wasabout 70 cm/s and decreased towardshigher latitudes.

TeleconnectionsEl Niño affects the climate drasticallyaround the equatorial Pacific but canalso impact all parts of the world. Theeffects of El Niño beyond the equatorialPacific are called ‘teleconnections’.Although the mechanisms for certainteleconnections are not yet fully understood, some effects have beenrecorded each time, or have at least ahigh statistical occurrence probabilitywhen an El Niño events occurs. Thepredictions of the consequences of El Niño have the highest probability forzones closest to the tropical Pacific,such as the droughts in northernAustralia, Indonesia and northern Brazilassociated with big fire events[Buongiorno et al. 1997] and torrentialrains floods along the coast of Ecuadorand Peru.

Using ATSR-2 products it was possibleto show the vegetation increase in adesert area (northern Chile) and snowcoverage after or during the El Niñoevent. This was caused by the unusualprecipitation over the area (Fig. 5).

ConclusionERS missions have proved to be able toprovide useful information for monitoringan El Niño event. Thanks to the RA andthe ATSR instruments, it has beenpossible to observe the first signs of El

Niño and to follow its evolution. In orderto further exploit the great potential ofthese data for ENSO dynamic research,they were made available to oceanmodellers for model assimilation.

The ground data-processing of Envisat-1, ESA’s next-generation Earthobservation mission now under devel-opment, will permit observations of thetropical Pacific with high-quality fast-delivery RA and ATSR data released inless than three hours. This is an assetsince the early detection of the onset ofan ENSO event allows governmentsand individuals to react timely and takeappropriate measures to minimise thepossible drastic impact, for instance:stocking-up on foodstuffs to compen-sate penurious fishing and farming,preparing for drought- or flood-relateddiseases, advising the best course ofaction for livestock, intensifying fire prevention and fire-fighting capacities.

AcknowledgementsWe would like to thank the RutherfordAppleton Laboratory for the ATSR dataprocessing.

ReferencesBuongiorno A, O Arino, C Zehner, PColagrande & P Goryl: ERS-2 MonitorsExceptional Fire Event in South-EastAsia, EOQ 56-57, 1997.

Zavody AM, CT Mutlow, DT Llewellyn-Jones: A radiative transfer model forsea-surface temperature retrieval for theATSR, JGR 100, C1, 937-952, 15 Jan.1995.

ESA: ERS-1 System, ESA SP-1146,1992.

ESA: ERS User Handbook, Rev.1, ESASP-1148. 1993.

ESA: ESA Bulletin 65, 1991.

ESA: ESA Bulletin 76, 1993.

Mason 1981: Satellite Measurement ofSea-Surface Temperature, DoctorateThesis, Univ. of Oxford, 1981.

RAL: SADIST-2 v200 Products, P BaileyER-TN-RAL-AT-2164, Ruth. Appl. Lab.,6.9.1995.

ESAMS 1999

European Symposium onAtmospheric

Measurements from Space

ESTEC, Noordwijk, The Netherlands

18-22 January 1999

In anticipation of the launch ofEnvisat in the year 2000, with itsgroup of three atmospheric chemistryinstruments, it has been decided thatthe time is opportune to encourageinteractions between the variousgroups of scientists working on theseinstruments and those alreadyexploiting data from the GOME(Global Ozone MonitoringExperiment) instrument currently flying on ERS-2.

With this general objective in mind,the European Space Agency (ESA) istherefore organising a meeting atESTEC, namely ESAMS ’99, whichwill have two specific objectives:

– to describe the three ‘chemistry’instruments which will be flown onEnvisat and to present for review thealgorithms being developed for them;

– to provide a working forum for scientists to exchange information onactivities linked to GOME and thechemistry instruments on Envisat.

Further information on arrangementsfor this workshop and the variousinstruments will be found on theESTEC Conference Web site: http://www.estec.esa.nl/CONFANNOUN/.

el niño33