Comparative assessment of agricultural uses of ENSO-based climate forecasts in Argentina, Costa Rica and Mexico: project summary

Material Information

Comparative assessment of agricultural uses of ENSO-based climate forecasts in Argentina, Costa Rica and Mexico: project summary
Added title page title:
Agricultural uses of ENSO-based climate forecasts in Argentina, Costa Rica and Mexico
Jones, James Wigington, 1944-
Place of Publication:
Gainesville, Fla.
University of Florida
Publication Date:
Physical Description:
26 p. col. ill., col. maps ;28 cm.


Subjects / Keywords:
Ocean-atmosphere interaction ( LCSH )
Caribbean ( LCSH )
University of Florida. ( LCSH )
Farming ( LCSH )
Agriculture ( LCSH )
Farm life ( LCSH )
Climatic change -- Forecasting -- Argentina ( LCSH )
Climatic change -- Forecasting -- Costa Rica ( LCSH )
Climatic change -- Forecasting -- Mexico ( LCSH )
El Nino Current -- Forecasting ( LCSH )
Southern oscillation -- Forecasting ( LCSH )
Spatial Coverage:
North America -- United States of America -- Florida


General Note:
"An IAI Initial Science Program. (ISP)-III Project." "This summary is an abridged, general audience version of a more technical report submitted to the IAI in October 2000."
Electronic resources created as part of a prototype UF Institutional Repository and Faculty Papers project by the University of Florida.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact Digital Services ( with any additional information they can provide.
Resource Identifier:
60338631 ( OCLC )

Full Text
Comparative Assessment of Agricultural Uses of ENSO-Based Climate Forecasts in Argentina, Costa Rica and Mexico
tl )I I D
An IAI Initial Science Program (ISP)-III Project
Coordinated by the University of Florida Dr. James W. Jones, Principal Investigator

About this summary
This summary is an abridged, general audience version of a more technical report submitted to the IAI in October 2000. The full report, which is almost 200 pages long, contains numerous graphs, charts, and statistical analyses for each of the three countries which participated in the project, as well as a comparative section which evaluates the project as a whole. Full bibliographical references are included. A copy of the full report in Adobe Portable Document Format (PDF) can be downloaded from the Florida Consortium web site:

Flood control retention area in Irrigation District 10, Sinaloa, Mexico.
Introduction .......................................... 4
James W Jones, Principal Investigator
Acknowledgements ..................................... 7
Argentina: Instituto Nacional de Tecnotogia Agropecuaria .......... 8
Graciela 0. Magrin, Maria I. Travasso, Gabriel R. Rodriguez,
Diego R. Boulldn, Santiago Meira, Edgardo Guevara
Costa Rica: Instituto Metorol6gico Nacional ................... 12
Roberto Villalobos, Jose Retana
Mxico: Instituto Mexicano de Tecnologia del Agua .............. 16
Jaime Collado, Angel Villalobos
Comparative Assessment: University of Florida ................ 20
Fred Royce, James W Jones, Graciela 0. Magrin, Roberto Villalobos, Jaime Collado
Glossary of terms ..................................... 26
Funding for this project was provided by grants from the
Inter-American Institute for Global Change Research, and the
National Oceanic and Atmospheric Administration
Office of Global Programs.

Intod 6o
Dr. James W. Jones
Principal Investigator
Since the 1982-1983 El Nifio event, often referred to us to determine the overall usefulness for a wide as the "worst ever," the El Nifio Southern Oscillation range of ENSO-based climate prediction methods. (ENSO) has been a popular news item. The very As hoped, the participation and enthusiasm of scienstrong El Nifio and La Nifia events of the 1997-1998 twists and researchers was matched by feedback from season, especially the fires, flooding, and agricultur- end users-our third goal of establishing a relationally destructive freezes which occurred in the United ship of trust between researchers and end users. States, established the reputation of ENSO as a powerful force. Results
As the work of this project shows, a more complete The project summarized here showed the promise of
understanding of ENSO can be very valuable to ENSO-based climate prediction. Researchers working
farmers and agricultural decision makers in our at universities, government research centers, and for
hemisphere. I am proud to introduce this summary government agencies worked with farmers to model
of an Inter-American Institute for Global Change ENSO effects and educate those potentially affected.
Research (IAI) collaborative research project for In each country, project cooperators isolated the
investigating the application of ENSO-based climate features unique to their situation-in some cases prediction in agriculture. This IAI Initial Science developing very diverse techniques due to the wide Program (ISP)-III project began in May 1997 and variety of locations studied.
concluded in July 2000. In Argentina, farms are often mechanized and are the
Project goats most technically advanced of those studied. Argentine
farmers are the most ready to respond to pressures of Unlike some other climate prediction research efforts tegoa akt hygo ievreyo
which have investigated ENSO-based methods, this cr o a ma ize s ob a adesunflowers. project focused on end users-the people who could use the information gained from ENSO-based climate prediction to manage crops, set policies, and make other decisions. Our efforts began with farmers, but also included research scientists, political leaders, and government officials.
Five areas in three countries were selected for study: the Pampas and Pergamino regions of Argentina, the Los Chiles and Liberia regions of Costa Rica, and the State of Sinaloa, Mexico. The three sections of this summary each contain maps which show the study areas in detail.
The project was very diverse. Researchers worked in a variety of disciplines and professions. Many agricul- Back row, from Left: Brad Wilcox, former IAI science officer, tural systems, producers, climates, and institutions with project researchers James W. Jones, Jaime Coliado, were represented. Agricultural infrastructure varied Jos6 Retana, and Roberto Villalobos. Front row: project widely across study locations. This diversity enabled researchers GracieLa Magrin, Maria Travasso, and Fred Royce.
Photo taken at International Potato Research Center (CIP) Lima, Peru, May 2000.
4 Agricultdral Uses of ENSO-Elased Climate Forecasts in Argentina Costa Rica and Me co

Project Summary 5
About ENSO and
climate prediction
ENSO is short for El Nifio Southern
A crop advisor and Oscillation-the technical name for oceanic L
farmer discuss the and atmospheric conditions in the Pacific
current growing Ocean which have a tremendous effect on
conditions and weather patterns in our hemisphere and
expectations. The
crop shown is maize. around the globe. ..
ENSO phenomena are defined when sea Figure one: elevated SST
ENSO phases definitely alter precipitation patterns surface temperatures (SST) in the tropical in equatorial Pacific
during 1987 EL Nihio event
and crop yields. However, high variation in the loca- Pacific Ocean differ from normal. In the (Courtesy NOAA).
tion and intensity of changes caused by ENSO make warm phase, better known as El Nifio, SSTs
climate prediction's economic value somewhat tow are higher than usual. During the "La Nifia" cold phase, SSTs are
(about 5%). cooler than the norm. Awarenes of El Nifio is nothing new-hunIn Costa Rica, there are a wider variety of farm sizes, dreds of years ago, South American fishermen noticed weather changes around Christmas of some years, and named this warming most in the small to medium range. Much of the "lNio fe h hitcid
"Et Nifio" after the Christ child.
work is done by hand, and there is little irrigation.
ENSO neutral years are more common than either cold or hot phase The focus of project research was end ENSO years. Since 1950 there have been 22 warm, 18 cold, and 23
users-those who could use the neutral years. Ocean temperatures affect climate by providing a
source of heat and moisture. This affects air temperature, wind curinformation gained from ENSO-based rents, and perhaps most importantly, the timing and amount of
climate prediction to manage crops, rainfall. That's why ENSO can have huge impacts on crop production.
set policies, and make other decisions. What is climate prediction?
Climate prediction is the science of using historical weather inforRice and beans are the principal crops. In the north, mation and up to date weather observations to predict broad ctimatEl Nifio decreases rice yield, while La Nifia increases ic patterns-unlike weather forecasting which focuses on the short yield. But bean crops show tittle effect. Overall, term and is very localized. Most climate predictions look from three
ENSO-based climate prediction has helped to miti- weeks to a year into the future. Enhanced capabilities to monitor
gate losses caused by weather patterns. However, the oceans and atmosphere, and computer modeling of weather
infrastructure is weak, and communication is the data, continue to improve the accuracy of climate prediction.
greatest challenge. Why is ENSO-based cimate prediction important?
In Mexico, farms are small and farmers are well Because ENSO can drastically affect the amount and time of year
organized. ENSO effects are not large enough to jus- when rainfall occurs, climate prediction can be a very important tool
tify effective alteration of crop management prac- for agricultural decision making.
tices. However, irrigation district management infrastructure provides a unifying influence, and using For example, if climate prediction shows more rain in the early part
ENSO-based forecasts to better manage irrigation of a growing season, a farmer would select crops or hybrids which
resources seems very promising. benefit from that early rain. If less rainfall or higher temperatures
are predicted, a farmer could delay planting or choose a drought-tolThe comparative work of University of Florida and erant crop-or decide not to plant at all.
other scientists offers an in-depth analysis of results gthesienetd f s all hrereinsetu singf r t But climate prediction is not a magic bullet. Work is needed to improve data collection, modeling, and methods for getting data to final project reports (see page 2) and interviews with farmers. Those goals are at the heart of the current study. farmers as primary sources.

6 Agricultural Uses of ENSO-Based Climate Forecasts in Argentina, Costa Rica and Mexico
Introduction, continued
Future work will continue the development of climate forecasting techniques. But as this study demonstrated, ways in which climate forecast information is delivered to the end user are equally important.
Overall the study showed clearly that there is considerable potential for climate prediction applications in each of the three regions studied. It is likely that other regions could benefit from similar climate prediction. However, there are risks associated and more work is needed to make certain
Future work will continue the development South American schoolchildren ike these can benefit directly
of climate forecasting techniques. from the improved yields and Loss prevention possible with ENSObased climate prediction.
This study demonstrated the ways that raw information is delivered to end users are equally important. Learning more
To learn more about the work encompassed by this these risks are fully understood. This summary, and the po e t toe fuv th reportbwhi
full report which goes with it, are one step toward under- cta amt 200 paes of taileprt reslt
standing the ways researchers can work with end users to an als See page 2.
enable more effective decision making in every level of agricultural practice. There are many web sites which provide excellent
information on El Nifio Southern Oscillation (ENS0), Dr. James W. Jones is a Distinguished Professor in the ifraino lNfoSuhr silto ES)
Dr.aJamet one Agisultur a nd istinguid PoE ineg including historical data, up to the minute observations
Department of Agricultural and Biological Engineering of Pacific Ocean conditions, and references to relevant
at the Universityt of Florida. research projects both on and off the Web.
The NOAA Tropical Atmosphere Ocean (TAO) theme page contains graphical and charted representations of ocean conditions and links to many other useful sites: el-rno/
The Center for Ocean-Atmospheric Prediction Studies (COAPS) at Florida State University also offers valuable materials pertinent to ENSO and climatology in general. The COAPS team continually updates its site to reflect current research with relevance for scholars, agricultural decision makers, and the general public:

We wish to thank the following individuals and We are also grateful to:
organizations who provided financial support, advice, Enrique Mejia of the Colegio de Postgraduados,
information, and other assistance which contributed Texcoco, Mexico
to the success of this project.
* names Hansen, Columbia University Partial funding for the comparative assessmenteJa sHneCoubaUirit
International Research Institute for Climate portion of the project was provided by the National Prediction (IRI)
Oceanic and Atmospheric Administration (NOAA) Office of Global Programs (OGP). Anthony Hall, Universidad de Buenos Aires.
Funding assistance for the production of this report The amount of work project researchers put into this was provided by IAI. Thanks to Marcella Ohira study may not be represented by this brief summary.
Schwarz for her suggestions about the formatting of We want to thank all of them one more time, and this document. look forward to more work in this exciting area of
Many farmers and growers in Argentina, Costa Rica, global research. and Mexico spent hours working with researchers, answering questions, sharing information, and cooperating in extremely important ways. Without their commitment this work would not have been possible. Fred Royce provided many photographs, the best of which we are proud to include here.
Several members of the Florida Consortium,
especially Dr. Guillermo PodestA of the University of Miami, provided valuable
assistance and advice as well as conducting preliminary research which facilitated much of the work of this comparative analysis.

Instituto Nacional de Tecnotogi Agc-, cropecuaria I T
The diversity of crops cultivated in the Pampas Annual precipitation varies from 400 mm in the
region, the growing importance of agriculture in the southwest to 1100 mm in the northeast. Rainfall area, and the profound effects of ENSO on Pampas generally increases from southwest to northeast.
agriculture make study of ENSO in Argentina very The highest rainfall occurs in December and March
productive. This study focused on three sites in the (up to 120 mm) and the lowest in July and August area: Pergamino, Pilar, and Santa Rosa. On-farm trials (often less than 20 mm). Rainfall is the main factor augmented the historical analyses and modeling which determines crop yield by affecting the
common to the other participant countries, reproductive phases as well as the severity of
fungal diseases.
The study region
The sometimes severe climate of the Pampas creates The Pampas Region is a broad, fiat plain between high variability in crop production. Grain output
30-39 South latitude and 57-65 West longitude, averages 37 million
The area has occasional rolling hills and rises gener- tons (Mt) but has varally to the north and west. Generally speaking, soil ied from 23 to 59 Mt quality does not constrain agriculture, though recent between 1988 and expansion of land use into marginal areas has created 1998. Recently soybean the risk of land quality degradation in some areas, has been the most The climate is temperate and humid with no dry important crop,
season-rainfall normally occurs during all months- followed by wheat, Pampas farmers often attend farm
and a very hot summer. Mean annual temperatures sunflower, and maize. exhibitions to shop for the latest
range from 14'C in the south to 19C in the north. However, since land in machinery, such as this four
wheel drive artculated tractor.
Weather can be extreme, with droughts and excessive the Pampas is used for precipitation leading to both grazing and cropping, the area sown as well as flooding and losses the yield of a particular crop can fluctuate greatly
in planted areas. over time. Figure three (at the far right of the next page) shows annual production for the area. Crop production in Argentina has increased in recent years as the technological sophistication of farmers Figure two: land-cover has increased. Between 1991 and 1996 agricultural map of Argentina shows investment increased from US$300 to US$700 million, Pthe Pampas Region the use of fertilizers quadrupled, and irrigation instalstudy area. lations appeared for the first time. Crop management procedures modernized as well, with less aggressive
__ Referencias tilling, use of herbicides and insecticides, and drying
Boso, Chaqueno and storing of grain becoming more common.
Eosques Subantarticos
Bosques y Pastizales de la Serania Gentrales
Bosques y Sabanas del N. E Production systems vary from pasture and crop
E'stapa Patagoreca
Fstapas y Desienos de Montana del N. 0.
Monte Occidental rotation in the south and west to almost continuous
P Parque Chaquenio
Pa~izales Patep .....cropping in the northeast. Farm ownership has
Selva Misonera
increased from 43% in 1947 to 75% today. Farm size L> has also grown, with many medium size farms and a
8 Agrculturat Uses of ENSO-Based Climate Forecasts in Argentna Costa Pica and Mexico

Project Summary 9
smaller number of very large farms, though most farms (93%) 60 60
are smaller than 1,000 hectares (ha). 5 Maize
5 50 Wheat 50
ENSO effects on climate and agriculture 0 40 Sunflower 40
=30 Soybean
-0 30 30
Historical analysis of data from government statistics showed 2 Total
that El Nifio has a significant impact on yield at the national 0 20: 20
level. In 15 of 20 El Nifo events (75%) maize yield was equal 10 10
to or as much as 36% higher than the mean historical value, 0 0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 while in 17 of 23 La Nifia events (74%) it was as much as 56% 19 lower than the mean. Figure three: annual crop production for maize, wheat, soybean, and
sunflower (1900-1997) in the Pampas region.
In soybeans, an even greater impact from La Nifia was observed, with yield reductions in 71/0 of events. In El Nifio Wheat shows no significant difference at any site, though and neutral years, yield residuals were positive in 58% and trends do follow the general pattern established for maize. 62% of cases, respectively. (See figure four, below, and Agricultural decision makers
figure seventeen, on page 22.)
The high spatial and temporal variability of weather on the Sunflowers presented less association between yield and ENS0 apscniulycalnesae eiinmkrmkn
phass, ithonlyonetred: 59%proabiity f ahieing Pampas continually challenges area decision makers, making phases, with only one trend: a 59%I probability of achieving the availability of climatic information a necessity.
higher yield during La Nifia years. Wheat showed the least response of all, with only 57% of La Nifia years showing lower Many Argentine government agencies provide climatic inforyields. Neither sunflowers nor wheat showed consistent trends mation at monthly or weekly intervals. Historical information during El Nifio years. is also available. In climate forecasting, both the National
An analysis at the county level showed broad reproduction of Meteorological Service (SMN) and the INTA provide monthly forecasts which consider ENSO impacts on rainfall and temperthese national trends, but indicated several areas where ENS0 tue ature.
effects on crop production consistently differed from the norm.
A 1999 survey by Ignacio Llovet (University of Belgrano, Historical analyses of maize and wheat yield, harvested area, A res anave Letson (University of Mianvi
and ropprodctin wre cmpltedfor hre aras rpreen- Buenos Aires) and Dave Letson (University of Miami) investiand crop production were completed for three areas represen- gated the relationship between farmers, climate variability,
tative of humid (Pergamino), sub-humid (Pilar), and semi-arid and te reition the sure iniate-omwat (Santaand climate prediction. The survey indicated-somewhat unexpectedly for an area where climate is so important-that In Pergamino and Pilar, maize crop production is significantly 30% of farmers were "indifferent" to climate prediction. reduced in La Nifia years, due to yield reduction in Pergamino Product prices were the most important factor influencing and yield and harvested area reduction in Pilar. In Santa production decisions (50%), and agronomic factors such as
Rosa, maize production tends to decrease in La Nifia years and crop rotation were second (29%). Climatic forecast was the increase in El Nifio years, though results were not significant most important factor for only 8% of farmers. because of great variability.
E0 1 00 ...
Figure four: yield anomalies in Neutral mEl Ni'o uLa Nifta
A r g e n t i n e s o y b e a n c r o p 5 0 0 1 .l.'.'.'.l.
(kg/ha), 1941-1995. 1941 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995

10 Agricultural Uses of ENSO-Based Climate Forecasts in Argentina, Costa Rica and Mexico
Argentina, continued
When farmers do consider climate variability, 66% draw upon their own experience, and 20/o on the experience of "elders." Argentine
researchers count
Only 12/n rely on technical or formal knowledge. and weigh maize
However, a majority of farmers indicated they considered cli- kernels as part of
~validation of the
mate forecasts before all else during the 1997-1998 season- simulation model
likely due to the well-publicized El Nifio event which occurred for new hybrids.
that year. This event may have affected farmers in several trials of those options were carried out during the 1998-1999
ways. First, the accuracy of predictions from that year may and 1999-2000 growing seasons.
have created unrealistic expectations for many farmers. Survey data shows that 42% of farmers made decisions for For maize, advancing the planting date and increasing
1998-1999 based on a forecasted La Nifia. Second, the success fertilizer worked well during El Nifio years, while delaying of 1997-1998 resulted in several very different forecasts being planting and reducing fertilizer were good options for published in 1998-1999. This confusion, coupled with the La Nifia years.
inaccuracy of even the most reputable forecasts that year, Wheat was less uniform. In the north, better results were
may have caused some toss in the overall credibility of obtained by planting earlier during La Nifia years. In the
climate forecasting. south, advancing planting dates during El Niho was the best
Analysis of decision options option. Nitrogen applications needs varied widely.
The Decision Support System for Agrotechnology Transfer With good soil water availability, early sowings of soybean
(DSSAT) was used with crop models to optimize management during El Niho years and late swings during La Niha years
practices and net returns for crop (wheat, maize, soybean, were the best options. Significant exceptions to these results
and sunflower) as well as for ENSO phase (El Nifio or La Nifia). underscored the importance of climate variability. Climatic inputs were obtained from four meteorological sta- Analyses of sunflower growth indicate early sowing is the
tions (Pergamino, Santa Rosa, Junin, and Nueve de Julio). best option for El Nifo conditions, but offer no differences
under La Ninia conditions. More fertilizer should be used To test crop model performance at the field level, experimen- dur L Niniyears. tal trials were carried out on farms all over Argentina. The during Et Nifio years. necessary data were available from fields in Pergamino, Alberdi, Carmen de Areco, and Catrit6. Simulated yield showed good performance (figure five, below).
Management options including planting dates and nitrogen fertilization rates were defined for the farms selected, and
6000 F 15000 0 Mgure five: observed yield and predicted yield
O aa5000 (both kg/ha) simulated by computer models for
0 Soybean 0 maize, soybean and wheat from several farms
4000 1 .... selected for on-farm trials.
'5, 30000
2000 50
0 0
0 1000 2000 3000 4000 5000 0 5000 10000 15000
Observed yield (kg/ha) Observed yield (kg/ha)

Project Summary ii
The economic value of climate forecasting was calculated for maize, soy- 35
bean, and wheat, as the difference between net returns obtained using his- 30 -US $
torical data and returns obtained considering the three forecasted phases 25
(see maize forecast values in figure six, at right).
20 -Naturally, the value of forecasting depended to some extent on product 15
price, though maize results for Santa Rosa indicate clearly that in some 10
years more drastic reductions are quite possible. 5
Conclusions and recommendations 0
" El Nifio and La Nifia alter precipitation patterns and crop yields in Pergamino JunRn Santa
the Pampas Region. 6
" ENSO effects show high spatial and temporal variability (difference in 5 % m
strength depending on location and time of year). 4" Early planting dates and higher nitrogen fertilizer doses are likely 3
the best strategies during El Nifio years. During La Nifia years plant- 2 ing dates should be delayed and nitrogen fertilizer doses reduced. 1
" High spatial and temporal variability diminish the economic value of 0
forecasts based solely on ENSO phase (El Nifio, La Nifia or neutral).
Figure six: economic value of $85
" The economic value of the forecast increases substantially under ENSO-based climate forecast for
extreme economic conditions (low product prices or high input prices). maize in three locations U $100
(Pergamino, Junin and Santa El $175
" Based on tong-term probabilities, climate predictions can help Rosa), considering three product
farmers manage climate-related risk. prices (US$85, 100, and 175). The top graph
shows value in US$/ha; the bottom graph as
Generally speaking, Argentine agricultural decision makers are extremely a percentage of gross margin.
interested in climate forecasts, especially after the El Nifio event of 1997-1998. The collaboration, enthusiasm and goodwill typical of farmers involved in this project confirm the value of forecast information. However, farmers' expectations may exceed the capacities of climate forecasts due to the variability of ENSO effects. Though selecting a crop or altering management practices if it is a La Nifia, El Nifio, or neutral year wilt likely benefit farmers in the tong term, climate variation could cause discouraging short-term economic losses in any particular year or site.
Argentine researcher inspects a
soybean experiment

Instituto Meteorol6gico Nacional
Costa Rica is a noteworthy area for ENSO climate areas. Like Los Chiles it is a small region, measuring
forecasting for two reasons. First, agriculture consti- approximately 70 km along its widest (north-south)
tutes 19% of the GDP and employs 24% of the eco- axis. The prevailing vegetation is deciduous and
nomically active population. Second, Costa Rica semi-deciduous lowland forest, and it is classified as
shows a strong ENSO influence on climate variability, tropical dry forest transition. The proximity of the In Costa Rica, this study examined the ENSO impact Pacific Ocean affects Liberia, where the dry season
on the bean crop in the northern Caribbean area (Los lasts from December through May. Mean annual preChiles), on the rice crop in the northwestern Pacific cipitation is 1,653 mm with September averaging
area (Liberia), and on the climate in both areas. 343 mm and January averaging only 4 mm.
The study region Temperatures are fairly consistent; 33 C is the
annual mean.
The Los Chiles region is an hilly area located in the
extreme north of Costa Rica, near the border with This area is heavily agriculNicaragua. Its vegetation and soil are classified as tural and was once considlowland tropical rain forest. The climate of Los Chiles ered the granary of Costa is heavily influenced by the Caribbean Sea. The mean Rica, although a series of
annual precipitation is 2,223 mm, with March the dry spells have reduced the
rice yield in recent years.
driest month and July the wettest. The temperature Neveteless, rice reasa
varies little over the course of the year, with a mean
average temperature of 261C. basic commodity and its
production has historically
Land use in Los Chiles has traditionally been exten- received preferential treatsive, with permanent crops, livestock, and forestry ment from the state. A Costa Rican farmer
each playing an important role. Bean production is of makes a point regard
Rice production in the nis rice crop.
increasing importance in this area, both as a major LiberauReion in the
source of protein and a cash crop for local farmers.
hands of medium and large growers who cultivate Los Chiles is currently one of the most important an e frm 1t 7 a l and s ubsiste
bean production areas in the country. anywhere from 10 to 770 ha. Small and subsistence
growers make up only a small and dwindling percentLiberia is located west and slightly south of Los age. Cultivation is increasingly concentrated, with 1%
Chiles, in the Northwestern Pacific region. It is a geo- of growers utilizing 23% of the cultivated land. In
graphically diverse area that includes both gently contrast, bean production in the Los Chiles area
undulating river valleys and ruggedly mountainous remains in the hands of small and medium growers, with most plots between 20 and 200 ha. W _ile the crops and type of producer are very different in these two areas, both are heavily dependent upon rain distribution. In addition, proximity to the Caribbean lO and the Pacific also affects agriculture in both areas.
Figure seven: digital elevation map (DEM) of
[ the study area in Costa Rica, showing both the
- Los Chiles and Liberia areas.
-86o -85A -84o 83o
12 Agricultural Uses of ENSO-Based Climate Forecasts in Argentina Costa Rica and Mex'ico

Project Summary 13
ENSO effects on climate and 3000
crop production 2500In Liberia, rainfall figures are above 1- 2000 average in La Nifia phases and below E t5oo+ average during El Nifio phases (see_ IOOOG figure eight), both annually and
during the growing season. 500_Additionally, during El Nifio phases, 0 1
maximum air temperature increases, 1957 1960 1965 1970 1975 1980 1985 1990 1995 1998
although minimum temperatures are Figure eight: annual rainfall (mm) for Liberia region Cold (La Nia)
not consistently different from those of Costa Rica, 1957 to 1998. Neutral
- Warm (El Nifio)
of other phases. During La Nifia,
temperature were not significantly different from neutral years. Agricultural decision makers
In La Nifia phases non-irrigated rice yield surpassed the mean As a result of the impact of ENSO-related dry spells and yield, but during El Nifio, three out of five (60%) were below floods, decision makers are concerned about climate average (figure nine). This tendency is associated with rainfall variability. Actions implemented to mitigate dry spells and during the growing season, which increased during La Nifia floods are now based on ENSO forecasts and potential effects
and decreased during El Nifio. The major issue is rainfall dis- on the climate of Costa Rica. While this has made it possible tribution, not rainfall quantity. During flowering, water avail- to issue some timely forecasts, other problems, such as loan ability is critical to success in grain development. However, suspension, have arisen due to weather uncertainty and protoo much water can ultimately diminish yield and quality of duction hazards. Only recently (since 1982) has the ENSO
the grain, effect on agriculture been seriously studied. Most of these
Historical climate records for Los Chiles do not show ENSO- studies have focused on ENSO warm phases and the attendant
related statistical differences for either annual or seasonal dry spells in the Pacific watershed. These studies are based precipitation, or for mean temperature. Bean yield compar- largely on information submitted by cooperatives and farmers,
isons likewise showed little ENSO influence. These data indi- and lack scientific analysis. cate that ENSO phases are less important for this region. Prediction of the development of the 1997 ENSO warm
Other climate variations, like cyclonic activity and northeast- episode, as well as the spread of information and bulletins, ern wind, have a greater impact. helped the National Emergency Committee (CNE) coordinate
responses to the consequences of this episode. Recommendations for sowing time and resource exploitation
- -s -100-0
1980 1985 1990 1995 1998
Figure nine: rice yield anomalies (kg/ha) in Cold (La Nia)
Liberia region of Costa Rica, 1980 to 1998. -Warm (El No)

14 Agricultural Uses of ENSO 1Based Climate Forecasts in Argentina Costa Rica and Mexico
Costa Rica, continued
based upon information from
the National Institute of s advances in
Meteorology (IMN) helped n nrvtS
maintain yields of rice, losses i n
maize, citrus, melon, beef
cattle and sugar cane within While professional decision makers are usually very concerned
recent averages. This was with ENSO-based climate forecasting, political decision makers
considered a success both for are often ill-informed, and their attitude toward adverse
agriculture and for the deci- weather conditions is, at best, unclear.
sion makers involved, largely Researchers and rice farmers Weather variations during crop
because of the experience meet in the field near
gained and new emphasis on Liberia, Costa Rica. growth may affect production
climate data and forecasting. and yields, but weather is not
the only influence. If forecasts
Despite these advances in forecasting and planning, ENSO provide alerts of risks, flexible
events have caused notable losses in agriculture. Figure ten management mechanisms should (below) shows the losses caused by the ENSO droughts of be implemented. However, due to
1997-1998 (primarily in the Pacific and central regions). the problems ranging from a tack
Some of these losses could potentially be reduced. However, of support for alternate crops and obstacles which limit actions to prevent losses include: the behavior of grain importers,
" limited implementation of weather event contingency the potential for implementing
plans (often due to the lack of forecasting credibility) flexible management is limited,
" difficulty maintaining and expanding the national especially for small farmers (see this sugarcane cr neremeteorological network the Comparative Assessment from climate pred cfors.
which begins on page 20).
" bureaucratic obstacles, such as the lack of
national action plans Analysis of decision options
" a restructuring of insurance policy that has left many One option for ensuring production is crop rotation during producers without coverage adverse weather conditions. In spite of flooding associated
with La Nifia, the correlation between La Nifia and high rce a lack of research specific to certain crops or areas.
yields in Liberia indicates that rice cultivation should be encouraged during La Nifia phases. Conversely, 93% of dry. spells correspond to El Nifio years, as shovwn in figure eleven, Product Losses Cost at right. Lower rice yields are associated with these dry years,
(million US$) and crop rotation might appear to be an alternative, yet suitRice 40,425 tons 11.4 able crops such as maize and sugar cane are also affected by
Bean 13,598 tons 10.9 extended dry spells. A better alternative might be forage
Maize 8,083 tons 1.7 crops for cattle, or fruit trees, although these require a large
Sugar cane 200,000 tons 3.0 investment. Irrigation remains the best option for preventing
Others* n/a 6.2 damage to rice cultivation during dry spells.
Beef castle 2,000 head 7.2 The Decision Support System for Agrotechnotogy Transfer
Poultry n/a 1.4 (DSSAT) CERES-Rice computer model was used to study other
Figure ten: tosses in agriculture caused by drought, alternatives for rice crops. With management optimized by
1997-1998. "Others" includes orange, melon, paLmetto, ENSO phase, model results indicate that rice yelsnt
tuber, mango, papaya, and ptaintain.
Costa Rica

Project Summary 15
Liberia region during ENSO cold phases could reach
3.5 tons/ha with a growing cycle of 119 days. Warm phase yields could also be increased above current levels. In the northern (Los Chiles) region, although there was no relationship between bean yields and ENSO episode, bean crops in this region have been severely affected by weather events, especially dry spells and heavy rains. For the 19992000 bean cycle, growers in the northern region were told to change the sowing time from mid-November and midDecember to the first half of January. According to the conversations held with some growers, those who followed the
suggstins ereableto arvst n te drestpar ofthe Costa Rican farmers in the research area reported that economic cycle, which is ideal for maturing and drying the grain, conditions have decreased their capacity to purchase and maintain
Conclusions and recommendationsfamaciey
The atinalMetorolgicl Istiute(IMN curenly ses Climate variability has affected basic grain production in
information from various sources to predict rainfall, which is Costa Rica, though its effect varies strongly over time and in increasingly being used in agriculture as well as other sectors. difrnloaos.Ithaesexmeuigvreis
The nfomaton aailbleto ate as eenwellrecive by with shorter growing cycles and coordinating planting time decision makers at the technical and professional levels. wt xetdrifl a mrv ieadba rdcin
A real agricultural forecast is not yet available in Costa Rica. Political decision makers are often il-informed, However, demand continues to grow for this type of service.
and their attitudes toward adverse weather More studies of the actual ENSO influence over crops and
conditions are, at best, unclear, agricultural regions are needed.
However, work still will be needed to get political decision makers to develop real, sustainable, multi-sector plans for the medium and long term.
Year Liberia Caflas Filadeli a Santa Cruz
1972 D D D
1976 D D D D
1977 D D D D
1982 N N N
1983 N N N
1986 N N D
1987 D D D
1991 D D N
1992 N D N
1994 D N N
1997 D D D D
Figure eleven: annua[ precipitation during years influenced by the ENSO warm phase in four stations of the rice producing area of the Liberia region. N (grey): norma[; SR (blue): slightly rainy; R (green): rainy; SD (orange): slightly dry; D (red): dry.

Instituto Mexicano de 1
Tecnologia del Agua L1
The work of Mexican researchers included the only Reservoirs Rivers
B Sanabona 2 ~m
irrigated agriculture considered in this study. The Sonora C Jos Loaz P Tmaz
area investigated showed much weaker ENSO effects 4'" "ca
than other countries considered. The small but very Chihuahua
weU-organized farmers of the Mexican irrigation dis- Sinaloa
tricts showed that climate prediction data could be .
useful in planning even if it was not very relevant to A
farm-based decision making. t 2
Gulf of Calif ornia B
The study region
Several areas of Mexico were studied, principally ID-010
Irrigation District 010 (ID-010), Culiacan-Humaya-San Lorenzo, in the center part of the State of Sinaloa in Pacific Ocean
northwestern Mexico. The district is a 2,700 km2 Coastline *a Reservoir
area, mostly flat coastal plain with some low hills. ---- State line ID-010 Climate is diverse, but generally warm and temperate, River Culiacan River basin
withmos aveagetempratres rom24 t 26 C.San Lorenzo River basin with most average temperatures from 24 to 26 C. Figure twelve: detail of Sinatoa, the irrigation district, and
Precipitation is irregular, with rainfall generally high- the surrounding area, showing the two river basins, rivers, er in the northern mountains and lower along the and reservoirs.
producing 4.043 million tons annually. Crops include The irrigation districts are well basic grains and forages, produce and vegetables, and
organized, with elected farmer sugar cane and fruits. The average volume of producrepresentatives playing a large role in tion in the district is almost 14 tons/hectare (ha).
district management. Agriculture is the basis of Sinaloa's economy and
directly influences development including commerce southern coast. Most rainfall occurs between July and and construction. In Sinaloa, 1,333,450 ha are used October. Average precipitation in the study area for agriculture of which 743,500 ha are irrigated.
ranges from 586 to 710 mm. Current trends in land use show a gradual increase in
the proportion of irrigated crop land.
ID-010 is noted for the variety of crops in its irrigated agriculture fields. Many products are cultivated, The irrigation district is divided into three systems, six units and sixteen irrigation modules with a surI .face of 272,800 ha. There are 27,500 district users
30' with an average farm size of about 10 ha. Three
rivers and three reservoirs are interconnected through canals and other waterworks (see figure 25- twelve, above). The irrigation districts are very well
Gulf of Mexico organized, with elected farmer representatives play20 Ocean Pacific ing a large role in district management.
SrFigure thirteen: map of Mexico showing the 15 State of Sina[oa to the Left in green.
-1 15 -1 10 -10 5 o -10 0 -9 5 -9 0 o
16 Agricultural Uses of ENSO-Based Climate Forecasts in Argentina, Costa Rica and Mexico

Project Summary 17
ENSO effects on climate and agriculture Agricultural
For most weather stations in the district, analyses showed decision makers
that the ENSO effect is not significant in either precipitation A wide range of agricultural E A TR
or temperature. There is, however, a weak impact on stream- decision makers, from small LTA
flow during the fall-winter season. As shown in figure twelve farmers to national irrigation I A& &
(on left, previous page), streamflows originate in river basins district planners, stand to mostly outside the irrigation district itself, and therefore do benefit from climate predic- office: "The water you waste today not depend only, or even mainly, on precipitation in the dis- tions. District farmers are rel- will be missed tomorrow." trict. In particular, El Nifio events have coincided with an atively unconcerned about increase of water available for irrigation. Historical analyses on-field rainfall, but they must make a variety of planting were also performed for the most important crops in the irri- decisions. Module-level planning includes determining the gation district, and for rain-fed agriculture for Sinaloa as timing and amount of water to release from each reservoir, well. These analyses found little or no ENSO significance for estimating a crop mix that is feasible given available water, any agricultural indicator. Finally, statistical analyses were and allocating resources to specific maintenance needs within done for other Mexican states that are important producers of the system. corn, and for grain production in Mexico as a whole. The very limited ENSO effects found can be reviewed in the full report. ENSO influence increases the availability of Availability of water in the district is closely related to the irrigation water during El Niio events, yet streamflows in the Culiacan and San Lorenzo watersheds. An analysis of streamflow shows ENSO impacts on yearly, fall-win- decreases availability in the La Nihia phase. ter, October, November, December streamflows in the San
Lorenzo River and the watersheds. Figure fourteen (below) Several government agencies provide climate and weather
shows the impact on streamflow. information which aid decision making. Most of these forecasts are used in the short term only, and are not routinely used for tong-term agricultural management and planning.
This irrigation canal near
Culiacan, Sinaloa needs
maintenance, yet scarce
resources must be divided
between irrigation and
drainage works. Accurate
climate predictions will
permit allocations based on
expected conditions.
Average annual streamflows, million cubic meters (Mm3) Location La Nifia Neutral EL Nifio
Humaya River, Cuiacan watershed 1,772.8 (-8.55%) 1,938.6 2,212.0 (+14.11%)
TamazuLa River, Culiacan watershed 784.4 (+3.91%) 754.8 904.4 (+19.82%)
San Lorenzo River San Lorenzo watershed 1,517.9 (+0.360%) 1,512.4 2,067.7 (+36.71%)
Total ID-010 streamflow, Culiacan + San Lorenzo 4,075.0 (-3.11%) 4,205.8 5,184.1 (+23.26%)
Figure fourteen: effect of ENSO phase on average annual streamflows for three locations in ID-010. Percentages shown are differences from neutral years. The amount of usable irrigation water in each district is closely related to streamfLows.

18 Agriculturat Uses of ENSO-Based Climate Forecasts in Argentina, Costa Rica and Mexico
Mexico, continued
Mexican farmers and decision makers are broadly aware of The project plan proposed to use the Decision Support System
ENSO, due to newspaper reports and observed impacts in agri- for Agrotechnology Transfer (DSSAT) package to optimize crop culture and fishing. In 1991-1992 excessive rain and floods management practices based on ENSO phase, taking advantage damaged 55,000 ha with economic losses of US$54.5 million. of positive conditions and mitigating the impact of negative
For many farmers, the economic effects of these problems ones. Unfortunately, there were not enough maize experimenlasted well into 1997. In part because of awareness generated tal data available from this district to enable DSSAT validation by these losses, most look favorably on the use of climate and use. The suggested model-based methodology was ultiforecasting in agriculture planning. From irrigation district mately employed in Mexico however, using data from prelimiauthorities to individual farmers, decision makers at all levels nary experimentation in different regions of Mexico (see have expressed interest in working with researchers to devel- "A second irrigation district" on the far right, next page). op methods for using ENSO-based climate prediction.
AnaLysis of decision options Interviews indicate that Mexican producers
Although historical production data do not reveal a signifi- are eager to use climate forecasts to cant ENSO impact on Sinaloa agriculture, interviews indicate introduce modifications in agriculture that producers are eager to use climate forecasts to introduce management options. modifications in agriculture management options, such as changing planting and harvest dates, rotating crops, selecting On the other hand, climate forecasts could be used in resercrops and crops varieties for cultivation, consuming water for voir operation policies. Up to now, ENSO activity has not been irrigation, using fertilizers, and choosing tillage practices. a factor in the annual determination of reservoir operating rules. Since, as is demonstrated in this study, ENSO is an important factor in streamflow into some reservoirs, it may be possible to define operating policies based on ENSO activity. Streamfiow effects An example of how to define ENSO-based optimal operating
Following a year of slightly below normal or above rules has been presented in this work. (see "Streamflow
normal streamfiow into the reservoir, water extraction effects," below left).
will be substantially reduced during La Nifia years. In addition to ENSO, other persistent and perhaps predictable
2500 meteorological phenomenon affect Mexico's climate. The eventual integration of these into climate models should lead to 2000.. ........... .... ............ greatly improved predictions.
S1 500 /_____7-__.... Whole record
500 La Nina years
Neutral years
- El Niho years
0 1000 2000 3000 4000 5000
Previous year streamflow (Mm3)
Figure fifteen: water availability estimates for the Jose Lopez PortiLo Reservoir, Sinatoa, as a function of ENSO
phase and previous year streamflow.

Project Summary _19
A second irgation distrit The Alto Rio Lerma irrigation district in the north central Mexican state of Guanajuato is organized along the same lines as the Sinaloa district.
Unlike the Sinaloa farmers, those in the Alto Rio Lerma district count on rainfall to provide part of their spring-summer crop water requirements. There is usually sufficient reservoir water for one irrigation to start the crop, and sometimes enough for a second watering, but without late-season rains, there is no harvest. Furthermore, since the Alto Rio Lerma district is at higher elevation, over 1,700m above sea level, crops SoLis dam and reservoir in Guanajuato, Mexico, are sometimes damaged by freezes and hail.
The relationship between these climatic vulnerabilities, field Conclusions level management, and ENSO phase was explored using the
Although the ENSO signal is not statistically significant DSSAT crop simulation model. Although this maize model has for Sinaloa agriculture, producers are eager to use cli- been used in a number of studies in Mexico, no data from sysmate forecasts to introduce modifications in agricultur- tematic comparisons of its performance compared to actual al management. crop results were available. Therefore, these encouraging
Preliminary results from the Alto Rio Lerma irrigation results are only tentative.
district show a role for ENSO-based optimization of crop Best management varied substantially between La Nifia and management using simulation models, even under irri- El Nifio phases.
gated conditions. Simulated yield under La Nifia conditions is improved by earStreamflow in certain rivers was one of the few ENSO lier planting, more irrigation, and more nitrogen fertilizer.
effects established. This indicates that climate forecasts
* El Niflo climate provides the potential for substantially could be used in reservoir operation policies, higher profits.
* The value of adjusting management to ENSO phase may be over US$25 per hectare.
Engineers from the irrigation district in SinaLoa indicate the placement of a planned flow-monitoring station.

UFLORIDAThis section of the summary compares the methods opening markets to competition from imports, and
used and results obtained in the three countries stud- permitting local producers to compete globally. ied. The primary source for this comparative work is Overall, this liberalization appears to make econorithe full project report, augmented by surveys, a crop catty weaker producers more vulnerable to market management optimization methodology, and observa- fluctuations, and is concentrating ownership and tions from visits to the three countries studied. control of agricultural land.
Focus on the erd user' Participation in the world market is not new to
Pergamino farmers. Argentine grain producers have
This project emphasized understanding end-user atti- ee n di retl sub e t i w r m r o ces l n e tudes toward climate prediction and the application than grain farmers of other areas examined. Their of predictions to agriculture. Farmers were recognized large, highly mechanized farming systems have the as important end users. The inclusion of other deci- economic efficiency required to compete worldwide. sion makers from government and industry enriched Mot entin rduree to compete
the eserch esuts.Most Argentine producers seem to welcomneth
the research results. opportunities offered by reduced barriers to trade.
Locations co ar d aNearly the opposite situation exists in northwestern To focus the comparisons, a single site from Costa Rica. Liberalization has increased hardship and
Argentina and Costa Rica, and two sites from Mexico uncertainty for most rice growers in the Liberia are highlighted. Two Mexican sites are presented region. The weakening or removal of government
because different application methodologies were price guarantees and oversight of seed
used for two areas. import restrictions has affected the economic viability of many smaller rice producers, and most view
For Costa Rica, Liberia in the Chorotega region is the globalization as a severe threat. obvious choice, since its climate shows a strong ENSO correlation. On the Argentine Pampas, the zone During the 1990's nearly
around Pergamino in Buenos Aires province is the all Mexican irrigation
most extensively investigated, districts experienced a
transfer of management :.i
The situation in Mexico is quite different-none of the trs from the federauthorityfrmteedlocations examined show a strong ENSO influence. In al government to elected spite of these limitations, adjustments were made to representatives. In the the same methodology used in the Argentine and Costa two districts examined in Rican studies, and implemented for Santa Julia in the this study this transfer From te am to te fec state of Guanajuato. A different approach toward using appears to have been suc- outlet, Mexcan faanie s are climate predictions in agricultural planning was cessfut. Not only are the organized to manage e
demonstrated for the other Mexican site, the Culiacan- districts better managed Humaya-San Lorenzo Irrigation District in Sinaloa. at a local level, but the effort has help are HistorV a[ nd soiau contextorganize and deal with the problems of globalizaon.
Each of the agricultural areas under consideration has been powerfully affected by globalization, the gener- The cases selected span a considerable range of cal shift toward a uniform, enforceable set of rules mate and ENSO effects, but include an even greater governing international trade. In agriculture, these diversity of production systems. Usually comprizing rules focus on reducing government intervention, hundreds of hectares, the Pampas farms around
20 u U casts Costa Rca a ,

Project Summary 21
Pergamino are the largest of those considered. These lands 350
tend to be highly mechanized, though usually not irrigated. 300 =Liberia
Usage of hybrid seed and chemical inputs is common. E Sinaloa
E20 -Santa Julia / \\
At five to ten hectares, the average-sized farm in a Mexican 200
irrigation district is tiny by Pampas standards. Yet district 150
farmers have access to irrigation, thus reducing a key source of 100 uncertainty and risk. Tillage and harvesting of the usual grain 50 crops (wheat, maize, sorghum and beans) are mechanized, o
though securing affordable, timely equipment services is often Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
a challenge. Figure sixteen: monthly mean rainfalL at four locations studied.
Blue: Pergamino; Green: Liberia; Yellow: Sinaloa; Red: Santa Julia.
The rice growers of the Liberia region of Costa Rica occupy a
middle ground in size, between ten and 500 hectares. The In spite of nominal membership in cooperatives, small and
smatter farms have the lowest market interaction of any medium rice farmers in the Liberia region are not very well
examined here: little credit is available, seed is selected from organized. Furthermore, they face a very adverse economic the previous harvest, there are few chemical inputs, and the situation, including low credit availability, competition from farm family supplies most of the labor. Larger farmers own imports and reduced government support. Each of these facmuch of the usual field machinery, yet even their operations tors appears to limit the practical appeal of incremental, tend to be partially mechanized: planting, fertilizing and farm-level innovation.
chemical weed control are often performed by hand. Few of As a group, the Mexican irrigation district farmers are by far
these upland rice farmers have access to irrigation. the smallest landowners within the overall study. Relatively
few own field machinery, or computers. Yet their capacity for Overall, globaLization appears to make adopting and benefiting from innovations may rival much
economically weaker producers more larger Pampas growers due to their high level of organization.
vulnerable to market fluctuations, and is Although the system exists to operate and maintain the irrigation infrastructure, its effects go well beyond water control concentrating ownership and control of and delivery. At all levels, there was enthusiasm for better
agricultural Land. solutions, backed up by a growing local support infrastructure
including computers, automatic weather stations, and techniThe smaller farmers of Liberia do not commonly practice crop cal personnel. Farmer participation in collective decision-makrotation, although they may change crops in response to mar- ing, and communication between different decision-making ket conditions. Nor do livestock play a significant role in their levels both appeared to be functioning. operations. By contrast, farmers around Pergamino commonly ENSO influences on climate
rotate some combination of wheat, maize, soybeans and cattle, and on many large farms, operations are vertically inte- Rainfall was judged to be the ENSO-related climate parameter grated to include grain brokering, transformation of agricul- most relevant to agriculture in each location. As seen in figtural products into livestock feed, and livestock into food ure sixteen, above, all locations show some tendency toward products. Despite small farm size, more Mexican farmers are higher rainfall in summer than in winter months (winter in also vertically integrating through cooperative arrangements Pergamino, Argentina, is from late June through late within their irrigation districts. September).
End user attitudes Both Pergamino and Liberia rainfall data show strong ENSO
influence, with the months most affected by ENSO occurring
First impressions from Pergamino indicate a farming popula- during the season for the crop of interest. In Pergamino tion of "early adopters": conservation tillage, computer-based El Nifio is the wetter phase, while in Liberia La Nifia brings accounting systems, and genetically modified (GM) crops are above average rainfall. all becoming common.

22 Agricultural Uses of ENSO-Based Cimate Forecasts in Argentina Costa Rica and Mexio
ENSO influences on crop productio Among all the crops analyzed for this study Argentine maize yields showed the strongest correlation to ENSO phase (figure seventeen, below). ENSO phase precipitation differences fal squarely during the maize growing season, and the low rainfall that characterizes La Nifia years creates the likelihood of water stress during maize flowering (figure eighteen, top far right).
Rice yield data from Liberia show some ENSO influence, with yields generally above average in La Nifa years. The correlaDistrict engineers in Sinaloa demonstrate a new method for tion is much weaker than Pergamino maize, due to the comdelivering water to this field of maize paratively reduced risk of drought, market fluctuations, and
The timing of the ENSO effect differs between Pergamino and fewer years of data available for analysis. northeastern Costa Rica. The most important months for No significant ENSO signal was observed for crop production or
ENSO-related precipitation differences are July and August in yield data from either irrigation district in Mexico. This is not Liberia and November and December of the same year in surprising, given the short data series, the relatively weak
Pergamino. Therefore, the lead-time for decision-making based ENSO-rainfall correlation, and the availability of irrigation.
Rainfatll was judged to be the ENSO-re[ated Results of model-based analyses
climate parameter most relevant to For El Nifio years in Pergamino, higher water availability assoagriculture in each location. ciated with El Nifio during the mid-growing season permits
higher yield, particularly if plant density and nitrogen fertion a particular ENSO phase prediction, is about four months izer are increased. The late-season (January-February) longer in Pergamino than in Liberia. increase in La Nifia rainfall favors later planting dates that
result in more available water during the crop's grain ll[ing
Compared to Pergamino and Liberia, statistical analysis of result in more available water during the crop's gra rainfall data from Santa Julia, Mexico shows weaker ENSO phase.
influence. Even so, the most important phase-related rainfall The value of ENSO optimized management showed sy-mmetry differences occurs in August during the maize growing between the Nifio and Nifia phases: knowledge of either of
season, making ENSO knowledge a potentially useful criterion these two conditions has practically the same value, while for determining planting date, irrigation amount and timing, there is practically no value to knowledge of a neutral year. and possibly other management options. Like the Pergamino example, in Santa Julia, Mexico the most
widely divergent planting dates occur in El Nifio and La Nina optimizations, and greater economic value is associated with ENSO-based management during the El Niflo and La Nifia 1000
Figure seventeen: yield
anomalies (kg/ha) for maize in
Pergamino, Argentina, from -1000ra :
1902 to 1995. ENSO phase is
shown by color. -1500 I I 1
1902 1910 1920 1930 1940 1950 1960 970 1980 '9 9
Comparative assessment

Project Summary 23
phases, rather than the neutral phase. Best simulated man- 450
agement for the drier La Nifia phase included an additional 400 -- Rice: Liberia, Costa Rica
irrigation application, higher nitrogen application and _350
increased planting density. Since model results for Mexican V50
varieties had not been validated, the absolute yields and 200
profits cannot be meaningfully compared between the 5 150
Pergamino and Santa Julia results. 100
ENS0 phase optimization of rice for the Liberia region of Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar
Costa Rica was begun, but only preliminary results were Jul & Aug significant (P <.01) I Plant I C., D-eel .. I
achieved. Excess water from heavy La Nifia rains is probably less of a problem than water stress during El Nifio years. 180
The higher water availability during the La Nifia years 160 Maize: Pergamino, Argentina ....
(figure eighteen, at right) appears to permit the crop to uti- 140 c 120 -lize more nitrogen and attain higher yields. Also in Costa "2
" 100
Rica, simulation model runs using climate predictions led 80
researchers to recommend that farmers delay planting beans 2 6 .
At one Argentine site, the 1999-2000 20 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar
season was a typical La Niia year: very dry. ov &Decsignificant(P<.01) l i
The recommended treatment outperformed Figures eighteen and nineteen: monthly
both the El Niho management and the rainfall by ENSO phase with the crop Cold (La Nirha)
development calendars. Top: maize in Neutral
rs usual management. Pergamino, Argentina, 1931-1998. Bottom: Warm (El Nifio)
rice in Liberia, Costa Rica, 1937-1997.
during the 1999-2000 season from the usual period of mid- Growing seasons are shown in bars below the graphs. Variations in
November through mid-December, until the first half of average monthly rainfall between ENSO phases indicate potential
January. Indications from the field are that the farmers who effects on yield due to water stress during critical phases of plant
followed the recommendations have benefited. development.
Argentine researchers also tested simulated results at various Pampas locations during the 1999-2000 La Nifia. In most locations the recommendations corresponding to La Nifia did not perform the best, probably because in most Pampas areas the rainfall recorded for 1999-2000 was well above the average for La Nifia years. At a site near Pergamino, the 1999-2000 season was typical for a La Nifia year: very dry. In that case, the treatment recommended for La Nifia indeed outperformed both the El Nifio management, and the farmer's usual management.
Argentine researcher
demonstrates computer
simulation results at a
1999 trade show
"Expo Chacra."

24 Agricultural Uses of ENSO-Based Climate Forecasts in Argentina, Costa Rica and Mexico
User attitudes toward ENSO forecasts Climate predictions, even with a lead-time of up to a year, are not entirely new in the areas studied. In Liberia (Costa Rica) and around Pergamino (Argentina), farmers were eager to recount traditional methods of weather and climate prediction that are still used by some in those areas. The use of traditional forecasting methods may affect a farmer's ability to understand the uses and limitations of scientific forecasts. Conversely, local terminology, time frames, or other concepts of traditional systems may be useful guides in the application of our forecasts to agricultural activity. Grassland in Costa Rica, used primarily for grazing
Individuals' beliefs regarding specific ENSO effects have shifted rapidly in some locations. The 1997-1998 El Nifio event Potential applications
brought above normal rainfall to much of the Pampas, con- In each of the areas compared, decision makers expressed or
firming predictions issued by meteorologists, leading some to even demonstrated ways that climate forecasts could benefit believe that knowledge of the current ENSO phase was equiva- agriculture. At the field level, ideas varied with the nature of lent to a near-certain spring-summer rainfall forecast. But, the production system, and the options available to farmers. the ensuing "ENSO euphoria" was short-lived; however, as nei- Larger Pampas growers stressed adjusting the amount of land ther of the two subsequent La Nifia years was as dry as pre- in each crop; irrigation district members focused on timing of dicted. In the Mexican irrigation districts, several years of planting; only the small farmers of Liberia suggested that well-publicized, but not very accurate predictions have left given an unfavorable prediction, they would consider not serious doubts regarding the usefulness of ENSO as a climate planting at all. predictor. In Costa Rica awareness of the ENSO climate effect is high, but with limited options, forecasts have not led to Beyond the field scale, Argentine farmers and commodity broi y highdoptwitited otichns, kers were intrigued with the possibility of using climate forewidely adopted management changes. casts for other grain-producing regions to inform decisionmaking. In Costa Rica, some institutional responses were detrimental to farmers: insurance companies reduced coverage for ENS-related crop loss, grain importers increased sports, lowering prices, and some banks have begun to condition agricultural loans on predicted climate, further restricting scarce agricultural credit. In Mexico, irrigation district administrators suggested prediction-based allocation of resources for maintenance of irrigation infrastructure: if a drier season ENSO Phase Al Years is predicted, focus on water delivery canals; if a wetter season
ENSO Management Management ENSO is predicted, on drainage ditches.
Phase Yield Profit Yield Profit $/ha
kg/ha $/ha kg/ha $/ha
El Niho 7701 280.85 7150 260.18 20.67
La Nifia 5216 154.93 5395 134.67 20.26
6678 226.86 6677 226.32 0.54
El Nino 8967 138.85 8561 113.13 25.72 Figure twenty: value of ENSO
La phase for maize production using La Niria 9099 86.50 7667 48.12 I38.38 historical weather data.
8601 104.83 8349 97.72 7.11 Top: Pergamino, Argentina.
Bottom: Santa Julia, Mexico.
Comparative assessment

Project Summary 25
ImpLications of Research For climate forecasts to be applicable to smaller farmers like
What is the current state of climate predictions some of those in the Costa Rican
in agriculture? portion, they will have to be
well targeted, and probably cornAll researchers involved in the project expressed the need for bined with other organizational, improved forecasts, whether based on atmospheric/oceanic technical, or social initiatives.
simulation models, or on improved statistical models that besides ENSO will include other predictable phenomena such The irrigation districts stand out as Atlantic sea surface temperatures. Future advances since some decisions are based
notwithstanding, caution will always be required when mak- on events occurring over several ing recommendations based on climate predictions: they will months at a watershed scale. Maize experiment at
sometimes be wrong. Decision makers in irrigation dis- Pergamino, Argentina.
tricts can integrate predictions at
The integration of climate prediction into agriculture appears spatial and temporal scales that are well suited to most cito represent a useful technology for farmers and other agri- mate prediction methodologies. Furthermore, in the Mexican cultural decision makers. In particular, climate prediction case rrigtion d tt sd of ismall, twel ora n e Merscav research in agriculture should probably retain a strong focus the cra d nation ad supporenel tomus he
on sattr frmes. Mthos dvelpedwithsmaterfarers the central administration and support personnel to use the on smaller farmers. Methods developed with smaller farmers kidodaattclmepriton osofnpove.
willbe pproriaed y lagerfarmrs f demeduseul;kind of data that climate prediction most often provides. will be appropriated by larger farmers if deemed useful; whereas those developed exclusively with a large-farmer focus What partners should be sought for development may not be readily applicable on resource-limited small farms. and application of methodologies?
Where can climate predictions be most Researchers lamented the limitations imposed both by scarce
appropriately applied? international cooperation and shrinking national government
budgets. Part of the solution may be to form partnerships The accuracy of existing predictions for a particular agricul- with farmer organizations. This project received significant tural area, the scale of farming operations, and the agricultur- cooperation and input from a variety of organizations: the al and economic options available to growers are some of the irrigation district user associations in Mexico, and the farmer factors that affect the relevance of climate prediction to a organizations Argentine Association of Regional Consortiums given operation. Blanket statements regarding applicability for Agricultural Experimentation (AACREA) and Cambio Rural that do not account for specific conditions are bound to be in Argentina. The possibility of such organizations providing misleading. direct budgetary support for this type of research should be
Large, technologically sophisticated producers, such as many investigated for the future. of those in the Argentine portion of this study, often have the management flexibility to incorporate climatic forecasts. copen on the ost proisin
consequence of this multinational study, which could not have been accomplished without collaboration among researchers with diverse talents and among institutions with diverse mandates.
A fled worker with onions in
hand explains how tomatoes
are harvested in ID-011,
Guanajuato, Mexico.

anomaly-difference from observed or historical Instituto Metorol6gico Nacional (IMN)-Costa
norms. Rican National Meterological Institute.
climate prediction or climate forecasting-the Instituto Nacional de Tecnologia Agropecuaria
science of using historical and/or observed weather (INTA) -Argentine National Institute for Agri:tua data and climatic patterns to predict weather trends Research. on a long-term and large-scale basis. Instituto Mexicano de Tecnologia del Agua
Committee Nacional Emergencia (CNE)-Costa (IMTA)-Mexican National Hydrotechnology
Rican national emergency committee Institute.
crop management-the growing techniques used in La Nifia-ENSO cool (or cold) phase, characterized by
agriculture, including planting and harvest negative SST anomalies.
scheduling, variety or crop selected, crop rotation, National Atmospheric and Oceanic Administration and pesticide, herbicide, fertilizer, and irrigation (NOAA)-United States government agency usage. responsible for a large portion of ENSO research, as
crop modeling-computer based calculation of a well as the Tropical Atmosphere Ocean (TAO) project,
particular crop's life cycle from planting to harvest. which is critical for ENSO detection. Often used with historical and/or forecasted weather remote sensing-usage of satellite or other data to aid in agricultural decision making. technology for observation of natural phenomena.
Decision Support System for Agrotechnology Used frequently to collect weather data.
Transfer (DSSAT)-A computer-based decision sea surface temperature (SST)-temperature of
support system developed by a consortium of water on the surface of the ocean, which affects
universities and technical institutions, including ocean currents, wind patterns, rainfa, and many many involved in this project. other weather events. SST anomalies usually indicate
El Nifio-ENSO warm (or hot) phase, characterized by an ENSO event. positive SST anomalies. significance-the statistical importance, or lack
El Nifio Southern Oscillation (ENSO)-the global thereof, of the result of a scientific experiment or set
weather patterns characterized by SST anomalies near of observations. Often expressed as a number the equator in the Pacific Ocean, east of South between 0 and 1, with lower numbers representing
America between 10N and 100S latitude and 1400E higher significance.
and 100aW longitude. streamflow-the measure of the amount of water
globalization-the liberalization of trade through flowing through streams or rivers the removal of regulatory barriers such as tariffs and variety-the breed or hybrid of a crop. Different import/export restrictions, varieties can have very different growth
hectare (ha)-Metric unit of area equal to 10,000 characteristics, drought tolerance, expected yield,
square meters. One hectare is about 2.5 acres. insect resistance, etc.
Inter-American Institute for Global Change
Research (IAI)-an international intergovernmental organization, headquartered in Brazil, which
facilitates an increased understanding of global
change phenomena and their socio-economic
consequences for the Americas.
26 Agricultural Uses of ENSO-Based Climate Forecasts in Argentina. Costa Rica and Mexico

Printing by Warecraft Press, Jacksonville, FRorida.
Copyright 912001 University of Florida.

Left: Oil painting of researchers at IPergamino, Argentina measuring the results of a soybean experiment. Cover background: Irrigation canal in need of maintenance near Huacho, Peru. Cover: Costa Rican researcher, extensionist and farmer compare notes on ENSO cmate variability (top); detail from "SugaIndustry" by Diego Rivera, Cuernavaca, Mexico (middle); Mexican farmer transporting irrigated sugarcane for processing (bottom).
Agricuttlral & Biological Engineeing
PO Box 110570
Gainesvte FL 32611 0570