Soybean yield in the Pampas region (Argentina): relationship with climate variability and analysis in the context of future climate change

Argentina is the third major soybean producer in the world after United States and Brazil (http://faostat.fao.org/site/339/default.aspx). The 84.5% of soybean total production takes place in the Pampas region which is located in the center east of Argentina, southeastern South America, and has an extension of more than 600000 km2.

Crop growth and yield are largely determined by weather during the growing season. Local and daily scale variables depend, among other factors, on the large-scale atmospheric fields. Therefore, the study of the circulation structures related to local weather, their frequency, distribution and temporal variability are important elements for diagnosis and forecast, particularly in the context of future climate change.

In this context, the present work is structured to fulfill the following objectives: to identify daily circulation patterns in southern South America and to associate them with surface climate (precipitation and temperature) in the Argentine Pampas region; and to analyze soybean yield interannual variability in the region in relation to the identified atmospheric circulation patterns. As a complementary study, the simulations of these atmospheric structures by the GCMs during the 20th century and their projected changes at different time horizons of the 21th century were analyzed.

Different data-sets were used in this study: a) Daily mean sea level pressure (SLP) fields, corresponding to the NCEP reanalysis 2 data provided by the NOAA/OAR/ESRL PSD, Boulder, Colorado, USA: http://www.cdc.noaa.gov/, were used as representative of observed circulation for the period 1979-2000. The chosen domain extends from 15ºS to 60ºS and from 42.5ºW to 90º W on a 2.5º latitude-longitude grid. This area includes the Pacific and the Atlantic Oceans and the Andes Mountains, which have a significant influence on the atmospheric circulation over South America. b) Daily rainfall, maximum and minimum temperature series located in the Argentine Pampas region, provided by the Argentine National Meteorological Service. c) Soybean yield series in provincial districts in the Pampas region were used. These series were supplied by the Ministry of Agriculture, Livestock and Fishing of the Nation for 1979/1980–1999/2000. d) GCM outputs from the climate of the 20th century (20C3M) were used to describe present climate and the SRES A1B 720 ppm stabilization experiment were used to represent future climate. The set of 12 GCMs analyzed outputs is available at the Program for Climate Model Diagnosis and Intercomparison (PCDMI) and from the ENSEMBLES CERA archives. The analysis focused on austral summer (DJF) which is the season that coincide with key stages of the growing season of soybean (flowering: end of December to beginning of January; pod setting: January; grain filling: February).

Cluster analysis was performed coupled with PCA to determine the NCEP dominant circulation types (CT). The cluster analysis was carried out in the subspace given by the leading unrotated PCs. SLP fields from the GCM outputs, were classified using the cluster centroids from this (NCEP) original typing. The resulting classification has five circulation type categories.

The most persistent is CT4 that corresponds to an intensification of the southern Atlantic anticyclone, which interrupts the passage of the eastern perturbations and diverts them to the south. This anticyclone induces stability at low levels and can be significantly associated to dry days and high minimum temperatures in the region. CT5 corresponds to an intensification of both high pressure cells (over the eastern South Pacific and western South Atlantic), favoring warm and humid advection from the subtropical zones. It is significantly related to high maximum and minimum temperatures over the Pampas region. Negative significant correlation is found between frequencies of CT4 and CT5 and the anomalies of soybean yield over de Pampas region. This would indicate that the combined presence of these CTs during the summer would induce an adverse effect on the yield as it increases the thermal and water stress conditions during flowering and pod setting.

Local rainy days and low maximum and minimum temperatures are significantly benefited by patterns with a cyclonic disturbance at the centre of the continent (CT1), favoring cold advection and cloudy days. Heavy rainy days (precipitation greater than the 75th percentile) are significantly related with an intensification of the southern Pacific anticyclone, which favors the entrance of perturbations over the continent (CT2). CT3 is significantly related to low maximum temperatures as it could be related to cloudy conditions. The interannual variability of soybean yield is directly associated with frequencies of CT2 and CT3. Years with high values in yield anomalies are accompanied by high frequencies of these daily patterns. Both CTs induce anomalies in the flow which reduce thermal and water stress during flowering and pod set stages producing a positive effect on the final yield.

For the present climate, the GCMs underestimate the frequency of the CTs that favor heavy rainfalls of the region (CT2), while they estimate reasonably well the frequency of atmospheric situations that favor dry days and high minimum temperatures (CT4). An inter-model variability of the representation of CT1, CT3 and CT5 frequencies is observed.

Projections for the 21st century, indicate no clear tendency in the frequency of CT1, CT2 and CT5 while a trend of reduction in the frequency of CT4 is observed. This result coincides with other climate change results over the region that shows positive trends in the total seasonal precipitation and extreme temperatures as well as changes in the precipitation variability. The majority of the GCMs indicate an increment of CT3 frequencies for future climate.

These results would imply a change in the distribution of rainfall and extreme temperatures over the Pampas region and also a trend of conditions conducive to the development of soybean in the region.