Science Briefs

ENSO, Seasonal Rainfall Patterns and Zimbabwe Maize Yield

Figure 1
Maize is a staple food in Africa. (Photo courtesy CIMMYT, CGIAR)

The El Niño/Southern Oscillation (ENSO) is a recurring pattern of interannual oscillations in both sea-surface temperature and sea-level atmospheric pressure in the tropical Pacific which shows strong correlations with climate patterns around the globe. Because of recent advances in ocean circulation modeling, it is now possible to predict ENSO events by as much as one year in advance. This predictability in turn suggests the possibility of successful seasonal climate forecasting, which may have important practical applications.

One region where rainfall is strongly influenced by ENSO is southern Africa, almost halfway around the globe from the Pacific Ocean. Zimbabwe is a strong candidate for the application of ENSO predictions to reduce the risk in agricultural production associated with rainfall variability. In Zimbabwe, maize is the primary food crop grown by small-holder farmers and is preferred over more drought-resistant crops such as millet and sorghum. Jennifer Phillips (GISS/Columbia), Mark Cane (Columbia LDEO) and Cynthia Rosenzweig (GISS) tested the idea that knowledge of the ENSO phase could be used in planting date decisions to avoid the worst impact of drought or to take advantage of a good rainy season in Zimbabwe.

Using 39 years of recorded daily climate data from four sites in Zimbabwe, we ran simulations with a maize growth model parameterized for soils and maize varieties typical of each area. The simulations were run with three planting dates: two weeks earlier than normal, normal, and two weeks later. Records of sea-surface temperatures (SSTs) in the NINO3 region of the Pacific (extending from 5°N to 5°S and from 90°W to 150°W) during November, December and January were used to define which of the simulated years between 1951 and 1989 were El Niño years (years with SSTs well above normal), La Niña years (SSTs well below normal), or neutral years.

All four sites showed a decrease in seasonal precipitation and slight shortening of the rainy season with the El Niño phase, compared to both neutral and La Niña years. At the sites in the drier zones, there was some evidence of an increased likelihood of a short-season drought occurring in January for both El Niño and La Niña years compared to neutral years.

Figure 2
The 1987-88 El Niño: The vertical blue bars indicate rainfall at Masvingo in Zimbabwe. The horizontal green bars mark important stages in the maize growth cycle for different planting dates. Early planting would have shifted pollination to before the January dry spell.

While simulated maize yields were generally lowest in El Niño years, our results showed that early planting in both El Niño and La Niña years could improve yields. For the El Niño years, this improvement was related to the timing of the critical pollination period for the crop. If drought stress occurs during pollination, it can cause fewer kernels to develop on the ears; early planting helped shift the pollination period to December before the January dry spell. In the case of the wetter La Niña years, the relative decrease in rainfall during January was not significant enough to reduce yields, and the advantage of early planting was due to improved nitrogen fertilizer use. Under the normal planting date, the high December rainfall in La Niña years can cause nitrogen to leach out of the soil before plant growth is strong enough to take it up. With earlier planting, most of the plant biomass accumulation occurs before this rainy period, and plants are able to "store" the nitrogen in their leaves before it is washed away.

These initial results imply that predictions of seasonal climate patterns across this region will provide valuable information for crop management. We are continuing this work in collaboration with the government of Zimbabwe to assure that useful forecast information reaches local farmers.


Phillips, J., M.A. Cane, and C. Rosenzweig 1998. ENSO, seasonal rainfall patterns and maize yield variability in Zimbabwe. Agri. Forest. Meteor. 90, 39-50.

Cane, M.A., G. Eshel, and R.W. Buckland 1994. Forecasting Zimbabwean maize yields using eastern equatorial Pacific sea-surface temperature. Nature 370, 204-205.