OOS 91-3
Agro-climatic forecasts for managing risk in Colombian rice cropping systems
Climate variability introduces inter-annual yield fluctuations and economic risk into agricultural cropping systems, especially for farmers without irrigation or crop insurance. In Colombia, it is estimated that climate variability associated with the El Niño/ La Niña phenomenon contributes to ~40% of inter-annual yield fluctuations for rice, maize, coffee and other important commercial crops, with increasingly worse impacts in the most recent decade due to several flooding and severe drought events.
Seasonal climate forecasts for a 3 to 6 month time horizon have the potential to help farmers cope with climate variability by choosing crops, varieties and sowing dates appropriate for the expected conditions. Sea surface temperatures as well as other atmospheric variables provide some predictive power in regards to land-based temperature and precipitation patterns in tropical regions, although significant variables and predictor zones can vary with the sub-region and month of interest. Here, we present results of an evaluation study of seasonal climate forecasts, generated using the Climate Predictability Tool (http://iri.columbia.edu/our-expertise/climate/tools/cpt/), for the departments of Valle del Cauca and Córdoba in Colombia. Then, we present a case study using weather data generated from the forecasts in process-based crop models to identify optimal sowing dates and varieties for Colombia's rice-growing regions.
Results/Conclusions
Preliminary results of the seasonal climate forecast evaluation show that in many months and regions, there is a bias towards "normal", with the forecasts lacking the ability to accurately predict extreme precipitation events. However, other months and regions exhibit higher skill in capturing anomalous conditions. In general, there is higher skill in Valle del Cauca close to the Pacific Ocean, with reduced skill in Córdoba near the Caribbean coast, which is also influenced by Atlantic Ocean circulation patterns.
In the second half of 2014 in northern Colombia, the seasonal forecasts predicted below-normal rainfall associated with an El Niño event. We used the ORYZA2000 crop model with the forecast information to identify optimal sowing dates and varieties for the rice crop during this period. Working with FEDEARROZ, the Colombian rice-growers' union, and the Colombian ministry of agriculture (MADR), CIAT advised farmers to plant early in May. Given economic constraints, many farmers decided not to plant rice at all this year, which helped them to avoid economic losses associated with the drought that occurred later that year.