Recent scientific and technological advances have made it possible to predict, with usable skill, the occurrence of ENSO events with lead times of several months. Many assume that ENSO-related climate forecasts will benefit the agricultural sector by allowing farmers to mitigate potential negative consequences of climate variability or, alternatively, capitalize on potentially beneficial effects. However, we modify that the availability of climate forecasts is not sufficient to ensure that the agriculture sector will benefit from enhanced climate information. We propose a framework for the successful end-to-end application of ENSO-related climate forecasts. To illustrate this framework, we focus on the agricultural sector of Argentina.
The first element required is a significant ENSO signal on the regional climate and agriculture. We are exploring ENSO signals on climate variables of agronomic interest (precipitation, temperature, and radiation). Also, we are quantifying associations between ENSO phase and crop yields and net economic returns. Simply documenting an ENSO impact does not imply a benefit from adoption of climate forecasts. Climate forecasts must induce changes in the decision-making process, and on the actions taken by sector agents. The second required element, thus, is a decision-support system to evaluate the consequences of alternative management actions in response to an ENSO forecast. To this effect, we are using synthetic weather series, biophysical crop models and non-linear optimization techniques. Differences between results of optimal actions with and without climate information provide an ex ante estimate of the economic value of a forecast. We are evaluating model-based management recommendations through on-farm tests.
The third necessary element is a skillful climate forecast at a regional level, and in a format appropriate to feed process models of interest (e.g., crop models). We are developing and evaluating a high-resolution regional climate model nested within a global ocean-atmosphere model. To produce realistic daily weather sequences (a notorious problem for numerical models), we have developed stochastic weather generators re-scaled to match numerical model monthly output. This tool also provides a formalism to incorporate error estimates in model forecasts to the risk assessment process.
Finally the willingness to adopt the more complicated, flexible management in response to a climate forecast is a function of (a) the decision makers’ perceptions of ENSO-related risks and forecast quality, and (b) financial, technological, institutional, or cultural constraints. To identify impediments for forecast adoption, and to communicate climate risk more effectively, we are assessing decision-makers’ mental models of ENSO-related climate risks.