2000 Articles

Generating stochastic daily weather constrained to target monthly means.

Hansen, J.W.

A number of research groups around the globe are seeking to apply seasonal climate forecasts to improve management of food production systems and security of farmer livelihoods in the face of climatic risk. One of the tools frequently employed by these efforts is dynamic crop simulation models. Promising results have been obtained using categorical indices of climate teleconnenctions (e.g., ENSO phases). Growing interest in incorporating climate forecasts from dynamic atmospheric models into these efforts has been hampered methodological barriers, particularly the mismatch between the coarse spatial and temporal resolution of climate prediction model outputs, and the fine resolution of crop model input requirements and predictions. Participants of the CLIMAG Geneva Workshop (Geneva, Switzerland, 28-29 September 1999) identified the appropriate methodology for linking climate prediction and crop simulation models as a critical knowledge gap.

John Ingram (GCTE) conveyed a request to the IRI, on behalf of the CLIMAG community, to host a one-day workshop that would bring climate prediction scientists, crop model users and other interested scientists together to clarify and address the relevant technical issues. The IRI recognizes the relevance of these issues and, in line with its mission, is interested in advancing efforts within the agricultural community to apply seasonal climate prediction. We therefore organized the workshop in conjunction with the International Forum on Climate Prediction, Agriculture and Development. START (on behalf of CLIMAG) shared the cost of the workshop with the IRI.

The workshop was quite successful in terms of clarify issues; communicating relevant crop model requirements climate model characteristics, capabilities and limitations; and highlighting some existing and potential approaches for addressing scale mismatches – the first three workshop goals. The fourth goal – developing a strategy for continued progress – was more elusive. Some progress was made in planning a limited collaborative study focused on Australia. However, participants generally felt constrained by resource limitations. The IRI remains committed to improving utility of dynamic model-based forecasts for use with crop and other simulation models used for impact prediction and decision support. We have initiated a limited, exploratory study within the IRI. As we gain experience and identify additional resources, we plan to again engage the broader community in this endeavor.

James Hansen Coordinator