Forecasting plant disease in a changing climate: a question of scale

The subtle changes in climate attributed to climate change can affect plant-disease development. These changes are not easily determined, and consequently, the ability to forecast how disease changes under altered growth conditions is not simple. One method is the use of forecast climate change derived from global-change models that are analogous to general-circulation models used for weather forecasts. However, these models predict conditions on such a gross scale that they are unacceptable for most disease forecasting. Downscaling provides a method whereby weather and climate conditions estimated at a very large scale can be transferred to a fine resolution (∼200-m grid points). This process is explained, in particular, in the context of disease forecasting. An example is presented of how estimates of extreme low temperature at a local scale have been derived from a mesoscale (mid-range) weather forecast model, which in turn was derived from a general-circulation model. Similarly, the derivation of forecasts at local scale from mesoscale weather forecast models have been demonstrated for grapevine downy mildew. Important considerations of scale definition and information transfer across different scales are discussed