Seasonal forecast of agricultural impacts of droughts in Mexico through a principal component regression based approach

Drought monitoring and forecasting allows to adopt mitigating actions in early stages of an event to reduce the vulnerability of a wide range of environmetal, economical and social sectors. In Mexico, various drought monitoring systems on national and regional scale perform a follow up of these events, such as the Drought Monitor in Mexico, and the North American Drought Monitor, but seasonal drought forecasting is still a pending task. This study aims at fill this gap applying a methodology that uses data derived from a globally available atmospheric reanalysis product and a principal component regression based model oriented to predict drought impacts in rainfed crops associated to deficits in the soil moisture, estimated by means of the standardized soil moisture index (SSI). Using the state of Guanajuato (Center-North of Mexico) as a study case, the model generated yielded RSME values of 0.74 using regional and global hydrological, climatic and atmospheric variables as predictors with a lead-time of 4 months.