Drought Recurrence and Seasonal Rainfall Prediction in the Río Yaqui Basin, Mexico

Abstract

A statistical approach is used to explore the variability of precipitation and meteorological drought in Mexico?s Río Yaqui basin on seasonal-to-decadal time scales. For this purpose, a number of custom datasets have been developed, including a monthly 1900?2004 precipitation index for the Yaqui basin created by merging two gridded land surface precipitation products, a 349-yr tree-ring-based proxy for Yaqui wintertime rainfall, and a variety of large-scale climate indices derived from gridded SST records. Although significantly more rain falls during the summer (June?September) than during the winter (November?April), wintertime rainfall is over 3 times as variable relative to the climatological mean. Summertime rainfall appears to be unrelated to any large-scale patterns of variability, but a strong relationship between ENSO and Yaqui rainfall during the winter months offers the possibility of meaningful statistical prediction for this season?s precipitation. Analysis of both historical and reconstructed rainfall data suggests that meteorological droughts as severe as the 1994?2002 Yaqui drought occur about 2 times per century, droughts of even greater severity have occurred in the past, and such droughts are generally associated with wintertime anomalies. Whereas summertime reservoir inflow is larger in the Yaqui basin, wintertime inflow is more variable (in both relative and absolute terms) and is much more strongly correlated with same-season rainfall. Using the identified wintertime ENSO?rainfall relationship, two simple empirical forecast models for possible use by irrigation planners are demonstrated.