88
Statistics of multi-season precipitation deficits across the contiguous United States
David S. Gutzler, University of New Mexico, Albuquerque, NM; and D. Kann
We develop an extremely simple definition of regional multi-season drought based on persistent runs of 3-month precipitation anomalies. This definition is designed so that it is closely related to the set of 3-month running averages of tercile-based precipitation outlooks that form the basis for the NOAA Climate Prediction Center's seasonal forecast. Using this definition, multi-season drought episodes can be defined from 20th Century historical data when a specified number of consecutive 3-month anomalies lie in the lowest tercile, corresponding to the "below normal" tercile in the CPC's seasonal outlooks. The initial set of results is derived using a threshold of six consecutive running 3-month anomalies as the definition of drought. The sensitivity of the results to this specification, and to consideration of isolated "gaps" in the run of consecutive dry seasons, is currently being assessed. Results are presented for the CONUS based on NCDC's multi-state regions to allow comparison between the statistics of drought in different parts of the country.
We enumerate the characteristics of multi-season drought defined this way across the seasonal cycle to explore seasonal tendencies for drought onset or demise. We find that multi-season droughts tend to start and end in extreme seasons (summer or winter), with seasonal minima in onset or demise during the transition seasons. We will examine relationships between multi-season drought and large-scale climate forcings associated with SST anomalies (such as those related to ENSO). Thus this analysis complements the U.S. CLIVAR Drought Working Group approach, which specifies different SST forcing functions and examines the response. Here we specify the structure of the drought response and then examine the corresponding hypothesized climate forcings.
Poster Session , Seasonal to Interannual Variability: Observations and Predictions
Tuesday, 19 January 2010, 9:45 AM-11:00 AM
Previous paper Next paper