Simulating Multi-Season Past Droughts
Abstract

Abstract Author: Emilia Jin and David Straus

Abstract Title: Ocean Forced Predictability of North American Drought

Abstract: A suite of numerical experiments have been conducted to understand the SST-forced mechanisms by which North American droughts on monthly and seasonal time scales occur, using as a tool the Coupled Forecast System (CFS) of NCEP.
One suite of experiments uses the “pacemaker” strategy, in which the eastern tropical Pacific SST was specified from observations, while the remainder of the global ocean is represented by a slab mixed-layer model interacting with the atmosphere. A ensemble of long (~55-year) simulations with this configuration has been performed. Predictable patterns of drought and pluvial conditions are investigated on the interannual timescale by analyzing seasonal, monthly and intraseasonal features related to SST. The mechanisms by which stationary wave propagation connects the forcing and response regions are explored by analyzing the SST and circulation patterns associated with both droughts and pluvials.
The relationships of droughts and pluvials with the flow dynamics is further assessed by with the use of intra-seasonal circulation regimes that are defined explicitly on the basis of both low frequency planetary wave flow and high frequency eddy feedback. An important question is the degree to which the extreme monthly and seasonal rainfall anomalies are associated with well-defined regimes and are thus partly predictable.
A second diagnostic experiment consisted of ensemble of simulations in which the slab ocean model was replaced by specified, climatologically varying, SST. A comparison of the two ensembles emphasizes the role of ocean temperature variations outside of the tropical Pacific which are forced from the tropical Pacific including North Pacific and North Atlantic SST anomalies. An ensemble of 32-year simulations with the full CFS is analyzed to study the natural occurrence of droughts on monthly and seasonal time scales within the fully coupled model context. Inconsistencies with observations are shown to be a result of slow coupled dynamics in the coupled GCM.