IMPACTS OF PRECIPITATION IN THE UPPER OCEAN IN THE WESTERN PACIFIC WARM POOL DURING TOGA-COARE

An ocean mixed-layer model with a modified Kraus-Turner parameterizati on scheme is used to investigate the impacts of precipitation in the u pper ocean in the western Pacific warm pool during Tropical Ocean Glob al Atmosphere-Coupled Ocean-Atmosphere Response Experiment (TOGA-COARE ). Heat and salt budgets calculated in the upper ocean indicate local balance between surface forcing and the ocean response. Thus the mixed -layer model captures the dominant processes governing heat and salt v ariability. The model responses are analyzed and compared with the obs erved upper ocean in three distinctive layers determined by Monin-Obuk hov length scales. In the top layer (the top 5 m), about 90% of the su rface buoyancy flux is absorbed, and strong diurnal and intraseasonal variations are excited. The second layer, 5-20 m, contains intraseason al variability that is characterized by nearly neutral stratification during strong westerly wind events, strong thermal stratification duri ng clear-sky days, and strong saline stratification during heavy preci pitation. The dominant effect of precipitation is to generate a stable stratification and to form a barrier layer. The third layer, 20-50 m, has intraseasonal variations due to mixing during westerly wind event s. Heavy precipitation amplifies mixed-layer temperature fluctuations by a cycle of strong surface cooling and entrainment warming through t he following processes. Heavy precipitation causes a shallower mixed l ayer and a larger cooling rate. Surface temperature drops rapidly, and the upper ocean becomes thermally unstable. The salinity maintains a- weak density stability, which causes strong entrainment warming. Surfa ce freshwater flux is the key factor controlling saline structure when advection is excluded. However, experiments without the entrainment p rocess show a significant bias toward a lower salinity, indicating tha t the entrainment process must be properly treated in the model to pre vent a biased trend.