Prediction of Central Arizona Convection. Part II: Further Examination of the Eta Model Forecasts

Abstract

Output from simulations of the Eta model are compared to special observations collected during the 1990 Southwest Area Monsoon Project (SWAMP). The emphasis is on the model's prediction of the preconvection air mass over Phoenix, Arizona, and on the model's representation of the low-level jet and moisture surge observed over southwest Arizona. At times the model fails to capture the rapid increase in low- and mid-level moisture that is observed in the hours prior to the onset of convection. Subsequent convection is not predicted by the Eta model. In one event the model very accurately predicts the evolution of the air mass over Phoenix during the period just prior to the outbreak of severe convection. However, no convection is predicted by the model. The model seems unable to generate convection over the high terrain or lower deserts of central Arizona regardless of whether the air mass is simulated correctly. A low-level jet feature observed over southwest Arizona during SWAMP is not correctly simulated by the Eta model. The model produces a very strong sea-breeze circulation from the Gulf of California into western Arizona in each simulation. The moisture and stability profiles associated with the sea-breeze are inconsistent with observations over southwest Arizona, which leads to a misrepresentation of the low- and midlevel moisture field over the region. Poor initial conditions in the sea surface temperature field over the Gulf of California are, at least in part, responsible for the model error.