Experiments of LAPS Wind and Temperature Analysis With Background Error Statistics Obtained Using Ensemble Methods

Abstract

Main purpose of this study is to show that if background error statistics obtained using ensemble methods can improve forecast as well as analysis by ingesting background wind and temperature error statistics in the LAPS (Local Analysis and Prediction System) data assimilation. Wind and temperature analysis are verified by comparing the current LAPS analysis that has spatial and temporal constant weighting of background errors (OLD) with the upgraded LAPS analysis incorporating ensemble-derived weighting as background error statistics (NEW). Using the ensemble-derived weighting for LAPS analysis (NEW) produces more realistic and even more detailed analysis field that would be affected by topography than does the constant weighting for LAPS analysis (OLD). In particular, the wind speed analyses are clearly improved at most pressure levels by ingesting background error statistics, whereas the wind direction and temperature analyses with ingested background wind variance are improved at many levels but deteriorated at others. Two forecasts initialized with both LAPS OLD and NEW analysis are compared each other and then verified against the independent observations. The errors of wind speed/direction and temperature from NEW forecasts are similar to or less than those from OLD forecasts. In particular, wind direction show little improvement in the NEW forecast at some levels, but those could not be statistically significant. Temperature improvements appear better than the wind improvements.