Effective Assimilation of Global Precipitation

Assimilating precipitation observations by modifying the moisture and sometimes temperature profiles has been shown successful in forcing the model precipitation to be close to the observed precipitation, but only while the assimilation is taking place. After the forecast start, the model tends to "forget" the assimilation changes and lose their extra skill after few forecast hours. This suggests that this approach is not an efficient way to modify the potential vorticity field, since this is the variable that the model would remember. In this study, the ensemble Kalman filter (EnKF) method is used to effectively change the potential vorticity field by allowing ensemble members with better precipitation to receive higher weights. In addition to using an EnKF, two other changes in the precipitation assimilation process are proposed to solve the problems related to the highly non-Gaussian nature of the precipitation variable: a) transform precipitation into a Gaussian distribution based on its climatological distribution, and b) only assimilate precipitation at the location where some ensemble members have positive precipitation. The idea is first tested by the observing system simulation experiments (OSSEs) using SPEEDY, a simplified but realistic general circulation model. When the global precipitation is assimilated in addition to conventional rawinsonde observations, both the analyses and the medium range forecasts are significantly improved as compared to only having rawinsonde observations. The improvement is much reduced when only modifying the moisture field with the same approach, which shows the importance of the error covariance between precipitation and all other model variables. The effect of precipitation assimilation is larger in the Southern Hemisphere than that in the Northern Hemisphere because the Northern Hemisphere analyses are already accurate as a result of denser rawinsonde stations. Assimilation of precipitation using a more comprehensive global model and with high-resolution TRMM Multi-satellite Precipitation Analysis (TMPA) data is ongoing work. The climatological probability distribution of these precipitation data is calculated in preparation for the use of the proposed transformation algorithm. The impact of using an efficient precipitation assimilation in real global analyses and forecasts will be investigated.