Assimilation of rainfall and precipitable water to study the sensitivity of precipitation processes.

Dong-Wook Shin and Joong-Bae Ahn

Korean Journal of Amospheric Sciences, 3, 1-14. 2000.

Abstract

To improve our understanding of how 4DVAR data assimilation works in determining the optimal initial conditions for a numerical weather prediction model and assessing precipitation processes, the MM5 model and its adjoint model are run for rainfall and (total) precipitable water (PW) assimilation. In this study, four experiments are conducted, which consist of NOCON (rainfall assimilation excluding convective rainfall), NOLSP (rainfall assimilation excluding large-scale rainfall), NOCON_PW (rainfall/PW assimilation excluding convective rainfall), and NOLSP_PW (rainfall/PW assimilation excluding large-scale rainfall). Experiment NOCON shows that the model, without any cumulus parameterization scheme, has a potential for producing rainfall comparable to the observed field with the help of optimized initial conditions. For this case study, the non-convective rainfall process is found to be negligible since the exhibited changes in the initial fields are very small. The NOCON and NOLSP experiments confirm that the 4DVAR data assimilation method is able to produce an optimal initial condition. By assimilating PW as well as rainfall in the experiments (NOCON_PW and NOLSP_PW), we find that similar results are obtained with somewhat different initial conditions. However, PW assimilation may produce a better initial condition for the humidity distribution and the model rainfall prediction.