An Iterative Method for Assimilating Precipitation Observations

Despite all the advances in Numerical Weather Prediction (NWP) and in Data Assimilation (DA) methods, Quantitative Precipitation Forecasting (QPF) still remain a challenging task. One of the reasons is the poor specification of the initial state of the atmosphere due to limitations in the current DA methods and insufficient number of observations. In the latter matter, precipitation estimations from radar observations become a valuable source of information for improving prediction of convective events in NWP because they provide detailed information on the spatio-temporal evolution of these events.

Although all precipitation DA techniques have demonstrated some success, precipitation improvement has a limited persistence, ranging from a few hours up to 12 hrs. As there is no obviously superior scheme, in the last decade, simple "nudging" techniques have received considerable attention These methods are based on the insertion of small forcing terms into the prognostic equations of the model in order to "nudge" or relax the model state towards the observations in a gradual manner.

The present study presents a new iterative nudging method which compares the observed and the model precipitation accumulation at a given time and modifies the model buoyancy through humidity, but nudge the profiles at the beginning of the accumulation interval ("a priori''). The humidity profile is adjusted at the beginning of the time interval iteratively until the model accumulation resembles the model accumulation. To characterize the presented assimilation method an Observing System Simulation Experiment (OSSE) type strategy was implemented in a study case. The main goal of these experiments is to evaluate the performance of the scheme and address its limitations. A possible explanation for the limited improvement persistence will be presented.