A Three-Dimensional Variational Radar Data Assimilation Scheme Developed for Convective Scale NWP

Abstract

A three-dimensional variational data assimilation scheme (3DVAR) has been developed for Convective scale NWP. In the scheme, a cost function is defined by a background term, an observation term, and a weak constraint term. The function is minimized through a limited memory, quasi-Newton conjugate-gradient algorithm. The background error covariance matrix, though simple, is modeled by a recursive filter. Furthermore, the square root of this matrix is used to precondition the minimization problem. In its original development, only radar radial velocity data could be assimilated. Recent developments for 3DVAR include the use of a model-derived diagnostic pressure equation constraint (DPEC) as a weak constraint, and the capability to assimilate reflectivity directly in the 3DVAR framework. The original radial-velocity-only 3DVAR method is applied to assimilate radial velocity observations considering beam broadening and earth curvature for an idealized supercell storm case, and real supercell storm cases. It is shown that the horizontal circulations, both within and around the storms, as well as the strong updraft and the associated downdraft, are well analyzed. The results also indicate that the method is quite insensitive to the effect of beam broadening, but very sensitive to the effect of earth curvature. So in the real data case studies, the effect of earth curvature is considered while beam broadening is not. Based on this 3DVAR framework, a real-time, weather-adaptive analysis system has been developed for the NOAA Warn-on-Forecast (WoF) project to incorporate all available radar observations within a moveable analysis domain. The system performed very well within the NOAA Hazardous Weather Testbed Experimental Warning Program during preliminary testing in recent years when many severe weather events were successfully detected and analyzed. The impact of DPEC on radar data assimilation is examined primarily in the context of storm forecasts. It is found that the experiments using DPEC generally predict higher low-level vertical vorticity near the time of observed tornados than the experiments not using DPEC. Finally, the impact of assimilating both radar reflectivity and radial velocity data with an intermittent 3DVAR system is explored using an idealized thunderstorm case. It is found that by assimilating reflectivity data using simple hydrometer classification while also assimilating radial velocity data, the model can reconstruct the supercell thunderstorm quickly and the quality of analyses are improved compared to two other experiments without reflectivity and hydrometer classification. This paper represents the author’s research efforts in radar data assimilation for convective scale NWP during the past several years.