The impact of applying three-dimensional variational data assimilation (3D-Var DA) on convective-scale forecasts is investigated by using two mesoscale models, the Weather Research and Forecasting model (WRF-ARW) and the Hirlam and Aladin Research Model On Non-hydrostatic-forecast Inside Europe (HARMONIE-AROME). One month (1 to 30 December 2013) of numerical experiments were conducted with these two models at 2.5 km horizontal resolution, in order to partly resolve convective phenomena, on the same domain over a mountainous area in Iran and neighboring areas. Furthermore, in order to estimate the domain specific background error statistics (BES) in convective scales, two months (1 November to 30 December 2017) of numerical experiments were carried out with both models by downscaling operational ECMWF forecasts. For setting the numerical experiments in an operational scenario, ECMWF operational forecast data were used as initial and lateral boundary conditions (ICs/LBCs). In order to examine the impact of data assimilation, the 3D-Var method in cycling mode was adopted and the forecasts were verified every 6 hours up to 36 hours for selected meteorological variables. In addition, 24 h accumulated precipitation forecasts were verified separately. Generally, the WRF and HARMONIE-AROME exhibit similar verification statistics for the selected forecast variables. The impact of DA on the numerical forecast shows some evidence of improvement in both models, and this effect decreases severely at longer lead times. Results from verifying the 24 h convective-scale precipitation forecasts from both models with and without DA suggest the superiority of the WRF model in forecasting more accurately the occurred precipitation over the simulation domain, even for the downscaling run.

通过使用两个中尺度模型，天气研究和预报模型 (WRF-ARW) 以及 Hirlam 和 Aladin 研究模型，研究了应用三维变分数据同化 (3D-Var DA) 对对流尺度预报的影响。欧洲内部静水预测 (HARMONIE-AROME)。为了部分解决对流现象，对这两个模型在 2.5 公里水平分辨率下进行了一个月（2013 年 12 月 1 日至 30 日）在伊朗山区和邻近地区的同一域上进行的数值实验。此外，为了估计对流尺度的域特定背景误差统计数据 (BES)，通过缩减操作 ECMWF 预测，对两种模型进行了两个月（2017 年 11 月 1 日至 12 月 30 日）的数值实验。为了在操作场景中设置数值实验，ECMWF 操作预测数据被用作初始和横向边界条件 (ICs/LBCs)。为了检查数据同化的影响，采用了循环模式下的 3D-Var 方法，并对选定的气象变量每 6 小时到 36 小时验证一次预测。此外，24小时累计降水预报分别进行了验证。一般来说，WRF 和 HARMONIE-ARME 对选定的预测变量表现出相似的验证统计数据。DA 对数值预报的影响显示了两种模型都有改进的证据，并且这种影响在更长的提前期会严重下降。