No-precipitation reflectivity observations from Doppler radar contain valuable information on areas lacking precipitation; however, they are often ignored in four-dimensional variational (4DVar) radar data assimilation (DA). This study incorporated a neighborhood-based scheme to assimilate no-precipitation observations as a mechanism to suppress spurious convection. The impact of the scheme on convective forecasting using 4DVar was evaluated by comparing the performance of experiments with and without assimilation (ExpCTL) using eight diverse storm cases that occurred over the central United States during summer 2016. Three no-precipitation assimilation experiments with different neighborhood radiuses of 10 (ExpR10), 30 (ExpR30), and 50 km (ExpR50) were conducted to examine the sensitivity of the scheme to neighborhood size. Results indicated that all the no-precipitation assimilation experiments significantly improved quantitative precipitation forecast skill with large reduction of the bias and the false alarm ratio from ExpCTL, as well as improving representation of the intensity and coverage of the precipitation. The horizontal wind, temperature, and water vapor were also improved, especially the latter. The scheme was found sensitive to neighborhood size and greater benefit was found in ExpR30 in comparison with ExpR10 and ExpR50. Analysis revealed that ExpR30 reduced low-level cooling and mid-level warming corresponding to decreased water vapor in areas of overpredicted and false precipitation, and it was more effective in conserving the total water content balance during cycled radar DA. The findings of this study could provide reference information for assimilation of no-precipitation observations into the Weather Research and Forecasting model using 4DVar, which would be valuable for severe weather prediction.

多普勒雷达观测到的无降水反射率包含了有关降水不足地区的有价值的信息。但是，它们在四维变分（4DVar）雷达数据同化（DA）中通常被忽略。这项研究采用了一种基于邻域的方案来吸收无降水的观测结果，以此作为抑制虚假对流的一种机制。通过比较使用和不使用同化（ExpCTL）的实验的性能（使用2016年夏季在美国中部发生的八种暴风雨情况下）的性能进行了评估，评估了该方案对使用4DVar的对流的影响。三个不同邻域的无降水同化实验半径分别为10（ExpR10），30（ExpR30）和50 km（ExpR50），以检查该方案对邻域大小的敏感性。结果表明，所有无降水同化实验均显着提高了定量降水预报技能，大大降低了ExpCTL的偏差和虚警率，并提高了降水强度和覆盖率。水平风，温度和水蒸气也得到了改善，尤其是后者。发现该方案对邻域大小敏感，并且与ExpR10和ExpR50相比，在ExpR30中发现了更大的收益。分析显示，ExpR30减少了过度预测和虚假降水区域中的水汽减少，相应地降低了低层冷却和中层变暖，并且在保持循环雷达DA期间的总水分平衡方面更有效。