**6. Conclusions**

This study explores the effect of radar reflectivity and GTS data assimilation from assimilation frequency using WRF-3DVar for rainfall forecasting. Four heavy storm events at the Daqinghe catchment in the Beijing Tianjin Hebei region of northern China are selected to be regenerated by the WRF model. We employed three nested domains, and adopted the GFS data for driving the WRF model. From two aspects of cumulative rainfall and spatial distribution of rainfall, two observational data types (radar reflectivity and GTS data) assisted in investigating how WRF rainfall forecasts were potentially improved in space and time through data assimilation. We designed four data assimilation schemes considering various possible combinations of the two data assimilation frequency types in the three nested domains. We compared the analysis with data assimilation and that without data assimilation, finding that the assimilation results partly fit observations in the case and that WRF-3DAVR with radar reflectivity and GTS data better represents the rainfall forecasts in space and time.

Precipitation simulated by the WRF model is always much lower than observed rainfall, but assimilation systems can increase rainfall. The improved initial conditions in WRF-3DVAR system via radar data assimilation and GTS data achieved better shortterm and convective strong precipitation in the both temporal dimension and spatial dimension. The high assimilation frequency significantly helps to trigger and maintain the convective activities in the 3DVAR framework as well as the storm case applied. Forecasts of events indicate that the temporal rainfall distributions of convective storms can be much better predicted with high assimilation frequency, compared with the 6 h assimilation time interval run. At the same time, employing the high assimilation frequency to the assimilation showed improved skill of precipitation forecasting in WRF-3DVAR on spatial

rainfall distributions. The only exception happened in Event II. In this case, the impact of false rainfall forecasting fields is enlarged because of the inaccurate radar observed data, so that the negative impact was found after assimilation. However, this does not seem to pose a threat to the hourly assimilation frequency. In addition, the data assimilation of outside domain has small impact on output of inside domain in not only temporal but also spatial dimension. In general, the hourly data assimilation frequency together with strict outputs from domain resolutions is closer to actual precipitation. In this study, the assimilation by combining the radar reflectivity and Global Tele-communication System (GTS) data with high assimilation frequency is helpful for further enhancing the temporal and spatial distribution of the short-term precipitation forecast. The results can be used as a reference for areas with similar climatic conditions as well as rainfall characteristics. The methodology is of guiding significance for WRF-3DAVR rainfall forecasting. In this case, in order to explore universally applicable data assimilation guidelines for rainfall forecasting, research should be conducted over more storm events in different study areas.

**Author Contributions:** Conceptualization, Y.L. and J.L.; Methodology, Y.L., J.L. and F.Y.; Software, Y.L. and Y.L.; Validation, Y.L. and W.W.; Formal analysis, Y.L. and J.L.; Investigation, W.W., C.L. and F.Y.; Writing-original draft preparation, Y.L. and J.L.; Writing-review and editing, Y.L., J.L. and C.L.; Visualization, Y.L.; Funding acquisition, J.L. and C.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Major Science and Technology Program for Water Pollution Control and Treatment (2018ZX07110001), the National Natural Science Foundation of China (51822906), the National Key Research and Development Project (2017YFC1502405), and the IWHR Research & Development Support Program (WR0145B732017).

**Conflicts of Interest:** The authors declare no conflict of interest.
