The accuracy of flood forecasting in small and medium-sized basins in arid and semi-arid regions needs to be improved, primarily due to a limited understanding of critical factors such as topographic features, critical rainfall, time of concentration, and recurrence intervals. In this paper, Xiahe County, Gansu Province, which is located in the semi-arid region, and Yongchang County, Gansu Province, which is located in the arid region, are selected as the research objects. Field investigations were carried out on flood-related factors across 34 small basins in 86 riverine villages of Xiahe County (2015) and 240 cross-sections in 395 riverine villages of Yongchang County (2014). Flood characteristics were assessed using the instantaneous unit hydrograph method, the regional empirical formula method, and the method proposed by the China Railway First Survey and Design Institute Group Co., Ltd. These methods were employed to calculate the time of concentration, design storms, and design floods for the study areas, and flood warning thresholds were estimated. Based on the calculation results, machine learning methods(linear regression, random forest and neural network) were used to establish a flood warning model for arid and semi-arid areas, and each model was evaluated and analyzed. The results show that there is a linear correlation between prepared transfer rainfall and factors such as rainfall during storms, warning rainfall threshold, and main channel slope, and the linear regression model can accurately calculate the warning rainfall. In order to further verify the applicability of the regression model, the model is used to invert and analyze the prepared transfer rainfall of 34 survey basins in Hezuo City, Gansu Province, the mean absolute error is only 0.56 mm.

There was a rare rainstorm weather process in the North of China on July 21，2012，which led to the most severe flooding disasters．The weather system of the process was the cold front with the upper trough，but such a system would not lead to more than 200 mm rainfall in 24 hours according to the statistical results of the last decade ( 2000 － 2011) ． After the precipitation in Beijing ，the torrential rain occurred in the middle of Hebei Province ( Langfang and the northeast of Baoding) ，where heavy rainfall occurred relatively less in the past． Thus，some reasons of the torrential rainfall in Hebei and the forecasting focus were analyzed in the article，based on the conventional observation data，reanalysis data from NCEP at 1° × 1° spatial resolution，regional automatic meteorological observation data and Doppler weather radar data． The comparative study of weather patterns，physics conditions and meso － and micro－ scale systems were carried out with another rainstorm process which occurred on the 24th July 2011． The results show that there were great differences between two weather processes in extreme value of precipitation and rainfall area，although their precipitation efficiency and meso － and micro － scale systems which directly resulted  in two precipitation processes were similar under the same synoptic scale circulation situation ( low － cold front type) ． The extreme precipitation in 7． 21 process had a closely relation to strong moisture transport，intense dynamic condition and high precipitation efficiency，but the train effect for 6 hours was the leading cause of the heavy rainfall． The configuration of low frontal zone of se and low － level jet on the south of shear line led to the train effect in the middle of Hebei Province，which was concluded as a forecast point of the kind heavy rain events． And，the ground mesoscale convergence line ahead of the cold front was a key point for accurately forecasting heavy rainfall area，which was better than based on the southeast quadrant of the lower vortex，because strong precipitation occurred along with lines．

The climatic feature of precipitation over 10 mm in Ningxia region in recent 49 years were analyzed by using daily precipitation data of 20 meteorological stations from 1961 to 2009. The results show that weather events with precipitation over 10 mm increased gradually with the variety of terrain from north to south, and the amount of the events in the loess hilly region of the south was double more than that in the irrigated area. Moderate rain contributed greatly to rainfall and rainstorm increased gradually. By the statistics of rainfall intensity, the frequency of rainfall more than 1.0 mm/h but less than 2.0 mm/h was the most, and more than 4.0 mm/h was the least in Ningxia, both the frequency of rainfall more than 3.0 mm/h but less than 4.0 mm/h and that less than 1.0 mm/h were gradually increased since the 1980s. The years with extremely more weather events with precipitation over 10 mm in Ningxia are 1961,1964,1985 and 1990.