利用额尔古纳牧业气象试验站降水量与土壤水分数据，通过降水与土壤水分动态变化及转化过程分析，确定土壤水分响应的降水临界值与不同降水级别引起土壤水分响应的概率，构建了降水过程量与土壤水分增量函数关系。结果表明：（1） 研究区降水量呈“先降后升”变化趋势，年内降水量呈单峰型分布。（2） 研究区以无降水天气为主，降水又以小降水事件占主导，大降水事件发生频次低、过程降水量大，小降水事件则相反。（3） 可以引起研究区0~50 cm各层土壤水分响应的降水临界值分别为8.1 mm、10.1 mm、19.0 mm、27.9 mm和31.6 mm，小雨仅能引起0~10 cm土壤水分响应的概率为28.6%，中雨不能引起40~50 cm土壤水分的响应。（4） 降水量与0~10 cm和10~20 cm土壤水分达到最大值时的滞后时间呈现出极显著负相关关系，与20~30 cm呈显著负相关关系，0~30 cm各层土壤水分达到最大值时的滞后时间与降水量符合幂函数关系。（5） 降水量和0~50 cm土壤水分增量均呈现出极显著正相关关系，降水量与0~10 cm和10~20 cm土壤水分增量符合线性关系，与20~30 cm、30~40 cm和40~50 cm土壤水分增量符合多项式关系。检验结果表明，构建的函数模型可以较好地模拟研究区0~30 cm各层水分增量。研究结果为地方政府抗旱减灾提供了科学依据。

The regulation of structure and function in an ecosystem is largely affected by soil moisture, and the grassland ecosystem is more sensitive to soil moisture than other ecosystems. The change of soil moisture is determined by precipitation, so how does precipitation affect grassland soil moisture? What is the response law to the precipitation of grassland soil moisture? These are the key problems in the benign development of grassland ecosystems. However, there have been few reports on the response of soil moisture to precipitation in the meadow steppe in north China. To reveal the characteristics of precipitation and the response of soil moisture to precipitation pulse, the Hulunbuir meadow steppe was taken as the research object, dynamic change, transformation process of precipitation, and soil moisture were analyzed. The influence characteristics of different scale precipitation processes on soil water were explored and the critical values of rainfall, which could cause each layer of soil moisture (0-50 cm) of the meadow steppe were determined based on the data of precipitation and soil moisture from Ergun Animal Husbandry Meteorology Experiment Station. The relationship models between precipitation and soil water increment were established to invert soil water content, which could significantly reduce the spatial uncertainty of refined drought assessment and provide a scientific basis for local government to fight drought and reduce the disaster. The response regularity of soil moisture to precipitation pulse at different depths was revealed in the meadow steppe. The results showed that (1) the precipitation showed a trend of “decreasing first and then rising”, and the annual precipitation showed a single-peak distribution. (2) The weather was dominated by no precipitation, and the precipitation was dominated by small-precipitation events in a meadow steppe. The heavy precipitation events were characterized by low occurrence frequency and large precipitation, whereas the small-precipitation events were the opposite. (3) The critical values of precipitation that can cause the response of 0-50 cm soil moisture were 8.1 mm, 10.1 mm, 19.0 mm, 27.9 mm, and 31.6 mm, respectively. The probability of light rain only causing 0-10 cm soil moisture response was 28.6%, whereas moderate rain could not cause 40-50 cm soil moisture response. (4) Precipitation was extreme significantly negatively correlated with the lag time when soil moisture reached the maximum value in 0-10 cm and 10-20 cm, and was significantly negatively correlated with the lag time when soil moisture reached the maximum value in 20-30 cm. The lag time of soil moisture reaching the maximum value in each layer of 0-30 cm accords with the power function relationship with precipitation. (5) There was a significant positive correlation between precipitation and soil moisture increments of 0-50 cm. Precipitation was linear with soil moisture increments of 0-10 cm and 10-20 cm, and polynomial with soil moisture increment of 20-30 cm, 30-40 cm, and 40-50 cm. The model test results showed that the simulation effect of soil moisture increment models in each layer of 0-30 cm was better.