Analysis On Precipitation Characteristics Of Guizhou Mountain Area Based On Multi-source Data

Precipitation is the core component of the global hydrological and material energy cycle,it is of great significance to regional social and economic development,agricultural production,ecological balance,optimal allocation of water resources,environmental protection to master the distribution of precipitation appropriately.Due to the interaction of monsoon climate and complex terrain,the spatial and temporal variation of precipitation in Guizhou mountain area is quite different.Accurately obtaining the distribution characteristics of precipitation at different spatial-temporal scales is one of the key issues in the development and utilization of water resources in this region.However,few results have been paid attention to the analysis of the characteristics and causes of precipitation changes at different time scales in the mountainous area of Guizhou based on multi-source data,and due to the limited availability of data,the precipitation data used in the study usually have a lower spatial and temporal resolution and cannot finely depict the spatial and temporal heterogeneity of precipitation in the mountains.The thesis comprehensively utilizes multi-source precipitation observation information such as ground gauge station network,meteorological radar and satellite to explore its effectiveness on the basis of establishing a quantitative model of precipitation data merging,and conducts research on the characteristics and influencing factors of multi-scale precipitation.First,the Random Forest precipitation fusion scheme based on the satellite-ground information was constructed to evaluate its fusion effect and error sources at different time scales.Then,the long-term precipitation characteristics of the complex terrain mountain area based on the merging precipitation products are analyzed to reveals different regular pattern of precipitation;Finally,the hourly precipitation data and meteorological radar data are selected to analyze the spatial-temporal variation characteristics of small-scale under complex terrain conditions.The main conclusions of this thesis are as follows:(1)The multi-source precipitation information merge scheme including ground data,CHIRPS2.0,PERSIANN-CDR and DEM constructed based on the Random Forest algorithm,which can combine the ground measured values and the spatial distribution information of satellite products effectively and improve precipitation estimate accuracy.On the monthly scale,the average absolute error can be reduced to less than 25 mm,and the Kling-Gupta Efficiency(KGE)is increased to 0.8640.On the daily scale,it can also effectively reduce the relative deviation of satellite observation daily precipitation and improve the measured value from the ground of relevance.(2)The perennial spatial distribution of precipitation in the karst mountain area of Guizhou is uneven,and the Southwest Monsoon and terrain factors are important influencing factors.There are three rainy centers and a lowprecipitation rain area in the distribution of precipitation.The distribution of rainy centers in the flood season and the non-flood season is inconsistent,which is mainly affected by the Southwest Monsoon and the difference in water vapor transport.In the past 37 years,precipitation has increased not significantly.Among them,abrupt changes occurred in 2014,after the abrupt change,the center of high and low precipitation shifted from west to east,and the difference in precipitation change rate also increased.The precipitation change showed a trend of decreasing in the southwest and middle northern parts,and increasing in the southeast part.This change is related to the weakening of the Southwest Monsoon.The perennial precipitation decreases with increasing altitude in the Guizhou region of the Yangtze River Basin,increases first and then decreases with the elevation in the Pearl River Basin,and the maximum precipitation height is 1500 m.There is a positive correlation between precipitation and slope,the slope of maximum precipitation is 41°,the aspect of maximum precipitation is SE,followed by WSW.(3)The small-scale precipitation has frequency-intensity variation characteristics,the night rain characteristics are prominent,and the occurrence and development of severe convective cloud clusters have obvious rules.The intensity of precipitation gradually increases from north to south,the frequency of precipitation is high in the middle and low in the north and south,and the distribution of high-frequency high-intensity precipitation areas and lowfrequency low-intensity precipitation areas is consistent with the annual precipitation.The daily precipitation amount,precipitation frequency and precipitation intensity in the flood season all have obvious single-peak characteristics,the daily precipitation peak time appears in the second half of the night(1:00~6:00),and is gradually delayed from northwest to southeast,the intensity peak time appears in the afternoon to evening(13:00~19:00),and the frequency peak appears at midnight to early morning(1:00~8:00).Severe convective rain clusters have obvious banding characteristics.The formation orientation is mainly in the West(W/WNW/WSW).The nascent time is from noon to early morning,its duration is about 1 hour,and it has a significant positive correlation with the maximum radar reflectivity factor.Through this study,the internal mechanism of precipitation in mountainous areas was recognized,and the pattern of precipitation change in Guizhou mountain areas in was clarified,which can provide a reference for improving the accuracy of monitoring and forecasting of rainstorms and floods as well as water resources planning and management.