Spatial-temporal Variation Characteristics And Influencing Factors Of Drought In Karst Region Based On Multi-source Data

In recent years,the frequency and intensity of drought in southwest China have shown a trend of increasing,seriously endangering the economic,social and ecological development of the region.Therefore,studying drought in the southwest karst region is crucial for local sustainable development.Most of the existing remote sensing agricultural drought indexes only consider and characterize drought from one or two aspects,making it difficult to obtain good results when applying each index to complex geomorphic areas such as the karst region in the southwest.In this paper,based on the Euclidean distance principle,an agricultural drought index-Vegetation Evapotranspiration Precipitation Drought Index(VEPDI)--suitable for karst areas in southwest China is constructed using crop water stress index,normalized vegetation water supply index,and precipitation state index,and its applicability is analyzed.Using MODIS and GPM as the main data sources,Guangxi,Guizhou,and eastern Yunnan as the study areas,and 2001-2020 as the study period,the VEPDI was calculated,and the spatiotemporal evolution characteristics of drought at different time scales were studied.Finally,geographical detectors were used to study the drought driving factors in the study area at different time scales.The main research conclusions are as follows:(1)The Pearson correlation coefficient between VEPDI and soil moisture content(SM)in different months is higher than the other four indexes,with the weakest reaching 0.707;On the seasonal scale,the R~2 of linear fitting between VEPDI and SM is also the largest,reaching0.7024,indicating that VEPDI has a high applicability in the southwest karst region.(2)From the perspective of spatial distribution,in the past 20 years,drought has mainly occurred in non karst areas,karst mountains,and karst fault basins in the west,while karst trough areas have shown relatively humid performance.It is worth noting that the spatial distribution of drought varies at different time scales.On a monthly scale,non karst regions,karst mountain regions,and karst canyon regions are the most arid,while karst trough regions are not.On a seasonal scale,karst fault basins are most prone to drought.On an annual scale,karst mountains,karst hills,and karst fault basins are more prone to drought.(3)In terms of time trend,on a monthly scale,the southeast tends to become wetter at a faster rate;Drought presents periodic changes throughout the year,with severe drought in March each year.From a seasonal perspective,the regions with higher wetting rates in spring,autumn,and winter are located in the southeast,and the overall wetting rate in autumn is the fastest;Spring is the driest.Overall,over the past 20 years,the drought in the study area has tended to moderate,with the entire study area tending to become wet at a rate of 0.0081/20a,with the highest rate of karst geomorphology being peak cluster depressions and the lowest rate of karst fault basins.(4)Using geographical detectors to study the explanatory power of various factors on VEPDI,the results of factor detectors show that the driving factors of annual scale,spring,summer,and autumn dominant droughts are all bedrock exposure rates,while the dominant driving factor in winter is elevation.The results of interaction detectors show that the interactive combination of annual scale,spring,summer,and autumn dominant droughts is precipitation∩bedrock exposure rates.The dominant interaction combination in winter is elevation∩bedrock exposure rate.The Pearson correlation coefficient results show that there is a positive correlation between bedrock exposure rate,elevation,and drought severity.Areas with high bedrock exposure rate and elevation are more prone to drought attacks.There is a negative correlation between precipitation and drought,and the more precipitation,the less likely drought is.All three factors indicate that karst fault basins,karst canyons,and karst mountains are more prone to drought.