Remote Sensing Detection Method Of Farmland Irrigation Events In Arid Areas Of Northwest China

China is a country with a large population,and the basic national conditions of large population and small land determine the important position of cultivated land protection.The red line of cultivated land is not only quantitative,but also qualitative.Carrying out the evaluation of cultivated land quality is an important measure to implement the strategy of"storing grain on the ground and storing grain in technology",keep the red line of cultivated land quality firmly and ensure food security.Irrigation capacity is an important index of farmland quality evaluation,so it is of great significance to obtain irrigation information quickly for farmland quality monitoring and protection.At present,the existing monitoring methods of agricultural water used in ground stations in China can not provide comprehensive technical support for the fine management of agricultural water resources and the comprehensive monitoring of agricultural irrigation water.As a fast and efficient means,remote sensing has been widely used in the study of farmland irrigation area.However,the current research lacks the research on irrigation events,and mostly focuses on the use of single data source and single index method,which lacks the comparison and combination of methods.Therefore,the remote sensing detection method for farmland irrigation events needs further research.In this study,Ganzhou District and Daman Super Station in Zhangye City,Gansu Province are taken as the research areas.Based on the point-scale Daman Super Station,the remote sensing detection method of irrigation events in 2016 in dry years is selected,using the near-infrared band reflectivity method of MOD09GA optical remote sensing data,the green index GI response curve method and the segmented PDI differential threshold model method,and combining microwave moisture products with PDI differential threshold method.The main conclusions are as follows:(1)Comparison and optimization of different irrigation detection remote sensing methods in Daman Super StationThe detection rate of microwave AMSR-2 soil moisture product data growth rate threshold method is 0.8,and it is missed once.The reason that affects the detection accuracy is the lack of data in time series.The detection rate of near-infrared reflectance method is 1.2,and it is detected three times more.The detection rate of green index GI response curve method is 0.6,and the missing detection is twice.The detection rate of segmented PDI difference threshold model method is 1.6,and it is detected three times more.The detection rate of microwave moisture products combined with PDI difference threshold method is 1,and there is no false detection.Microwave moisture products combined with PDI differential threshold method has the highest accuracy in detecting irrigation,which is the best irrigation detection method for Daman Super Station in 2016.(2)Building the segmented PDI threshold modelThe correlation R~2between PDI difference and NDVI is 0.8789,showing a strong negative linear correlation,and the detection result of the segmented PDI threshold model constructed on this basis is better than PDI index.(3)Inter-annual test of microwave moisture products combined with PDI difference threshold method.The detection rate of microwave moisture products combined with PDI differential threshold method is 1 in dry year 2016,0.875 in wet year 2017 and 1 in normal year 2020.Microwave moisture products combined with PDI differential threshold method have the best effect in detecting irrigation in dry year and ordinary year,and the inspection accuracy in wet year is also reliable.(4)Spatial distribution of annual irrigation frequency in Ganzhou DistrictThe spatial distribution of irrigation events in Ganzhou District in 2016 was studied.In Ganzhou District,the farmland irrigated four times or more is the most,accounting for52.96%,and the farmland irrigated once or less is the least,accounting for 8.32%.The annual irrigation times of farmland in Daman Town and Xiaoman Town are the most,and Dangzhai Town is the least.The duration of each round of irrigation and the spatial distribution of irrigated farmland in five rounds of irrigation in Ganzhou District were studied.The factors restricting farmland irrigation in Ganzhou District are:limited river drainage and lack of surface water resources;overexploitation of groundwater and groundwater level decline;the existing reservoirs have insufficient storage capacity and limited runoff regulation capacity;the efficiency of water use is not high and the efficiency of water resources utilization is low;the field facilities are poor and the irrigation technology is backward.The results of this study provide a theoretical basis for the objective and efficient acquisition of irrigation information in arid areas and the detection and protection of cultivated land quality.