Research On Spatial Scaling And Transformation Model Of Gully Erosion Parameters Extraction

With development of remote sensing technology and improvement of spatial resolution,the method of extracting geographic things using remote sensing has been mature.And remote sensing technology has been more and more used in the method of gully erosion parameters extraction.However,the method and precision of extracting gully erosion and topographic parameters by remote sensing images and topographic data still need to be improved.Although the precision of traditional methods such as field measurement and visual interpretation is high,it takes too much time and money.And there exist uncertainties and variability in the extraction of gully erosion parameters by remote sensing and topographic data with different spatial scales.So these all limit the quantitative study of gully erosion.Therefore,this paper selected the small watershed in the western margin of the Anli Reservoir in Qixia City as the study area and DEM terrain data was data source.The combination of SA terrain model and catchment area threshold treatment was used to extract gully erosion network system in the study area.And the scale effects of topographical parameters and gully erosion parameters at different spatial scales were analyzed and the appropriate scale of using DEM terrain data to extract gully erosion network was determined.At the same time,the fractal theory was applied to analyze the fractal features of topographic parameters and gully erosion parameters under different spatial scales and the fractal transforming models of the topographic parameters and gully erosion parameters were constructed.The research had important theoretical reference and practical application significance for broadening the study method of gully erosion and improving the research precision of it.The main conclusions are as follows:(1)In term of topographic parameters,the average height didn't show obvious spatial scale effect under the simulated single mountain peak and the real subsurface.While the average slope always showed a very significant exponential function as the spatial resolution changed.The flat curvature and the vertical curvature had no obvious scale effect under the simulated single peak,but showed a significant power function under the real surface with the spatial resolution change.The number of nodes and irregular triangles in the TIN topographic data showed a very significant power function relationship with the spatial scale.In addition to the average elevation,there were some spatial scale effects on other topographic parameters.And there also existed obvious fractal features in each topographic parameter.(2)In terms of extracting the number of gully heads,the variation law of the number of gully heads extracted by different SA thresholds showed a very significant power function relationship,R~2>0.99.However,the variation law of the number of gully heads extracted under different spatial resolutions showed a significant exponential function.In terms of other gully erosion parameters,the gully area,gully length,gully density and gully heads density showed significant power function under different SA thresholds and different spatial resolutions.For spatial dimension analysis of gully erosion parameters in the study area,there were obvious fractal features under different SA thresholds treatment.In terms of spatial resolution,there were obvious fractal features of the gully area,gully length and gully density in the study area.But,the fractal features of the number of gully heads and gully heads density were not obvious.(3)In this paper,Beiyanzi small watershed was used as verification area.Compared with the measured gully erosion parameters,the gully area had relatively poor accuracy under different SA thresholds treatment and different spatial resolutions.It was related to the extracted gully effect by DEM data resolution and SA terrain model.However,the precision of other gully erosion parameters were relatively high,and the accuracy of gully heads number was the highest.When SA threshold is 45 and spatial resolution is 5m,the accuracy of gully erosion parameters was the highest and they were appropriate standard.(4)Through the accuracy of spatial scale conversion model testing,the error of the estimation value of the terrain parameter spatial scale conversion model was relatively small,and the relative error of average slope was the smallest,the error of topographic parameters in TIN was the second,and the relative errors of plane curvature and vertical curvature were the largest.In terms of the gully erosion parameters conversion model,the estimated value errors of the spatial scale conversion models under different SA thresholds were generally smaller than those of the conversion models with different spatial resolutions.This indicated that the applicability of spatial scaling of gully erosion parameters under different SA thresholds was better.The errors of spatial scale conversion model of gully area,gully length and gully density were relatively small.However,the errors of the conversion model of gully heads number and gully heads density were larger.Among them,the error of the spatial scale conversion model with gully heads number was the highest,the fitting accuracy of the model was relatively low,and the applicability was poorer.