Monitoring Soil Erosion And Modeling The Sediment Transport In Small Loess Catchments Based On Digital Elevation Model

The Loess Plateau of China is one of the regions with the most frequent surface material exchange and the most dramatic morphological changes.Small catchments are not only the natural geographical unit,but also the basic unit of watershed management.A large number of researches have focused on the mechanism and modelling of soil erosion in small loess catchments.However,simulating and quantifying the sediment transport in space is still challenging.Monitoring and simulating sediment transport are crucial and significant for soil and water conversation,agricultural production,river management,and so on.Hence,this study aims to develop a set of methodological frames to monitoring soil erosion and quantify sediment transport in space.In order to monitoring soil erosion in small loess catchments,an optimized work frame for UAV photogrammetry was proposed at first.Then,based on the highprecision Digital Elevation Model(DEM)established by the optimized Unmanned aerial vehicles(UAV)photogrammetry,a topographic change detection method that considers the spatial autocorrelation of DEM errors was constructed for effectively evaluating the erosion and deposition budget.Finally,according to the principle of mass conservation,the sediment transport rates were inferred from the topographic changes at pixel level.This study explored the research paradigm of inferring surface processes from topographic changes.The main conclusions are as follows:(1)This study summarized the principle of mass conservation in surface flow and analyzed its application conditions in loess catchments.High precision DEM is an important base when using mass conservation to infer sediment transport.Moreover,it is necessary to clarify the primary and secondary relationship between wind erosion,hydraulic erosion,and wind deposition in the study area.(2)This study explored the effects of camera inclination,flight height,direct geographic georeference and control point quality on vertical accuracy of DEM.Then,an optimized work frame for UAV photogrammetry was proposed.The results from three loess catchments showed that,in the case of no control measurement,oblique photography(especially the camera angle is greater than 20 degrees)can reduce the correlation of camera distortion parameters,and then mitigate systematic error.The influence of flight height on DEM accuracy is related to the camera inclination angle.By using oblique photography,the sensitivity of DEM accuracy to flight height is reduced.In the case of control measurement,compared with vertical photography,oblique photography needs slightly more control points to reach high accuracy.In the field application,the measurement scheme can be flexibly selected according to the above conclusion to improve the measurement accuracy.(3)A novel method considering spatial autocorrelation of DEM errors is proposed to detect topographic changes.The influence of the significance threshold segmentation and spatial autocorrelation of random error on change detection is explored.The results show that the significance threshold segmentation is very important for the calculation of gross erosion and gross deposition in the stable terrain area,but it has little effect on net erosion.Given the ordinary catchments usually consists of stable,erosion and deposition areas,it is necessary to use the threshold segmentation to detect terrain changes.The spatial distribution of DEM error produced by UAV photogrammetry can be quantified by Monte Carlo bundle adjustment simulation.Then,root mean square error(RMSE)should be replaced by the spatial distribution of DEM error for error propagation and change detection.(4)According to the runoff characteristics of loess catchments,the method for quantifying sediment transport from topographic change was established.It can effectively evaluate the sediment transport pathway and flux(i.e.the spatial distribution of sediment transport rate).The method consists of two sub-methods: one-dimensional(1D)and two-dimensional(2D)methods.The 1D method accumulates the amount of terrain change in each cross-section and propagates it to the downstream,that is,to view the changes of sediment transport rate from the perspective of the longitudinal section(upstream to downstream).The 1D method can provide an overall trend of sediment transport rate from upstream to downstream.The variation of sediment transport rate with the upstream or downstream distance can reflect the development of branch gullies.The 2D method is to infer the sediment transport from terrain changes at pixel scale.How to design sediment routing algorithms according to different surface runoff processes is the key of this method.This study explored the effectiveness of different routing algorithms in hillslope and gully areas.The results from a laboratory catchment show that the multiple direction algorithm based on slope index should be used in the slope area;the routing algorithm based on hydraulic simulation should be used in the gully area;the coupling method of the former two should be used to simulate the spatial distribution of sediment transport rate for the whole catchment.The proportion of negative sediment transport rate(mass non-conservation)is only 0.67%-9.97% in the study area,which indicates that the proposed method is successful.In addition,this study also discussed the application conditions of the proposed methods,as well as the influence of topographic change detection,DEM temporal and spatial resolution,the spatial difference of soil bulk density,and model parameterization on the proposed method.(5)The method for quantifying sediment transport from topographic changes was validated by four field loess catchments.This study also analyzed the relationship between sediment transport rates and the characteristic indices of catchments.The application of the spatial sediment transport rate was demonstrated by taking sediment connectivity as an example.The results from four catchments showed that the method can effectively quantify sediment transport in space.The proportion of negative sediment transport rate in the four areas is only 2.53%-7.85%,and the negative transport area is mostly affected by human activities.The correlation analysis shows that the transport rate is positively correlated with upstream catchment area,aspect ratio,and three hydrological parameters(i.e.SPI,TWI,and LS).In terms of application,the spatial distribution of sediment transport rate and sediment connectivity has brought new development to the study of surface processes.In this study,a set of methodological frames for monitoring soil erosion and quantifying sediment transport in space was proposed.The research of sediment transport rate is promoted to the pixel level.The spatial distribution of sediment transport rates reflects the exchange intensity of surface materials,which is likely to bring new insights into geomorphology in the immediate future.