| Rill erosion as one of the main forms of hillslopes erosion is an important component ofwater erosion physical process on hillslopes. As the rill erosion process with dynamic, randomand temporal variation of the three major characteristics of nonlinear dynamics, traditionalerosion model is difficult to accurately dynamic simulate and reproduce rill initiation anddevelopment process on hillslopes. Most of the existing erosion and sediment yield models focusonly on the general purpose of soil erosion prediction, which are lacking of space-time simulationof dynamic rill morphology evolution processes. Cellular Automata (CA) is a form up to downdynamic modeling framework, being capable to simulate the spatial-temporal evolutionaryprocess of a complex geographical system. In this thesis the CA numerical model of rill erosionprocess on loess slope is built combining the theories and methods of cellular automata, soilerosion, hydraulics and GIS.The main points are as follow:(1) The model of the simulation of the process of rillevolution is built. The model is defined including cellular station, cellular states, neighbors andtransformation rules. The model calibration and validation are performed with the indoor artificialprecipitation experiment.(2) The influence of cell size on the results and effectiveness of themodel is analyzed and the reasonable cell size is determined.(3) The model is verified. Theresults of artificial precipitation experiment and CA model are analyzed by comparison from theaspects of the process and parameters of morphological development, the process of longitudinalerosion of erosion and the accumulative amount of runoff and sediment.The conclusions were as follow:(1) CA is an efficient method in simulation the process ofrill erosion, the process that the gully headward erosion gradually forming rill is fully exhibit.(2)Compared with other models, the CA and GIS model does not need complicate solution and hassimple calculation process. The coefficient of Nash Sutcliffe of cumulative runoff and erosion is0.898and0.589, the coefficient of error is14%and-22%. The model has better modeleffectiveness.(3) It can be calculated after analysis the CA and GIS model and result ofexperiment that in the process of rill erosion rill length grows faster at erosion stationary phase,rill width and depth grow faster at initial stage and after rill mature, it is more intense to the rilllongitudinal direction erosion at the bottom and middle part of slope, so the measure of soil andwater should be enhanced.(4) The result of model is influenced by the cell size, when the size of cell becomes bigger, the process of rill erosion will accelerate, the process of rill erosion willaccelerate, at the same time, the cumulative runoff will decrease and the cumulative erosion willincrease. Only the ratio of cell length and time step approached to the flow velocity of overlandcan improve the effectiveness of model. |
PDF Full Text Request