Research On Improvement Of The DLA Prediction Model For Multifractal River Network

Multifractal theory is an important technical tool for analyzing and deconstructing the mechanism of complex systems.The scientific quantification of watershed morphology using this theory is one of the hot research topics in geomorphology.Meanwhile,computer simulations are introduced into the river dynamics field to provide a basis for studying fractal river networks in watersheds.Diffusion-limited aggregation(DLA)model has been widely used to simulate fractal aggregation processes.This improved DLA model can reasonably plan and manage rivers within the basin scale.This thesis took China's northwestern basin(Daxia River)as the research object.First,GIS image processing technology was used to process remote sensing DEM image data of this watershed.Then,combined with the multi-fractal algorithm,the scale-free and multi-fractal characteristics of the topographical water system in the basin were verified.The DLA river network simulation algorithm was further designed based on this feature.Finally,through CPS fusion modeling theory,the improved DLA algorithm was applied to predict the fractal river network of a real watershed for practical applications.The main work is as follows:(1)By processing remote sensing images of the Daxia River basin.Using the multiple fractal models,the multifractal features and quantified feature parameters of the water system distribution(Daxia River and 7 sub-basins)were extracted and analyzed holistically and distributively.The geomorphic significance of the spatial feature differentiation in the watershed was explored.Results show that the multiple fractal spectrum of the water system in the basin is right hook-shaped((?)f?0),the span(?)a of the singularity index is bounded by 0.860?1.160 with relatively small values.This indicates that the Daxia River system watershed shows a uniform distribution in most of the area,the water system is less well developed.Meanwhile,(?)and (?) satisfy a certain power-law relationship,indicating that the distribution characteristics of the river network water system is scale-free within a certain range of scale variation.There is a non-linear relationship between(?)and q,which indicates that the basin geomorphological features(Daxia River system)have obvious multiple fractal characteristics.This has important implications for scientific solutions to complex problems in geomorphology.(2)The scale-free nature of the water system(Daxia River basin)and the multiple fractal characteristics of the geomorphology are used as entry points.By improving the defects of the particle aggregation and growth method in the diffusion-limited aggregation(DLA)model,a fractal algorithm scheme was designed to simulate the evolution of river networks.Based on the environmental factors in the study area,runoff classification,slope probability,and contour features were analyzed.Establish the calculation method of unit-gradient growth,anisotropic Brownian motion and the optimization mechanism of other related parameters(probability-position,step-size).A two-dimensional fractal river network was simulated under different conditions.The simulation calculated the bifurcation ratio of the simulation effect map by Horton's law,and its values were all bounded between the values of the natural river channel(3?5).Therefore,the improved model more truly reflects the structural characteristics of the river network through quantified indicators.Meanwhile,the sensitivity test of the model was analyzed and evaluated based on the relationship between the number of particles and the fractal dimension.Its sensitivity coefficient gradually increased from 2.75 to8.38 based on a suitable data set.This will bring the predicted watershed hydrology closer to the level of river network development.(3)According to the designed scheme for improving the DLA simulation river network,GIS remote sensing image processing was applied to extract topographic features of the watershed based on environmental factors of the study area.Combined with Cyber-Physical Systems(CPS),the improved DLA model factors and the extracted geomorphic feature factors were transformed and integrated.An improved growth DLA model was developed based on landform features.Thus,the fractal river network in the study area can be predicted and analyzed.Meanwhile,the structural similarity between the predicted river network and the extracted real river network was verified and analyzed by applying the nipple cosine method,the values were all greater than 0.99.This indicates that the improved DLA model has obvious effect in simulating fractal river networks.Two-dimensional river networks with different morphologies can be truly reproduced and predicted.This will ultimately promote the scientific simulation and prediction of river network management in the basin,and realize the sustainable development of river basin ecological environment management.