Landscape Ecological Pattern Temporal And Spatial Changes And Evapotranspiration In The Ulan Mulun River Basin

Under the dual pressure of socioeconomic development and increasing population density,watershed landscape ecological patterns are prone to differential change trends,which also affect moisture variables such as surface evapotranspiration.To understand the spatial and temporal variation characteristics of watershed landscape ecological patterns,the commonly adopted methods are to improve the accuracy of key input parameters of the model and to integrate the influencing factors and environmental variables.In this study,the coal-rich Ulan Mulun River basin is used as the study area,and an FCN-CRF deep learning classification algorithm adapted to the whole basin is proposed.The spatial distribution characteristics of landscape ecological risk in the watershed are studied based on the landscape ecological risk index.The habitat quality index of the Ulan Mulun River basin was estimated by the In VEST-HQ model under the comprehensive consideration of multiple influencing factors,and the RSEI values of the basin were calculated.Considering the limitations of macroscopic scale for surface studies,corresponding validation experiments were also designed in this paper.Based on the improved Faster R-CNN,the remote sensing index values of trees in recent years were identified and calculated.In the final section,this study explores the characteristics of the influence of landscape ecological patterns on surface evapotranspiration.The main findings derived are as follows:(1)The main land use types in the Ulan Mulun River basin are seven categories,including water bodies,woods,grasslands,bare land,urban land,industrial and mining enterprise land,and agricultural land.The FCN-CRF deep learning feature classification algorithm proposed in this study works best with an overall accuracy of 93.67% and a kappa coefficient of 0.92.In the last 20 years,the area of water bodies increased from 1.38%in 2000 to 2.10% in 2020,the area of forest land decreased from 0.96% in 2000 to 0.51%in 2020,the area of grassland increased from 52.17% in 2000 to 62.89%,and the proportion of bare land decreased from 30.05% to 14.42%.Urban land increased from 3%to 9.76% in 2000,and coal mine land increased from 0.98% to 2.24%.Agricultural land use decreased from 11.44% to 8.06%.The main type of transfer out is bare land,and the type of transfer in is industrial and mining enterprise land and grassland,and the ecology of the whole watershed shows a healthy development trend.(2)In the study of landscape pattern of Ulan Mulun River watershed,different feature types showed different trends of change.The landscape loss of water bodies,bare land,trees and farmland is increasing and becoming less stable.The high landscape ecological risk areas are distributed along the valley of Ulan Mulun River,and the low landscape ecological risk areas are mainly distributed in the areas with low intensity of human activities.(3)Between 2000 and 2020,the ecological quality index of the entire watershed showed an increasing trend,and the ecological quality of the entire watershed tended to become better.Among the seven feature types,water bodies have the highest habitat index values.Habitat indices of bare land,towns and industrial and mining enterprises sites are lower.The western part of the watershed showed a degradation trend during 2005-2010.The habitat index showed a positive correlation with natural factors,while the habitat index showed a negative correlation with human activity factors.The RSEI index also showed an increasing trend over the 20-year period.(4)During the 20 years,the evapotranspiration in Ulan Mulun River basin showed an increasing trend.The evapotranspiration in the southeastern part of the basin was large,while the evapotranspiration in the northwestern part was small.In the whole basin,the potential evapotranspiration has been decreasing year by year in the last 20 years,and the rainfall has been increasing year by year.The percentage of evapotranspiration varies from season to season,from large to small in summer,spring,autumn and winter.The coefficient of variation of evapotranspiration is larger in the southern part of the basin,with a maximum value of 0.26.The coefficient of variation is smaller in the middle and upper parts of the basin,with little interannual variation.The order of evapotranspiration for different land use types is,from largest to smallest,watershed > grassland > farmland >woods > urban land > industrial and mining enterprises > bare land.The landscape grid with low landscape ecological risk index showed a positive correlation with evapotranspiration.Areas with high ecological quality index showed significant positive correlation with evapotranspiration,and areas with low ecological index showed significant negative or non-correlation with evapotranspiration.