| The landscape ecology risk assessment aims to reveal landscape pattern changes caused by climate change and human activities in a region.It provides a basis for identifying vulnerable areas of ecosystems and implementing targeted management through identifying the ecological risks.Existing studies have explored the evolution mechanism of regional ecological value and risk through a serial of methods for assessing ecological risks.However,most existing methods represent human activities using land use and land cover changes,with insufficient application of social and economic indicators such as population and GDP.The richness of human activity information could be improved in ecological risk assessments.Given the above,to address this problem,this study proposes an ecological risk assessment method that integrates climate characteristics,land use changes,population,and GDP indicators.Additionally,to better reflect the nonlinear features of climate change and human activities driving regional ecological risks over longer time scales,a combination of geographically weighted regression and deep learning models is used to establish a regional ecological risk prediction model.The model is then applied in the Hanjiang River Basin.The Hanjiang River basin is an important area in central China.Its ecological protection is crucial for promoting sustainable development of the Yangtze River Economic Belt.In recent years,due to changes in the natural environment and the impact of human activities,disturbances to the ecological environment of the Hanjiang River basin have intensified,and potential environmental problems have become increasingly prominent.Against this backdrop,this study establishes an assessment model for landscape ecological risk in the basin by applying a geographical weighted regression model that comprehensively considered climate characteristics,land use changes,population,and GDP indicators,using four sets of 30-meter resolution remote sensing data in the years 2000,2005,2010,and 2015.The relationship between the aforementioned indicators and the changes in landscape ecological risk are identified at the grid scale.Moreover,an artificial neural network is established to re-learn the assessment results for a more direct and comprehensive explanation of the relationship between influencing factors and changes in landscape ecological risk,basing on the advantages of deep learning in handling non-linear relationships globally.The artificial neural network is then used to predict changes in ecological risk in the basin under climate change scenarios.The following conclusions were drawn:(1)Among the land use type changes in the Han River Basin from 2000 to 2015,forest land changes accounted for the largest proportion.The changes in grassland area,forest,lakes and reservoirs showed a decreasing trend from 2000 to 2015.(2)Landscape pattern change showed a relatively small change in landscape dominance,while there was an upward trend in landscape loss degree index.The maximum patch index showed a decreasing trend,while patch density showed an increase.The contagion index showed a decreasing trend,while Shannon’s diversity and evenness indices showed an increasing trend.(3)In terms of ecological risk assessment,the high ecological risk area in the Han River Basin increased by 25.72% from 2000 to 2015.The spatial autocorrelation Moran’s I index value of the high ecological risk areas ranged from 0.933 to 0.936,indicating that the ecological risk in the basin has significant spatial clustering.Risk hotspots were mainly concentrated in the area downstream of Danjiangkou,while cold spots were mainly concentrated in the Tangbai River Basin area.(4)In predicting ecological risk under climate change,risk assessments were conducted using Shared Socioeconomic Pathways(SSP126,SSP245,SSP370,and SSP585)and Representative Concentration Pathways(RCP).Under SSSP126 and SSP245,the area of lower ecological risk level increased.Under SSP370 and SSP585,the area of high ecological risk level increased. |
PDF Full Text Request