Landscape Dynamic Change Of East Dongting Lake Wetland And Its Impact On Lesser White-fronted Goose

Wetland is one of the most productive ecosystems on the earth,which plays an important role in maintaining the balance of the ecosystem.As the second-largest freshwater lake in China,Dongting Lake is one of the key reserves for wetland biodiversity conservation.Additionally,Dongting Lake is an important wintering migratory bird habitat on the East Asia-Australian migratory bird migration line,which is one of the most critical wintering grounds for the globally vulnerable species(Anser erythropus).The change of Dongting Lake wetland has an important influence on the lesser white-fronted goose.Therefore,it is necessary to analyze and clarify the ecosystem characteristics and landscape pattern changes of the Dongting Lake wetland and the impact mechanism on the small white-fronted geese population at spatial and temporal scales.The 2011-2020 Landsat series multi-temporal remote sensing images in winter(October of the year to March of the following year)are selected as the main data source in this study.The effect of different machine learning algorithms(random forest,support vector machine,and gradient boosting decision tree)and remote sensing feature combinations in the East Dongting Lake wetland classification is analyzed on the Google Earth Engine platform.Based on the accurate extraction of wetland information,the dynamic change characteristics of the East Dongting Lake wetland are analyzed using the centroid spatial migration model and the trend analysis method.The landscape change characteristics of the East Dongting Lake wetland are probed through different landscape pattern indexes in the landscape level.The response mechanism of the population structure of small white-fronted geese to potential impact factors such as vegetation and landscape changes is explored used the univariate linear correlation model.The main conclusions of the study are as follows:(1)Integrating remote sensing features such as spectral information,vegetation index and topographic factors can significantly improve the accuracy of wetland classification.When only using spectral information,the classification accuracy OA and Kappa coefficients were 85.86%and 0.82,respectively;after adding vegetation index,OA and Kappa coefficients increased to 86.71%,0.83.The classification performance is the best(OA=94.14%,Kappa=0.92)when using spectrum,vegetation index,and topographic factors.Among the three classification algorithms,the random forest algorithm achieves the best classification performance,with an overall accuracy of 94.14%(OA)and a Kappa of 0.92.The classification accuracy of gradient boosting decision tree and SVM algorithms are OA=92.71%,Kappa=0.90;OA=79%,Kappa=0.72,respectively.(2)The NDVI(Normalized Difference Vegetation Index)of the East Dongting Lake Wetland in the winter of 2010-2020 has obvious fluctuation characteristics,and the overall trend is decreasing(θslope=-0.0064,p<0.01).The landscape centroid migration characteristics of the East Dongting Lake wetland are temporal and spatial heterogeneity.Different wetland types show violent centroid spatial migration characteristics from 2019 to 2020.The center of the mass migration path of East Dongting Lake wetland is from southeast to northwest and from west to east.The migration characteristics are consistent with the dynamic characteristics of the NDVI region.The spatial migration characteristics of the centroid of each wetland landscape type are different.Both the mudflat and meadow centroid coordinates have undergone significant migration,of which landscape centroid migrate from east to west;the position of the reed’s centroid is relatively stable and only migrates at latitude;the centroid of water change back and forth from southwest to northeast.The difference in the migration of different wetland landscape centroids reflects the changes in the ecosystem stability of each wetland landscape type under the background of driving forces.(3)During 2011 to 2020,the area of each landscape type of East Dongting Lake wetland has changed significantly,but the total area of wetland has basically remained stable.The main landscape types of the East Dongting Lake wetland are waters,mudflats,and reeds,followed by grass islands,other(including cultivated land,construction land,etc.),and arbor forests.During the study period,the fragmentation of the East Dongting Lake wetland landscape has decreased,and the shape of the landscape became simpler.In addition,landscape connectivity and landscape richness are increasing.(4)From 2015 to 2019,the number of wintering geese in Dongting Lake Wetland showed a stable trend.As landscape shape complexity increased,the number of lesser white-fronted geese increased.The increase of complexity of wetland landscape is conducive to the wintering habitat of lesser white-fronted goose.There was a weak positive correlation between the number of lesser white-fronted goose and the area change of water area,grassland,and reed.A weak negative correlation was shown between the number of lesser white-fronted goose and the area change of arbor forest,mudflat,and other types,and the NDVI of lesser white-fronted goose habitat.