| Under the dual role of rapid economic development and complex global climate changes,in order to support the provinces of the Yellow River Basin to win the fight against poverty and solve the ecological security problems concerned by the people in the basin,China has put forward the major national strategy of the ecological protection and high-quality development of the Yellow River Basin,and strive to promote wetland protection and ecological management in the lower reaches of the Yellow River.The Yellow River Delta is located at the estuary of the Yellow River.It is a typical ecological function area in the Yellow River Basin.The sea and land changes are active,and it is currently facing three severe ecological challenges.In order to deeply study the characteristics,laws and patterns of wetland ecosystem changes in the Yellow River Delta in recent years,this study took the wetland ecosystem in the Yellow River Delta as the research object,and used the Sentinel-2 remote sensing images from 2015 to 2022 as the data source.The classification schemes were constructed by comprehensive utilization of spectral features,texture features,vegetation and water index,red-edge index.The wetland classification information in the study area was extracted based on Random Forest algorithm.The dynamic changes of wetland ecosystem in the study area in recent years were analyzed by geographic information map,patch gravity center model and landscape pattern index.Based on Patch-generating land use simulation model,the spatial distribution of wetlands in the Yellow River Delta in 2030 was predicted under three scenarios of inertial development,cultivated land protection and ecological priority development.The specific research results are as follows:(1)Random Forest algorithm have the advantages of removing redundancy of feature variables and improving classification accuracy.Red-edge index could improve the accuracy of wetland information extraction and classification.The classification schemes composed of feature optimization based on Random Forest algorithm have the highest accuracy,and the overall accuracies in eight periods were all higher than 90%,indicating that the Random Forest algorithm could optimize the feature variables,and the classification system was suitable for the Yellow River Delta.The overall accuracy of the feature variable combination containing the red edge index was 85.59%,which was higher than that of the feature variable combination without the red edge index.And the contribution of red-edge index in the optimal feature combination was 37.5%.(2)In 2015-2022,the total area of wetlands in the Yellow River Delta increased,spatially developing along the Yellow River basin in southwest-northeast direction.The area of natural wetlands increased by 223.23 km~2,of which the river increased by 324.39 km~2,and 30.60%was converted from unused land.The area of shallow sea decreased by 129.36 km~2,including 53.15%converted to rivers and 22.79%converted to reservoirs and ponds.The area of reservoirs and pond increased by 238.38 km~2,while the area of paddy field and salt pans decreased by 55.93km~2 and 94.72 km~2 respectively.And the area of artificial wetlands net increased by 87.73 km~2.At the same time,the landscape shape index and Shannon’s diversity index were the largest in2021,the contagion index was the smallest,and the landscape distribution was scattered and fragmented.(3)In different periods,the dominant factors and interaction factors affecting the spatial distribution and area change of wetlands in the Yellow River Delta were different.In 2015-2022,the dynamic change of constructed wetland was greatly affected by vegetation coverage.The dominant factors of natural wetlands were precipitation(2015,2016)→DEM(2017,2018)→vegetation coverage(2019)→DEM(2020)→vegetation coverage(2021)→temperature(2022).In terms of interactive dominant factors,natural wetlands were dominated by precipitation∩vegetation coverage(2015,2016)→temperature∩vegetation coverage(2017)→temperature∩sunshine hours(2018)→temperature∩vegetation coverage(2019-2022).And the interactive dominant factors of constructed wetlands were precipitation∩vegetation coverage(2015)→sunshine hours∩vegetation coverage(2016)→slope∩vegetation coverage(2017)→temperature∩vegetation coverage(2018-2022).(4)Compared with the current situation of land use in 2022,the area of paddy field and dry land in the Yellow River Delta in 2030 decreased by 92.92 km~2 under the constraint of inertial development scenario,while increased by 30.25 km~2 under the constraint of cultivated land protection scenario.Under the constraint of ecological priority development scenario,the natural landscape area is 2463.92 km~2,which was significantly larger than the other two scenarios. |
PDF Full Text Request