Remote Sensing Of Spatio-temporal Change For African Wetland And Simulation Analysis Of The Response To Climate

The numerous ecosystem services provided by wetlands are critical for protecting biodiversity,mitigating climate change and promoting human well-being.However,as an important part of the world’s wetland ecosystem,there is still no thematic wetland map to provide information on the location,distribution,area and change of African wetlands,so as to help decision makers formulate effective protection policies and management plans to better protect and develop wetlands.In order to extract the distribution and change information of wetlands in Africa accurately and in detail,a remote sensing classification system of African wetlands was established.Wetlands across Africa are divided into two categories:inland wetland and coastal wetland.Inland wetland includes inland surface water,inland salt pan,inland marsh,and inland swamp.Coastal wetland includes coastal swamp,coastal marsh,tidal flat,and shallow marine water.After completing the large-scale uniform sampling of African wetlands and the construction of multi-source remote sensing features(spectral,topographic and hydrological)of wetlands,a random forest classification method based on multi-source remote sensing features and a seasonal wetland range delineation method based on Otsu algorithm were developed for the unique temporal dynamics,spatial heterogeneity,and land-water transition characteristics of wetlands.The random forest classification method based on multi-source remote sensing features can effectively improve the accuracy and mapping effect of African wetland classification.Compared with only using spectral and topographic features,the addition of hydrological features improves the classification accuracy by 8.31%,and effectively reduces the salt and pepper phenomenon in the wetland map.The seasonal wetland delineation method based on Otsu algorithm solves the problem of dynamic wetland mapping.The range of seasonal wetlands will expand and shrink with the change of local hydrological conditions,resulting in errors in the identified wetland sampling points,thus affecting the accuracy of supervised classification.The classification method based on the Otsu algorithm does not require wetland training data,and it can quickly and accurately realize the classification and mapping of African seasonal wetlands in different periods by automatically selecting thresholds for image segmentation based on the reference of Landsat time series and African climatic information.Based on the wetland remote sensing classification study,eight categories of wetlands across Africa were classified and mapped from 1984 to 2021(10 periods)using 270,000 sampling points,810,000 Landsat images and 96 feature bands.This study generated the first systematic and comprehensive set of thematic wetland maps(10 maps)in Africa,quantifying the distribution of existing wetlands in Africa at a spatial resolution of 30 meters,as well as the spatial and temporal changes in Africa’s wetlands during the past 38 years.The results show that the total area of existing wetlands in Africa is about 1.3415 million km~2,excluding shallow waters,wetlands account for 3.98%of Africa’s total land area.Of this,inland wetlands account for 81%of the total wetland area and coastal wetlands account for 19%.African wetlands are mainly concentrated in the tropics,and the wetland area between 20°N and 20°S can account for 97%of the total area.Among all wetland types,inland swamps account for the largest area of 411,500 km~2.Marsh is the second largest wetland type in Africa,and its area is highly dynamic.In the dry season,the area of marshes is only 202,200 km~2.With the arrival of the wet season,the area of marshes in Africa can increase by 186,400km~2,with an increase of 92.22%.The spatio-temporal change assessment of African wetlands from 1984 to 2021shows that there was no large-scale loss of wetlands in Africa as a whole in the past 38years(0.51%loss and an area of about 6,000 km~2).The area of wetland loss was concentrated in the coastal area(9.64%),and the area of inland wetland increased slightly(0.5%).By comparing the time series of wetland area and related driving factors,it is found that the correlation between temperature and wetland change is not strong,and the change of precipitation explains the violent fluctuation of wetland time series in a specific year.Soil moisture is affected by climate factors such as temperature and precipitation,which reflects the state of wetlands to a certain extent.It is the most relevant climate factors to wetland changes(Spearman correlation coefficient is 0.73).Among all the driving factors of inland wetlands,the human activity is the least important and not significant.The direct impact of agricultural reclamation on African wetlands is also significant.African marshes are the most serious wetland type affected by agricultural land reclamation.15.29%of the loss area of coastal marshes is caused by farmland reclamation,and the proportion of reclamation is increasing year by year.By comparing the time series,calculating Spearman correlation coefficient and analyzing the importance of random forest,we conclude that human activities aggravate the loss and degradation of coastal wetlands,and the impact of climate change on inland wetlands is more obvious.Based on the study of wetland loss and driving forces in historical periods,CMIP6climate simulation data and TOPMODEL method were further used to predict future changes in inland wetlands to assess the response of African wetland ecosystems to future climate change.The results suggest that future climate change impacts on African wetlands tend to be positive.Although it causes the loss of wetlands locally,it will further increase the area of inland wetlands in Africa as a whole.The models predict show that the area of African wetlands will increase by 11.74%,23.50%,42.35%,and41.26%,respectively,by 2100 under the SSP126/245/370/585 scenarios.The study has deepened the understanding of the past,present and future of African wetlands and will greatly contribute to the conservation,management and development of existing wetlands in Africa,which will be conducive to the mitigation of global climate change,the conservation of biodiversity and the improvement of the living standards of the African people.