| Desertification is one of the most serious environmental problems facing humanity.China is a country severely affected by desertification,which has brought serious environmental and economic pressure to our country,especially in the arid areas of the northwest.The Qaidam Basin,located in the ecologically fragile area of the northwest and extremely sensitive to climate change,has the characteristics of concentrated distribution,complete types,and high degree of development in terms of desertification.Due to the high salt content of the soil in the basin,saline desertification is severe,further exacerbating environmental and survival issues.Therefore,it is necessary to analyze the spatiotemporal distribution characteristics and evolution trends of basin desertification,and perform remote sensing inversion on typical saline desertification areas,in order to provide scientific basis for the prevention and control of basin desertification and ecological environment evaluation.In recent years,there has been a significant trend of warming and wetting in the basin,and human intervention has intensified.For this reason,considering the influence of climate factors,natural environmental factors,and human factors comprehensively,the driving forces of spatiotemporal variability of desertification in the basin are analyzed,and the future climate change and desertification in the basin are estimated.This has indicative significance for revealing the driving mechanisms of desertification evolution in the basin and the Qinghai-Tibet Plateau.This study uses the vegetation coverage index(FVC)as the monitoring indicator for desertification,and uses spatiotemporal statistical analysis methods such as geographic detectors to comprehensively analyze the spatiotemporal distribution characteristics,spatiotemporal evolution characteristics,and driving forces of basin desertification under the background of climate change;Under the 6th International Coupled Model Comparison Program(CMIP6),the temporal and spatial trends of desertification in the Qaidam Basin under different scenarios were simulated and predicted.Based on the spatial-temporal distribution characteristics of desertification in the basin,the typical saline desertification areas that have the greatest impact on human production and life were selected,and the quantitative inversion of saline desertification was carried out using decision tree(DT),K-nearest neighbor regression(KNN),random forest(RF),gradient elevator(GBM),extreme random forest(EF)and gradient lifting tree(GBDT)algorithms,combined with a variety of spectral indexes and measured salt content data.On this basis,the spatiotemporal variation characteristics and influencing factors of saline desertification in the study area were analyzed.The main conclusions drawn are:(1)The average annual temperature of the Qaidam Basin from 1961 to 2021increased significantly at a rate of 0.446℃/10a,much higher than the average levels of the world,China,and the Qinghai-Tibet Plateau,with the most obvious increasing trend in the central and western regions of the basin.The increase in precipitation in the basin is higher than the average increase in the Qinghai-Tibet Plateau,with an overall significant increase at a rate of 10.28 mm/10 years.Among them,the precipitation in the eastern part of the basin,the Qilian Mountains in the northeast,and the Kunlun Mountains in the south show a significant increasing trend.Overall,the Qaidam Basin is the region with the fastest increase in temperature and precipitation in China and the Qinghai-Tibet Plateau,and is a climate sensitive zone with significant warming and humidification.(2)In CMIP6 climate system model,the optimal model of Qaidam Basin temperature is Tai ESM1,and the optimal model of precipitation is Nor ESM2-LM.The trend of basin warming in the future from 2015 to 2100 is significant,and it increases with the increase of radiation forcing,while precipitation shows a decreasing trend with the increase of radiation forcing.Comparing the increasing trend of precipitation and temperature in the basin in recent years,it was found that the basin is currently in line with the scenario of moderate social vulnerability and moderate radiative forcing in SSP2-45.Under the SSP2-45 scenario,the precipitation growth rate of the basin from 2015 to 2100 is 15.27 mm/10a,and the warming rate is0.45℃/10a.(3)The spatial differentiation of basin desertification is obvious.The southern Kunlun Mountains,the northeastern Qilian Mountains,and the northern alluvial fan margin of the Kunlun Mountains have a low degree of desertification,mainly because these regions are relatively rich in water resources.The higher degree of desertification in the central and western regions is due to less precipitation and greater evaporation.The analysis of geographical detectors shows that the driving forces of different factors affecting desertification are in the following order:evaporation(0.481),precipitation(0.460),vegetation type(0.36),soil type(0.293),temperature(0.251),slope(0.091),river buffer zone(0.08),wind speed(0.077),road buffer zone(0.074).Therefore,the main controlling factors of spatial differentiation of desertification in the basin are evaporation and precipitation in climate factors,and vegetation type and soil type in natural environment factors.Although human activities have a relatively small impact(using road buffer zones as indicators),they still have a significant impact on desertification that cannot be ignored.The interaction between factors shows mutual enhancement and nonlinear enhancement,indicating that considering the combined effects of multiple driving factors is more conducive to explaining the spatial heterogeneity of desertification.(4)From 2000 to 2021,the increasing rate of FVC index in Qaidam Basin was0.029/10 a,and the overall improvement trend of desertification in recent 22 years was obvious,among which the most obvious improvement was in the Kunlun Mountains in the south of the basin,Qilian Mountains in the northeast,eastern and alluvial fan front areas.The areas with significant improvement in desertification in the Qaidam Basin from 2000 to 2010,and from 2010 to 2021 accounted for 29%and6%,respectively.The improvement trend of 2011-2021 is weaker than that of 2000-2010 because the increasing trend of temperature rise and precipitation in 2000-2010is more significant.(5)Under the four scenarios,the average annual FVC index of the basin shows a decreasing trend in the future,indicating that desertification in the basin is intensifying from 2015 to 2100.In the SSP2-45 scenario,the future FVC index has the smallest decrease,with a decrease of 0.002/10a.The trend of desertification exacerbation is the weakest in this scenario.In the future FVC index time series of the basin,the upward fluctuation trend of the FVC index tends to be stable from 2015 to2040,and the decreasing trend is obvious after 2040,indicating that the improvement trend of desertification in different scenarios stops around 2040 and then intensifies significantly.(6)The soil type in the middle and lower reaches of Golmud River in the typical saline desertification research area is dominated by meadow saline soil.The salt content of meadow saline soil reaches 22.04%.The soil with different salt content has similar multi spectral reflectance characteristics,and its reflectance increases with the increase of soil salt content.Research has found that using surface temperature(LST)can effectively identify high salinity saline alkali land and dry salt beach distribution areas.Among the six inversion models of DT,KNN,RF,GBM,EF and GBDT,the one with the largest determination coefficient(R~2)and relative analysis error(RPD)and the smallest average absolute error(MAE)belongs to the GBDT model,that is,the GBDT model has the strongest prediction ability and the smallest error.(7)Using the optimized GBDT model to invert the salt content of the study area from 2000 to 2020,it was found that there was no significant fluctuation in the soil salt content of Gexi Farm and Baoku Village,indicating that agricultural production activities did not exacerbate saline desertification.However,in the mountainous groundwater overflow zone,which is the area with concentrated grassland distribution,the soil salt content fluctuates significantly and belongs to a strong variation zone of saline desertification,showing an overall trend of intensification weakening intensification weakening.This is related to the fluctuation of groundwater level in the downstream groundwater overflow zone,which is controlled by the amount of upstream ice and snow melting and precipitation.The fluctuation of water level causes salt minerals to migrate to the surface through capillary action,changing the content of 0-10 cm of topsoil.Compared with the years 2000 to 2010,the saline desertification in the region showed a decreasing trend from 2011 to 2020,which is also related to the slowing down of the increasing rate of upstream temperature and precipitation in recent years.Overall,this study comprehensively explored the spatiotemporal evolution characteristics of desertification in the Qaidam Basin under the influence of multiple factors,and explored its driving mechanisms.Based on the CMIP6 climate model,the possible spatiotemporal change trends of future desertification in the basin were estimated.A typical saline desertification area in the basin was selected,and six machine learning algorithms were used to model and optimize the optimal model.The dynamic change inversion at the interannual scale was achieved,and the spatiotemporal variation characteristics and influencing factors of saline desertification in the study area were explored.This study has reference significance for the study of desertification in other extreme arid areas. |
PDF Full Text Request