Feedback Effect Of Land Cover Change On Land Surface Temperature A Case Study Of The Agricultural Pastoral Ecotone Of Northern China

Researching the interactions between terrestrial ecosystem and atmospheric interface is a critical focus for global change studies.The feedback and response effect of terrestrial ecosystem on global climate change are the focus of global change studies.Land cover change?LCC?,as an important carrier of human activities on the earth's surface,has a profound effect on climate change at the local,regional and global scale.Global warming is an indisputable fact,so revealing how to separate feedback effect of LCC on climate change from the background of global warming and how LCC affects regional climate change through biogeophysical processes will help us to better understand the contribution of human activities to climate change and further study the climate effect of LCC on a larger scale which will provide a theoretical basis for the formulation of global strategy.The Agricultural pastoral ecotone of northern China?APENC?,as one of the four major agro-pastoral ecotones in the world,is a typical ecologically fragile and climate change sensitive area,an important grain and livestock production base and also a natural ecological barrier of the middle and east plain area.However,The APENC has experienced dramatic LCC owing to intensive human disturbances.To curd land degradation,a series of ecological restoration projects have been carried out,particularly since 1999,leading to dramatic LCC in the APENC.The feedback effect of LCC on regional climate change needs to be quantitatively studied and deeply analyzed.In this study,the land cover distribution of APENC were mapped using the optimal DTW approach and time-series MODIS-NDVI?MOD13Q1?data during the period of2001²017,and the evolution characteristics of LCC were revealed at 1a scale.The results of LC mapping were verified from the perspective of traditional method and spatial agreement?compared with GlobLand30-2010 data?.Meanwhile,local verification of land cover mapping results were carried out and evolution characteristics of land cover change were revealed at interannual scale based on the sampling points collected from three times of field investigations in Mu Us sandy land from 2015 to 2017.Then,on the basis of data quality analysis,high quality data reconstruction and accuracy verification,the spatial,temporal and trend characteristics of APENC-LST were analyzed at interannual,seasonal and monthly scales using MODIS-LST?MYD11A2?data.Finally,linear transformation method was used to eliminate or weaken the influence of background climate on LST and diurnal temperature range?DTR?was selected as the index to measure the change of LST.Based on this,we analyzed the feedback effect of land cover type?cropland,forest and grassland?and its change,grassland restoration and degradation on DTR in the APENC.Meanwhile,we explored how LCC affects LST changes through biogeophysical parameters?Albedo and ET?.The results of this study are expected to provide some references for strategies-making of land resource management,scientific control of human land use behavior to mitigate and adapt the global climate change in the APENC.The results showed that:?1?We proposed a land cover mapping scheme based on time-series MODIS-NDVI data and DTW algorithm,which could achieve high accuracy in extracting land cover information in the APENC.The results of MODIS-NDVI reconstruction were optimal when NF and FET were set to 4 and 0.08,respectively.The real gap pattern of time-series NDVI data would significantly increase root mean square error and maximum absolute deviation in high vegetation cover area while it might be ignored for the moderate and low vegetation cover areas.The overall accuracy of LC-2010 mapping is 77.89%,and the Kappa coefficient is 0.71.The producer accuracy of the classification map ranged from 69.83%?grassland?to 100%?waterbody?and the user accuracy ranged from 58%?built-up?to 86%?forest?.The overall spatial agreement between LC-2010 and GL-2010 was 64.44%.Among different LC types,grasslands had the highest spatial matching degree?71.19%?,followed by water body?64.97%?,croplands?61.22%?,forests?48.58%?,and bare land?32.51%?,and value for built-up was the lowest?27.32%?.The results of local verification in Mu Us sandy land showed that the accuracy of land cover mapping is82.16%,and the kappa coefficient is 0.75.?2?The total area of forests,croplands,water body and built-up exhibited an increasing trend while the contrary trend were found for the grasslands and bare land during the 2001²017.Among them,the total area of sparse and moderate grassland decreased,while that of dense grassland area increased.The alternative evolution of cropland and grassland is the main LCC in the APENC.From 2001 to 2017,the total area of cultivated land increased by 61173.18km² with the annual change rate of 1.93%,which showed increasing of the single cropping agriculture and double cropping agriculture.However,the cropland has experienced the change process of increasing from 2001 to 2003 and decreasing from 2003 to2008 and then increasly slowly from 2008 to 2017 at 1a time scale.Grassland area decreased by 35401.12km² with the annual change rate of-0.48%,which showed the decreasing of sparse and moderate grassland.Dense grassland area increased by77877.02km² with the annual change rate of 3.16%,and the increased areas were distributed in the alpine grassland area of Huangnan-Gannan,the typical grassland area of Xilingol League and the alpine meadows of Greater Khingan-Yinshan Mountains.At 1a time scale,the trend of grassland change was opposite to that of cropland.The total area of forest increased by 55111.94km² with an annual change rate of 14.61%.The increase of forest was dominated by mixed forest,which was mainly located in the Lianhua-Taizi-Jishi Mountains in the southwest,the Luliang-Yanshan Mountains in the middle and the Greater Khingan Mountains in the east of APENC.The forest increased gradually and fluctuated at the same time and reached the minimum and the maximum value in 2001 and 2014 respectively.The waterbody and built-up area increased by 249.47 km² and 4290.28 km² with the annual change rates of 0.34%and 14.87%respectively while the bare land decreased by 81955.62 km² with the annual change rate of 4.97%.?3?The temporal,spatial and trend characteristics of LST were different at inter-annual and seasonal and monthly scales.The spatial variation characteristics of APENC-LST were as follows:vegetation cover was closely related to low temperature area,high/low temperature transition area and high temperature area on inter-annual scale.Among them,low temperature zone was mainly located in alpine meadow with high vegetation coverage,forest-grass transition zone and traditional cropland with high productivity;high temperature zone was mainly located in grassland area with low vegetation coverage,sandy area and cropland with low productivity;other areas were high/low temperature transition zone.At seasonal scale,the spatial distribution of LST was different due to the solar radiation,sunshine duration and underlying surface properties.During the day time,the spatial pattern of LST was similar in spring and summer,and was also similar between autumn and winter.At nighttime,the distribution pattern of LST was similar in spring and autumn,while that was quite different in summer and winter.The temporal variation characteristics of APENC-LST were as follows:at inter-annual scale,the LST variation ranged from 7.25 to 9.82?,the mean value was 8.92?,and the relative variability was between-18.73%and 10.01%.The fluctuation of relative variability was greater at night than at daytime.On seasonal scale,the highest mean LST was found in summer,followed by spring,autumn and winter.The sequence of relative variability variation was winter,autumn,spring and summer at seasonal scale and the seasonal scale in daytime while the order from high to low was spring,autumn,winter and summer at nighttime.At monthly scale,LST showed a single peak change pattern with the maximum/minimum value in January and July,respectively.From November to February,the mean LST values were below than 0?while that in other months were all above 0?.The trend variation characteristics of APENC-LST were as follows:at inter-annual scale,the LST of afforestation showed a trend of cooling and the cooling rate was-0.002?/a,while the area proportion of cooling zone was 50.71%.During the day time,LST showed a trend of cooling with the rate of-0.037?/a,while that showed a trend of warming with the rate of0.032?/a during the night time.The area proportion of warming zone reached85.99%.At seasonal scale,the change trends of LST were increased both in spring and summer.The warming rate and area proportion were 0.031?/a,0.02?/a and67.41%,61.82%,respectively.The change trends of LST were quite different in autumn and winter.The cooling rate and area proportion were-0.056?/a,-0.004?/a and 75.38%,50.83%,respectively.?4?The cropland,grassland?including sparse/moderate/dense grasslands?and grassland degradation?grassland to bare land,moderate grassland to sparse grassland and dense grassland to moderate grassland?tended to increase DTR during vegetation growth season.The warming effect caused by the decrease of Albedo was greater than the cooling effect caused by the increase of ET which leading to a rise of T_max.The increase of DTR was attributed to the increase of T_max.In the area of converting cropland to grassland,the significant increase of T_max was the main reason for the increase of DTR due to the simultaneous decrease of Albedo and ET.?5?During vegetation growth period,forest,afforestation?grassland to forest?,cropland reclamation?grassland to cropland?and grassland restoration?bare land to grassland,sparse grassland to moderate grassland and moderate grassland to dense grassland?tended to reduce DTR.The reason was that the cooling effect caused by the significant increase of ET was greater than that caused by the decrease of Albedo,which results in the decrease of Tmax.The increase of Albedo and ET resulted in a significant decrease of Tmax,which was the reason for the decrease of DTR in afforestation areas?grassland to forest?.During the non-vegetation growth period,the cooling effect caused by the significant increase of Albedo was greater than that caused by the decrease of ET in forest land and afforestation areas,which resulted in the decrease of Tmax.That was the reason for the decrease of DTR in these two areas.