| Grasslands play an important role in the terrestrial carbon cycle,water conservation and food supply.China is a major country in grassland resources and the world’s largest carbon emitter.Human efforts to mitigate climate change have prioritized carbon stored in trees while the role and value of grasslands is often underestimated.Grassland productivity or biomass accumulation is an important way to sequester atmospheric CO2 and enhance carbon sequestration,with the net primary productivity（NPP）is the most commonly used indicator to reflect the carbon sequestration capacity.Therefore,accurate estimation of grassland NPP is crucial.Over the past 20 years,many scholars have conducted a series of studies on grassland NPP across China,however,the results differed widely and many uncertainties remained in the estimation.This is due to differences in the extent of grassland,estimation methods and data sources used in different studies.At the same time,there is a lack of in-depth discussion on the processes and causes of NPP change.To address the above limitations,this paper evaluates the accuracy of grassland distribution in 10 authoritative land cover data in domestic and abroad based on 13129 validation samples,and clarifies the most accurate current grassland extent;we used 5455 grassland NPP samples collected in nationwide field surveys between 2002 and 2019,and combining vegetation,topography,soil and then a comprehensive remote sensing estimation model for grassland NPP was established.We develop a national grassland NPP dataset with 1 km spatial resolution year by year from 2001 to 2020;accounted for grassland NPP and biomass（vegetation）carbon stocks in China since the 21st century;analyzed the characteristics of spatial and temporal changes over the years,and further analyzed the causes of grassland NPP changes systematically and quantitatively by considering natural and anthropogenic factors.The following main conclusions were drawn:（1）The MCD12Q1 data most accurately reflect the distribution of grasslands across China.The area of these ten grassland products ranges from 107.80×104 to332.46×104 km2,with CLCD and MCD12Q1 have high area agreement with other grassland distribution maps.The spatial and sample consistency is highest in the regions of east-central Inner Mongolia,the Qinghai-Tibet Plateau and northern Xinjiang,while the distribution of southern grasslands is scattered and differs considerably among the ten products.MCD12Q1 is significantly more accurate than the other nine products,with an overall accuracy（OA）reaching 77.51%and a kappa coefficient of 0.51;CLCD is slightly less accurate than MCD12Q1（OA=73.02%,kappa coefficient=0.45）and is more conducive to the fine monitoring and management of grassland because of its30-meter resolution.The highest accuracy of grassland was found in the Inner Mongolia-Ningxia region and Qinghai-Tibet Plateau,while the accuracy was worst in the southeastern region.（2）The area of grassland in China has been declining steadily over the past 20years,mainly due to land reclamation and artificial forestation.Grassland decreases from 285.48×104 km2 in 2001 to 280.57×104 km2 in 2020.spatially,grassland shows a clear tendency to decrease in the agro-pastoral zone and southern regions,while an increase in grassland is clearly visible in regions such as the Qinghai-Tibet Plateau and the Loess Plateau.Grassland is mainly transformed（transferred in or out）into cropland,forest land and bare land.Land reclamation and artificial afforestation are the main causes of the decrease in natural grassland,and urbanization has also caused the occupation of grassland by construction land.Ecological restoration projects such as Grain for Green Project,Sand Control Project,and grassland grazing prohibition fences have promoted the restoration of bare land to grasslands.（3）An optimal comprehensive remote sensing estimation model for grassland NPP was constructed based on remote sensing,climate,topography and soil characteristics and the random forest algorithm.Four types of characteristic variables play an important role in the construction of the grassland NPP model,among which the remotely sensed vegetation index is the most critical index;in addition,the model is significantly streamlined and its accuracy is improved after variable screening.Compared to statistical models such as multiple linear regression（MLR）,support vector machine（SVM）,Cubist and artificial neural network（ANN）,the random forest algorithm（RF）is the most effective and efficient way to model NPP in the country.Random forest（RF）is currently the most optimal statistical method for estimating NPP in grasslands nationwide,with an accuracy of r=0.71 and RMSE=280.55 g·C/m2.In addition,the accuracy of the RF model simulated results（r=0.70）was higher than that of the CASA model（r=0.45）and the MOD17 product（r=0.46）.（4）From 2001 to 2020,the annual NPP of Chinese grasslands increased significantly,the biomass carbon stock was 1.14 Pg C,and the grassland ecosystems was generally a carbon sink.The average annual NPP（biomass carbon density）of grasslands in China is 403 g·C/m2,showing a spatial distribution pattern of high in the east and low in the west,and high in the south and low in the north.There is a clear trend of increasing annual NPP across the country and in different grassland types,with significant increases mainly in the hinterland of the Loess Plateau,the middle and lower reaches of the Yellow River,the Hexi Corridor in Gansu and the eastern part of Qinghai Province.The decrease in NPP is mainly a slight trend,accounting for 30%of the national grassland area,mainly distributed in the central-eastern part of Inner Mongolia,northern Xinjiang and central Tibet.From 2001 to 2020,China’s grassland biomass（vegetation）carbon stocks were 1.14±0.03 Pg C(1 Pg C=1015 g C),with the largest stocks in the northern regions,accounting for about 56%of the national total.The amount of biomass carbon stock varies greatly among different grassland types,with temperate grasslands and alpine meadows accounting for 64%of the national total.Over the past 20 years,the total net ecosystem productivity（NEP）of grasslands in China has reached 610 Tg C(1 Pg C=1012 g C).Except for the northwestern part of the Qinghai-Tibet Plateau,grassland ecosystems generally show carbon sink characteristics.（5）Land cover conversion,climate change,human activities and ecological restoration projects have a significant impact on changes in grassland NPP.The transfer in/out of grassland is the main cause of direct changes in grassland NPP,with grassland transfer in（or out）accounting for over 95%of the total NPP obtained（lost）by grassland conversion types,including forest,cultivated land,and bare land.In the areas where grasslands remained unchanged,climate and humans together drove the main reason for the increase in grassland NPP,accounting for 39.57%of the total grassland area;human activities were the main factor for the decrease in NPP,accounting for 14.19%of the total grassland area;ecological restoration projects played an obvious positive role in the increase for grassland NPP.In grassland conservation project,sand control project and river shelter forest project,a total of 73.66×104 km2 of grassland NPP showed an increasing trend,and the contribution of ecological projects was 42.36%.（6）Precipitation is the most important factor dominating the grassland NPP changes across the country.Of the seven drivers,precipitation dominated grassland NPP change over nearly half（48%）of the total grassland area,with nearly 80%of grassland NPP change consistent with precipitation trends;temperature dominated grassland NPP change accounted for the next largest area（11%）,with half（51%）of the grassland NPP increase being dominated by higher temperatures;o Solar radiation dominates 10%of grassland NPP,with nearly 70%of NPP increasing due to a decrease in solar radiation;GDP and animal husbandry output value dominated 7～9%of the grassland NPP change,of which about 80%of the NPP increase was dominated by an increase in economic output,concentrated in the western Tibetan region and the Loess Plateau region;population and livestock numbers dominate NPP changes in 17×104 and19×104 km2 of grassland areas,with half of the grassland NPP changes in the opposite direction to these two factors. |
PDF Full Text Request