Remote Sensing Estimation Of Forest GPP In Yunnan Province And Analysis Of Its Temporal And Spatial Characteristic

The carbon cycle of terrestrial ecosystems is an integral part of the global carbon cycle,and forests,as one of the largest terrestrial ecosystems in the world,have an important influence on the carbon cycle.Gross primary production（GPP）is an important parameter to characterize the carbon cycle,so accurate estimation of forest GPP is important for understanding the global carbon cycle,studying climate change and protecting the ecological environment.To accurately estimate forest GPP in mountainous areas of Yunnan province,three studies are conducted in this paper from2007 to 2015:Firstly,the typical light use efficiency models（VPM model,EC-LUE model and CASA model）are used to estimate the GPP of mountain forests in Yunnan province,and the flux site data and MODIS GPP products are combined to evaluate their applicability;Secondly,the optimization model suitable for mountainous forests in Yunnan province is constructed based on the basic structure of the light use efficiency model,and the flux site data and estimation results of typical light use efficiency models are combined to evaluate their applicability;Finally,the annual and monthly GPP estimated by optimization model for mountain forests in Yunnan province from 2007 to 2015,is analyzed using multivariate analysis for spatial and temporal characteristics changes.The primary conclusions and contributions of this paper are listed as follows:（1）The GPP estimation accuracy of VPM model,EC-LUE model and CASA model in mountainous forests of Yunnan province was much higher than that of MODIS GPP product,with R² increasing by 0.15,0.22 and 0.14,respectively.However,compared with the MODIS GPP product,the estimation error was larger,RMSE of VPM model was reduced by 22.47 g C·m^-2·mon^-1,while EC-LUE model and CASA model were increased by 43.89 g C·m^-2·mon^-1 and 41.99 g C·m^-2·mon^-1,respectively.（2）By combining seven photosynthetically effective radiation absorption ratios,four temperature stress factors and five water stress factors,140 recombinant models were obtained,among which the most suitable combinations of light use efficiency model parameters for estimating GPP in mountainous forests of Yunnan province were FPAR_EVI,T_TEM and W_EC-LUE.The optimization model constructed by the above three parameters compared with the VPM model,EC-LUE model,CASA model and MODIS GPP product,the improvement of R² was about 50%and the decrease of RMSE was about 58%;for MODIS GPP products,the increase of R² was about 371%and the decrease of RMSE was about 35%.（3）The multi-year average value of GPP in mountainous forests of Yunnan province was 951.01 g C·m^-2·yr^-1,with an overall trend of high in the south and low in the north,and high in the east and low in the west.It showed a trend of increasing and then decreasing during the year,reaching the annual maximum in July and decreasing to the annual minimum in December.（4）From 2007 to 2015,66.22%of forest areas in Yunnan province showed an increasing trend and 33.78%of areas showed a decreasing trend.In general,the trend of GPP was increasing in the south and decreasing in the north,and the decreasing trend of GPP is more obvious in the dry season than in the rainy season.During the year,regions with a increasing trend of GPP increased first and then decreased,while regions with a decreasing trend of GPP decreased first and then increased,reaching the highest percentage of the year in July and December,respectively.（5）From 2007 to 2015,62.37%of forest areas in Yunnan province were in a stable state and 37.63%were in an unstable state,and the stable areas were generally strong in the south and weak in the north.During the year,GPP stable areas first increased and then decreased,while unstable areas first decreased and then increased,reaching the highest proportion of the year in July and December respectively.（6）From 2007 to 2015,66.18%of the forest areas in Yunnan province had a continuous increase in GPP and 33.82%of the forest areas had a continuous decrease in GPP.In general,most of the areas sustained increase,of which the southern increase persistence is stronger,the rest of the areas are relatively weak.The western regions continued to decreases,most of which were weakly persistent or have shown a reverse trend,so GPP in these regions is likely to continue to rise in the future.