Simulation Of Spatio-temporal Dynamic Of Net Primary Productivity And Its Response To Meteorological Factors In Forest Ecosystems Of Poyang Lake Basin

Subtropical forest ecosystems play a crucial part in the global carbon cycle,and accurately simulating the spatio-temporal distribution of Net Primary Productivity（NPP）is essential for a comprehensive grasp of the carbon balance dynamics in this region.Identifying the main climatic controls of productivity differences in different forest ecosystems is beneficial for revealing the patterns of carbon cycle changes in terrestrial ecosystems.Moreover,exploring the differences in their productivity responses to temperature and precipitation at interannual and inter-monthly scales is conducive to identifying the mechanisms of carbon cycle changes in forest ecosystems,which can enhance the study of the forest ecosystem carbon cycle.The typical climate,topography,and hydrological features in Poyang Lake Basin represent the fundamental characteristics of China’s subtropical regions,and it is an important area for studying the carbon cycle of China’s subtropical ecosystems and its response to climate change.We employ the Biome-BGC model,based on measured parameters,at the site scale to dynamically simulate the Net Primary Productivity（NPP）of typical forest types in the Poyang Lake Basin.It reveals the interannual and inter-monthly correlations of NPP for typical forest types with temperature and precipitation.Utilizing remotely sensed NPP data,coupled with a global sensitivity analysis method and parameter optimization,specific vegetation physiological parameters required for simulating NPP of four forest types using the Biome-BGC model are effectively calibrated.The calibrated model overcomes the limitations of previous models,which tended to either overestimate or underestimate NPP in certain months.Subsequently,extending NPP assessment from site-level to basin-wide forest ecosystems,the calibrated model is utilized to simulate the spatiotemporal distribution of forest ecosystem NPP in the subtropical representative region of the Poyang Lake Basin from1960 to 2022,at both annual and monthly scales.Furthermore,the partial correlation between forest NPP and multiple meteorological factors is investigated.The results indicate:（1）At the site scale,the NPPs showed a fluctuating downward trend after 2000 in evergreen needleleaf forest（ENF）,evergreen broadleaf forest（EBF）,and evergreen needleleaf–broadleaf mixed forest（ENBMF）.Precipitation had an effect on evergreen coniferous forest NPP at the interannual scale（R²=0.33,P<0.01）,and at the inter-monthly scale its NPP was primarily driven by precipitation（R²=0.78,P<0.01）.The impact of precipitation and accumulated temperature（AT）on NPP of EBF is not significant interannually;the NPP of bamboo forest（BF）demonstrated a minor correlation with AT and no significant correlation with precipitation.However,the NPP variation of EBF and BF was significantly affected by the temperature at the inter-monthly scale（P<0.01）,with R² of 0.85 and 0.92,respectively.At the interannual scale,the NPP of ENBMF was mainly driven by precipitation（P<0.01）;at the inter-monthly scale,it was driven by precipitation（P<0.01）and AT（P<0.01）,with R² of0.74 and 0.62,respectively.Under designed climate scenarios,the precipitation changes will have a greater impact on NPP of ENF;the NPP changes in the EBF,BF,and ENBMF will be mainly controlled by temperature increase,and only a 10%change in precipitation leads to its smaller impact on their NPP changes.（2）Over the last 63 years,the average NPP of forest ecosystems in the area exhibited a spatial distribution characterized by higher values in the south and lower values in the north at the interannual scale.However,when scrutinized at the inter-monthly scale,the monthly average NPP from April to October showed a spatial distribution trend comparable between the south and the north.Contrastingly,from November to March,NPP was significantly lower than that from April to October,with a spatial distribution trend of higher values in the south and lower values in the north.The multi-year average NPPs of ENF,EBF,deciduous broadleaf forests（DBF）,and Shrub were 656.83,660.83,660.00,and 403.07 g C m^-2 year^-1,respectively.For ENF,EBF,and Shrub,their trends of the average NPPs were essentially the same as those of the months,and the monthly average NPP of Shrub was significantly lower than that of the other two forest types.However,the trend of NPP fluctuation with month was significantly larger for DBF,with the mean NPP being highest from May to August(80–102 g C m^-2 month^-1)and lowest from December to March(≈0 g C m^-2 month^-1).（3）At the interannual scale,most of the forest NPP in the Poyang Lake basin showed significant positive correlations with temperature and solar radiation,with correlation coefficients ranging from 0.2 to 0.8（P<0.05）and 0.2 to 0.6（P<0.05）,respectively,and with spatial distributions of high in the north and low in the south.The correlation coefficients between forest ecosystems and saturated water-air pressure differences were mostly negative,ranging from-0.8 to-0.2（P<0.05）in most of the regions.However,there were significant correlations between forest NPP and precipitation in only a few regions.Meanwhile,the spatially biased correlations between NPP and meteorological factors were more consistent for all four forest types.On the inter-monthly scale,temperature and solar radiation showed significant positive correlations with forest NPP in most regions,with most of the correlation coefficients in the range of 0.8–1（P<0.05）,but the spatial characteristics of the correlations that were comparable between the north and south were significantly different from their spatial characteristics that were higher in the north and lower in the south on the inter-annual scale.However,the biased correlation between forest NPP and precipitation had significant spatial heterogeneity;there was a significant negative correlation with saturated water-air pressure difference.In addition,the bias correlations between NPP and meteorological factors were spatially different for different forest types on the monthly scale.In this paper,we extended from the typical sample point to the whole Poyang Lake Basin,accurately simulated the forest ecosystem NPP in the basin through the calibration of model parameters,found the difference between the spatial distribution of forest ecosystem NPP in the inter-monthly scale and that in the inter-annual scale,and revealed the difference in the correlation between the forest NPP and different climate factors.Therefore,the method of this paper can provide a reference for other regions to assess the spatial and temporal patterns of forest ecosystem carbon balance,and also provide a certain theoretical basis for government departments to implement ecological protection projects.