Predicting Long-term Forest Carbon Balance In Hunan Province:Spatial-temporal Patterns And Its Responses To Climate Change

Forest carbon balance refers to the carbon sequestrated in gross primary productivity (GPP), net primary productivity (NPP) and net ecosystem productivity (NEP) of forest ecosystem in unit time and unit area. It not only plays an important role in the global climate change, but also is susceptible to the climate change. Therefore, the study on the regional forest carbon balance would help to enhance the understanding of matter cycle and energy flowing of ecosystem and reveal the corresponding rules of ecosystem to the global climate change and impacts of global climate change to the ecosystem and their sensitivity to increases in atmosphere CO2concentration, changes in temperature and precipitation. Accurate estimation of forest carbon balance is important for understanding the dynamics of the terrestrial carbon resulting from human-induced climate change, and makes the foundation for land surface, ecological and hydrological modeling, carbon-and water-cycle studies, and research on global climate change. TRIPLEX is a process-based hybrid mode, it focuses on the physiological processes of C assimilation, C contrast with most existing models, also supports component overload, where two or more components that have the same or a similar role in the customized model, can work together. This further increases the flexibility of model reuse. TRIPLEX has the potential and flexibility to allow users to piece together a complete, customized forest simulation model from components that were built for other well-established models. The model could be expanded dynamically and used for research on climate change, forest carbon budgets and dynamics, forest succession, and ecosystem disturbances (e.g. harvest, fire, and insect outbreaks).This study aims to produce accurate estimation of forest carbon balance and potential response to future climate change with process-based hybrid model TRIPLEX1.6in Hunan province, subtropical China. The objectives of this study were to:(1) at stand level, calibrate and validate TRIPLEX for applicability to large scale C. lanceolata and P. massoniana forests,(2) simulate C. lanceolata and P. massoniana forest stand production in Hunan Province, southern China, and make quantitative analysis of relationships between simulated forest stand growth on a provincial scale and spatial patterns of controlling factors;(3) regional validate TRIPLEX for applicabilities to four forest types;(4) predict long-term spatial-temporal patterns of forest carbon balance in Hunan province with the climate inputs from multi-CMIP5emsembled outputs and its responses to climate change, and make sensitive analysis of relationships between projected potential of forest carbon balance on multi-scale spatial-temporal impacts.In the support of Remote Sensing and Geographic Information System, based on the meteorological records of temperature and precipitation data measured at68stations, year1956-2009; the multi-CMIP5model ensembled half-Century projections of future climate trends over Hunan province under the RCP4.5and RCP8.5scenario; and the DEM data of the Hunan province region, the raster of climate inputs were intepolated, which would be the environment inputs of the carbon storage and NPP. MODIS data with250m spatial resolution, climate data,30m DEM, forest map and2173permanent sample plot (PSP2004) and1973plots of PSP2009data of National Forest Inventory (NFI) were used to the carbon model. To get a more accurate model of estimating forest carbon storage, the forest was divided into four types (pine forest, PF; fir forest, FF; deciduous broadleaved forest, DBF; evergreen broadleaved forest, EBF).Our results indicate that:By the use of monthly climate data from Jan.2000to Jan.2009at each stand level, the spatial pattern and dynamics of carbon storage for Cunninghamia lanceolata and Pinus massoniana forest stand production in Hunan Province, China. The process-based hybrid model is a promising tool for predicting forest stand roduction on regional scales. TRIPLEX1.6is capable in predicting forest growth and biomass dynamics of subtropical coniferous forests. Moreover, independent validations determined that TRIPLEX1.6demonstrated competence in extrapolating outcomes on regional scales as well as withstanding rigorous testing in predicting C storage in subtropical forest ecosystems.net primary production (NPP) flux for atmospheric carbon dioxide has varied slightly from year-to-year, but was predicted to have increased over short multi-year periods in the regions of the forested area, and the western since the year2000. These results for were found to be in contrast to other recently published modeling trends for terrestrial NPP with high sensitivity to regional drying patterns. Nonetheless, periodic declines in regional NPP were predicted by TRIPPLEX for the southern and western, the southern Hunan, and southern and eastern Hunan. NPP in subtropical forest zones was examined in greater detail to discover lower annual production values than previously reported in many global models across the subtropical forest zones, likely due to the enhanced detection of lower production ecosystems replacing primary subtropical forest.Nevertheless, TRIPLEX model validation is as yet incomplete. This is primarily the result of the absence of observed soil C data in the plots selected for this study. More rigorous testing of the ability of the model to simulate soil C, N, and water dynamics for various forest ecosystems is a priority in its ongoing development. There are, however, several sources of uncertainty associated with model estimates such as PSP measurement uncertainty, input data uncertainty, model structural uncertainty, and uncertainties pertaining to how strongly limited validation sources represent forest dynamics. Parameter values derived from different sources also contain substantial uncertainties as do impacts from natural and anthropogenic disturbances such as land use change, forest fires, insect-induced mortality and harvesting as well as uncertainties related to the techniques used to fill data gaps. However, despite these limitations, TRIPLEX simulation results fall into the reasonable bounds of measured and published values. Furthermore, additional modules related to the effects of CO2fertilization, ecosystem disturbances (e.g., fire, harvesting, insects, and disease), and soil water stress on forest growth and C and N dynamics must be included in the future development and application of TREPLEX.