Study On Space-time Heterogeneity Of Ecological Process Model's Sensitive Parameters

Global climate change has been the concern of governments and research institutions in recent years,including greenhouse gas,especially CO2 emissions and fixed as one of the core issues.As the most active and most economically disturbed carbon pools,The balance and stability of terrestrial ecosystems is significant for the Earth's ecosystem.Scientists from all over the world expect the carbon cycle of terrestrial ecosystems to be estimated,simulated and predicted by means of field measurements,model simulations,isotope markers and other methods in order to detect and address climate problems in a timely manner.With the deepening of research,the advantages of model simulation of revealing the inherent law of terrestrial ecosystem and processing time and space deduction gradually reflected which can be applied and developed widely,especially taking Biome-BGC as the representative of the ecological process model,which has the characteristics of strong mechanism and wide adaptability.However,in the process of application,due to the large number of ecological process model parameters,The characteristics of model parameters need to be analyzed in order to obtain relatively accurate and reasonable results.First of all need to determine the sensitivity of the parameters,the sensitivity of the parameters have been carried out much research and gotten a lot of useful results.But with the sensitivity parameters,different parameters and have space-time heterogeneity,especially some of the parameters of large space-time heterogeneity,the value of the parameters,but also need to accurately rate and put more effort.However,there are relatively few studies on the spatial and temporal heterogeneity of sensitive parameters,and it has not yet formed a mature theoretical method and system.so the Biome-BGC eco-process model is used as an example to study the spatiotemporal heterogeneity of sensitive parameters in order to enrich the theory and method of parameter sensitivity and spatiotemporal heterogeneity research,and improve the parameters of ecological process model to further understand.The paper screened out the sensitive parameters of the Biome-BGC model in the evergreen broad-leaved forest,deciduous broad-leaved forest and C3 grassland were selected by constructing the parameter sensitivity index,and two experimental sites were selected under each vegetation type,Using the simulated annealing algorithm combined with the measured flux data to build the objective function and optimized the selected sensitive parameters for month-by-month.At the same time,the Spatial and temporal heterogeneity discriminant index is used to make quantitative analysis of the space-time heterogeneity of the model sensitive parameters.The results showed that the sensitivity parameters of Biome-BGC under the three vegetation types were different.The evergreen broad-leaved forest consisted of 12 parameters including the conversion ratio of the inner leaves and fine roots.The deciduous broad-leaved forest included the proportion of the current growth and other 12 parameters,C3 grassland,including fine root carbon and nitrogen ratio of 12 parameters.At the same time,the sensitivity parameters of Biome-BGC model have different temporal and spatial heterogeneity.Different parameters under same vegetation types show different temporal and spatial heterogeneity,and the performance of the same parameters under different vegetation types is different also.The spatial and temporal heterogeneity of the distribution of root and carbon in the evergreen broad-leaved forest vegetation type and deciduous broad-leaved forest vegetation was the highest,and the discriminant index reached 0.3776 and 0.3928,respectively.The spatial and temporal heterogeneity of the C3 grassland vegetation type is the highest,and the discriminant index is 0.3506.Under the three vegetation types,the parameters such as fine nitrogen and nitrogen ratio,specific leaf area and so on were closely related to the seasonal and environmental changes,while the leaf nitrogen content in the Rubisco leaves,the spatiotemporal heterogeneity is not obvious.In addition,the temporal heterogeneity and spatial heterogeneity of sensitive parameters mostly show obvious linear correlation characteristics.The spatial heterogeneity of time heterogeneity is often also large,as the spatiotemporal heterogeneity of the product also relatively large,and vice versa.The sensitivity of the model is divided into five characteristic spaces by using the sensitivity as the vertical dimension and the spatial and temporal heterogeneity as the lateral dimension to cluster the sensitive parameters under the three vegetation types.The parameters with small sensitivity are the fifth quadrant,and the other parameters are divided into four quadrants according to the plane coordinate system.The parameters with large sensitivity and temporal and spatial heterogeneity are the first quadrant and the space-time heterogeneity is large while the sensitivity is smaller for the second quadrant,the sensitivity and temporal and spatial heterogeneity are smaller for the third quadrant,the sensitivity is larger and the space-time heterogeneity is smaller for the fourth quadrant.In the future optimization and determination of ecological process model parameters,the model can be sensitive parameters of the spatiotemporal heterogeneity of the different parameters of the parameters to take different optimization strategies.For a parameter class with a high sensitivity but a small spatial-temporal heterogeneity,it is only necessary to determine it at a time of a site,and it can be extended to other times and regions.For higher sensitivity,the temporal and spatial heterogeneity is also higher of the parameters,you need to be based on specific time and space conditions for dynamic optimization.