Spatiotemporal Heterogeneity Analysis Of Parameter Sensitivity Of Ecological Process Model

The carbon cycle of terrestrial ecosystems is a key process that drives ecosystem changes and is closely related to ecosystem water,nutrient cycling,and biodiversity.Therefore,studying the carbon cycle of terrestrial ecosystems and its response to global climate change has become an important part of today’s global research on change ecology.In order to estimate and predict the carbon cycle of terrestrial ecosystems reasonably and efficiently,and to reveal the inherent laws of terrestrial carbon cycle change,scientists have successively developed a large number of carbon models of terrestrial ecosystems,hoping to describe the process mechanism of the carbon cycle from different perspectives.Among them,the Biome-BGC model is widely used in the simulation of terrestrial ecosystem carbon cycle because of its strong mechanism,wide applicability,easy localization and strong operability.However,there are many models and many different forms.Different models have different simulation capabilities for different research areas and research objects.In the actual application process,the selection and evaluation of the model is very important.At present,the research on model simulation adaptability and accuracy has been carried out successively in various countries and regions,especially the study of model parameter sensitivity.However,the research on sensitivity of model parameters mainly focuses on how the reasonable value of model sensitive parameters influences the model simulation results.However,whether the model parameter sensitivity analysis results are driven by meteorological data at the same site and different time periods has a difference? In the same vegetation type and different spatial regions,are there any differences in the results of parameter sensitivity analysis? And can the model parameter sensitivity results obtained in a space-time background be deduced to other space-time backgrounds? There are few research reports on such issues.However,the answers to such questions have important guiding significance for the application of the model on a specific space-time background.Based on this,this paper takes the Biome-BGC model as an example to study the time heterogeneity,spatial heterogeneity,and spatiotemporal heterogeneity of its parameter sensitivity in order to answer these questions and propose the overall parameters through the model itself or The method of indirectly evaluating the universality,robustness,reliability,and effectiveness of the model is based on the heterogeneity of the temporal and spatial heterogeneity of the average sensitivity,and a preliminary verification of the feasibility of the method is carried out.In this paper,two flux observation stations of Qianyanzhou and Dinghushan are selected under the vegetation type of evergreen broad-leaved forest,two flux observation stations of Changbaishan and Harvard Forest are selected under the vegetation type of deciduous broadleaved forest,and two flux observation stations of Dangxiong and Haibei are selected under the vegetation type of C3 grassland.By constructing the time heterogeneity index,spatial heterogeneity index,and spatio-temporal heterogeneity index of the model parameter sensitivity,the sensitivity of the model parameters at different time periods of the same flux observation site is different between different sites under the same vegetation type.The sensitivity differences and the spatial and temporal differences in sensitivity under the three vegetation types were quantified.Studies have shown that using the Biome-BGC model as an example,the sensitivity of the parameters of the ecological process model is not a constant,but it will show some differences with the transformation of the space-time background,and has certain spatial and temporal heterogeneity;Different parameters have different spatial and temporal heterogeneity of sensitivities;the same parameters,under different vegetation types,have different spatial and temporal heterogeneity of sensitivities;Under the three vegetation types,the spatial-temporal heterogeneity of parameter sensitivity in the Biome-BGC model was relatively small(all below 10?).The results showed that the Biome-BGC model has good suitability and reliability to simulate ecosystem water-carbon cycle in the evergreen broad-leaved forest,deciduous broadleaved forest,and C3 grassland.In this paper,the overall or average spatial-temporal heterogeneity of parameter sensitivity in C3 grassland vegetation is the highest,the deciduous broad-leaved forest and evergreen broad-leaved forest are followed respectively.Overall,the temporal heterogeneity of model parameter sensitivity is significantly less than that of spatial heterogeneity under the same vegetation type.