基于Biome-BGC模型和集合卡尔曼滤波方法的阔叶红松林生态系统水碳通量模拟

Simulation of water and carbon fluxes in a broad-leaved Korean pine forest in Changbai Mountains based on Biome-BGC model and Ensemble Kalman Filter method

数据同化为模型与遥感观测结合提供了一条有效的途径,通过在模型运行过程中融入遥感观测数据,调整模型运行轨迹从而降低模型误差,提高模拟精度。本文利用集合卡尔曼滤波(En KF)算法同化生长季中分辨率成像光谱仪(MODIS)叶面积指数(LAI)与Biome-BGC模型模拟的LAI模拟长白山阔叶红松林的水碳通量。同时,通过改进模拟的雪面升华与土壤温度计算方法的参数,旨在降低冬季生态呼吸的模拟误差。结果表明,相对于原始模型,数据同化与模型改进后使得生态系统总初级生产力(GPP)的模拟值与观测值之间的相关系数提高0.06,中心化均方根误差(RMSE)降低0.48 g C·m^(-2)·d^(-1);生态系统呼吸(RE)的相关系数提高0.02,中心化均方根误差降低0.20 g C·m^(-2)·d^(-1);净生态系统碳交换量(NEE)相关系数提高0.35,中心化均方根误差降低0.50 g C·m^(-2)·d^(-1)。同时,数据同化对蒸散发(ET)的模拟精度没有显著影响,改进的模型提高了其相关系数。基于En KF算法的数据同化提高了长白山阔叶红松林碳通量模拟精度,对于精确估算区域碳通量有着重要的意义。

Data assimilation provides an effective way to integrate the model simulation and re- mote sensing observation, through the integration of remote sensing data in the run of the model, adjusting the model trajectory to reduce model error and improve simulation accuracy. This paper uses the ensemble Kalman filter （EnKF） assimilated MODIS LAI into the Biome-BGC model in growing season to simulate the water and carbon fluxes in a broad-leaved Korean pine forest in Changbai Mountains. At the same time, the simulated snow sublimation and the parameters of the calculation method of soil temperature are improved, which can effectively reduce the error of the ecological respiration in winter. The result shows that as compared with the original model simula- ted without data assimilation, the improved Biome-BGC model with the assimilation of the MODIS LAI makes the correlation coefficient between the simulated values and the observed values of the gross ecosystem primary productivity （GPP） increased by 0.06, and reduced the centered root- mean-square error （RMSE） by 0.48 g C ·m^-2·d^-1, ecosystem respiration （RE） correlation co- efficient increased by 0.02, centered root-mean-square error decreased by 0.20 g C·m^-2·d^-1 ; the correlation coefficient of net ecosystem exchange tered root-mean-square error decreased by 0.50 of carbon （NEE） increased by 0.35, ceng C ·m^-2·d^-1. Meanwhile, data assimilation has no significant effect on the simulation precision of evapotranspiration （ET）, but the model improves the correlation coefficient of ET. The data assimilation based on EnKF improves the accuracy of the carbon flux simulation in the Changbai Mountains, and has an important significance on flux at regional scale. broad-leaved Korean pine improved algorithm forest in more accurate estimation of carbon