应用集合卡尔曼滤波算法对土壤呼吸速率同化及NEP估算

Assimilation and NEP Estimation of Soil Respiration Rate by Ensemble Kalman Filter Algorithm

为了对净生态系统生产力(NEP)进行准确估算,以长白山通量观测站观测数据为基础,构建土壤温度、湿度耦合因子的更新模型(线性函数、指数函数、二次式函数),结合集合卡尔曼滤波算法(EnKF)获取高精度土壤呼吸速率数据,应用陆地生态系统碳循环综合模型(InTEC模型)准确估算NEP。结果表明:二次式模型的EnKF算法同化结果估算效果最好,决定系数(R^(2))为0.782,均方根误差为52.90 g·m^(-2)·a^(-1);指数模型EnKF算法同化结果估算值的R^(2)为0.755,均方根误差为56.47 g·m^(-2)·a^(-1);线性模型EnKF算法同化结果估算值的R^(2)为0.742,均方根误差为62.80 g·m^(-2)·a^(-1)。选取二次式模型优化后的土壤呼吸速率数据,InTEC模型模拟长白山通量观测站长时间序列净生态系统生产力的R^(2)为0.900,均方根误差为61.77 g·m^(-2)·a^(-1);InTEC模型模拟东北三省森林生态系统2003—2010年的净生态系统生产力年均值,由初始模拟的30.07 g·m^(-2)·a^(-1),经EnKF算法更新后提升到176.87 g·m^(-2)·a^(-1)。因此,采用EnKF更新土壤温度-湿度耦合因子获取的土壤呼吸速率数据,能够提高InTEC模型估算NEP的精度,为大区域尺度森林生态系统NEP估算提供技术支持。

In order to accurately estimate the Net Ecosystem Productivity(NEP),a model for updating the coupling factor of soil temperature and humidity(linear function,exponential function,quadratic function)was constructed based on the observational data from the Changbaishan Flux Observation Station.A coupled Ensemble Kalman Filtering(EnKF)algorithm was used to obtain high-precision soil respiration rate data and to apply the Integrated Terrestrial Ecosystem Carbon(InTEC)cycle model to accurately estimate NEP.The results showed that the EnKF assimilation results of the quadratic model had the best estimation effect,with a coefficient of determination(R^(2))of 0.782 and a root mean square error(RMSE)of 52.90 g·m^(-2)·a^(-1).The EnKF assimilation results of the exponential model had an R^(2)of 0.755 and RMSE of 56.47 g·m^(-2)·a^(-1),while the EnKF assimilation results of the linear model had an R^(2)of 0.742 and RMSE of 62.80 g·m^(-2)·a^(-1).After optimizing the soil respiration rate data using the quadratic model,the InTEC model simulated a long-term time series of NEP at the Changbaishan Flux Observation Station,achieving an R^(2)of 0.900 and RMSE of 61.77 g·m^(-2)·a^(-1).The InTEC model simulated the annual average NEP of forest ecosystems in the three northeastern provinces of China from 2003 to 2010,which increased from the initial simulation of 30.07 g·m^(-2)·a^(-1)to 176.87 g·m^(-2)·a^(-1)after the update by the EnKF algorithm.Therefore,using the EnKF to update soil temperature-humidity coupling factor for obtaining soil respiration rate data can improve the accuracy of NEP estimation by the InTEC model,and provide technical support for the estimation of NEP in forest ecosystems at the regional scale.