摘要
[目的]探究土壤呼吸的空间变异特征及协同克里金法预测土壤呼吸准确性,从而为精准预测陆地不同生态系统土壤碳收支研究提供数据参考。[方法]选取黄淮海平原北部典型耕地、草地和林地作为研究样区,基于表征土壤理化性质的辅助指标用协同克里金法对土壤呼吸进行了预测,以评估其与普通克里金法预测的精准度。[结果]土壤呼吸在不同土地利用方式的研究样区均呈现中等程度变异,变异系数分别为33%(耕地),24%(草地)和31%(林地)。土壤呼吸空间相关距离由大到小依次是林地(15.12 m)、耕地(3.83 m)和草地(2.47 m)。基于相关分析和协同克里金法模拟确定土壤呼吸空间变异最优辅助指标,在耕地为碳磷比(C/P)、氮磷比(N/P)和土壤pH值;草地和林地除共有辅助指标土壤水分(SM)和土壤温度(ST5)外,土壤pH值和气温(TEMP)依次为其二者特有的辅助指标。土壤呼吸空间变异预测表明协同克里金均比普通克里金法插值精度高,其在草地提升最为明显,平均绝对误差(MAE)、平均相对误差(MRE)、均方根误差(RMSE)和平均标准误差(MSE)分别提高10.14%,9.82%,9.10%,10.87%,在林地中各误差精度提高介于4.78%(MAE)~12.09%(MRE),在耕地中预测精度提升最小,协同克里金法比普通克里金法预测精度仅平均提高3.27%。[结论]土壤呼吸的影响因子作为辅助变量的协同克里金法可以更准确地预测林地、耕地和草地土壤呼吸的空间变异特征。
[Objective]The aims of this study are to explore the spatial variation characteristics of soil respiration and the accuracy of soil respiration prediction with Kriging method,so as to provide data reference for accurate prediction of soil carbon budget in different terrestrial ecosystems.[Methods]The typical farmland,grassland and forestland in the northern part of the Huang-Huai-Hai Plain were selected as sample areas,and the co-Kriging method was used to predict soil respiration based on the auxiliary indicators of soil physical and chemical properties,so as to evaluate the accuracy of prediction with the common Kriging method.[Results]Soil respiration showed moderate variation in different land use patterns,and the coefficients of variation were 33%(farmland),24%(grassland)and 31%(forest land),respectively.The correlation distances of soil respiration space in descending order were forest land(15.12 m),farmland(3.83 m)and grassland(2.47 m).Based on correlation analysis and co-Kriging simulation,the optimal auxiliary indexes of soil respiration space variation were determined,which were carbon/P ratio(C/P),nitrogen/P ratio(N/P)and pH value in farmland.In addition to soil moisture(SM)and soil temperature(ST5),pH value and temperature(TEMP)were the unique auxiliary indexes of grassland and forestland.Soil respiration spatial variation prediction showed that the interpolation accuracy of the collaborative Kriging method was higher than that of the ordinary Kriging method,and the improvement was most obvious in the grassland.The mean absolute error(MAE),mean relative error(MRE),root mean square error(RMSE)and mean standard error(MSE)increased by 10.14%,9.82%,9.10%and 10.87%,respectively.The improvement of each error accuracy was between 4.78%(MAE)and 12.09%(MRE)in forestland,and the improvement of prediction accuracy was the smallest in farmland,and the prediction accuracy of collaborative Kriging method was only 3.27%higher than that of ordinary Kriging method.[Conclusion]The co-Kriging method with the influence f
作者
李键
薛澄
杨扬
谢梦姣
彭正萍
王洋
Li Jian;Xue Cheng;Yang Yang;Xie Mengjiao;Peng Zhengping;Wang Yang(School of Resources and Environmental Sciences,Hebei Agricultural University/Key Laboratory of Farmland Eco-Environment of Hebei Province,Baoding,Hebei 071000,China;School of Land and Resources,Hebei Agricultural University,Baoding,Hebei 071000,China)
出处
《水土保持研究》
CSCD
北大核心
2024年第2期101-109,121,共10页
Research of Soil and Water Conservation
基金
国家重点研发计划项目“黄淮海粮田多因子障碍消减与产能提升技术模式集成及应用”(2021YFD1901005)。
