文献类型：期刊文献

中文题名：长白落叶松解析木数据参数化3-PG模型

英文题名：3-PG Model Parameterization Using Destructive Sampling Data of Larix olgensis

作者：白羽[1,2] 庞勇[1,2] 夏晓运[1,3] 贾炜玮[4]

第一作者：白羽

机构：[1]中国林业科学研究院资源信息研究所,北京100091;[2]国家林业和草原局林业遥感与信息技术重点实验室,北京100091;[3]安徽农业大学林学与园林学院,合肥230036;[4]东北林业大学林学院,哈尔滨150040

年份：2022

卷号：58

期号：1

起止页码：98-110

中文期刊名：林业科学

外文期刊名：Scientia Silvae Sinicae

收录：CSTPCD;;EI(收录号：20221311852487);Scopus;北大核心:【北大核心2020】；CSCD:【CSCD2021_2022】；

基金：十三五国家重点研发项(2017YFD0600404);中央级公益性科研院所基本科研业务费专项资金项目重点项目(CAFYBB2016ZD004)。

语种：中文

中文关键词：3-PG模型;解析木数据;参数化;长白落叶松

外文关键词：3-PG(physiological principles in predicting growth)model;destructive sampling data;parameterization;Larix olgensis

分类号：S154.5;Q958.15

摘要：【目的】验证解析木数据参数化3-PG模型的可行性,利用解析木数据标定的模型预测长白落叶松人工林生长变化,为扩展参数化3-PG模型数据源提供依据,为模型数据选择提供参考。【方法】以孟家岗林场长白落叶松人工林为研究对象,基于2019年解析木数据模拟胸径连续观测数据,结合相关公式计算叶、干、根生物量和蓄积量。根据参数敏感性分析结果,采用直接计算、参考文献、迭代拟合和默认参数等方法对3-PG模型进行参数化,利用密度试验林和固定样地数据对模型输出进行精度验证,并对模型输出与地面观测值进行回归分析。【结果】验证结果表明,模型拟合精度很高(P<0.01,n=138),可以较好反映解析木样地中林分生长变化。胸径、叶干生物量比、干生物量、总生物量和蓄积量的决定系数(R^(2))在0.95以上,根生物量拟合精度略低(R^(2)=0.88)。密度试验林数据(n=140)和固定样地数据(n=87)与模型输出之间具有较高相关性,各计算量的R^(2)在0.81~0.97之间(P<0.01)。敏感性分析结果表明,当胸径为20 cm时,叶与干生物量分配比(pFS20)及初级生产力分配给根最大值(pRx)2个参数均对胸径、根、干生物量和蓄积量显示出较高敏感性。【结论】解析木数据参数化3-PG模型的拟合精度和预测精度均较高,研究结果扩展了参数化3-PG模型的可用数据源,可为利用3-PG模型模拟长白落叶松人工林生长变化提供新的依据。
【Objective】In order to verify the usability of destructive sampling data in 3-PG(physiological principles in predicting growth)model parameterization,this study predicted the growth of Larix olgensis using the 3-PG model calibrated by destructive sampling data,which can help to expand the data source of 3-PG model parameterization and provide a reference for model data selection.【Method】Taking L.olgensis plantation in Mengjiagang forest farm as the study object,the destructive sampling data collected in 2019 was calculated to simulate the continuous observation data.The biomass of foliage,stem,and root,and stand volume were calculated based on the relevant biomass formulas.According to the sensitivity analysis results,the 3-PG model was parameterized by direct calculation,reference,and iterative fitting.Then the modeling results were verified by continuous observation data and fixed plot data,and the model output was regressed with observed data.【Result】The results showed that the fitting accuracy was very high,and the model output data could well reflect the growth of the stand(n=138).The determination coefficient(R^(2))values of DBH,the ratio of foliage to stem biomass,stem biomass,total biomass and volume were all above 0.95(P<0.01).In contrast,the root biomass had a relatively lower fitting accuracy(R^(2)=0.88).The model’s predicting accuracy was verified by continuous observation data and fixed plot data.High correlations were observed between the model output and continuous observation data(n=140)and fixed plot data(n=87).The R^(2) of all values ranged from 0.81 to 0.97(P<0.01).According to the sensitivity analysis results,foliage∶stem partitioning ratio at DBH=20 cm(pFS20)and Maximum fraction of NPP to roots(pRx)showed high sensitivity to DBH,root biomass,stem biomass,and volume.【Conclusion】3-PG model calibrated by destructive sampling data had a relatively high fitting accuracy and prediction accuracy.The results of this study would expand the data source of 3-PG model parameterization and provide a new basis for the 3-PG model to simulate the growth of L.olgensis plantation.