基于随机森林模型的天然林立地生产力预测研究
|
摘要
从汪清林业局大荒沟林场、大柞树林场等11个林场的森林资源二类调查数据中选取优势树种为蒙古栎、阔叶混交林及针阔混交林的小班。以海拔、土层厚度、坡位、坡向、腐殖质层厚度、坡度6个立地因子以及年平均气温、月平均气温差等19个气候因子为输入变量,以树种年平均蓄积生长量为输出变量,应用随机森林回归算法分别建立蒙古栎、阔叶混交林及针阔混交林的立地质量评价模型,对不同立地条件下的造林地进行生产潜力预测。同时,分析了各环境因子对树种生长的影响权重。结果表明:1)所建立的3种回归模型的RMSE的值分别为:0.22、0.54、0.52,R~2值分别为:0.79、0.79、0.72,模型的拟合效果较为理想。2)研究区域内,对蒙古栎生长影响较大的因子依次为月平均气温差、温度季节性变化、坡度、年降水量、年平均气温差;对针阔混交林生长影响较大的因子依次为:坡度、腐殖质层厚度、月平均气温差、最湿季度降水量、最暖季度降水量;对阔叶混交林生长影响较大的因子依次为:坡度、坡位、坡向、温度季节性变化、最干旱季平均气温。3)通过对比同一立地3种类型的生产力,针阔混交林、阔叶混交林的年平均蓄积生长量均高于蒙古栎纯林,针阔混交林略高于阔叶混交林。4)因此,应客观考虑环境因子对于林木的影响程度,使其生长环境条件尽可能地处于最佳组合状态。
The data of dominant tree species were selected from the second-class survey data of 11 forest farms such as Dahuanggou Forest Farm and Damong Quercus Forest Farm of Wangqing Forestry Bureau as the subcompartment data of Mongolian Quercus, broadleaved mixed forest and coniferous-broad mixed forest. Taking 19 climatic factors, such as annual average temperature, monthly average temperature difference etc. and 6 site factors(altitude, soil thickness, slope position, slope aspect, humus layer thickness, slope degree)as the input variables, annual average increment as the output variable, three regression models such as site quality evaluation model for Quercus mongolica, for conifer-broadleaf forest and for broadleaf forest were set up respectively by adopting random forest regression algorithm, and productive potential of afforestation land under different site conditions was predicted. At the same time, the weights of environmental factors affecting tree species growth were analyzed. The results are as follows. 1) The RMSE values of the three regression models were 0.22, 0.54, and 0.52, and the R~2 values were 0.79, 0.79, and 0.72, respectively; the ?tting effect of the model was ideal. 2) In the study area, the major factors affecting the growth of Q.mongolica were monthly average temperature difference, seasonal temperature change, slope, annual total water volume and annual average temperature difference, and the major factors affecting the growth of coniferous-broad-leaved mixed forest were slope, humus layer thickness, monthly average temperature difference, wettest season precipitation and warmest season precipitation, and the factors affecting the growth of broadleaf mixed forest were slope,slope position, slope direction, seasonal variation of temperature and average temperature in the driest season. 3) By comparing the productivity values of three kinds of non-forest land types, the average annual growth values of coniferous + broad-leaved mixed forest and broad-leaved mixed forest were higher than that of pure Mongolian oak, and that of the coniferous + broad-leaved mixed forest was slightly higher than that of broad-leaved mixed forest. 4) Therefore, the in?uences of environmental factors on trees growth should be considered objectively so as to make their growth environment in the best combination state as far as possible.
The data of dominant tree species were selected from the second-class survey data of 11 forest farms such as Dahuanggou Forest Farm and Damong Quercus Forest Farm of Wangqing Forestry Bureau as the subcompartment data of Mongolian Quercus, broadleaved mixed forest and coniferous-broad mixed forest. Taking 19 climatic factors, such as annual average temperature, monthly average temperature difference etc. and 6 site factors(altitude, soil thickness, slope position, slope aspect, humus layer thickness, slope degree)as the input variables, annual average increment as the output variable, three regression models such as site quality evaluation model for Quercus mongolica, for conifer-broadleaf forest and for broadleaf forest were set up respectively by adopting random forest regression algorithm, and productive potential of afforestation land under different site conditions was predicted. At the same time, the weights of environmental factors affecting tree species growth were analyzed. The results are as follows. 1) The RMSE values of the three regression models were 0.22, 0.54, and 0.52, and the R~2 values were 0.79, 0.79, and 0.72, respectively; the ?tting effect of the model was ideal. 2) In the study area, the major factors affecting the growth of Q.mongolica were monthly average temperature difference, seasonal temperature change, slope, annual total water volume and annual average temperature difference, and the major factors affecting the growth of coniferous-broad-leaved mixed forest were slope, humus layer thickness, monthly average temperature difference, wettest season precipitation and warmest season precipitation, and the factors affecting the growth of broadleaf mixed forest were slope,slope position, slope direction, seasonal variation of temperature and average temperature in the driest season. 3) By comparing the productivity values of three kinds of non-forest land types, the average annual growth values of coniferous + broad-leaved mixed forest and broad-leaved mixed forest were higher than that of pure Mongolian oak, and that of the coniferous + broad-leaved mixed forest was slightly higher than that of broad-leaved mixed forest. 4) Therefore, the in?uences of environmental factors on trees growth should be considered objectively so as to make their growth environment in the best combination state as far as possible.
引文
[1]吴菲.森林立地分类及质量评价研究综述[J].林业科技情报,2010,42(1):12-14.
[2]郭艳荣,刘洋,吴保国.福建省宜林地立地质量的分级与数量化评价[J].东北林业大学学报,2014,42(10):54-59.
[3]张伏全,魏汗功,陈远材.滇西南地区立地龙竹质量评价的研究[J].林业科学,1994,30(2):104-110.
[4]CURT T,BOUCHAUD M,AGRECH G.Predicting site index of Douglasfir plantations from ecological variables in the Massif Central area of France[J].Forest ecology and management,2001,149(1):61-74.
[5]LOUWA J H,SCHOLES M.Forest site classification and evaluation:a South African perspective[J].Forest Ecology and Management,2002,171(1-2):153-168.
[6]迟健,李桂英,陈佳明,等.浙江省马尾松人工林立地质量的数量化研究[J].林业科学研究,1995,8(3):303-308.
[7]朱光玉,康立,何海梅,等.基于树高-年龄分级的杉木人工林多形立地指数曲线模型研究[J].中南林业科技大学学报,2017,37(7):18-29.
[8]黄家荣,马天晓,王艳梅,等.基于BP网络的无林地立地质量评价模型研究[J].山地农业生物学报,2006,25(6):479-483.
[9]高若楠,苏喜友,谢阳生,等.基于随机森林的杉木适生性预测研究[J].北京林业大学学报,2017,39(12):36-43.
[10]IVERSON L R,PRASAD A M,MATTHEWS S N,et al.Estimating potential habitat for 134 eastern US tree species under six climate scenarios[J].Forest ecol.Manage,2008(254):390-406.
[11]PRASAD A M,IVERSON L R,LIAW A.Newer classification and regression tree technique:Bagging and Random forests for ecological prediction[J].Ecosystems,2006,9(2):181-199.
[12]DONG L J,LI X B.Prediction of rockburst classification using Random Forest[J],Transaction of Nonferrous Metals Society of China.2013,23(2):472-477.
[13]余坤勇,姚雄,邱祁荣,等.基于随机森林模型的山体滑坡空间预测研究[J].农业机械学报,2016,47(10):338-345.
[14]张雷,王琳,张旭东,等.随机森林算法基本思想及其在生态学中的应用:以云南松分布模拟为例[J].生态学报,2014,34(3):650-659.
[15]赖成光,陈晓宏,赵仕威,等.基于随机森林的洪灾风险评价模型及其应用[J].水利学报,2015,46(1):59-65.
[16]蒋育昊.基于随机森林模型的红松潜在分布预测及适宜性评价[D].北京:中国林业科学研究院,2018.
[17]Hijmans,Robert J,Cameron,et al.Very high resolution interpolated climate surfaces for global land areas[M].2005.
[18]BREIMAN L.Random Forests[J].Machine Learning,2001,45(1):5-32.
[19]Liaw A,Wiener M.Classification and Regression by random Forest[J].R News,2002,23(23).
[20]林成德,彭国兰.随机森林在企业信用评估指标体系确定中的应用[J].厦门大学学报(自然科学版),2007,46(2):199-203.
[21]李贞子,张涛,武晓岩,等.随机森林回归分析及在代谢调控关系研究中的应用[J].中国卫生统计,2012,29(2):158-163.
[22]李丽辉,朱建生,强丽霞,等.基于随机森林回归算法的高速铁路短期客流预测研究[J].铁道运输与经济,2017,39(9):12-16.
[23]曾斌,翟学昌,彭丽,等.赣南不同立地杉木人工林生产力研究[J].安徽农业科学,2010(14):7625-7626.
[24]单长卷.黄土高原不同立地刺槐林水分关系研究[D].杨凌:西北农林科技大学,2004.
[25]刘丹,高永刚,顾红,等.伊春林区影响3种主要林木生长量的气候因子分析[J].黑龙江气象,2007(4):21-22.
[26]贾翔,马芳芳,周旺明,等.气候变化对阔叶红松林潜在地理分布区的影响[J].生态学报,2017,37(2):464-473.
[27]张雷,王琳琳,张旭东,等.随机森林模型算法基本思想及其在生态学中的应用--以云南松分布模拟为例[J].生态学报,2014,34(3):650-659.
[2]郭艳荣,刘洋,吴保国.福建省宜林地立地质量的分级与数量化评价[J].东北林业大学学报,2014,42(10):54-59.
[3]张伏全,魏汗功,陈远材.滇西南地区立地龙竹质量评价的研究[J].林业科学,1994,30(2):104-110.
[4]CURT T,BOUCHAUD M,AGRECH G.Predicting site index of Douglasfir plantations from ecological variables in the Massif Central area of France[J].Forest ecology and management,2001,149(1):61-74.
[5]LOUWA J H,SCHOLES M.Forest site classification and evaluation:a South African perspective[J].Forest Ecology and Management,2002,171(1-2):153-168.
[6]迟健,李桂英,陈佳明,等.浙江省马尾松人工林立地质量的数量化研究[J].林业科学研究,1995,8(3):303-308.
[7]朱光玉,康立,何海梅,等.基于树高-年龄分级的杉木人工林多形立地指数曲线模型研究[J].中南林业科技大学学报,2017,37(7):18-29.
[8]黄家荣,马天晓,王艳梅,等.基于BP网络的无林地立地质量评价模型研究[J].山地农业生物学报,2006,25(6):479-483.
[9]高若楠,苏喜友,谢阳生,等.基于随机森林的杉木适生性预测研究[J].北京林业大学学报,2017,39(12):36-43.
[10]IVERSON L R,PRASAD A M,MATTHEWS S N,et al.Estimating potential habitat for 134 eastern US tree species under six climate scenarios[J].Forest ecol.Manage,2008(254):390-406.
[11]PRASAD A M,IVERSON L R,LIAW A.Newer classification and regression tree technique:Bagging and Random forests for ecological prediction[J].Ecosystems,2006,9(2):181-199.
[12]DONG L J,LI X B.Prediction of rockburst classification using Random Forest[J],Transaction of Nonferrous Metals Society of China.2013,23(2):472-477.
[13]余坤勇,姚雄,邱祁荣,等.基于随机森林模型的山体滑坡空间预测研究[J].农业机械学报,2016,47(10):338-345.
[14]张雷,王琳,张旭东,等.随机森林算法基本思想及其在生态学中的应用:以云南松分布模拟为例[J].生态学报,2014,34(3):650-659.
[15]赖成光,陈晓宏,赵仕威,等.基于随机森林的洪灾风险评价模型及其应用[J].水利学报,2015,46(1):59-65.
[16]蒋育昊.基于随机森林模型的红松潜在分布预测及适宜性评价[D].北京:中国林业科学研究院,2018.
[17]Hijmans,Robert J,Cameron,et al.Very high resolution interpolated climate surfaces for global land areas[M].2005.
[18]BREIMAN L.Random Forests[J].Machine Learning,2001,45(1):5-32.
[19]Liaw A,Wiener M.Classification and Regression by random Forest[J].R News,2002,23(23).
[20]林成德,彭国兰.随机森林在企业信用评估指标体系确定中的应用[J].厦门大学学报(自然科学版),2007,46(2):199-203.
[21]李贞子,张涛,武晓岩,等.随机森林回归分析及在代谢调控关系研究中的应用[J].中国卫生统计,2012,29(2):158-163.
[22]李丽辉,朱建生,强丽霞,等.基于随机森林回归算法的高速铁路短期客流预测研究[J].铁道运输与经济,2017,39(9):12-16.
[23]曾斌,翟学昌,彭丽,等.赣南不同立地杉木人工林生产力研究[J].安徽农业科学,2010(14):7625-7626.
[24]单长卷.黄土高原不同立地刺槐林水分关系研究[D].杨凌:西北农林科技大学,2004.
[25]刘丹,高永刚,顾红,等.伊春林区影响3种主要林木生长量的气候因子分析[J].黑龙江气象,2007(4):21-22.
[26]贾翔,马芳芳,周旺明,等.气候变化对阔叶红松林潜在地理分布区的影响[J].生态学报,2017,37(2):464-473.
[27]张雷,王琳琳,张旭东,等.随机森林模型算法基本思想及其在生态学中的应用--以云南松分布模拟为例[J].生态学报,2014,34(3):650-659.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |