首页分享云南省针叶林地上生物量时空分布及动态分析研究

云南省针叶林地上生物量时空分布及动态分析研究

来源：花匠小妙招时间：2025-05-01 03:43

摘要:

基于云南省2002年、2007年、2012年、2017年森林资源连续清查4期数据及同期Landsat影像，结合气候和地形数据，利用随机森林分类器（RFC）获取4期云南省针叶林面状数据，并对比了2类Bagging和Boosting技术共5种模型的拟合效果，确定了随机森林回归（RFR）为最优模型，用于4期全省针叶林地上生物量（AGB）反演，进而探讨云南省针叶林AGB的时间动态变化及空间分布。结果表明：利用RFC模型对2002—2017年土地利用进行分类，分类结果OA均在0.76～0.78，Kappa系数在0.65～0.69，具有较高的准确性。在对针叶林AGB进行遥感建模估测时，基于Bagging的决策树模型在拟合稳定性和效果上优于基于Boosting的决策树模型。云南省针叶林2002年、2007年、2012年、2017年的总AGB分别是1.49、1.33、1.35、1.58 Gt，呈先减少后增加趋势，2017年总AGB较2002年增长6%。2002—2017年除部分州市外，整体AGB呈增长趋势。

关键词: 针叶林 / 生物量 / 时空分布 / 随机森林

Abstract:

Based on the four-phase (2002, 2007, 2012, and 2017) continuous forest inventory (CFI) data of Yunnan Province and contemporaneous Landsat imagery, in conjunction with climate and topographic data, this study employed the Random Forest Classifier (RFC) to derive coniferous forest spatial distribution data for the four time periods. The fitting performance of five models representing two categories of ensemble learning techniques—Bagging and Boosting—was compared. Random Forest Regression (RFR) was identified as the optimal model and subsequently utilized for the inversion of aboveground biomass (AGB) of coniferous forests across the four periods at the provincial scale. The temporal dynamics and spatial distribution of coniferous forest AGB in Yunnan Province were further analyzed.The results indicate that land use classification using the RFC model for the period 2002–2017 achieved an overall accuracy (OA) ranging from 0.76 to 0.78, with Kappa coefficients between 0.65 and 0.69, demonstrating high classification accuracy. In the remote sensing-based modeling and estimation of coniferous forest AGB, decision tree models based on the Bagging technique exhibited superior fitting stability and performance compared to those based on Boosting. The total AGB of coniferous forests in Yunnan Province was 1.49 Gt, 1.33 Gt, 1.35 Gt, and 1.58 Gt in 2002, 2007, 2012, and 2017, respectively, showing an initial decline followed by an increase. The total AGB in 2017 increased by 6% compared to 2002. From 2002 to 2017, despite declines in certain prefectures, the overall trend of AGB exhibited an increase.

图 1 2002—2017年样地实测AGB分布

Figure 1. The distribution of field-measured AGB from 2002 to 2017

图 2 土地利用分类模型变量重要性

Figure 2. Variable importance in land use classification models

图 3 预测模型变量重要性

Figure 3. Variable importance in predictive models

图 4 模型评价结果

Figure 4. Results of Model Evaluation

图 5 2002—2017年各州市针叶林AGB总量

Figure 5. Total AGB of coniferous forests in each prefecture of Yunnan Province from 2002 to 2017

图 6 2002—2017年各州市针叶林平均AGB占比

Figure 6. Proportion of average AGB in coniferous forests across each prefecture of Yunnan Province from 2002 to 2017

表 1 2002—2017年针、阔叶林样地个数

Table 1 The number of coniferous and broadleaf forest plots from 2002 to 2017

树种样地数/个 2002年 2007年 2012年 2017年针叶林 884 937 563 578 阔叶林 894 934 272 314 　注：阔叶林样地仅用于土地利用分类阶段，不参与后续的遥感建模与估测分析。

表 2 建模因子概况

Table 2 Overview of modeling factors

变量类型因子描述植被指数　归一化差异植被指数（NDVI）、归一化水体指数（NDWI）、比值植被指数（RVI）、差值植被指数（DVI）、基于波段6～7的NDVI（ND67）、大气阻抗植被指数（ARVI）、增强型植被指数（EVI）、土壤调节植被指数（SAVI）、亮度植被指数（B）、绿度植被指数（G）、温度植被指数（W）、中红外温度植被指数（MV17）、多波段线性组合（ALBEDO）、Kauth–Thomas变换（KT1、KT2、KT3）、改进型归一化差异水体指数（MNDWI）、归一化差异建筑指数（NDBI）、基于指数的建筑指数（IBI）、第二修正比值植被指数（MSRI）、湿度植被指数（MVI）、归一化差异指数（NDI）、归一化差异指数（NGBDI）、优化土壤调整植被指数（OSRVI）、转换植被指数（TVI）、抗大气指数（VARI）、可见光波段差异植被指数（VDVI）、基于波段43的NDVI（ND43）、红外植被指数（II）、非线性指数（NLI）、基于波段65的RVI（RVI65）、基于波段75的RVI（RVI75）纹理　均值（Mean）、变异性（Variance）、同质性（Homogeneity）、对比度（Contrast）、相异性（Dissimilarity）、信息熵（Entropy）、角二阶矩（Angular Second Moment）和相关性（Correlation）生物气候　年平均温度（Bio_1）、平均气温日较差（Bio_2）、等温性（Bio_3）、温度季节性（Bio_4）、最热月份最高温度（Bio_5）、最冷月份最低温（Bio_6）、年温差（Bio_7）、最潮湿季节平均温度（Bio_8）、最干燥季节平均温度（Bio_9）、最热季节平均温度（Bio_10）、最冷季节的平均温度（Bio_11）、年均降水量（Bio_12）最潮湿月份降水量（Bio_13）、最干旱月份降水量（Bio_14）、降水季节性（Bio_15）、最湿季降水（Bio_16）最干旱地区降水（Bio_17）、最暖季降水量（Bio_18）、最冷季降水量（Bio_19）地形海拔（Elevation）、坡度（Slope）、坡向（Aspect）　注：纹理因子包含3 × 3、5 × 5、7 × 7窗口的因子。

表 3 各土地类型样本分布

Table 3 Distribution of Samples Across Land Types

土地利用类型样本数量针叶林350阔叶林250农田50水体50冰雪50其他用地100

表 4 2002—2007年土地利用分类评价

Table 4 Evaluation of land use classification from 2002 to 2007

年份OAKappa 20020.770.6920070.770.6920120.780.6920170.760.65 [1]

Alam A, Kilpeläinen A, Kellomäki S. Impacts of initial stand density and thinning regimes on energy wood production and management-related CO2 emissions in boreal ecosystems [J]. European Journal of Forest Research, 2012, 131(3): 655−667. DOI: 10.1007/s10342-011-0539-8

[2]

Tamiminia H, Salehi B, Mahdianpari M, et al. Decision tree-based machine learning models for above-ground biomass estimation using multi-source remote sensing data and object-based image analysis [J]. Geocarto International, 2022, 37(26): 12763−12791. DOI: 10.1080/10106049.2022.2071475

[3]

Peng D L, Zhang H L, Liu L Y, et al. Estimating the aboveground biomass for planted forests based on stand age and environmental variables [J]. Remote Sensing, 2019, 11(19): 2270. DOI: 10.3390/rs11192270

[4]

Abdel-Rahman E M, Mutanga O, Adam E, et al. Detecting Sirex noctilio grey-attacked and lightning-struck pine trees using airborne hyperspectral data, random forest and support vector machines classifiers [J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2014, 88: 48−59. DOI: 10.1016/j.isprsjprs.2013.11.013

[5]

Li Y C, Li M Y, Li C, et al. Forest aboveground biomass estimation using Landsat 8 and Sentinel-1A data with machine learning algorithms [J]. Scientific Reports, 2020, 10(1): 9952. DOI: 10.1038/s41598-020-67024-3

[6]

Huang T B, Ou G L, Wu Y, et al. Estimating the aboveground biomass of various forest types with high heterogeneity at the provincial scale based on multi-source data [J]. Remote Sensing, 2023, 15(14): 3550. DOI: 10.3390/rs15143550

[7]

Dong T F, Liu J G, Qian B D, et al. Estimating crop biomass using leaf area index derived from Landsat 8 and Sentinel-2 data [J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2020, 168: 236−250. DOI: 10.1016/j.isprsjprs.2020.08.003

[8]

Lu D S, Chen Q, Wang G X, et al. A survey of remote sensing-based aboveground biomass estimation methods in forest ecosystems [J]. International Journal of Digital Earth, 2016, 9(1): 63−105. DOI: 10.1080/17538947.2014.990526

[9]

Hernández-Stefanoni J L, Castillo-Santiago M Á, Mas J F, et al. Improving aboveground biomass maps of tropical dry forests by integrating LiDAR, ALOS PALSAR, climate and field data [J]. Carbon Balance and Management, 2020, 15(1): 15. DOI: 10.1186/s13021-020-00151-6

[10]

Qadeer A, Shakir M, Wang L, et al. Evaluating machine learning approaches for aboveground biomass prediction in fragmented high-elevated forests using multi-sensor satellite data [J]. Remote Sensing Applications: Society and Environment, 2024, 36: 101291. DOI: 10.1016/j.rsase.2024.101291

[11] 杜春雨. 基于主被动遥感数据协同的森林AGB估算及其饱和点分析[D]. 哈尔滨: 东北林业大学, 2021. [12] 张亦然, 刘廷玺, 童新, 等. 基于XGBoost算法的草甸地上生物量的高光谱遥感反演 [J]. 草业学报, 2021, 30(4): 1−12. DOI: 10.11686/cyxb2020190 [13]

Song J, Liu X L, Adingo S, et al. A comparative analysis of remote sensing estimation of aboveground biomass in boreal forests using machine learning modeling and environmental data [J]. Sustainability, 2024, 16(16): 7232. DOI: 10.3390/su16167232

[14] 陈思婷, 黄衍, 何霄. 中国森林碳汇发展潜力分析 [J]. 中国林业经济, 2022(4): 68−72. [15] 周瑞伍, 彭明春, 张一平. 云南主要森林植被碳储量及固碳潜力模拟研究 [J]. 云南大学学报(自然科学版), 2017, 39(6): 1089−1103. DOI: 10.7540/j.ynu.20160767 [16]

Fang P F, Ou G L, Li R N, et al. Regionalized classification of stand tree species in mountainous forests by fusing advanced classifiers and ecological niche model [J]. GIScience & Remote Sensing, 2023, 60(1): 2211881.

[17] 胥辉, 张子翼, 欧光龙, 等. 云南省森林生物量和碳储量估算及分布研究[M]. 昆明: 云南科技出版社, 2019. [18]

Xiong N N, Yu R X, Yan F, et al. Land use and land cover changes and prediction based on multi-scenario simulation: a case study of Qishan county, China [J]. Remote Sensing, 2022, 14(16): 4041. DOI: 10.3390/rs14164041

[19]

Huang H J, Wu D S, Fang L M, et al. Comparison of multiple machine learning models for estimating the forest growing stock in large-scale forests using multi-source data [J]. Forests, 2022, 13(9): 1471. DOI: 10.3390/f13091471

[20]

Zou H, Hastie T. Regularization and variable selection via the elastic net [J]. Journal of the Royal Statistical Society Series B: Statistical Methodology, 2005, 67(2): 301−320. DOI: 10.1111/j.1467-9868.2005.00503.x

[21]

Zhao P P, Lu D S, Wang G X, et al. Examining spectral reflectance saturation in landsat imagery and corresponding solutions to improve forest aboveground biomass estimation [J]. Remote Sensing, 2016, 8(6): 469. DOI: 10.3390/rs8060469

[22] 茶枝义. 云南省针叶林碳储量及固碳潜力分析 [J]. 西部林业科学, 2019, 48(4): 7−12. [23]

Ding T, Gao H. The record-breaking extreme drought in Yunnan Province, southwest China during spring-early summer of 2019 and possible causes [J]. Journal of Meteorological Research, 2020, 34(5): 997−1012. DOI: 10.1007/s13351-020-0032-8

[24] 王以惠, 牛香, 王兵, 等. 云南省天保工程区森林生态系统服务功能及驱动力分析 [J]. 林业资源管理, 2023(3): 21−28. [25] 陈先刚, 张一平, 詹卉. 云南退耕还林工程林木生物质碳汇潜力 [J]. 林业科学, 2008, 44(5): 24−30. DOI: 10.3321/j.issn:1001-7488.2008.05.007 [26]

Poorter L, Bongers F, Mitchell Aide T, et al. Biomass resilience of neotropical secondary forests [J]. Nature, 2016, 530(7589): 211−214. DOI: 10.1038/nature16512

[27] 陈云, 李玉强, 王旭洋, 等. 中国生态脆弱区全球变化风险及应对技术途径和主要措施 [J]. 中国沙漠, 2022, 42(3): 148−158. [28] 李子辉, 苏湘媛, 田甜, 等. 基于格局-质量-功能的高寒草甸区生态脆弱性分析: 以云南迪庆为例 [J]. 中国环境科学, 2024, 44(4): 2273−2285. DOI: 10.3969/j.issn.1000-6923.2024.04.048 [29] 艾文竹, 陈棋, 田湘云, 等. 近20年间云南省石漠化演变过程及特征分析 [J]. 西部林业科学, 2023, 52(2): 40−47. [30]

Wang B, Chen Y, Yan Z J, et al. Integrating remote sensing data and CNN-LSTM-attention techniques for improved forest stock volume estimation: a comprehensive analysis of Baishanzu forest park, China [J]. Remote Sensing, 2024, 16(2): 324. DOI: 10.3390/rs16020324