Residual trend method based on regional modeling and machine learning for attribution of vegetation changes
-
摘要: 现有的残差趋势法采用逐像元建模策略, 利用普通最小二乘法构建模型, 存在着一定的局限性: 一方面, 逐像元建模策略使每个模型都包含了局部空间内的人类活动信号干扰; 另一方面, 普通最小二乘法不利于模拟普遍存在的非线性特征。因此, 该文提出一种全新的基于区域建模策略和机器学习算法的残差趋势法, 并对比了用于表达空间异质性的2种环境变量: ①地形、水文和土地利用等直接环境变量(direct-environmental variables, DEVs); ②植被和气候时空序列组合的代理环境变量(proxy-environmental variables, PEVs)。首先, 采用区域建模策略, 分别引入DEVs和PEVs, 使用机器学习算法构建植被-气候关系模型; 其次, 根据残差趋势法的定义得到残差值; 最后, 评估气候和人为因素对植被变化的贡献。结果表明: ①相比以往的逐像元普通最小二乘残差趋势法, 所提方法的优势不仅表现为机器学习能够模拟植被-气候关系的非线性特征, 还表现为区域建模具备更强的抗人类信号干扰能力; ②新方法中, 使用PEVs明显优于使用DEVs, 前者充分利用了原有建模数据, 没有增加数据获取难度, 也避免了引入额外的数据误差。该文提出的区域机器学习残差趋势法可以实现更有效的植被变化归因。Abstract: Existing residual trend methods utilize a pixel-by-pixel modeling strategy, in which the ordinary least squares method is employed. These methods suffer certain limitations. On the one hand, the pixel-by-pixel modeling strategy causes each model to contain signal interference from human activities in local space. On the other hand, the ordinary least squares method is unfavorable for simulating commonly observed nonlinear characteristics. This study proposed an entirely new residual trend method based on regional modeling and machine learning. Besides, this study compared two types of environmental variables used to express spatial heterogeneity: ①direct-environmental variables (DEVs) such as terrain, hydrology, and land use; and ②proxy-environmental variables (PEVs) that combine the spatiotemporal series of vegetation and climate. First, a regional modeling strategy was adopted. After DEVs and PEVs were introduced individually, models for the vegetation-climate relationship were built using machine learning. Second, residuals were determined based on the definition of the residual trend method. Finally, the contributions of anthropogenic and climatic factors to vegetation change were assessed. The results indicate that compared to the previous pixel-by-pixel residual trend method that utilizes ordinary least squares, the new residual trend method can simulate the nonlinear features of the vegetation-climate relationship and exhibits enhanced resistance to human signal interference. For the new method, significantly higher performance can be achieved using PEVs compared to DEVs. PEVs can fully utilize the original modeling data, without increasing difficulties with data acquisition and avoiding additional data errors. The residual trend method based on regional modeling and machine learning proposed in this study allows for more effective attribution of vegetation changes.
-
-
[1] 孙红雨, 王长耀, 牛铮, 等.中国地表植被覆盖变化及其与气候因子关系──基于NOAA时间序列数据分析[J].遥感学报, 1998, 2(3):204-210.
Sun H Y, Wang C Y, Niu Z, et al.Changes of surface vegetation coverage and its relationship with climate factors in China : Analysis based on NOAA time series data[J].National Remote Sensing Bulletin, 1998, 2(3):204-210.
[2] Intergovernmental Panel on Climate Change (IPCC).Climate change 2013:The physical science basis working group I contribution to the fifth assessment report of the intergovernmental panel on climate change[M].Cambridge:Cambridge University Press, 2014.
[3] Liu Y, Li Y, Li S C, et al.Spatial and temporal patterns of global NDVI trends:Correlations with climate and human factors[J].Remote Sensing, 2015, 7(10):13233-13250.
[4] Evans J, Geerken R.Discrimination between climate and human-induced dryland degradation[J].Journal of Arid Environments, 2004, 57(4):535-554.
[5] Guo E, Wang Y, Wang C, et al.NDVI indicates long-term dynami-cs of vegetation and its driving forces from climatic and anthropogenic factors in Mongolian Plateau[J].Remote Sensing, 2021, 13(4):688.
[6] Radda I A, Kumar B M, Pathak P.Land degradation in Bihar, In-dia:An assessment using rain-use efficiency and residual trend analysis[J].Agricultural Research, 2021, 10(3):434-447.
[7] Li Z D, Wang S, Li C J, et al.The trend shift caused by ecological restoration accelerates the vegetation greening of China’s drylands since the 1980s[J].Environmental Research Letters, 2022, 17(4):044062.
[8] Qu L L, Huang Y X, Yang L F, et al.Vegetation restoration in response to climatic and anthropogenic changes in the Loess Plateau, China[J].Chinese Geographical Science, 2020, 30(1):89-100.
[9] Zhao C L, Yan Y, Ma W Y, et al.RESTREND-based assessment of factors affecting vegetation dynamics on the Mongolian Plateau[J].Ecological Modelling, 2021, 440:109415.
[10] Liu Z J, Liu Y S, Li Y R.Anthropogenic contributions dominate trends of vegetation cover change over the farming-pastoral ecotone of Northern China[J].Ecological Indicators, 2018, 95:370-378.
[11] 金凯, 王飞, 韩剑桥, 等.1982-2015年中国气候变化和人类活动对植被NDVI变化的影响[J].地理学报, 2020, 75(5):961-974.
Jin K, Wang F, Han J Q, et al.Contribution of climatic change and human activities to vegetation NDVI change over China during 1982-2015[J].Acta Geographica Sinica, 2020, 75(5):961-974.
[12] Zhao Y B, Sun R H, Ni Z Y.Identification of natural and anthropogenic drivers of vegetation change in the Beijing-Tianjin-Hebei megacity region[J].Remote Sensing, 2019, 11(10):1224.
[13] Zhou S L, Zhang W C, Wang S H, et al.Spatial-temporal vegetation dynamics and their relationships with climatic, anthropogenic, and hydrological factors in the Amur River Basin[J].Remote Sensing, 2021, 13(4):684.
[14] Chen L F, Zhang H, Zhang X Y, et al.Vegetation changes in coal mining areas:Naturally or anthropogenically Driven?[J].Catena, 2022, 208:105712.
[15] Leroux L, Bégué A, Lo Seen D, et al.Driving forces of recent vegetation changes in the Sahel:Lessons learned from regional and local level analyses[J].Remote Sensing of Environment, 2017, 191:38-54.
[16] Yang L, Shen F X, Zhang L, et al.Quantifying influences of natural and anthropogenic factors on vegetation changes using structural equation modeling:A case study in Jiangsu Province, China[J].Journal of Cleaner Production, 2021, 280:124330.
[17] 刘志红, Tim R.McVicar, Van Niel, 等.专用气候数据空间插值软件ANUSPLIN及其应用[J].气象, 2008, 34(2):92-100.
Liu Z H, McVicar T R, Niel V, et al.Introduction of the professional interpolation software for meteorology data:ANUSPLINN[J].Meteorological Monthly, 2008, 34(2):92-100.
[18] 姚楠, 董国涛, 薛华柱.基于GoogleEarthEngine的黄土高原植被覆盖度时空变化特征分析[J].水土保持研究, 2024, 31(1):260-268.
Yao N, Dong G T, Xue H Z.Analysis on the characteristics of the spatiotemporal change in vegetation coverage on the Loess Plateau using the google earth engine[J].Research of Soil and Water Conservation, 2024, 31(1):260-268.
[19] Drewa P B, Platt W J, Moser E B.Community structure along elevation gradients in headwater regions of longleaf pine savannas[J].Plant Ecology, 2002, 160(1):61-78.
[20] Pei H W, Liu M Z, Jia Y G, et al.The trend of vegetation greening and its drivers in the Agro-pastoral ecotone of Northern China, 2000-2020[J].Ecological Indicators, 2021, 129:108004.
[21] Chaaban F, El Khattabi J, Darwishe H.Accuracy assessment of ESA WorldCover 2020 and ESRI 2020 land cover maps for a region in Syria[J].Journal of Geovisualization and Spatial Analysis, 2022, 6(2):31.
[22] Hasan M A, Mia M B, Khan M R, et al.Temporal changes in land cover, land surface temperature, soil moisture, and evapotranspiration using remote sensing techniques:A case study of Kutupalong Rohingya refugee camp in Bangladesh[J].Journal of Geovisualization and Spatial Analysis, 2023, 7(1):11.
[23] 敖泽建, 王建兵, 蒋友严, 等.2000-2017年甘南牧区植被变化特征及其影响因子[J].沙漠与绿洲气象, 2020, 14(1):95-100.
Ao Z J, Wang J B, Jiang Y Y, et al.Characteristics of vegetation change and influencing factors in Gannan pastoral area from 2000 to 2017[J].Desert and Oasis Meteorology, 2020, 14(1):95-100.
[24] 李文龙, 薛中正, 郭述茂, 等.基于3S技术的玛曲县草地植被覆盖度变化及其驱动力[J].兰州大学学报(自然科学版), 2010, 46(1):85-90, 95.
Li W L, Xue Z Z, Guo S M, et al.Vegetation coverage changes and analysis of the driving forces in Maqu County based on 3S technology[J].Journal of Lanzhou University (Natural Sciences), 2010, 46(1):85-90, 95.
[25] 李丽丽, 王大为, 韩涛.2000-2015年石羊河流域植被覆盖度及其对气候变化的响应[J].中国沙漠, 2018, 38(5):1108-1118.
Li L L, Wang D W, Han T.Spatial-temporal dynamics of vegetation coverage and responding to climate change in Shiyang River basin during 2000-2015[J].Journal of Desert Research, 2018, 38(5):1108-1118.
[26] 胡春艳, 卫伟, 王晓峰, 等.甘肃省植被覆盖变化及其对退耕还林工程的响应[J].生态与农村环境学报, 2016, 32(4):588-594.
Hu C Y, Wei W, Wang X F, et al.Change in vegetation cover as affected by grain for green project in Gansu[J].Journal of Ecology and Rural Environment, 2016, 32(4):588-594.
[27] Zhou W, Gang C, Zhou L, et al.Dynamic of grassland vegetation degradation and its quantitative assessment in the Northwest China[J].Acta Oecologica, 2014, 55:86-96.
[28] 张建永, 李扬, 赵文智, 等.河西走廊生态格局演变跟踪分析[J].水资源保护, 2015, 31(3):5-10.
Zhang J Y, Li Y, Zhao W Z, et al.Tracking analysis on changes of ecological patterns in Hexi Corridor Region [J].Water Resources Protection, 2015, 31(3):5-10.
[29] Wang Q, Adiku S, Tenhunen J, et al.On the relationship of NDVI with leaf area index in a deciduous forest site[J].Remote Sensing of Environment, 2005, 94(2):244-255.
-
计量
- 文章访问数: 32
- PDF下载数: 5
- 施引文献: 0
下载: