1.College of Ecology and Environment,Nanjing Forestry University,Nanjing,Jiangsu 210037,China
2.National Observation and Research Station of Fujian Wuyishan Forest Ecosystem,Wuyishan,Fujian 354300,China
3.Wuyishan National Park Fujian Research and Monitoring Center,Wuyishan,Fujian 354300,China
4.Information Technology Research Center,Beijing Academy of Agriculture and Forestry Sciences,Beijing 100097,China
5.National Engineering Research Center for Information Technology in Agriculture,Beijing 100097,China
Show less
文章历史+
Received
Published
2025-08-18
2026-09-10
Issue Date
2026-06-11
PDF (1705K)
摘要
目的 分析武夷山垂直植被带谱上不同植被类型生长的年际变化规律及其对气候因子的响应机制和时滞效应,为揭示亚热带山地植被对气候变化的响应提供科学依据。 方法 以武夷山国家公园为研究对象,基于2001-2024年中分辨率成像光谱仪(MODIS)MOD13Q1产品的归一化植被指数(normalized difference vegetation index,NDVI)数据,采用卡尔曼滤波算法重构缺失数据,通过决定系数(R²)和均方根误差(root mean square error,RMSE)验证重构精度;运用STL(seasonal and trend decomposition using loess)时间序列分解方法提取去季节性NDVI数据集;结合同期降水、气温数据,系统分析武夷山垂直带谱上5种典型植被类型(常绿阔叶林、针阔混交林、针叶林、矮林、草甸)生长的年际变化趋势及其对气候因子的响应机制和滞后时间。 结果 基于卡尔曼滤波重构的缺失和异常NDVI数据的精度均较高(R²>0.90)。常绿阔叶林和针阔混交林均呈现三阶段变化模式,分别于2017和2019年由增长转为下降;针叶林在2011年后进入持续下降;矮林于2019年由增转降;草甸则在2010年由降转升,增长最为显著。5种植被类型的生长状况均与年均温存在显著正相关关系(P≤0.05),而与降水无显著相关性,温度是影响武夷山垂直带谱植被生长的主要气候因子。生长季节内,不同植被类型NDVI对气候因子的滞后响应呈分异特征,常绿阔叶林对温度滞后1个月,针阔混交林和针叶林对温度滞后2个月,矮林和草甸对温度即时响应但对降水滞后3个月。 结论 年均温度是驱动武夷山垂直带谱植被生长年际变化的主导气候因子,降水主要通过时滞效应间接影响植被生长,不同植被类型对气候因子的响应存在显著时空异质性。
Abstract
Objective This research aims to analyze the interannual variation of different vegetation types along the vertical zonation of Wuyi Mountain and their response mechanisms and time-lag effects to climate factors,to provide scientific basis for understanding how subtropical montane vegetation responds to climate change. Method The normalized difference vegetation index (NDVI) data derived from MODIS MOD13Q1 from 2001 to 2024 was collected.The missing data was reconstructed by using a Kalman filter,and the reconstruction accuracy was assessed by using R² and root mean square error (RMSE).Seasonal and trend decomposition using loess (STL) was then applied to obtain deseasonalized NDVI series.Combined with concurrent temperature and precipitation data,interannual growth trends of five representative vegetation types (evergreen broad-leaved forest,coniferous and broad-leaved mixed forest, coniferous forest,dwarf forest,and meadow) and their response and time-lag to climate factors were assessed. Result The data reconstruction using the Kalman filter achieved high accuracy (R²>0.90).Evergreen broad-leaved forest and mixed forest showed three-phase change,shifting from an increasing to a decreasing trend in 2017 and 2019.Coniferous forest showed a persistent decline after 2011.Shift from an increasing to a decreasing trend was observed in dwarf forest in 2019, while shift from a decreasing to an increasing trend was observed in meadow in 2010,which exhibited the most pronounced increase.The growth conditions of all the vegetation types were significantly and positively correlated with annual mean temperature (P≤0.05),whereas no significant correlation with precipitation was detected.Temperature,therefore,emerged as the dominant climatic factor.During the growing season,time-lag effects varied among vegetation types:evergreen broad-leaved forest responded to temperature change with a 1-month lag,mixed forest and coniferous forest with a 2-month lag,whereas dwarf forest and meadow responded with no time lag to temperature but responded with a 3-month lag to precipitation. Conclusion Annual mean temperature is the primary climatic factor driving interannual vegetation dynamics along the vertical zonation of Wuyi Mountain,while precipitation affects vegetation mainly through time-lag effect.Vegetation types exhibit significant spatiotemporal heterogeneity in their responses to climatic factors.
本研究以武夷山国家公园植被垂直带谱为研究区,采用卫星遥感影像、卡尔曼滤波数据重构、STL(seasonal and trend decomposition using loess)季节性分离-趋势提取相结合的方法,研究2001-2024年该区5种典型植被类型(常绿阔叶林、针阔混交林、针叶林、矮林和草甸)NDVI的年际变化趋势,探讨不同植被类型对气候因子(降水、气温)的即时性和滞后性响应,以期厘清亚热带山地垂直带谱上各植被类型对气候变化响应的机制差异,进而为区域生态环境保护和气候变化适应策略制定提供科学依据。
式中:xk 为k时刻的真实NDVI状态向量; Fk 为状态转移矩阵,描述系统动态特性;为(k-1)时刻的状态向量;wk 为过程噪声,服从零均值高斯分布N(0, Qk ); zk 为k时刻的NDVI观测值向量; Hk 为观测矩阵,用于建立状态向量与观测值的映射关系;vk 为观测噪声向量,服从零均值高斯分布N(0, Rk )。为基于(k-1)时刻信息对k时刻状态的先验估计值;为(k-1)时刻状态的后验估计值;为k时刻的先验估计误差协方差矩阵;为(k-1)时刻的后验估计误差协方差矩阵;T表示矩阵转置; Qk 为过程噪声协方差矩阵; Rk 为观测噪声协方差矩阵,反映NDVI观测值的不确定程度,其值越大,表示观测数据的噪声越大,可信度越低; Kk 为卡尔曼增益; P(k|k)为k时刻的后验估计误差协方差矩阵; I 为单位矩阵。
1.3.2 NDVI数据平滑处理
本研究使用R语言“zoo”包的“na.StructTS()函数”实现卡尔曼滤波,该函数基于结构化时间序列模型(structural time series model),将时间序列分解为趋势、季节和不规则成分,通过最大似然估计自动优化参数,实现缺失值的最优估计。有效补全了原始 NDVI 数据因云、雨、雾等干扰产生的缺失,构建了时空连续数据集。但插值未消除原始观测中大气干扰、传感器误差等残留的高频噪声,可能会影响后续分析的可靠性。因此,进一步采用 Savitzky-Golay(S-G)滤波器进行平滑处理[16],以消除非植被生理过程的异常波动,确保重构数据既完整覆盖研究时段,又真实反映植被生长的连续特征。平滑处理表达式为:
LIANGM C, QINY Y, FANX,et al.Interpretation of the main conclusions and suggestions of IPCC AR6 working group Ⅲ report[J].Environmental Protection,2022,50(13):72-76.
[3]
LIS Q, ZHANGX, LUZ Y,et al.Progress of vegetation modelling and future research prospects[J].Science China Earth Sciences,2024,67(9):2718-2738.
[4]
WANGB X, CHENGW M, XUH,et al.Vegetation differentiation characteristics and control mechanisms in the Altay region based on topographic gradients[J].Ecological Indicators,2024,160:111838.
CAOH H, CHENY P, ZHANGH L,et al.Spatio-temporal variations of drought and its impact on vegetation NPP in southwest China[J/OL].Desert and Oasis Meteorology,2025:1-10.[2025-08-18].
LIUH, TANGD W, SUNY,et al.Spatiotemporal variation of NDVI in the vegetation growing season of Wuling mountainous area and its response to climate change during 2000-2019[J].Research of Soil and Water Conservation,2021,28(5):245-253.
QIG Z, BAIH Y, ZHAOT,et al.Sensitivity and areal differentiation of vegetation responses to hydrothermal dynamics on the southern and northern slopes of the Qinling Mountains in Shaanxi province[J].Acta Geographica Sinica,2021,76(1):44-56.
LIH X, LIUG H, FUB J.Response of vegetation to climate change and human activity based on NDVI in the Three-River Headwaters region[J].Acta Ecologica Sinica,2011,31(19):5495-5504.
ZHOUM T, LIJ, ZHUK W.Spatial-temporal dynamics of different types of vegetation NDVI and its response to climate change in Xinjiang during 1998-2012[J].Arid Land Geography,2015,38(4):779-787.
[15]
VAN DER PLASG W, RUCINAS M, HEMPA,et al.Climate-human-landscape interaction in the eastern foothills of Mt.Kilimanjaro (equatorial East Africa) during the last two millennia[J].The Holocene,2021,31(4):556-569.
SONGC Y, SHIS Y, WANGF,et al.Farmland classification based on Google Earth Engine and Sentinel-2 data[J].Journal of Northwest A&F University (Natural Science Edition),2024,52(11):132-141,154.
[20]
XUD D, HARDERJ K, XUW X,et al.Evaluating the impact of grazing cessation and reintroduction in mixed prairie using raster time series analysis of landsat data[J].Remote Sensing,2021,13(17):3397.
LIW G, HOUM T.A review of reconstruction methods for remote-sensing-based time series data of vegetation and some examples[J].Remote Sensing for Natural Resources,2022,34(1):1-9.
[23]
LIQ, CHENGX L, LUOY Q,et al.Consistent temperature sensitivity of labile soil organic carbon mineralization along an elevation gradient in the Wuyi Mountains,China[J].Applied Soil Ecology,2017,117:32-37.
[24]
SEDANOF, KEMPENEERSP, HURTTG.A Kalman filter-based method to generate continuous time series of medi-um-resolution NDVI images[J].Remote Sensing,2014,6(12): 12381-12408.
[25]
CHENJ, JÖNSSONP, TAMURAM,et al.A simple meth-od for reconstructing a high-quality NDVI time-series data set based on the Savitzky-Golay filter[J].Remote Sensing of Environment,2004,91(3/4):332-344.
[26]
HODSONT O.Root-mean-square error (RMSE) or mean absolute error (MAE):when to use them or not[J].Geoscientific Model Development Discussions,2022,2022:1-10.
JIANGX B, HUY M, LIUZ H,et al.Reconstructing NDVI time-series data using a linear-interpolation with extended Kalman filter[J].Bulletin of Science and Technology,2017,33(2):137-142.
YANGH T, XUH Q.Assessing fractional vegetation cover changes and ecological quality of the Wuyi Mountain National Nature Reserve based on remote sensing spatial information [J].Chinese Journal of Applied Ecology,2020,31(2):533-542.
[33]
PIAOS L, WANGX H, PARKT,et al.Characteristics,drivers and feedbacks of global greening[J].Nature Reviews Earth & Environment,2019,1(1):14-27.
[34]
RENH Y, WENZ M, LIUY Y,et al.Vegetation response to changes in climate across different climate zones in China[J].Ecological Indicators,2023,155:110932.
AND S, XUD D, PUY H,et al.Responses of subalpine meadow to climatic factors and the time lag effects in Wuyi Mountains from 2000 to 2019[J].Chinese Journal of Applied Ecology,2021,32(12):4195-4202.
[37]
GOTTFRIEDM, PAULIH, FUTSCHIKA,et al.Continent-wide response of mountain vegetation to climate change [J].Nature Climate Change,2012,2(2):111-115.
YAOY H, ZHANGB P, ZHAOC.Geographical distribution of cripple tree forest and its importance for forest line in China [J].Progress in Geography,2017,36(4):491-499.
ZHANGX L, HUANGL M, QUANQ,et al.Relationship of vegetation cover change with climate factors in the Source Region of the Yellow River based on ITPCAS forcing data[J].Journal of Northwest A&F University(Natural Science Edition),2019,47(9):55-68.
ZHANGJ H, FENGZ M, JIANGL G,et al.Analysis of the correlation between NDVI and climate factors in the Lancang River Basin[J].Journal of Natural Resources,2015,30(9):1425-1435.
LUQ Q, JIANGT, LIUD L,et al.The response characteristics of NDVI with different vegetation cover types to temperature and precipitation in China[J].Ecology and Environmental Sciences,2020,29(1):23-34.
[46]
WUD H, ZHAOX, LIANGS L,et al.Time-lag effects of global vegetation responses to climate change[J].Global Chan-ge Biology,2015,21(9):3520-3531.