基于局部回归模型的森林生物量动态变化分析
Analysis of Forest Biomass Dynamics Based on Local Regression Model
卢士欣 1贾炜玮 1孙毓蔓 1张小勇 1吴思敏 1肖锐2
作者信息
- 1. 东北林业大学林学院,森林生态系统可持续经营教育部重点实验室,黑龙江哈尔滨 150040
- 2. 黑龙江省林业科学研究所,黑龙江哈尔滨 150040
- 折叠
摘要
基于丰林县地区4期Landsat影像和对应气象站点数据,结合该地区248块固定样地数据,利用全局回归模型(多元线性模型)和2种局部回归模型(地理加权回归模型、时空地理加权回归模型)建立研究区乔木地上生物量和遥感因子之间的关系,选出最优模型来研究丰林县乔木地上生物量时空变化.结果表明:根据3种模型的模拟结果数据与实测值的分析对比可以发现,局部回归模型的拟合效果要优于全局模型,加入时间特征的时空地理加权回归模型的拟合效果最好,模型评价指标与地理加权回归模型相比更为理想.统计得到研究区4个时期内总的乔木地上生物量分别为1.63 × 107、2.05 × 107、2.32 × 107、3.37 × 107 t,4个时期的平均乔木地上生物量分别为54.82、68.98、77.87、113.46t/hm2,乔木地上生物量呈现出逐期增加的趋势.利用遥感因子估测丰林地区的地上生物量,为未来该地区生物量分布的估计提供了依据.
Abstract
This study based on the 4-period Landsat remote sensing images and the meteorological station data in Fenglin County,the global regression model(multiple linear model)and 2 local regression models(geo-graphically weighted regression model and geographically and time weighted regression model)were used to es-tablish the relationship between above-ground biomass of trees and remote sensing factors in the study area.The optimal model was selected to study the spatial and temporal variation of above-ground biomass in Fenglin County.The results showed that the simulation results of the 3 models were better than the global model,and the geographically temporal weighted regression model with the addition of temporal characteristics had the best fit-ting effect,and the model evaluation indexes were better compared with the geo-weighted regression model.The total above-ground biomass of trees in the study area was 1.63 x 107,2.05 x 107,2.32 x 107,3.37 x i07 t.The average above-ground biomass of trees in the 4 periods was 54.82,68.98,77.87,113.46 t/hm2.The total above-ground biomass of trees in the study area showed a trend of increasing from period to period.The use of re-mote sensing factors to estimate the above-ground biomass in the rich forest area provides a basis for estimating the future biomass distribution in the area.
关键词
时空地理加权回归模型/地上生物量/时空动态变化/遥感估算Key words
geographically and time weighted regression model/above ground biomass/spatiotemporal dy-namics/remote sensing estimation引用本文复制引用
基金项目
国家重点研发计划项目(2022YFD2201003-02)
中央高校基本科研业务费专项资金项目(2572019CP08)
中央高校基本科研业务费专项资金项目(2572022DT03)
出版年
2024
国家科技期刊平台
NSTL
万方数据