Aboveground Forest Biomass Modeling Using Remote Sensing and FIA Data in Tennessee, USA
Abstract
In Tennessee, more than half of land area is covered by forest; however, total available aboveground forest biomass is unknown. Mapping spatial distribution of aboveground forest biomass is important as it is the prerequisite for assessing carbon stock, evaluating forest health as well as quantifying actual feedstock availability for bioenergy production. Traditional field-based forest inventory is costly, labor-intensive and time-consuming. Thus, remote sensing images which capture the spectral characteristics of vegetation could be important predictor variables for estimating aboveground forest biomass from pixel to landscape level. The main objective of this study was to identify the important predictor variables derived from remote sensing data while mapping total aboveground forest biomass at a scale that is relevant for operational forest management. Remote sensing data and their derivatives such as Landsat -5 thematic mapper, digital elevation model (DEM), national land cover data (NLCD) and climate data were used in this study. The vegetation indices from Landsat data were derived for leaf-on and leaf-off seasons in ArcGIS environment. Plot-level data from Forest Inventory and Analysis (FIA) from 2007 to 2011 were obtained and paired with feature layers derived from remote sensing data. A nonparametric approach i.e. Random Forests was used to build a predictive model for aboveground biomass in R statistical computing environment. The model explained 38.57% of variability with RMSE of 18.46 tons acre-1. A continuous gridded biomass map across the State of Tennessee was generated, which can be useful for long-term planning of forest-based woody biomass.
Subject Area
Statistics|Plant sciences|Remote sensing
Recommended Citation
Man Kumari Giri,
"Aboveground Forest Biomass Modeling Using Remote Sensing and FIA Data in Tennessee, USA"
(2018).
ETD Collection for Tennessee State University.
Paper AAI10842519.
https://digitalscholarship.tnstate.edu/dissertations/AAI10842519