Mapping Forest Aboveground Biomass Using Multisource Remotely Sensed Data

Dekker Ehlers; Chao Wang; John W. Coulston; Yulong Zhang; Tamlin Pavelsky; Elizabeth Frankenberg; Curtis Woodcock; Conghe Song

doi:10.3390/rs14051115

Treesearch

Mapping Forest Aboveground Biomass Using Multisource Remotely Sensed Data

Formally Refereed

Authors:

Dekker Ehlers, Chao Wang, John W. Coulston, Yulong Zhang, Tamlin Pavelsky, Elizabeth Frankenberg, Curtis Woodcock, Conghe Song

Year:

2022

Type:

Scientific Journal

Station:

Southern Research Station

DOI:

https://doi.org/10.3390/rs14051115

Source:

Remote Sensing

Download (PDF 4.69 MB)

Download (PDF 4.69 MB): https://research.fs.usda.gov/download/treesearch/64611.pdf

Abstract

The majority of the aboveground biomass on the Earth’s land surface is stored in forests.Thus, forest biomass plays a critical role in the global carbon cycle. Yet accurate estimate of forestaboveground biomass (FAGB) remains elusive. This study proposed a new conceptual model to mapFAGB using remotely sensed data from multiple sensors. The conceptual model, which providesguidance for selecting remotely sensed data, is based on the principle of estimating FAGB on theground using allometry, which needs species, diameter at breast height (DBH), and tree height asinputs. Based on the conceptual model, we used multiseasonal Landsat images to provide informationabout species composition for the forests in the study area, LiDAR data for canopy height, and theimage texture and image texture ratio at two spatial resolutions for tree crown size, which is relatedto DBH. Moreover, we added RaDAR data to provide canopy volume information to the model. Allthe data layers were fed to a Random Forest (RF) regression model. The study was carried out ineastern North Carolina. We used biomass from the USFS Forest Inventory and Analysis plots to trainand test the model performance. The best model achieved an R2 of 0.625 with a root mean squarederror (RMSE) of 18.8 Mg/ha (47.6%) with the “out-of-bag” samples at 30 30 m spatial resolution.The top five most important variables include the 95th, 85th, 75th, and 50th percentile heights of theLiDAR points and their standard deviations of 85th heights. Numerous features from multiseasonalSentinel-1 C-Band SAR, multiseasonal Landsat 8 imagery along with image texture features fromvery high-resolution imagery were selected. But the importance of the height metrics dwarfed allother variables. More tests of the conceptual model in places with a broader range of biomass andmore diverse species composition are needed to evaluate the importance of other input variables.

Citation

Ehlers, Dekker; Wang, Chao; Coulston, John; Zhang, Yulong; Pavelsky, Tamlin; Frankenberg, Elizabeth; Woodcock, Curtis; Song, Conghe. 2022. Mapping Forest Aboveground Biomass Using Multisource Remotely Sensed Data. Remote Sensing. 14(5): 1115-. https://doi.org/10.3390/rs14051115.