Quantifying aboveground vegetation water storage combining Landsat 8 OLI and Sentinel-1 imageries. Issue 9 (3rd May 2022)

• Record Type: Journal Article
• Authors: Luo, X. R.; Li, S. D.; Liu, L.; Yang, W. N.; Zhang, Y. H.; Chen, G.; Qiu, S. Y.; Tang, Q. L.; Tang, X. L.
• Is Part Of: Geocarto international. Volume 37:Issue 9:(2022)
• Page Start: 2717
• Page End: 2734
• Publication Date: 2022-05-03
• Subjects: Aboveground vegetation water storage -- Landsat 8 OLI -- Sentinel-1 -- ANN -- RF; Remote sensing -- Periodicals; Geographic information systems -- Periodicals; Geology -- Periodicals; Cartography -- Periodicals; 621.3678
• DOI: 10.1080/10106049.2020.1861662
• Languages: English
• ISSNs: 1010-6049

Abstract

Abstract: Although optical remote sensing has been widely used to monitor vegetation characteristics, its use on aboveground vegetation water storage (AVWS) is quite scare. Therefore, we combined the Landsat 8 OLI and Sentinel-1 imageries to quantify AVWS using generalized linear regression (GLM), artificial neural network (ANN) and random forest (RF) with the linkage of field observations in Mao Country, Southwest China. Field observations showed that the AVWS varied significantly among different ecosystems ( p < 0.001). In terms of model efficiency and root mean square error, ANN (0.66, 56 Mg ha −1 ) performed best compared to RF (0.52, 66 Mg ha −1 ) and GLM (0.48, 69 Mg ha −1 ). Total AVWS was 3.8 × 10 7 Mg for the whole study area with 76% contributions from coniferous forests. Strong spatial patterns of AVWS were observed among different ecosystems, which were highly consistent with the spatial distributions of vegetation types. This research highlights a potential way to estimate AVWS through the combination of multispectral remote sensing and SAR data by linking machine learning algorithms, particular in mountainous areas.