Estimating Aboveground Biomass Using Sentinel-2 MSI Data and Ensemble Algorithms for Grassland in the Shengjin Lake Wetland, China

Abstract

Wetland vegetation aboveground biomass (AGB) directly indicates wetland ecosystem health and is critical for water purification, carbon cycle, and biodiversity conservation. Accurate AGB estimation is essential for the monitoring and supervision of ecosystems, especially in seasonal floodplain wetlands. This paper explored the capability of spectral and texture features from the Sentinel-2 Multispectral Instrument (MSI) for modeling grassland AGB using random forest (RF) and extreme gradient boosting (XGBoost) algorithms in Shengjin Lake wetland (a Ramsar site). We use five-fold cross-validation to verify the model effectiveness. The results indicated that the RF and XGBoost models had a robust and efficient performance (with root mean square error (RMSE) of 126.571 g·m−2 and R2 of 0.844 for RF, RMSE of 112.425 g·m−2 and R2 of 0.869 for XGBoost), and the XGBoost models, by contrast, performed better. Both traditional and red-edge vegetation indices (VIs) obtained satisfactory results of AGB estimation (RMSE = 127.936 g·m−2, RMSE = 125.879 g·m−2 in XGBoost models, respectively), with the red-edge VIs contributed more to the AGB models. Moreover, we selected eight gray-level co-occurrence matrix (GLCM) textures calculated by four processing window sizes using the mean value of four offsets, and further analyzed the results of three analysis sets. Textures derived from traditional and red-edge bands using a 7 × 7 window size performed better in biomass estimation. This finding suggested that textures derived from the traditional bands were as important as the red-edge bands. The introduction of textures moderately improved the accuracy of modeling AGB, whereas the use of textures alo ne was not satisfactory. This research demonstrated that using the Sentinel-2 MSI and the two ensemble algorithms is an effective method for long-term dynamic monitoring and assessment of grass AGB in seasonal floodplain wetlands, which can support sustainable management and carbon accounting of wetland ecosystems.

Estimating the Leaf Water Status and Grain Yield of Wheat under Different Irrigation Regimes Using Optimized Two- and Three-Band Hyperspectral Indices and Multivariate Regression Models

“…Previous studies have reported that the SRIs constructed from the three-band wavelengths were more effective and accurate to assess the measured plant traits than those constructed from the two-band ones. This is because the three-band SRIs display less saturation and are less sensitive to several plant characteristics, such as the internal leaf structure and leaf biochemical compounds [12,[77][78][79][80][81][82][83][84][85][86][87][88][89]. Furthermore, the mathematical formulas of the SRIs, which are often a simple ratio, normalized difference, or mixed between both, play also a vital role in the efficiency of SRIs for accurately estimating plant traits under different growth conditions [12,80,81].…”

“…In this study, there are many three-band SRIs that combine the VIS, red-edge, and NIR wavelengths, and were effective for estimating the plant water status indicators and GY under different growth conditions. Interestingly, many studies have reported that there are several satellite-derived indices, especially those based on VIS/red-edge, NIR/VIS, and NIR/red-edge, that can achieve accurate estimation of different field measured agronomic traits, such as AGB [44,88], GY [89,90], and plant water status indicators [91,92]. Additionally, Herrmann et al [93] reported that the satellite-derived indices have been shown to provide almost comparable results as hyperspectral data.…”

Remote Sensing of Wetlands in the Prairie Pothole Region of North America

“…Such approaches may prove valuable in identifying waterbodies, which may subsequently be interrogated to determine more detailed wetland characteristics such as class, form, and type. Other decision tree methods have also been used to map wetland classes, flora, and fauna, including: Classification Tree (CT) Analysis [282][283][284][285], Gradient Boosting (GB) [106,[286][287][288], and Classification And Regression Tree (CART) [6,[289][290][291]. Baker et al [106] noted GB to be preferable to CT approaches for mapping wetland, non-wetland, and riparian land cover classes; however, Tulbure et al [282] obtained an overall accuracy of 96% when classifying water bodies from other land cover types.…”

Estimating Above-Ground Biomass of Potato Using Random Forest and Optimized Hyperspectral Indices

“…The spectral parameters are widely used to estimate AGB at local, regional, and global scales [12][13][14]. For decades, spectral indices (SIs) are common parameters for the non-destructive estimation of AGB in crops [15,16]. Much research has reported that the SIs were used to estimate the AGB of maize [6,17,18], wheat [10,19,20], and rice [15,21].…”