Determining Optimal SAR Parameters for Quantifying Above-Ground Grass Carbon Stock in Savannah Ecosystems Using a Tree-Based Algorithm

Remote Sensing in Earth Systems Sciences https://doi.org/10.1007/s41976-024-00170-8 RESEARCH Determining Optimal SAR Parameters for Quantifying Above‑Ground Grass Carbon Stock in Savannah Ecosystems Using a Tree‑Based Algorithm Reneilwe Maake1,2 · Onisimo Mutanga2 · Johannes George Chirima1,4 · Mahlatse Kganyago3 Received: 5 June 2024 / Revised: 15 November 2024 / Accepted: 28 November 2024 © The Author(s) 2024 Abstract The quantification and monitoring of above-ground grass carbon stock (AGGCS) will inform emission reduction policies and aid in minimising the risks associated with future climate change. This study investigated the sensitivity of Synthetic Aperture Radar (SAR)-derived parameters to predict AGGCS in a savannah ecosystem in Kruger National Park, South Africa. Particularly, we investigated the capabilities of Sentinel-1 derived parameters, including backscatter coefficients, intensity ratios, normalised radar backscatter, arithmetic computations, and the XGBoost tree-based algorithm, to predict the AGGCS. We further tested if incorporating texture matrices (i.e. Gray Level Co-Occurrence Matrix) can enhance the predictive capability of the models. We found that the linear polarisation (i.e. VV) and the intensity ratio (i.e. VH/VV) achieved similar results (R2 = 0.38, RMSE% = 31%, MAE = 6.87) and (R2 = 0.37, RMSE = 31%, MAE = 8.80) respectively. The Radar Vegetation Index (RVI) performed marginally (1%) better (R2 = 0.39, RMSE = 30% and MAE = 6.77) compared to the other variables. Nevertheless, the incorporation texture matrix into the model enhanced prediction capability by approximately 20% (R2 = 0.60, RMSE% = 20%, MAE = 3.91). Furthermore, the most influential predictors for AGGCS estimation were RVI, VHcor and VVcor order of importance. These findings (R2 values of 0.35–0.39) suggest that SAR data alone does not fully capture the variability in above-ground grass carbon stock, particularly in the complexly configured savannah ecosystems. Nevertheless, the results further suggest that the prediction accuracy of SAR-based above-ground grass carbon stock models can be enhanced with the incorporation of texture matrices. Keywords Above-ground grass carbon stock · Savannah ecosystems · Sentinel-1 · XGBoost 1 Introduction The savannah ecosystem is one of the many vital biomes that exist on Earth. It is characterised by a rich biodiversity dominated by grasslands co-existing with patches of * Reneilwe Maake 1 Agricultural Research Council - Natural Resources and Engineering, Pretoria, South Africa 2 School of Agricultural, Earth and Environmental Sciences, Geography Department, University of KwaZulu-Natal, Pietermaritzburg, South Africa 3 Department of Geography, Environmental Management and Energy Studies, University of Johannesburg, Johannesburg, South Africa 4 Department of Geography, Geoinformatics & Meteorology, University of Pretoria, Pretoria, South Africa woodlands and shrublands [1, 2]. The grass layer serves as a feeding resource for grazing animals, while the woody layer serves as a feeding stock for browsing animals. From a climate perspective, the savannah grass and woody layers account for an estimated 25% of total gross primary production (GPP), rendering this ecosystem a vital sink of carbon [3]. They help regulate the Earth’s climate and maintain a balance of greenhouse gases in the atmosphere. The woody vegetation layer is known for storing large quantities of carbon; however, the sustainability of this capability is threatened by extreme weather conditions such as drought, heatwaves, and increased fire activity [4]. Studies have uncovered a proliferation of mortality in woody plants, which is associated with extreme weather events [5], fire [6, 7], and wildlife interactions [8, 9]. In the face of climate change, the grass layer, due to its biological configuration, is emerging as a sustainable terrestrial carbon sink [4]. The grass layer is resilient to various Vol.:(0123456789) Remote Sensing in Earth Systems Sciences disturbances, including grazing, fire activity, animal stampedes, and extreme weather conditions [10]. This implies that the grass layer can sustainably maintain a relatively stable carbon stock over time, an attribute that is vital as we approach the era of above-normal climate change. As such, the quantification and monitoring of aboveground grass carbon stock (AGGCS) will inform emission reduction policies and aid in minimising the risks associated with future climate change. This will align with the urgency emphasised in Sustainable Development Goal 13, which calls for immediate action to address climate change and lessen its repercussions. A further benefit of quantifying AGGCS is the generation of baseline information for scientific research and environmental management. The Paris Agreement is one environmental management initiative mandating countries to disclose their greenhouse gas emissions along with the reduction efforts [11]. In addition, statistics on grass carbon stock are essential for meeting reporting demands and ensuring transparency and accountability in climate actions taken by multiple nations. A myriad of grass carbon stock quantification missions rely on space-borne sensors due to their ability to capture data over large geographical and temporal footprints [12–17]. Primarily, optical sensors have been in use for decades, regardless of their limitations. For example, optical sensors rely on the visible and infrared bands, which are obstructed by clouds and smoke, and their interaction is limited to the canopy structure of the vegetation [18]. Consequently, optical sensors are unable to ensure a consistent distribution of data throughout multiple seasons. While SAR data has been in existence for decades, its application has been hindered by its price tag [18]. Sentinel-1 offers new opportunities for quantifying and monitoring above-ground carbon stock beyond the visible range of the electromagnetic spectrum, at no cost, and across smoky, hazy and overcast environmental conditions. Sentinel-1, constructed by the European Space Agency (ESA), is a Synthetic Aperture Radar (SAR) satellite mission [19]. The mission comprises a duo of identical sensors, namely Sentinel-1A and Sentinel-1B. These sensors can capture data under several weather conditions, an attribute that ensures a consistent flow of data, allowing for continuous quantification of carbon stock. Furthermore, Sentinel-1 can penetrate beyond the canopy structure of the vegetation, making it particularly valuable for capturing intra-canopy properties of vegetation [20]. SAR data affords several metrics which can be explored for quantifying above-ground carbon stock, some of which have been documented to be optimal for forests [21–23] and agricultural fields [24–27]. Studies showed that incorporating texture matrix with SAR data could improve biomass estimation [28–30]. However, fewer studies coupled SAR-derived parameters from senti (...truncated)