Suitable Satellite Sensor for Elephant grass Above-Ground Biomass Estimation from Field Spectro-Radiometry Data

Abstract

A satellite sensor is a key instrument that remotely collects data about an object or scene. However, since different sensors have varying spatial, temporal, spectral and radiometric properties, it is very necessary for vegetation cover mapping to identify and select suitable sensor for specific purposes. This study analyses seven of the most widely used satellite sensors for vegetation mapping; and evaluate their performance on elephant grass Above-Ground Biomass (AGB) estimation. Spectro-radiometry and AGB data of 40 grass samples were used for modelling and validation. The site for the experiment was Daware grazing land, Nigeria. The satellites analysed were Landsat products (OLI and ETM), Sentinel 2 MSI, MODIS 09Q1, IKONOS, Worldview and SPOT 5. The spectral window for each sensor was identified. Red and NIR reflectance were extracted from the Spectro-radiometric measurements. Variations in the distribution of the Red and NIR spectral responses for each satellite window was evaluated. A ratio of NIR and Red was calculated as Vegetation Index Number (VIN). The calculated VIN and the measured AGB were correlated. The result indicates that Sentinel 2 MSI has a good data distribution in the Red band and the NIR band. The level of correlation between the field AGB and the VIN was also good (R2 = 0.927). The AGB calculated from Sentinel 2A MSI was validated at a good accuracy (RMSE = 0.326kg/pixel size and P value < 0.001) with the field measured AGB. The study concludes that Sentinel 2 MSI is the most suitable for estimating AGB for elephant grass. This provides a scientific contribution for accurate estimations of AGB specifically in grazing lands where grass information is vital.