Estimating mineral content of indigenous browse species using laboratory spectroscopy and Sentinel-2 imagery

Abstract

Understanding variation in the foliar nutrient among indigenous species of the bushveld is crucial for rural livelihoods, in particular the integration of trees into agroecosystems. The study explored nutrient composition of common browse species with regard to nitrogen (N), phosphorus (P), potassium (K) and calcium (Ca) using leaf spectra (400–2500nm) and chemical data collected from nine bushveld species, along with partial least squares (PLS) analysis. The work further explored the relationship between canopy reflectance of Sentinel-2 image and foliar properties of the identified species. Spectroscopic analysis reveals useful information about nitrogen at leaf and canopy scales whereas modelling reflectance using satellite image did not yield satisfactory results. At the leaf level, nitrogen was highly correlated with leaf spectral reflectance (R2=0.72, p<0.05) for winter and (R2=0.88, p<0.05) for summer. The coefficient of determination for the relationships between leaf reflectance and concentrations of phosphorus, potassium and calcium were low. Modelling the relationship using Sentinel-2 data also showed higher correlations (R2=0.44, p<0.05) for nitrogen compared with the other nutrients investigated.