Assessment of Sentinel-2 spectral correlation with grass nutrient concentrations in the semi-arid Telperion Nature Reserve, South Africa

Understanding the correlation between spectral reflectance and plant nutrient concentrations is critical for advancing remote-sensing applications in grassland ecosystems. This study evaluated the association between Sentinel‑2 spectral reflectance and nutrient concentrations in grassland ecosystems in the Telperion Nature Reserve, South Africa. Sentinel‑2 bands (B2–B12) were analysed in relation to macro- and micronutrients, including nitrogen (N), phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), iron (Fe), sodium (Na), zinc (Zn), manganese (Mn), and copper (Cu). Pearson correlation, analysis of variance (ANOVA), and regression analyses were applied to assess the relationships and predictive capability of Sentinel‑2 bands. The study results revealed generally weak associations between spectral reflectance and nutrient concentrations across most bands. However, statistically significant moderate correlations were observed between Ca and Mg and the shortwave infrared (SWIR) bands (B11 and B12), while Fe showed a positive relationship with B12, and P correlated with visible bands (B2 and B3). Analysis of variance showed significant variation in spectral reflectance among species, particularly in red-edge, near-infrared (NIR), and SWIR regions, highlighting their importance for species differentiation. Despite this, regression models using B8 (NIR) and B11 (SWIR) demonstrated low predictive power for most nutrients (R² < 0.15), with only Ca and Mg showing slight explanatory potential. Overall, the study findings suggest that Sentinel‑2 spectral reflectance has limited sensitivity for accurately predicting nutrient concentrations in the study species, although certain bands show moderate utility. These results highlight the challenges of estimating nutrient concentrations from multispectral data and the need for advanced modelling approaches or higher-resolution sensors to improve accuracy.