Spectral imaging of grass species in arid ecosystems of Namibia

A groundbreaking study reveals that data fusion products derived from Sentinel-2 and EnMAP imagery are significantly more effective for mapping arid grasslands compared to standalone multispectral or hyperspectral data, with validation showing R² values of 0.51 (NDVI) and 0.66 (SAVI). This research, conducted in Namibia's NamibRand Nature Reserve (NRNR) and ProNamib Nature Reserve (PNNR), employs a multi-sensor approach combining Mixture Tuned Matched Filtering (MTMF) with field-collected spectral data to analyze grassland distribution. The study also introduces a unique spectral library of grass species, though dry conditions limit inter-species spectral separability. These findings underscore the potential of advanced remote sensing techniques for monitoring and conserving arid ecosystems, aligning with global restoration efforts.

Bantelmann, P., Wyss, D., Pius, E. T., & Kappas, M. (2024). Spectral imaging of grass species in arid ecosystems of Namibia. Frontiers in Remote Sensing, 5. https://doi.org/10.3389/frsen.2024.1368551