Oblique field spectral measurements for enhanced vegetation cover mapping and sand vegetation discrimination in semi-arid gobi landscapes

The Gobi Desert, although often appearing barren in nadir satellite imagery, contains ecologically significant grasslands with sparse yet vital vegetation. This study investigates how oblique field spectral measurements can enhance vegetation detection and improve discrimination between vegetation and sandy surfaces in semi-arid Gobi landscapes. A key observation is that nadir measurements primarily capture the spectral characteristics of bright sand, whereas oblique views increasingly reveal clear vegetation spectral signatures, even at very low biomass levels. Field measurements were conducted during two separate campaigns in 2023 and 2024, using a multiband Liquid Crystal Tunable Filter (LCTF) camera that captured 33 and 65 spectral bands, respectively, between 460 and 780 nm, across view zenith angles from 0° to 87° and azimuth angles from 0° to 360°. Together, these datasets represent the first multi-angular, field-based Bidirectional Reflectance Factor (BRF) measurements of sparse Gobi vegetation spanning a complete azimuth–zenith domain. The 2023 measurements primarily examined spectral variation and angular dependence, whereas the 2024 campaign was designed to confirm these relationships and to construct complete BRF maps. The results reveal pronounced anisotropy in surface reflectance, particularly in the near-infrared (NIR) region, where reflectance increased by up to 20% under oblique angles compared to nadir observations. Vegetation indices such as NDVI (Normalized Difference Vegetation Index) rose by up to 64.7% at a view zenith angle of 60°, demonstrating enhanced vegetation signal detection from multi-angular observations. These results provide the first direct evidence that viewing geometry fundamentally controls vegetation detectability in newly vegetated Gobi surfaces. These findings highlight the potential of oblique spectral imaging to enhance vegetation mapping accuracy in arid regions, suggesting that peripheral satellite data, often discarded due to high view zenith angles, may contain valuable information for vegetation monitoring. Such advancements are crucial for developing nature-based solutions that promote sustainable land management and biodiversity conservation in desert ecosystems.