Answers are grounded in this paper and cite their sources.
High Arctic tundra vegetation is highly fragmented, reflecting micro-topographical heterogeneity that governs local patterns of snow accumulation and soil moisture, and thus the distribution of plant communities. In this case study from Svalbard, we show the potential of using high-resolution earth observation data obtained with an uncrewed aerial system (UAS) to map and classify tundra vegetation communities, highlighting the importance of fine-scale hydro-topographic features. We used multispectral and Light Detection and Ranging (LiDAR) UAS data with a spatial resolution of 10 cm to map the land cover in a 140 ha study area. The resulting land cover map has a mean F1-score of 0.79 and shows the distribution of 14 land cover types. We further derived the Topographic Position Index (TPI) and Topographic Wetness Index (TWI) to analyze the terrain exposure and wetness of heath, moss tundra, and wetland vegetation communities. Our results reveal hydro-topographic vegetation gradients at centimeter-scale, demonstrating the importance of high-resolution data in understanding vegetation distribution in High Arctic tundra ecosystems. We observed a moderate negative correlation between TPI and TWI in highly exposed sites dominated by heath vegetation (plateau: ρ=-0.39, p<.001; ridges: ρ=-0.54, p<.001), which aligns with previous research about the effect of small-scale differences in snow depth on soil moisture and thereby the distribution of heath vegetation communities. Moreover, we found a moderate positive correlation between TWI and in situ soil moisture measurements (ρ=0.52, p<.001), indicating that TWI can be used as a rough approximation of soil moisture in High Arctic ecosystems at a reasonably high level of confidence. Our findings highlight the potential of high-resolution UAS data in understanding micro-topographic properties of vegetation distribution in High Arctic tundra ecosystems.
No takes yet. Share an insight, caveat, or question.
Rauth et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: