The overlooked dark-colored lichens: Blind spots in Antarctic vegetation mapping with remote sensing

Lichens are highly tolerant of extreme climates and are widely distributed in Antarctica. As pioneer species, they can fix nitrogen, break down rocks, absorb water, and stabilize soil, which helps create conditions for higher plants and supports biodiversity and ecosystem stability in Antarctica.1 However, conventional remote sensing based on spectral indices often struggles to detect lichens.2 The normalized difference vegetation index (NDVI), widely used to estimate vegetation density, relies on red and near-infrared (NIR) reflectance, with larger differences indicating higher values. However, dark-colored lichens exhibit relatively low NIR compared with red reflectance, resulting in persistently low NDVI (Figure 1A). Other indices, such as the enhanced vegetation index or the chlorophyll/carotenoid index, which target vegetation structure or pigment composition, also exhibit limited sensitivity for dark-colored lichens due to weaker red-edge and green band signals compared with vascular plants.2 Consequently, the extent and distribution of Antarctic vegetation are often underestimated and remain incompletely quantified. Recently, Walshaw et al. presented a satellite-derived map of vegetation, including vascular plants, bryophytes, green algae, and lichens across Antarctica based on Sentinel-2 imagery and spectral indices. They also noted the limitations of their method in detecting dark-colored lichens.3 Building on their results, here we examine how these limitations affect Antarctic vegetation mapping using multiple field investigation datasets.

We evaluated the accuracy of Walshaw et al.’s results in two representative Antarctic vegetated areas: the South Shetland Islands (62.2°S, 58.9°W) and the Northeast Peninsula (64.2°S, 57.3°W). Survey data were collected from 22 plots in the South Shetland Islands during 2018–2020 and from 26 plots in the Northeast Peninsula during 1988–2017 (Figure 1B). Given the slow growth rate of lichens (0.5–2.0 mm year−1), the different time spans between our study and Walshaw et al.’s are unlikely to have a significant effect on the relative proportion of dark-colored lichens. At each site, one to two landscape photographs were taken to capture vegetation types together with non-vegetated surfaces such as rocks, bare soil, snow, and water. In addition, two to five close-up photographs were obtained to quantify vegetation cover and to distinguish among vascular plants (grasses and herbs), mosses, lichens (light- and dark-colored), and algal patches. Each photograph was analyzed using a grid of 100 dots to determine percentage cover for each category (Figure 1B). We then extracted the corresponding classification data from Walshaw et al.’s results for each plot and computed the cover information for each vegetation type. Finally, we compared these statistics with field-based estimates while accounting for the contribution of dark-colored lichens to total lichen cover.

To enhance the robustness of the evaluation, we utilized an additional dataset for lichen coverage estimation, which includes 90 samples collected from 4 sites including Badaling, West Coast, South Coast, and Adelay Island within the South Shetland Islands, during the austral summer of 2009–2010.5 At each site, samples were taken from four habitat types (rock, gravel beach, thin soil on stone, and thin soil), with four to six samples per habitat type. Lichen coverage was visually estimated, and species identity was confirmed through microscopic identification. The proportion of dark-colored lichens relative to total lichens was then calculated based on species composition and coverage.