Suitability and limitations of ground-based imagery and thermography for long-term monitoring of vegetation changes in Victoria Land (continental Antarctica)

Abstract

Antarctic vegetation has been recognized to be a valuable bio-indicator to track climatic and environmental changes through an accurate long-term monitoring. The extremely harsh climatic conditions of Antarctica, its limited logistical accessibility and remoteness encourage the substitution or integration of field surveys with remote sensing monitoring.Here we assess the applicability and limitations of ground-based remote sensing (visible imagery and thermography) for accurate long-term monitoring of vegetation changes in continental Antarctica with reference to: a) total vegetation coverage; b) cover of the dominant species; c) vegetation seasonality. We selected the three most widespread continental Antarctic vegetation types (high cover moss; low cover moss; low cover moss and/or lichen).For the total vegetation cover the best fitting with the field data was achieved by the fishnet grid analysis performed by the expert and by the RGB_sup analysis, the two methods providing the highest feasibility, especially for the high cover moss, while for the other vegetation types the remote sensing methods provided over- and/or under-estimations (including GEI, differently from the Arctic).Regarding the dominant species cover (%), none of the remote sensing methods provided suitable results, while we demonstrated that seasonality affects the quantification of total vegetation cover through remote sensing due to changes of vegetation temperature, hydration and activity, especially for moss vegetation, even analyzing mono-specific vegetation plots. This finding underlines the importance of the timing of the digital image acquisition, an issue that has never been addressed before in continental Antarctica.

Keywords

• Climate change
• Vegetation as bio-indicator
• Field survey data
• Ground-based remote sensing
• Fishnet grid
• RGB supervised