Assessing the Utility of Alternate Digital Image Color Space for Deriving Phenological Dynamics in a High-Arctic Tundra Ecosystem

Abstract:

The need to improve the spatial and temporal scaling and extrapolation of plot level ecosystem properties and processes to the landscape level remains a persistent research challenge in the Arctic. Plant and landscape phenology is sensitive to a number of spatiotemporally variable environmental factors such as soil moisture, temperature, and radiation. Seasonal and inter-annual differences in phenology can affect surface energy balance and land-atmosphere carbon flux. Considering the relative importance of the Arctic to global carbon balance, improved scaling and extrapolation of phenological dynamics from the plot level to the landscape level is important for advancing our understanding of the impact of climate and other environmental change in arctic terrestrial ecosystems.

Seasonal and interannual landscape phenology was observed over the Mobile Instrumented Sensor Platform (MISP) grid (2 x 50 meters) located in Barrow and Atqasuk, Alaska using imagery acquired from kite aerial photography (KAP), a hyperspectral ground-based spectrometer, and a phenocam. These data were analyzed in RGB and non-traditional HSV and l*a*b*color spaces to determine site, plant community seasonal, and inter annual phenological dynamics. Results were also compared to high spatial resolution satellite imagery to determine optimal indices for scaling vegetation dynamics from plot to landscape level.

These results show that greenness indices similar to those acquired from hyperspectral remote sensing platforms can be derived using low-cost and low-tech techniques that could be deployed at multiple sites at low cost.