Estimating Peatland water table depth and net ecosystem exchange: A comparison between satellite and airborne imagery
Remote Sensing , Volume 10 - Issue 5
Peatlands play a fundamental role in climate regulation through their long-term accumulation of atmospheric carbon. Despite their resilience, peatlands are vulnerable to climate change. Remote sensing offers the opportunity to better understand these ecosystems at large spatial scales through time. In this study, we estimated water table depth from a 6-year time sequence of airborne shortwave infrared (SWIR) hyperspectral imagery. We found that the narrowband index NDWI1240 is a strong predictor of water table position. However, we illustrate the importance of considering peatland anisotropy on SWIR imagery from the summer months when the vascular plants are in full foliage, as not all illumination conditions are suitable for retrieving water table position. We also model net ecosystem exchange (NEE) from 10 years of Landsat TM5 imagery and from 4 years of Landsat OLI 8 imagery. Our results show the transferability of the model between imagery from sensors with similar spectral and radiometric properties such as Landsat 8 and Sentinel-2. NEE modeled from airborne hyperspectral imagery more closely correlated to eddy covariance tower measurements than did models based on satellite images. With fine spectral, spatial and radiometric resolutions, new generation satellite imagery and airborne hyperspectral imagery allow for monitoring the response of peatlands to both allogenic and autogenic factors.
|Bog, CASI, Hydrology, Hyperspectral, Landsat 8 OLI, Landsat TM5, Mer Bleue, Northern peatland, SASI, Sentinel-2|
|Organisation||Department of Geography and Environmental Studies|
Kalacska, M. (Margaret), Arroyo-Mora, J.P. (J. Pablo), Soffer, R.J. (Raymond J.), Roulet, N.T. (Nigel T.), Moore, T.R. (Tim R.), Humphreys, E, … Inamdar, D. (Deep). (2018). Estimating Peatland water table depth and net ecosystem exchange: A comparison between satellite and airborne imagery. Remote Sensing, 10(5). doi:10.3390/rs10050687