Assessment of Surface Water Depth using Hyperspectral Reflectance: Case study from a Large-scale Hydrological Manipulation Experiment in the Arctic

As part of an NSF supported Biocomplexity project at Barrow, AK, this study focused on estimating surface water depth using hyperspectral reflectance data. Water tables have been manipulated (drained, flooded, control) in a vegetated Arctic thaw lake basin to investigate the effect of soil moisture on land-atmosphere carbon balance. Throughout the 2008 growing season, hyperspectral reflectance data were collected in the visible and near IR using a robotic tram system that operated along a 300m tramline situated in each treatment. Water table depths were also collected from water wells placed along transects. Our analysis shows a strong correlation between the 970nm reflectance band and the natural log of the water table depth (R2=0.72) for the flooded section, while this was not true for the drained (R2=0.36) or control (R2=0.38) sections, which had water tables at or below the ground surface. We discuss these results in light of using 970nm reflectance for the assessment of surface water depth.