Multispectral and LiDAR-Derived Vegetation Indicators of Water Table Dynamics in Forested Wetlands

We examined the potential to use remotely sensed measures of vegetation composition and structure, including spaceborne multispectral and airborne light detection and ranging (LiDAR) variables, as indicators of forested wetland hydrology. First, we identified statistically significant associations between a suite of remote sensing variables (multispectral and LiDAR-derived vegetation indices) and field-measured metrics of wetland vegetation structure and composition. We then used redundancy analysis (RDA) to evaluate the relationships between remotely sensed vegetation indices and growing season water table dynamics in 13 north-temperate forested wetlands in southwestern Quebec, Canada. Our results show that ratio measures of near-infrared (NIR) and red/green band combination (SR-Simple Ratio, NDVI-Normalized Difference Vegetation Index and NDVIG-Normalized Difference Green Vegetation Index) are associated with many hydrometric variables describing the water table position (WTP), water level fluctuations and hydroperiod. Individual band brightness measures are associated with a smaller subset of these hydrometric variables. However, the incorporation of LiDAR derivatives with image variables in multivariate RDA did not significantly contribute to the explained variance in hydrometric variables. Overall, our results demonstrate that image-derived vegetation indices, more so than LiDAR-derived metrics of vegetation structure, can be used as indicators of water table dynamics in forested wetlands.