Hydrological effects of tree invasion on a dry coastal Hawaiian ecosystem

In ecosystems invaded by non-native plants invasion effects are often spatially variable, and this variability is difficult to capture via plot-scale sampling. We used airborne high-resolution LiDAR (Light Detection and Ranging) to generate spatially explicit and contiguous information on hydrological effects of invasive trees (Prosopis pallida (Humb. & Bonpl. ex Willd.) Kunth). We developed regression relationships between LiDAR metrics (i.e., ground elevation and tree canopy height) and plot-scale measurements of vegetation stem water δ18O, to assess groundwater use, and transpiration rates. We used electrical resistivity imaging to assess subsurface geology and hydrology and their relationships to P. pallida stand structure. P. pallida biomass and transpiration varied greatly across the study area; both were controlled by depth to groundwater. Stem water δ18O values (-8.6 to 3.7‰) indicated a threshold ground elevation of ca. 15 m above sea level, above which P. pallida could not access groundwater; this threshold corresponded to declines in tree biomass and height. Transpiration modelled across the study area was 0.034 ± 0.017 mm day−1, but over 98% of transpiration came from the ca. 25% of the total study area where groundwater depths were less than 15 m. Our combination of methods offers a new way to incorporate fine-scale spatial variation into estimation of plant invasion effects on hydrology, increase our understanding of interactions of geology, hydrology, and biology in such invasions, and prioritise areas for control in well-advanced invasions.