Abstract
Barrier island forests are sensitive to changing precipitation characteristics as they typically rely on a precipitation-fed freshwater lens. Understanding and predicting significant rainfall losses is, therefore, critical to the prediction and management of hydrometeorological processes in the barrier island forest ecosystem. This study measures and models one such loss, canopy rainfall interception, for a barrier island forest common across subtropical and tropical coastlines: epiphyte-laden Quercus virginiana on St. Catherine’s Island (Georgia, United States). Reformulated Gash analytical models (RGAMs) relying on static- and variable-canopy-storage formulations were parameterized using common maximum water storage (minimum, mean, maximum, and laboratory submersion) and evaporation (Penman–Monteith, saturated rain–throughfall regression, and rain–interception regression) estimation methods. Cumulative interception loss was 37% of rainfall, and the epiphyte community contribution to interception loss was 11%. Variable-storage RGAMs using inferred evaporation and maximum water storage estimates performed best: mean absolute error of 1–2 mm, normalized mean percent error of 15%–25%, and model efficiency of 0.88–0.97, resulting in a 2%–5% overestimate of cumulative interception. Static- and variable-storage RGAMs using physically derived evaporation (Penman–Monteith) underestimated observed interception loss (40%–60%), yet the error was significantly lowered for submersion estimates of maximum water storage. Greater apparent error when using Penman–Monteith rates may result from unknown drying times, evaporation sources, and/or in situ epiphyte storage dynamics. As such, it is suggested that future research apply existing technologies to quantify evaporative processes during rainfall (e.g., eddy covariance) and to develop new methods to directly monitor in situ epiphyte water storage.
Rainfall interception loss by corn crop