The White method and its many adaptations have been widely used to estimate evapotranspiration (ET) from shallow groundwater (e.g., riparian areas) and surface water (e.g., lakes and wetlands) level fluctuations. A plot of water level over a single day is termed a diurnal curve. In both modes of application (surface or groundwater), an estimate of specific yield is required. Specific yield controls the magnitude of water level fluctuations for a given input (precipitation, groundwater seepage) or output (ET). For groundwater applications, the specific yield is often assumed constant. For surface water applications, specific yield varies with depth. The relationship between specific yield and depth is termed a specific yield function. For this study, a previously published specific yield function shown to perform well for a daily time step is tested and adapted for a subdaily (hourly) time step. Both daily and subdaily estimates of ET will be presented for a forested headwater wetland that is seasonally flooded. Rapid drawdowns of surface water, due to groundwater seepage and ET, occur during the growing season. During these rapid drawdowns of surface water, groundwater seepage varies considerably throughout the day. An existing technique is applied to estimate diurnal variability in seepage using a piecewise linear function. This is combined with the specific yield function to estimate hourly ET total. Hourly estimates of ET based on the Penman-Monteith (PM) Equation are scaled with a cover, or crop, coefficient and compared to hourly diurnal curve estimates.
Results/Conclusions
Results were obtained for six drawdown periods during the growing season. Hourly estimates of ET from diurnal curves compared favorably in magnitude and displayed similar temporal trends (number of peaks), but the PM estimates were found to consistently lag behind the diurnal curve estimates by 2-3 hours. The diurnal curve method also predicted more nighttime ET than the PM method. The results demonstrate that reasonable estimates of hourly ET are possible for surface water applications with variable specific yield. Although significant variation in specific yield occurs over a full drawdown of surface water (~6-10 days for the study wetland), the diurnal variation in specific yield can be neglected and assumed constant over a single day.