where R_i is incoming solar energy (measured using a pyranometer), S is the depth of water on the land surface, and C₁, C₂, and C₃ are coefficients determined by using least-squares regression to minimize the difference between measured latent heat and latent heat calculated using equation (1). The site models typically have coefficients of variation (CV) of 15 to 30 percent for the half-hourly ET totals, which are equivalent to about 2 to 4 percent for daily totals.

Regional models were then developed for the natural Everglades system. The seven individual site models for the vegetated sites were combined using least-squares regression into a single relation that can be used to estimate ET anywhere in vegetated areas as a function of available energy, solar intensity, and stage. This regional model is applicable to sawgrass marshes, wet prairies, and cattail areas. In open-water areas, the regional model consists of assigning a in equation 1 a constant value of 1.26, the theoretical maximum value. Neither of these regional models is intended for use in saltwater areas, forested areas, or agricultural areas.

Preliminary measured and simulated mean annual total ET for 1996-97 was calculated for all sites. In calculating the measured ET totals, simulated ET from the site models was used to estimate ET for intervals when air temperatures and vapor pressure gradients were missing because of sensor malfunction, or when the gradients were too small for accurate measurement. Simulated ET mean annual totals were computed entirely from the regional models.

The measured average annual total ET at the nine sites ranged from 42.3 inches at a vegetated site (C111) to 56.0 inches at an open-water site (Enr). These differences mostly reflect availability of water and density of vegetative cover. At site C111 (data available for 1997 only), the water was below land surface nearly half of the days, whereas at the open-water sites water was always above land surface with no emergent vegetation. Among the predominately-wet vegetated sites, the measured annual total ET ranged from 43.7 to 50.6 inches. These differences seem to be related to vegetative density and depth of water on the land surface.

ET at another site, dominated by cattails, was noticeably lower than at the other sites where water level was always above land surface. This relatively low ET at the cattail site could be related to the carpet of duck weed that often completely covered the water surface, perhaps isolating the water surface from the atmosphere to some degree and thus retarding evapotranspiration. Another possibility is that accumulations of dead cattail stalks could have intercepted significant amounts of solar radiation and increased the transport of convective heat relative to latent heat.

The ability of the regional models to simulate ET is indicated by the comparison of the measured and simulated annual ET. Absolute values of the differences between simulated and measured mean annual total ET at the nine sites ranged from 3.7 inches to 0.4 inch, and averaged 1.8 inches for all sites. Improved estimates of ET allow for more accurate simulation of water flows, a very important factor in restoration of the Everglades.

REFERENCE

German, E.R., 1996, Regional evaluation of evapotranspiration in the Everglades: U.S. Geological Survey Fact Sheet, FS-168-96, 4 p.

(This abstract was taken from the Proceedings of the South Florida Restoration Science Forum Open File Report)

Back to Project Homepage