|Home||Archived October 29, 2018||(i)|
James E. Saiers
Our co-located sampling networks allow us to track the interaction of hydrology, sediment, and vegetation over time, and will provide the opportunity to monitor the progress of the Everglades restoration and to gauge its success.
Our research addresses processes relevant to the following restoration and related questions:
1. How will increasing freshwater flow affect wetland primary production? 2. Will increasing freshwater inflow alter nutrient availability? 3. Does recovery following disturbance in mangroves depend on freshwater inflow? 4. Will the position of vegetation ecotones change in response to upstream water management? 5. What will be the influence of global climate change, such as sea-level rise, on the Everglades restoration? 6. Will processes of wetlands soil formation be altered by sea-level rise and changed freshwater inflow?
Smith, Thomas J., III
Genereux, David P.; Saiers, James E.
The full article is available via journal subscription or single article purchase. The abstract may be viewed on the Wiley InterScience website
Saiers, James E.
The full article is available via journal subscription or single article purchase. The abstract may be viewed on the Science Direct website
Smith, T. J., III
Lynch, J. C,; Hensel, P.; Boumans, R.; Perez, B. C.; Segura, B.; Day, J. W. Jr
Powell, M. D.
The full article is available via journal subscription or single article purchase. The abstract may be viewed on the JSTOR website.
Davis, G.; Loope, L.; Roman, C.; Smith, T. J. III; Tilmant, J.
Cahoon, D. R.
Hudson, J. H.; Robblee, M. B.; Powell, G. V. N.; Isdale, P. J.
Robblee, M. B.; Wanless, H. R.; Doyle, T. W.
Kadlec, R. H.
Kouwen, N.; Soulis, E. D.
The full article is available via journal subscription or single article purchase. The abstract may be viewed on the American Geophysical Union website
We selected the model boundary conditions to ensure agreement between simulated flow patterns and general flow directions observed in the field. We assumed no-flow conditions across the southeastern boundary (SEB) and across the segment of the northwest boundary (NWB) located downgradient of the NP203 site. We assigned values of hydraulic head along the entire northeastern boundary (NEB) on the basis of head measurements collected from the NP202 and NE5 sites. Boundary heads were set to decrease linearly between NP202 and NE5, while the boundary heads between NE5 and the corner of the model domain were set equal to the head measured at NE5. Values of the boundary heads between NP202 and NP203 were determined from linear interpolation of head measurements made at these sites. Specified head conditions also were adopted for the southwest boundary (SWB), such that all boundary heads were set equal to the hydraulic head measured at the P35 site. During the time periods of model application, heads along the upgradient boundary varied between 1.9 and 2.5 m (relative to National Geodetic Vertical Datum 29 (NGVD-29)), while heads along the downgradient boundary varied between 0.4 and 0.8 m. Initial conditions were estimated through interpolation of hydraulic head values measured at the monitoring sites along the edges of the domain (i.e. NP202, NP203, NE5, and P35) and at the monitoring sites within the interior of the domain (i.e. P33, P36, and S1).
We obtained time-series data on hydraulic head at sites NP202, NP203, NE5, P33, P35, P36, and S1 from databases maintained by Everglades National Park and the US Geological Survey. For each site, we used daily measurements of hydraulic head (relative to NGVD-29) that were computed by averaging 15-min interval data.
In order to run the flow model and simulate the spatiotemporal variability in hydraulic heads, it is necessary to specify the evapotranspiration rates (E), the rainfall rates (P), the ground-surface slopes (dz/dx and dz/dy), the wetland porosity, the surface water conductivity coefficient (Kf), and the exponential constants (beta and lambda). We obtained E and P from field measurements, wetland porosity on the basis of literature values, and Kf and the exponential constants from calibration.
We assumed that evapotranspiration was uniform across the model domain and that the rates of evapotranspiration could be determined on the basis of measurements made at the P33 site. Ed German of the US Geological Survey supplied evapotranspiration rates for the P33 site. Based on estimates of energy fluxes, German (1999) calculated evapotranspiration rates at 30-min intervals according to the Bowen-ratio method. We averaged the 30-min interval evapotranspiration measurements to obtain average-daily evapotranspiration rates. For the time periods studied, evapotranspiration rates ranged from 0.30 to 8.0 cm d-1.
We employed a simple zonation procedure, in coordination with rainfall data collected from gages at the P33, P35, and P36 sites, to represent the spatial distribution in rainfall at the study area. We divided the model domain into three separate zones: rainfall rates could vary between zones, but were considered uniform within a zone. The zonal boundaries were positioned at the midline between the locations of adjacent rainfall gages. We assumed that the daily average rainfall rate within a zone was equal to the rainfall rate recorded in the gage situated within that particular zone. Results of preliminary analyses demonstrated the necessity of incorporating this rainfall zonation, as the flow model performed poorly when uniform rainfall rates, based on precipitation measurements at P33, were assumed for the model domain. All rainfall data used in the work reported here were obtained from the hydrological databases maintained by the Everglades National Park.
We modeled flow in the slough by incorporating constant values for wetland porosity. We do not have measurements of wetland porosity at our study site, and a review of the literature demonstrates that little effort has been devoted to determining values of this parameter for overland flow. In previous work, investigators have assumed values of wetland porosity ranging from 0.8 to 0.95 (Hammer and McKillop), which reflects that plant material typically occupies a small fraction of the water column. For this work, we adopt a value of 0.9. Although measurements necessary to tightly constrain wetland porosity a priori are unavailable, precise knowledge of the value of this parameter does not appear to be critically important because modeled results are relatively insensitive to changes in this parameter within the range 0.8-1.0 (Hammer and Kadlec, 1986).
We invoked the simplifying assumption that surface-water flow within the slough can be quantified in terms of uniform ground-surface slopes. Reliable data on land-surface elevation within Shark River Slough are scarce; however, the few measurements that are available indicate that the gradient in land-surface elevation is approximately parallel with the southeast and northwest boundaries of our model domain. Based on differences in measured surface elevations at the P33 and P35 sites, we estimated the ground-surface slope in the direction parallel to the long dimension of our model domain (dz/dy) as 5×10-5. We assigned a value of zero to dz/dx, the component of the surface slope in the x direction.
Values of Kf, beta, and wetland porosity were determined from model calibration; that is, optimal values of these parameters were identified by using a Levenberg-Marquadt algorithm to minimize an objective function.
We conducted three separate simulations: one calibration simulation, designed to estimate the flow parameters (Kf, beta, and wetland porosity), followed by two forward simulations, designed to assess the predictive capability of the calibrated model. The 243 d calibration period included data from both the wet season (May-October) and from the dry season (November-April); it began on 18 June 1996 and ended on 15 February 1997. The head data used in the first predictive simulation were collected over a 194 d period, beginning 17 January 1998 and ending 29 July 1998, while the data record for the second predictive simulation spanned a 139 d period, from 14 August 1998 to 30 December 1998. The modeling periods were chosen on the basis of the availability of continuous records for precipitation, evapotranspiration, and water levels at the sites representing the boundaries.
|Home||Archived October 29, 2018|