USGS - science for a changing world

South Florida Information Access (SOFIA)

publications > paper > a hydrological budget (2002-2008) for a large subtropical wetland ecosystem > results

A Hydrological Budget (2002-2008) for a Large Subtropical Wetland Ecosystem Indicates Marine Groundwater Discharge Accompanies Diminished Freshwater Flow


Equations, Figures & Tables

Rainfall A seasonal and bimodal pattern was observed in all 7 years of rainfall data (Fig. 2a) with high rainfall occurring in the wet season (June-October). While the rainfall for 2002-2008 averaged over the SRS was similar to a longer term average annual rainfall (1,366 mm, based on 1980-2009 at Royal Palm Ranger Station, ENP), inter-annual variability was observed (Table 3; Fig. 2 (top)) with 2002 receiving the lowest (1,142 mm) and 2005 the highest (1,448 mm). While daily rainfall varied considerably between the stations, indicating spatial patchiness in precipitation on a daily time scale, monthly rainfall amounts did not vary greatly between stations. Monthly rainfall did however vary from year to year (Fig. 5). Uncertainty in rainfall data is due to several factors: instrument precision (0.254 mm on a daily reading translates to 92 mm/year), spatial variability (insignificant on a monthly scale), and inter-annual variation.

Inflow and Discharge from SRS Over 2002-2008, surface water inflows commenced well into the wet season and lagged behind rainfall by 1-2 months (Fig. 2 (middle), Fig. 4, Fig. 5). There was considerable inter-annual variation in the inflows, with the high of 691 mm in 2005 coinciding with the highest precipitation in 2002 through 2008, part of which was contributed by the hurricanes Katrina and Wilma) while almost no inflow occurred in 2007 (Table 3). Inflows were 25-50% of discharges from SRS in each year (Table 3; Fig. 4) with the exception of 2003.Outflows varied from 1,399 mm in 2005, the year with the highest rainfall, to a low of 502 mm in 2003 (Table 3).

Evapotranspiration The Shuttleworth PM model yielded ET values that varied seasonally: 5-7 mm/day in summer and 1-4 mm/day in winter (Fig. 3, Fig. 4, and Fig. 5). Mangrove communities were associated with higher ET than sawgrass on account of greater LAI and vegetation height (Fig. 3). Other models showed a similar seasonal variation but with different magnitudes (Table 1); for instance the FAO PM model had higher ET values than the Shuttleworth PM model, possibly reflecting the agricultural bias of the FAO PM model that is designed to work for a single species under well-watered conditions; thus the FAO PM model yielded similar potential ET estimates for South Florida as the USGS ( The Priestley-Taylor equation yielded values twice as large; this model has been reported to overestimate ET in humid areas in summer (Yoder et al. 2005; Suleiman and Hoogenboom 2007). ET estimates obtained from latent heat flux data showed the same seasonal pattern as the other models but was lower in magnitude as has been noticed elsewhere (Bidlake et al. 1996) whereby latent heat based measurements are almost always lower than values yielded by meteorological vapor transport models.

On account of several hurricanes and tropical storms that passed through the area over 2004-2007, the weather towers were severely damaged, leading to large gaps in data over this period. Hence, while we calculated ET over the entire 2004-2008 data availability period, for the budget we consider 2008 as a "model" year with complete daily ET estimates calculated from SRS6 eddy flux tower data. The monthly ET values obtained by Abtew's simple radiation model for 2004 and 2008 did not differ significantly (p>0.99). ET calculated from SRS6 meteorological data can be considered representative of SRS, because net radiation, the single most important driver of ET in the Everglades (Abtew 1996) was found to be very similar at the four eddy flux towers spread over an area of almost 5,000 km2 in the month of August 2008 when we had data from all four weather towers. Error in ET measurements stems from instrumental precision (taken as 5% from Price et al. 2006) as well as from differences in vegetation composition and soil moisture at a given instant of time (Fig. 3).

raw data plots of: monthly rainfall averaged over Shark River Slough and the 30-year average monthly rainfall at Royal Palm Ranger Station; monthly inflows and outflows (five major rivers into the Gulf of Mexico); and, daily water level or stage averaged over the Shark River Slough
Fig. 2 (above) Raw data for selected water budget components for Shark River Slough, ENP. Top, monthly rainfall averaged over Shark River Slough; also shown is the 30-year average monthly rainfall at Royal Palm Ranger Station, ENP (with less than 4% data missing). Middle, monthly inflows (S12s and S333) and outflows (five major rivers into the Gulf of Mexico). Bottom, daily water level or stage averaged over the SRS [larger image]

plot of monthly evapotranspiration estimates obtained by the Shuttleworth Penman-Monteith model for mangrove and sawgrass
Fig. 3 (above) Monthly evapotranspiration estimates obtained by the Shuttleworth PM model for mangrove and sawgrass, the dominant vegetation communities in the Shark River Slough. The difference in vegetation communities were parameterized by LAI values of 3 and 2 as well as by vegetation height of 9 and 1 m for mangroves and sawgrass, respectively [larger image]

plot of average monthly water budget components (2002-2008)
Fig. 4 (above) Average monthly water budget components (2002-2008). Negative y-axis values signify outputs from the Shark Slough [larger image]

glyph plot of the monthly water budget components (2002- 2008)
Fig. 5 (above) Glyph plot of the monthly water budget components (2002- 2008). The spokes of each glyph (polygon) represent the magnitude of each variable (independently standardized by variable from 0 to 1) for the indicated month. For example, the glyph for Sept 2008 represents high rainfall, moderate ET, high inflow, high outflow, low GWD, and very low salinity [larger image]

Table 2 Parameters for the Penman-Monteith equation (Shuttleworth 1992)
Variable Description Notes:
es Saturated vapor pressure at the measurement height in kilopascals (kPa). T is the mean air temperature in degrees Celsius (°C) saturated vapor pressure equation
Δ The gradient of the actual vapor pressure with the mean air temperature es (in kPa°C-1) gradient of the actual vapor pressure equation
A Energy budget for a unit area. A is the available energy A = Rn - G - S - P - Ad
Ad Loss of energy associated with horizontal air movement. Significant in an oasis situation, generally neglected Units are MJ m-2 day-1
Rn Net incoming radiant energy Units are MJ m-2 day-1. Obtained at TSPh7b
G Outgoing heat conduction into the soil G = cs*ds(T2 - T1)/Δt. Units are MJ m-2 day-1
cs Soil heat capacity 2.1 MJ m-3°C-1
ds Estimated effective soil depth (m) For daily temperature fluctuations, a value of 0.18 can be assumed
S Energy temporarily stored within the volume Often neglected except for forest
P Energy absorbed by biochemical processes in the plants Typically taken as 2% of net radiation
D Vapor-pressure deficit D = [(es*Tmax + es*Tmin)/2]*(1 - RH)/100 RH = relative humidity
ea Actual vapor pressure (kPa)  
ra Aerodynamic resistance (s m-1) aerodynamic resistance equation
rs Surface resistance of the land cover (s m-1) One approximation is 200/L s-1 m-1
Zu Height of the wind speed measurements (m)  
Ze Height of the humidity measurements (m)  
Uz Wind speed (ms-1)  
Z0m Roughness length for momentum transfer (m) Z0m = 0.123 hc
Z0v Roughness length for vapor transfer (m) Z0v = 0.0123 hc
d zero-plane displacement height (m) d = 0.67 hc
K Van Karman constant (=0.41)  
hc Vegetation height (m)  
ρa Sensible heat (kg m3), density of air sensible heat equation
Cp Specific heat of air at constant pressure, taken as 1.01 KJ kg-1 K-1  
P Atmospheric pressure in kPa  
T Mean air temperature in degrees Celsius  
Γ Psychrometric constant (kPa°C-1) psychrometric constant
E Ratio of the molecular weight of water vapor to that for dry air 0.622
λ Latent heat of vaporization of water (MJ kg-1) λ = 2.501 - 0.002361Ts

Table 3 Annual water budget components (2002-2008) in millimeters/year for Shark River Slough, Everglades National Park, including average and standard error of the mean (SEM) computed over the entire 7-year period
Budget component (mm) 2002 2003 2004 2005 2006 2007 2008 Average SEM
Rain 1,142 1,285 1,224 1,448 1,325 1,269 1,375 1,295 38
ET 1,367 1,367 1,367 1,367 1,367 1,367 1,367 1,367 0
Inflow 509 688 396 691 303 24 258 410 119
Outflow 952 502 1,077 1,399 892 521 771 873 91
Seepage 166 178 177 168 148 115 149 157 8
Storage change -78 131 -109 117 -216 -116 303 5 69
Groundwater 449 -88 585 572 270 336 673 400 97
Total error 306 292 306 339 292 258 283 297 9
Negative values for storage mean a net decrease in storage over the year. We have taken the 2008 value of ET for all years. Inflow and outflows have been expressed as millimeters per year by dividing the total volume by the area of Shark River Slough. Total error refers to the difference between the residual (obtained by summing all known inputs and outputs) and groundwater discharge

Water Levels Water levels in SRS varied by less than 1.5 m on a seasonal basis (Fig. 2c). The highest water levels were observed in September-November of each year. Water levels were at their lowest in May of each year.

Water Budget and Net Groundwater Discharge The annual change in water storage in SRS varied from deficit to surplus, from a net gain of 303 mm at the end of 2008 to a net loss of 216 mm in 2006 (Table 3). The magnitude of annual change in water storage was orders of magnitude lower than the other water budget components (Fig. 4). Over the period studied, 2002-2008, rainfall was the largest input, averaging 1,295 mm/year while ET was the largest output averaging 1,367 mm/year (Table 3). Outflows exceeded inflows and all other inputs combined, which led to a net groundwater input into the SRS (termed Groundwater Discharge or GWD) on an annual basis. GWD varied seasonally, with the largest inputs occurring in May-July, and a net recharge or input from the slough surface water in Jan-April (Fig. 4). There was a large annual variation with the highest value of net GWD into the SRS (673 mm) in 2008 and a net recharge of groundwater (88 mm) occurring in 2003 (Table 3; Fig. 5). The estimate of GWD varied with the ET model used (Table 1). The values of GWD obtained in the water budget result from the selection of the modified Shuttleworth PM model to estimate ET.

Surface Water and Groundwater Salinity Surface water salinity varied seasonally between 0 and 23.5 psu. Groundwater salinity also varied seasonally, with the maximum annual values increasing from less than 5 psu in 2003 to about 15 psu in 2008 (Fig. 6). With monthly data, analysis by lag correlation indicated that surface water salinity in SRS had a significant positive correlation with GWD (Fig. 7) when leading by 1 month (p<0.03). The same pattern was demonstrated in the daily data, where the surface water salinity expressed a significant positive correlation with GWD when leading by 24-34 days (p<0.05).

Surface Water and Groundwater Level Both groundwater and surface water levels at site SH2 co-vary over monthly and yearly time scales (Fig. 8 (top)). Daily groundwater level is consistently higher than surface water in early wet season (end of May to August). Monthly averages indicated that at SH2, groundwater discharge to the surface water was dominant in June-August, with surface water recharging the groundwater being dominant during the other months (Fig. 8 (bottom)).

plot of salinity in surface water and groundwater at site SH-2 and SRS-4
Fig. 6 (above) Salinity in surface water and groundwater at USGS site SH-2, and FCE-LTER site SRS-4 [larger image]

plot of monthly groundwater discharge to Shark River Slough and monthly surface water salinity at SH2 averaged over 2002-2008
Fig. 7 (above) Monthly groundwater discharge to SRS and monthly surface water salinity at SH2 (the oligohaline ecotone) averaged over 2002-2008 [larger image]

Fig. 8 (right) Top, daily groundwater (black solid line) and surface water level (red dashed line) in cm relative to NAVD 1988 at USGS site SH2. Bottom, monthly average difference between groundwater level and surface water level calculated over 2000-2008 and expressed in cm at SH2. Positive values of this difference, as seen in June-Aug, signify that the groundwater was higher than surface water on average in this time every year. Error bars signify the standard error of the mean [larger image] plot of daily groundwater and surface water level at site SH2; and bar chart of monthly average difference between groundwater level and surface water level calculated over 2000-2008 at SH2

< Methods | Discussion >

| Disclaimer | Privacy Statement | Accessibility |

U.S. Department of the Interior, U.S. Geological Survey
This page is:
Comments and suggestions? Contact: Heather Henkel - Webmaster
Last updated: 04 September, 2013 @ 02:04 PM (KP)