**1. Introduction**

Crystal River/Kings Bay is a small but complicated, spring-fed estuarine system located on the Gulf coast of central Florida (Figure 1). It has a very small runoff basin, as spring water accounts for 99% of the freshwater flows entering Kings Bay. The estuarine system includes the 2.43 km2 Kings Bays as its head water and the 10 km long Crystal River that joins Kings Bay with the Gulf of Mexico. It is a first magnitude spring system, which is defined as having a discharge rate of 100 cubic feet per second (cfs) or greater [1]. In fact, it is the fourth largest spring system in Florida with an estimated discharge of about 1000 cfs or higher. Because SGD is an overwhelming part of the total freshwater inflow received by the estuarine system, the Crystal River/Kings Bay estuary serves as an excellent example demonstrating the importance of SGD in controlling physical, chemical, and biological processes in coastal waters.

The Crystal River/Kings Bay system is ecologically very important for some marine species such as West Indian manatees (*Trichechus manatus*), because a large amount of warm spring water with a relatively constant temperature of about 23 °C flows to the Kings Bay through numerous spring vents on a daily basis. This creates a large warm water pool in Kings Bay during the coldest days when the air temperature plunges to several degrees below 0 °C (32 °F) and the water temperature at the mouth of the Crystal River drops to 10 °C or lower. Because manatees need to be in water that is at least 20 °C (68 °F) or warmer to maintain a safe internal body temperature, this large warm water pool in Kings Bay becomes a critical refuge site for manatees to survive when water temperature in the area falls below 20 °C and attracts many manatees to the Crystal River/Kings Bay system in winter. With approximately 350 manatees inhabiting the spring-fed estuary during winter months, it is believed that the Crystal River/Kings Bay area is the largest natural refuge for manatees in the United States.

**Figure 1.** An aerial photo of the Crystal River/Kings Bay system located on the southwest coast of the Florida peninsula. Locations of USGS *in-situ* measurement stations are marked with triangles and locations of identified spring vents are marked with asterisks. The solid circle at the bottom right is the location of a well called ROMP TR21-3.

In addition to the obvious effect of the spring flow on thermo-characteristics of the Crystal River/Kings Bay system, SGD is also a key factor determining the salinity distribution in the system, which controls the ecological structure and biological productivities in the estuary. Maintaining a certain volume of fresh water or brackish pool in the Crystal River/Kings Bay system is crucial for many species. The tidal brackish ecosystem supports abundant fish and wildlife resources that are of great importance to the region both economically and ecologically.

Because of the importance of warm freshwater input to Crystal River/Kings Bay, it is necessary to have a sound management of spring flow to the estuary so that the natural warm water refuge for manatees and the health of the ecosystem are protected. Obviously, a good set of data of spring flows from all the spring vents in and around Kings Bay is critical in managing the system. A number of previous studies were conducted to study spring flow, water circulations, water quality, aquatic vegetation, water clarity, sediment characteristics, management of manatees, *etc.* in Crystal River/Kings Bay (e.g., [2–4]) Yobbi and Knochenmus [2] estimated the total spring discharge exiting Kings Bay to be about 975 cfs during 1965–1977. Their study reported a relatively low spring flow rate in summer and fall months, when rainfall and tides were higher, and a high spring

**130**

flow rate in winter and spring months, when rainfall and tides were lower. They attributed this anomalous timing of SGD in the Crystal River/Kings Bay system to the seasonality of tides.

Numerous SGD-related investigations have been performed over the last couple of decades, trying to quantify SGDs and study processes affecting SGDs at various geophysical settings. Most previous SGD measurements focused on the diffusive seepage through sediments [5–7], which is a relatively slow process and is in general measured with a time scale that is much longer than that of a tidal cycle. In a Karst landscape such as the Crystal River region, SGDs from localized submarine spring vents can be quite large in magnitude and normally vary swiftly with time because of the high-frequency tidal variability. The relatively large magnitude of point flow from a coastal spring allows the discharge to be measured with a regular velocity meter such as an acoustic Doppler velocimeter.

In order to quantify the freshwater input to the Crystal River/Kings Bay system and to study effects of tides on spring flows in Kings Bay, the Southwest Florida Water Management District contracted Vanasse Hangen Brustlin, Inc. to conduct field measurements in 2008–2009. Data collected in this field investigation were analyzed and an empirical formula relating spring flow to tides and groundwater level was obtained based on real-time SGD data collected from a small portion of spring vents. A 3D hydrodynamic model that simulates circulations, salinity transport, and thermodynamics in the Crystal River/Kings Bay system was used to find out if this empirical formula is applicable for all of the identified spring vents. A successful model calibration/verification against real-time data measured in Crystal River/Kings Bay during a 2.84-year period from 24 April 2007 to 23 February 2010 confirms that this empirical formula for estimating real-time spring flows out of the numerous vents in Kings Bay is reasonable.

In the following, details of the data collection during 2008–2009 are first described, followed by an analysis of the field data, which results in an empirical formula relating spring flows with tides and the groundwater level. The use of this empirical formula to estimate discharges out of each spring vent in a 3D hydrodynamic model application to the Crystal River/Kings Bay system is then presented, before conclusions are drawn at the end of the paper.

## **2. Field Data**

#### *2.1. Data Collection*

There exist only limited data collection activities that have tried to quantify flows out of submerged vents in Kings Bay. Rosenau *et al.* [8] measured instantaneous flows and water quality parameters from selected springs that flow into Kings Bay, with a total of 30 reported springs being identified and listed in their report. The same 30 springs in Kings Bay were also listed in a more recent bulletin of Florida Bureau of Geology [9]. Spring flow rates and water quality in the Crystal River were measured by Seaburn *et al.* [10] in April 1974 to support a water quality modeling study of the system. Yobbi and Knochenmus [2] estimated the average total spring discharge during 1965–1977 to be about 975 cfs for Kings Bay. In an effort to simulate circulation and flushing characteristics of Kings Bay [3], the United States Geological Survey (USGS) conducted a flow measurement during 7–8 June 1990 near Bagley Cove (Figure 1) in the Crystal River. The net

flux through this cross section during the tidal cycle was found to be 735 cfs. In a 2D hydrodynamic simulation by Hammett *et al.* [3], 28 major springs in Kings Bay were included in their model based on information from [8]. In a spring water quality study by the Southwest Florida Water Management District (SWFWMD) during 1993–2004, additional spring vents in Kings Bay were identified.

As none of the aforementioned previous studies of spring discharges to Kings Bay is spatially or temporally comprehensive, it is necessarily to conduct a more extensive data collection study to quantify spring flows and to find out how SGD is affected by tides and the groundwater level in the estuary. For this purpose, a two-phase field investigation in Kings Bay was conducted during 2008–2009. The first phase was a thorough inventory survey, in which all the identifiable spring vents were identified with their locations (latitudes and longitudes) were recorded and configurations, including dimensions (areas) and orientations, were documented. The second phase was to measure discharges out of each spring vent with divers diving to the vents to measure the velocities of spring flows. The discharge was simply the product of the vent area and the velocity. For most of spring vents, multiple field trips were made to measure discharges under various tidal conditions.

During the inventory survey, previously documented spring vents were first visited and validated via snorkeling and SCUBA equipped diving [11]. The entire Kings Bay was then searched for additional spring vents that were not previously documented. At several spring sites, multiple vents are located in a close proximity, forming a vent cluster that jointly contributes to the overall discharge for the spring. A single set of coordinates was recorded for the vent cluster, which is considered as a single spring. The inventory survey was able to identify a total number of 70 springs, which is more than double the previously documented number of springs. Figure 1 shows locations of these 70 springs (marked with asterisks). It should be noted that some asterisks appear to be overlapped, because several springs are very close to each other.

After the inventory survey of detectable spring vents was completed, flow measurements were conducted using acoustic Doppler current profiler (ADCP) type meters. Instantaneous discharge measurements for the detectable spring vents were carried out under various tidal conditions (e.g., spring and neap tides) during 28–31 July, 17–20 August, 21–25 September, and 5–8 October 2009. In addition to the flow measurement for each spring vent, a multi-parameter water quality monitoring sonde was used to measure specific conductance and temperature at the same time. Water quality data are not the focus of this paper and thus not discussed in detail in the following discussion.

In order to study effects of tides on spring discharges, two multi-beam ADCPs were deployed to measure real-time cross-sectional fluxes in two channels, each conveying discharges out of a group of spring vents discharge to Kings Bay. In Figure 2, G1 and G2 denote the locations where Groups 1 and 2 of the springs were gauged, respectively. Group 1 consists of three springs (#8–#10), while Group 2 consists of eight springs (#15, #16, #18–#23). The ADCP measurements of the cross-sectional fluxes through the channel were recorded every 15 minutes and were conducted during a 25-day period between 27 July and 20 August 2009, during which both surface water level data in Kings Bay and groundwater level data at a nearby well were available. The groundwater well is called ROMP TR21-3 and located roughly 2.5 km southeast of the center of Kings Bay (Figure 1).

**Figure 2.** Locations where Group 1 (G1) and Group 2 (G2) spring flows were gauged. Identified springs are marked by white circles with numbers (for mapping purposes, springs in a close proximity are combined together sharing a single number).
