Determination of Groundwater Flow Patterns from Cave Exploration in the Woodville Karst Plain, Florida

 (Reprinted from Florida Geological Survey Special Publication No. 46)

Christopher Werner

Woodville Karst Plain Project

 

Abstract

                The Woodville Karst Plain (WKP), located in the panhandle of Northern Florida, is characterized by a layer of unconsolidated sediments 3-20 m thick, predominantly sands, with shell and clay, overlying an extensive sequence of carbonate deposits, 150-600 m thick.  The surface of this area is distinguished by the presence of numerous sinkholes, karst windows, sinking streams, and large springs.  There is over 42 km of surveyed underwater cave passages present in several large systems within the WKP.  These include Indian Spring, Sally Ward Spring, Wakulla Springs, Shepherd Spring and the Leon Sinks Cave System.  Several physical controls are observed to operate, with varying degree, in cave passage development, including lithology, stratigraphy, regional and local groundwater flow patterns, and water table elevation directly influenced by sea-level fluctuations.  These parameters are considered in the context of cave-system development presently charted in the WKP.  Regional groundwater flow extends from southern Georgia through the WKP and south to the Gulf of Mexico.  The orientations of cave passages within the WKP suggest a nearly parallel alignment with regional flow.  Current drainage patterns primarily transport groundwater south through these conduits toward the Gulf.  However, notable exceptions to this trend occur within Wakulla Spring cave and Indian Spring cave.  Geomorphic features, cave passage orientation, current branchwork drainage patterns and flow directions suggest paleoflow directions during conduit formation in the above mentioned caves were most likely in contrast to present day observations.

 

Introduction

The exploration and survey of underwater cave systems south of Tallahassee by the Woodville Karst Plain Project (WKPP) has significantly improved the ability of hydrologists to understand the complexity of groundwater flow in a multiple-porosity medium such as the carbonates of the WKP.  The improvement in understanding has been hampered in the past by the lack of qualified scientists to visit the remote and hostile environment of deep underwater caves.  There is no substitute for detailed observations in improving the quality and quantity of information needed to advance scientific understanding within this multifaceted system of fluid flow.

While the recent observations are an important glimpse into the groundwater flow regime, it should be noted that they represent a significantly limited data set.  The small size of the data set and its relation to the regional groundwater flow pattern may be restricted, but this does not mean that it is insignificant.  Reasonable assumptions may be made, such that a greater overall understanding of the evolution of this complex groundwater drainage basin may emerge.  This paper is an endeavor to add accurate and thorough scientific observations to the current body of knowledge.  By using reasonable assumptions in conjunction with detailed observations, it is expected that significant conclusions may be drawn which aid in the understanding of this unique resource.

 Regional geology and hydrology

                The WKP developed in Leon, Wakulla, and Jefferson Counties, Florida, is characterized by a thin veneer of unconsolidated and undifferentiated Pleistocene quartz sand and shell beds, overlying a thick sequence of relatively horizontal carbonates (Hendry and Sproul, 1966).  The WKP is a gently sloping topographic region of low sand dunes and exposed carbonates rising northward from the Gulf of Mexico to approximately 20 m in elevation within Leon County, terminating at the Cody Scarp.  The loosely consolidated Pleistocene sands, being very porous and permeable, allow rapid infiltration of precipitation.  Important to our study, the St. Marks and Suwannee limestones, underlying the unconsolidated

Figure 1. Plan view map of the western Woodville Karst Plain illustrating of the Leon Sinks cave system (comprising northernmost Sullivan sink and Big Dismal sink to southernmost Turner sink), Chips Hole cave, Indian Springs cave, Sally Ward Spring cave, McBride’s Slough Spring cave and Wakulla Springs cave system.

 sands, comprise hydrostratigraphic units of the Upper Floridan Aquifer System (FAS).  These limestones, being very porous, permeable and soluble, have undergone considerable dissolution from groundwater movement (Hendry and Sproul, 1966).  Consequently, the topography is karstic in nature, with numerous sinkholes, karst windows, sinking streams, and large springs (Rupert & Spencer, 1988).

                The St. Marks is a predominately fine to medium-fine grained, partially recrystallized, silty to sandy limestone that has undergone degrees of secondary dolomitization (Hendry and Sproul, 1966).  It also contains extensive shallow conduits in portions of the Leon Sinks cave system, Chips Hole cave and Indian Springs cave.  It pinches out against the Suwannee limestone in southwestern Jefferson County and reaches a maximum thickness of approximately 60-m in western Wakulla County.

                The Suwannee limestone, Oligocene in age, reaches a maximum thickness of 160 m at approximately 30-150 m below land surface within Leon and Wakulla Counties (Davis, 1996).  The thickest portion of the Suwannee is found south at the Gulf of Mexico and the thinnest is located near the Georgia border (Hendry and Sproul, 1966).  It consists of two types of permeable rock: (1) a crystalline tan, highly fossiliferous limestone and (2) a white to cream, finely crystalline limestone containing foraminifer with micritic limestone pellets (Davis, 1996).  The Suwannee limestone is the principal lithology transporting much of the groundwater of the Upper FAS within the WKP.  The majority of dissolution conduits within the WKP are primarily developed in the Suwannee limestone.

                The regional recharge area for WKP extends north of the Georgia border for over 80 km and covers portions of over five Georgia Counties (Davis, 1996).  The regional groundwater flow pattern, taken from piezometric contour maps, shows overall south trending flow lines (Davis, 1996; Fig. 32 Scott et. al., 1991).  There is an interesting feature of the piezometric contour maps which show a saddle or potentiometric low area extending well into the WKP.  This indicates a convergent flow line pattern toward the south-central region of the karst plain.  In this area, there are several first-order magnitude springs including Wakulla Springs, St. Marks Springs and Spring Creek Springs.  The regional convergence of flow is thought to originate from the fact that the WKP confining unit is absent.  Thus, flow through Leon county, being confined by the Miccosukee and Hawthorn Formations, converges in the WKP, where the lack of confining units allows groundwater transport to result in artesian flow at the surface.

                We will turn our attention to the south-central region within the WKP.  Within this region, where the majority of springs discharge and the majority of caves have been explored, there are several interesting features.  The first is the numerous sinkholes.  These depressions are readily observed on 7½ minute quadrangles, as well as easily identifiable in the field.  As an example, the Leon Sinks cave system connects over 25 sinkholes through multiple underground conduits (Fig. 1). 

The second is the dimensions of the conduits.  Many of the large conduits, herein referred to as primary conduits, have interior dimensions that typically exceed heights of 15 m and widths of 20 m.  Most are easily recognized as having a phreatic origin, where a slight few appear to have been modified by vadose entrenchment.  These primary conduits typically extend several kilometers in length and transport great volumes of groundwater.  There are several smaller dimension passages, which in most cases serve as tributaries to the larger primary conduits.  The exception is where the primary conduit ceiling has collapsed blocking the conduit.  In these cases, of which there are many, flow has been captured or redirected along a smaller tributary. 

                The third and most interesting feature is the apparent convergence of groundwater flow lines in the vicinity of Spring Creek.  Here as many as ten springs continually discharge groundwater into the Apalachee Bay of the Gulf.  Spring Creek appears to be the convergence of a branchwork (dendritic) groundwater drainage pattern (Palmer, 1991).  Its may include waters from as close as newly discovered caves a few kilometers north of Spring Creek to those of the Leon Sinks cave system, Chip’s Hole cave, and Wakulla Springs cave. 

Cave passage orientation

Surveys and plan-view maps of the conduits, obtained from the WKPP divers (Irvine, 1998), were utilized to infer orientation of many of the primary conduits.  The maps clearly exhibit a roughly north-south trendline (Fig. 1).  In fact, a majority of the primary conduits show slight deviations from the regional hydraulic gradient.  With this in mind, it is reasonable to infer that the evolution of this drainage basin has resulted from a nearly continuos southward-directed regional hydraulic gradient.  In so much as this hydraulic gradient has not been entirely steady, it is reasonable to suggest that it has remained, for the most part, in its present state for some time.  It is obvious, from glacial and interglacial records of the Quaternary, that sea level has not been constant this entire time, but has fluctuated significantly.  Given these fluctuations and the proximity of the WKP to the present sea level in the Gulf, it is reasonable to suggest that the overall drainage pattern directions have not been significantly altered during this brief geologic time scale. 

To better visualize the cave passage trendlines, rose diagrams were created for some of the major cave systems (not shown).  The compass directions from the surveys were calculated to have an error no greater than ±4.6°.  Given the large error inherent in these underwater surveys, the orientation plots provided clearly illustrate the significant trendlines found in most of the cave systems of the WKP. 

                Sinkholes constitute the most prevalent geomorphic feature of the WKP.  Sinkholes form from the collapse of large cavernous voids within the carbonates underlying the land surface.  The cavernous voids are the result of dissolution of the carbonate rock by chemically aggressive waters.  Sinkholes form as the ceilings of these large voids become unstable and collapse.  The distribution of sinkholes is not without patterns or trends.

                To better understand the spatial distribution of sinkholes, it is important to realize that sinkholes are indicative of collapsed caves and/or caverns.  These void spaces are not formed in isolation from other void spaces, but are typically formed in response to the concentrated flow of aggressive waters.  Concentrated flow routes are generally interconnected forming the precursor toward primary conduit formation.  As aggressive waters are concentrated, extensive dissolution of limestone occurs initiating connections between the void spaces.  The subsequent increase in void space and/or porosity allows the influx of larger volumes of aggressive waters, reinforcing the interconnected flow route. As time progresses, this system may evolve to become a primary flow route transporting groundwater.  During this evolution, primary conduits are formed and enlarged by the imposed hydraulic gradient and the positive feedback loop.

                As the size of the primary conduits is systematically enlarged, many parts of the ceiling become unstable and collapse forming a sinkhole.  To envision this, we will refer to the earlier example of the Leon Sinks cave system (Fig. 1).  Here, there have been over 25 collapsed cave ceilings producing sinkholes, which are all direct entrances for divers into the cave system.  The sinkholes formed in response to enlargement of the primary conduits comprising the cave system.  These sinkholes are aligned, in most cases, directly above primary conduits.  This sinkhole alignment provides invaluable evidence for locating primary conduits.

                Topographic maps, aerial photographs and USGS orthoquadrangles were employed to locate additional sinkholes south of the Leon Sinks and Wakulla cave systems.  These sinks, as well as others not detectable on published maps, were located and surveyed in the field and plotted (Fig. 2).  The constructed map indicated several possible primary conduit locations.  One of the more interesting features of the sinkhole location map (Fig. 2) was the nearly north-south trendlines of the majority of the sinkholes.  This is consistent with the sinkhole alignment of the Leon Sinks cave system.  In addition, the sinkhole trendlines are also consistent with the regional hydraulic gradient within the WKP.

 Cave passage flow patterns

                The majority of subsurface conduit flow moves from north to south, which is in agreement with the regional flow pattern.  However, there are some exceptions to this trend.  The two most notable exceptions occur in Wakulla Springs cave and Indian Springs cave. 

                In Wakulla Springs cave, the primary conduit discharging water to the spring mouth, A-tunnel, flows north for upwards of approximately 1.5 km.  The interface between northward and southward flowing water varies considerably within the cave.  It appears to be dependent on head conditions.  During periods of low precipitation and low head conditions, the interface between the divergent flowing water occurs near the junction of A-tunnel and D-tunnel, a distance of 0.65 km from the spring mouth.  Conversely, following periods of extensive precipitation and high head conditions, the interface between northward and southward flowing water occurs at a penetration distance of 2.3 km inside the cave at the junction of O-tunnel and A-tunnel.  Here at the interface, the northward flowing water travels up A-tunnel 2.3 km to the spring mouth.  Both of these conditions transport water in the opposite direction to the regional and local flow regimes.

                In Indian Springs cave, a similar situation occurs.  Approximately 0.2 km inside the spring mouth, in a northward trending passage, there is a junction between the upstream and downstream tunnel.  Here the downstream tunnel continues north for 0.55 km where it terminates in a debris cone from a collapse.  Slightly before the debris cone there is a small siphoning southwestward trending passage.  The upstream tunnel trends westward for 0.85 km and then turns and trends northward for another 0.8 km. 

 

 

Figure 2.  Map of southwestern WKP at 1/300,000.  Red dots are sinkholes, blue lines surface streams and rivers, green lines are primary cave passages and pink arrows are inferred branchwork (dendritic) drainage lines.  Refer to Fig. 1 for specific cave and sinkhole names.

                During high head conditions, a majority of water flows from the upstream tunnel and turns south exiting the cave at the spring mouth.  A small proportion of water turns north into the downstream tunnel and flows into the downstream siphon.  During extended periods of low head, the majority of water flowing from the upstream tunnel turns north and flows toward the downstream siphon.  Usually in these conditions, there is no discharge of water from the upstream tunnel into the spring mouth.  In cases of extremely low head conditions, water flowing from the upstream tunnel is entirely diverted to the northward directed downstream tunnel and flows to the downstream siphon.  In addition, water from the spring basin is siphoned back into the cave and flows north 0.75 km to the downstream siphon.  Thus, depending on head conditions, the water in the front section of the cave, from the spring mouth to the junction of the upstream and downstream tunnels, can flow either north or south.

 Paleoflow patterns

                In order to gain further insight into the conduit system drainage patterns, it is necessary to envision the present drainage system during its Pleistocene evolution.  In addition to existing passage enlargement and increased connections (permeability) of large secondary-porosity voids, sea level height was fluctuating significantly.  There are indication that sea level was at the Cody Scarp 100 Ka BP and approximately 100 m lower 18 Ka BP (Rupert & Spencer, 1988; Chappell & Shackleton, 1986).  These extreme variations may have greatly affected fresh and salt water mixing zones, as well as alter drainage basin size and extent.  In as much as the conduit and geomorphic features are indicative of N-S paleoflow, it is important to realize that there were several mechanisms working to varying degrees and during particular times, in order to have produced the present drainage system.

                Following careful study of the geomorphic features and the current extent of explored passages a key observation becomes relevant.  There are four large sinks north of Wakulla Springs located parallel to FL highway 61.  These sinks are in linear alignment with the Leon Sinks cave system and Wakulla Springs Cave system.  It appears, after land and in-water dive surveys, that the surface area and volume of Cherokee sink, Wakulla Springs basin, and the large sink directly north of Wakulla on Rt. 61 are all of approximately the same magnitude.  Noting how these sinks trend linearly south, there is sufficient evidence to make a hypothesis as to their origin and evolution.

                First, it is proposed that Wakulla Springs A-tunnel passage formed from southward directed flow along the hydraulic gradient.  This indicates that paleoflow was at one time directed south from the spring mouth toward Cherokee sink.  This is reinforced by the passage dimensions of A-tunnel and O-tunnel, which have nearly the same overall box-canyon/ phreatic tube shape and size. 

                Second, it is proposed that there was a large conduit, the size of A-tunnel, connecting the large sink on Rt. 61 north of Wakulla and Wakulla Springs A-tunnel (Fig. 1).  This passage appears to have been a primary conduit of the paleodrainage basin and originated somewhere near the termination of Munson slough or Eight Mile Pond (Fig. 2).  This is in alignment with the four large sinks which parallel Rt. 61. 

                Last, it is proposed that there was a large collapse of this primary paleoconduit at the present Wakulla Spring location.  This collapse caused a complete blockage of the southward directed flow.  The collapse feature evolved into the large spring now present.

 Overflow valve

                It is necessary to place Wakulla Springs cave system in the context of the regional drainage basin patterns to understand its flow evolution to its present state.  The WKPP has believed for some time that Wakulla Springs cave, the Leon Sinks cave system, as well as many of the other caves south of these, were at one time connected or are still physically connected at present (Irvine, 1998).  This assumption has been a major driving force in the continued exploration of this area.  When the southwestern extent of the WKP is displayed in the context of the cave systems, the regional hydraulic gradient, and sinkhole alignment, it becomes clear that a drainage basin trend emerges. 

This trend, coupled with the presence of the sinking surface streams, indicates a very large branchwork (dendritic) drainage pattern (Palmer, 1991).  Each cave system, after proposed drainage lines connecting the systems along sinkhole trendlines are drawn (Fig. 2), appear to be part of the larger branchwork drainage system.  The terminal mouth of this system also appears to be Spring Creek Springs, at the edge of the Gulf. 

                With this in mind, following large precipitation events and/or a large head gradient within the basin, flow backs up at Spring Creek and other springs near the coast.  This causes an overall decrease in the ability of the conduit system to effectively transport water to the Gulf.  Increased flow at several springs, such as Shepherd Spring and Wakulla Springs, becomes the direct result of this inefficiency.  A very large head gradient appears to be responsible for the extreme fluctuations in discharge seen at Wakulla Springs.

                This intricate and complex feedback loop is typical of karst groundwater / surface water flow regimes (White, 1988).  There are several examples of these conditions in the literature, of which Mammoth Cave, Kentucky is the most notable (White & White, 1989).  The observations of Indian Springs made above are also consistent with the proposed model above. 

 Conclusions

                It is evident that the alignment of regional groundwater flow lines with primary conduits and sinkholes are not a coincidence.  They are intimately related surface and subsurface features.  Much of the subsurface flow through the cave passages are tied to surface water flow.  This indicates that the entire drainage basin, both surface and subsurface, must be considered in any further study.  It also becomes clear that the evolution of the drainage basin has been intricate and complex through its history.  The seemingly anomalous cave passage flow patterns may be explained by simple groundwater / surface water interactions during varying hydraulic head conditions.  These observations, models and conclusions should provide a foundation for continued exploration and research into this unique drainage basin.

 References

 Chappell, J. & Shackleton, N. J., 1986, Oxygen isotopes and sea level, Nature, V. 324, 137-140.

Davis, H., 1996, Hydrogeologic Investigation and Simulation of Ground-Water Flow in the Upper Floridan Aquifer of North-Central Florida and Delineation of Contributing Areas for Selected City of Tallahassee, Florida, Water Supply Wells: USGS Water-Resources Investigation Report 95-4296.

Hendry, C. W., and Sproul, C. R., 1966, Geology and groundwater resources of Leon County, Florida:  Florida Geologic Survey Bulletin 47.

Irvine, George, 1998, Woodville Karst Plain Project, www.wkpp.org, personal communication.

Palmer, A. N., 1991, Origin and morphology of limestone caves, GSA Bulletin, V. 103, p. 1-4.

Rupert, Frank, and Spencer, Steve, 1988, Geology of Wakulla County, Florida:  Florida Geologic Survey Bulletin 60.

Scott, Thomas M., Lloyd, Jacqueline, M, and Maddox, Gary, Florida’s Groundwater Quality Monitoring Program: Hydrogeological Framework, Florida Geological Survey Special Publication No. 32, 1991.

White, W. B., 1988, Geomorphology and hydrology of karst terrains, Oxford University Press, New York.

White, W. B. and White, E. L., 1974, Base level control of underground drainage in the Potomac River Basin, Proceeding of the 4^th Conference on Karst Geology and Hydrology, H. W. Rauch and E. Werner, Eds., West Virginia Geological Survey, pp. 41-53.

Wisenbaker, Mike, 1998, Woodville Karst Plain Project, www.wkpp.org, Figure 1 map, personal communication