FLOODING OF THE SINKING CREEK KARST AREA IN JESSAMINE … · UNIVERSITY OF KENTUCKY, LEXINGTON...

ISSN 0075-5591

KENTUCKY GEOLOGICAL SURVEYUNIVERSITY OF KENTUCKY, LEXINGTONDonald C. Haney, State Geologist and Director

FLOODING OF THE SINKING CREEKKARST AREA IN JESSAMINE AND

WOODFORD COUNTIES, KENTUCKYJames C. Currens and C. Douglas R. Graham

Report of Investigations 7Series XI, 1993

ISSN 007-5-5591

KENTUCKY GEOLOGICAL SURVEYUNIVERSITY OF KENTUCKY, LEXINGTONDonald C. Haney, State Geologist and Director



Report of Investigations 7Series XI, 1993

UNIVERSITY OF KENTUCKYCharles T. Wethington, Jr., PresidentLinda J. Magid, Vice President for Research and Graduate StudiesJack Supplee, Director, Fiscal Affairs and Sponsored Project AdministrationKENTUCKY GEOLOGICAL SURVEY ADVISORYBOARDSteve Cawood, Chairman, PinevilleLarry R. Finley, HendersonHugh B. Gabbard, RichmondKenneth Gibson, MadisonvilleWallace W. Hagan, LexingtonPhil M. Miles, LexingtonW. A. Mossbarger, LexingtonHenry A. Spalding, HazardJacqueline Swigart, LouisvilleRalph N. Thomas, OwensboroGeorge H. Warren, Jr., OwensboroDavid A. Zegeer, Lexington

KENTUCKY GEOLOGICAL SURVEYDonald C. Haney, State Geologist and DirectorJohn D. Kiefer, Assistant State Geologist for Adminis-trationJames C. Cobb, Assistant State Geologist for Research

ADMINISTRATIVE DIVISIONPersonnel and Finance Section:James L. Hamilton, Administrative Staff Officer IIRoger S. Banks, Account Clerk VClerical Section:Jody L. Fox, Staff Assistant VI IJoyce Belcher, Staff Assistant VIShirley D. Dawson, Staff Assistant VEugenia E. Kelley, Staff Assistant VJuanita G. Smith, Staff Assistant V, Henderson Office

Publications Section:Donald W. Hutcheson, HeadMargaret Luther Smath, Geologic Editor IIITerry D. Hounshell, Chief Cartographic IllustratorRichard A. Smath, Geologist III, ESIC CoordinatorMichael L. Murphy, Principal Drafting TechnicianWilliam A. Briscoe, III, Publication Sales SupervisorKenneth G. Otis, Stores WorkerGEOLOGICAL DIVISIONCoal and Minerals Section:Donald R. Chesnut, Jr., Acting Head

Garland R. Dever, Jr., Geologist VIICortland F. Eble, Geologist VDavid A. Williams, Geologist V, Henderson OfficeWarren H. Anderson, Geologist IVGerald A. Weisenfluh, Geologist IVStephen F. Greb, Geologist IIIRobert Andrews, Geologist IYue Jin Liu, Research Assistant

Petroleum and Stratigraphy Section:James A. Drahovzal, HeadTerence Hamilton-Smith, Geologist VPatrick J. Gooding, Geologist IVDavid C. Harris, Geologist IVBrandon C. Nuttall, Geologist IVMatthew Humphreys, Geologist IIX. Mara Chen, Post-Doctoral ScholarJames B. Harris, Post-Doctoral ScholarRobert R. Daniel, Laboratory Technician BAnna E. Watson, Staff Assistant IVFrances A. Benson, Staff Assistant IVLuanne Davis, Staff Assistant IVTheola L. Evans, Staff Assistant IVKimberly B. Stroth, Staff Assistant IV

Water Resources Section:James S. Dinger, HeadJames A. Kipp, Geologist VDaniel 1. Carey, Hydrologist IVJames C. Currens, Geologist IVDavid R. Wunsch, Geologist IVAlex W. Fogle, Hydrologist IIIPhilip G. Conrad, Geologist II0. Barton Davidson, Geologist IIDwayne M. Keagy, Geologist IIShelley A. Minns, Geologist IIEd Fortner, Jr., Geological TechnicianC. Douglas R. Graham, Geological Technician

Computer and Laboratory Services Section:Steven J. Cordiviola, HeadRichard E. Sergeant, Geologist VJoseph B. Dixon, Systems ProgrammerHenry E. Francis, Associate ScientistXenia P. Culbertson, Senior Research AnalystEric E. Lovins, Research AnalystSteven R. Mock, Research AnalystMark F. Thompson, Research AnalystTammie J. Heazlit, Senior Laboratory Technician



ABSTRACTTashamingo Subdivision in Sinking Creek Karst Valley, a tributary of the Garretts Spring Drainage

Basin in Jessamine and Woodford Counties, Kentucky, was flooded in February 1989. To determinethe cause of flooding, the boundary of the ground-water basin was mapped, discharge data weremeasured to determine intake capacity of swallow holes, and hydrologic modeling of the basin wasconducted. Swallow-hole capacity was determined to be limited by the hydraulic parameters of theconduit, rather than by obstruction by trash. Flooding from a precipitation event is more likely, andwill be higher, when antecedent soil moisture conditions in the watershed are near saturation.Hydrologic modeling shows that suburban development of 20 percent of the southeastern basin willcause an increase in flood stage at Tashamingo Subdivision.

INTRODUCTIONProject History and Objectives

From mid-February to early March of 1989 theTashamingo Subdivision and the Delaneys Ferry Roadarea of Jessamine County flooded when the SinkingCreek Karst Valley was unable to accommodate aprolonged and intense storm; 8.4 inches (21 cm) of rainfell from February 13to 16. Several homes wereisolated by blocked roads, and two homes wereflooded; one was damaged extensively (Figs. 1 and 2).In response to the concerns of area residents, theKentucky Geological Survey (KGS) initiated this study todetermine if future flooding could be prevented.

The first goal of the study is to predict what type ofstorm will cause flooding that threatens property, giventhe current degree of residential development. Thesecond goal is to determine, for any given storm, whatthe effect of continued development in the basin will be.These data will be essential for project design shouldan engineering solution be chosen to mitigate futureflooding. This study represents the first time a majorkarst ground-water basin in the Inner Blue Grass hasbeen monitored for an extended period of time.

Location and Physiographic SettingGarretts Spring Karst Drainage Basin lies in

Jessamine and Woodford Counties in the Inner BlueGrass region of central Kentucky (Fig. 3). The Inner

Blue Grass is a gently rolling upland with a subduedkarst topography formed on Ordovician limestones. Theupland is roughly bounded on the south and west by theentrenched Kentucky River flowing in a gorge as muchas 400 feet (130 m) deep. The gradients of streamsflowing off the upland steepen abruptly as theyapproach the gorge. Except along the gorge of theKentucky River, local relief is generally less than 150feet (50 m). Karst windows and sinkholes seldom havemore than 100 feet (30 m) of relief, and many sinkholesare too shallow to show up on topographic maps.Springs and caves are common, but the caves areusually very wet and most cannot be explored morethan a few hundred feet.

Garretts Spring is the headwaters of the northernbranch of Clear Creek, which flows approximately 14miles (22.5 km) to the Kentucky River. The drainagebasin covers approximately 4,766 acres (1,929 hectares[ha]). The basin is composed of two branches, theconfluence of which is underground near the resurgenceat Garretts Spring. Three major karst features andhundreds of smaller ones are within the drainage basin.Chenault Karst Window lies in the northwestern subbasin.Water emerges from a spring at the northern end of thekarst window, and flows 1,500 feet (457 m) to thesouthern end, where it sinks. The water then flows 1,900feet (580 m) to Garretts Spring. Owens Karst Window liesin the southeastern branch between Sinking Creek andGarretts Spring. Flow from Sinking Creek rises in OwensKarst Window at several springs along the eastern

upstream wall, and sinks in a series of swallow holes onthe western wall to flow to Garretts Spring, 3,800 feet(1,158 m) to the west.

Sinking Creek Karst Valley forms the headwaters ofthe southeastern branch. The topographically closedportion of the basin covers 197 acres (78 ha). SinkingCreek originates as surface runnels and small springsemerging just above the grade of the creek. The streamfollows a smooth, gradual gradient to the karst valley,without measurable flow loss, until it approaches thefootwall area. At the footwall, Sinking Creek divergesinto three distributaries that convey flow to the threeprincipal groups of swallow holes.

GENERAL GEOLOGY ANDHYDROGEOLOGY

GeologyThe Garretts Spring Basin is underlain by the Lexington

Limestone, which consists of thinly interbeddedcarbonates, argillaceous carbonates, and shales of

Middle Ordovician age (Cressman, 1965). Members ofthe Lexington Limestone exposed within the drainagebasin of Garretts Spring are, from bottom to top, theGrier Limestone Member, the Tanglewood LimestoneMember, the Brannon Member, and the Devils HollowMember (Fig. 4). The Grier Limestone is irregularly thinbedded, with occasional shale or silt interbeds. TheMacedonia Bed, a mappable argillaceous unit amaximum of 9 feet (2.7 m) thick, occurs within theGrier, 50 feet (15 m) below the top of the Grier.Overlying the Grier is the Tanglewood Limestone, athinly crossbedded, but relatively pure carbonate. TheTanglewood intertongues with the Grier and severalother lithologies within the Inner Blue Grass. TheBrannon Member, a thin-bedded, argillaceouscarbonate, is from 8 to 30 feet (2.4 to 9.1 m) thick andintertongues with the Tanglewood only 10 to 25 feet (3to 7.6 m) above the top of the Grier. The Devils HollowMember is a pure, highly fossiliferous limestone, 10 to15 feet (3 to 4.6 m) thick, which also intertongues withthe Tanglewood approximately 30 feet (9 m) above theBrannon.

2 Flooding of the Sinking Creek Karst Area in Jessamine and Woodford Counties, Kentucky

General Geology and

The study area is near the crest of the CincinnatiArch, and although the strata are relatively flat lying,they dip gently to the northwest at 15 feet per mile (2.8m/km). No faulting is mapped within the drainage basin,although prominent joints have been observed in thefield. A barite vein is mapped just north of GarrettsSpring. The vein strikes due north, and projects alongthe axis of the Chenault Karst Window. It mayrepresent an unmapped fault.

HydrogeologyThe general hydrogeology of the Inner Blue Grass

has been described by numerous authors (Hamilton,1950; Palmquist and Hall, 1960, 1961). More recentlyThrailkill and others (1982) have conducted moredetailed studies. Thrailkill and others defined twoprincipal types of karst aquifers in the region: interbasinsand ground-water basins. Interbasins occur betweenground-water basins where flow takes place in shallowconduits and channels that are eroded into bedrock, butroofed with soil. Some of these conduits are perched onargillaceous units, at least for short reaches. Thisshallow flow quickly returns to the surface, but thensinks again toward the interior of the basin. The interiorof the basin is the ground-water basin where conduitsmay breach and flow beneath the argillaceous units.This deep flow occurs in well-developed caves orconduits,

I

3

and resurges at a major spring near local base level.Both interbasin watershed boundaries andground-water basin boundaries are known to crosssurface watershed boundaries, although mostcommonly the interbasin boundary roughly coincideswith the surface watershed.

METHODOLOGYThree separate tasks were essential for

understanding flooding in the Garretts Spring Basin: first,mapping the watershed boundary; second, measuringthe intake capacity of the Sinking Creek swallow holesunder various stages of flooding; and third, collectinghydrologic data, which are input into a digital hydrologicmodel to estimate the basin response caused by variousstorm events and changes in land use.

Dye detectors were mounted on concrete anchors(called "gumdrops" by Quinlan, 1987) and consisted ofactivated carbon charcoal in fiberglass screen-wirepackets and surgical cotton later replaced with bleachedcotton broadcloth (Testfabrics, cat. no. 419). The use ofmanufacture and trademark names does not constitutean endorsement of the product by KGS or the Universityof Kentucky; they are included for reference only.

Ground-Water Dye TracingThe boundary for the Garretts Spring drainage basin

could not be unambiguously drawn because of the karsttopography. Sixteen ground-water dye traces wereconducted to determine where the divide lay (Fig. 5).Results for 14 other traces performed by Larry Spangler(personal communication, 1989) were also obtained andare shown on Figure 5. Standardized techniques,discussed in detailed by numerous authors (Aley andFletcher, 1976; Thrailkill and others, 1983; Jones, 1984;Davis and others, 1985; Quinlan, 1987; Mull and others,1988) were used.

Traces conducted by the authors utilized fourfluorescent dyes: Fluorescein (C.I. Acid Yellow 73),Rhodamine WT (Acid Red 388), Diphenyl BrilliantFlavine (C. I. Direct Yellow 96), and Tinopal CBS-X(optical brightening agent 351) (Smart, 1984).Straight-line distances traced ranged from 1,900 feet(580 m) to 6,300 feet (1,920 m). The Fluorescein andTinopal were introduced into swallow holes as drypowder, while the Direct Yellow was premixed withsufficient distilled water to dissolve the powdered dye.Dry quantities of dye used ranged from 1.1 pounds (500gm) of Tinopal to as little as 0.1 pound (50 gm) ofFluorescein. The Direct Yellow was 20 percent activeingredient by weight, the Rhodamine a 20 percentsolution, and the others 100 percent active ingredients.

4 Flooding of the Sinking Creek Karst Area in Jessamine and Woodford Counties,

Surgical cotton was occasionally lost because ofmechanical erosion and attack by crayfish. Wire framessupporting a cotton-fabric swatch were tried (Thrailkilland others, 1983). These proved much more durable;however, manufacturing these "bugs" was timeconsuming. Experimental traces were run using fabricboth on a frame and as a ribbon, 1 1/2 x 18 inches (3cm x 50 cm), tied into a "bow tie." Visual inspection ofthe positive fabric bugs showed equally good dyeadsorption for both methods, but the bow ties provedmuch more durable than surgical cotton. During ahiatus in the tracing program, bow ties were left in thefield for 3 months during the winter. Although severelydegraded, one-fourth to one-half of the fabric remained;

surgical cotton would have disappeared in a week or 10days.

Samples for quantitative traces were collected withISCO model 2900 programmable automatic samplers.Intake lines were purged before each sample andemptied after sampling to minimize mixing with residuefrom the previous sample. The ISCO model 2900 is onlyavailable with polyethylene bottles, which adsorbRhodamine dye readily. Therefore, glass culture tubeswere used that closely fit inside the mouth of the plasticISCO bottles and were of correct length to allow thesampler distributor to clear the mouth of each tube. The

Ground-Water Dye Tracing


small size of the culture tubes required careful calibration ofthe sample volume in the field. Dye concentrations weredetermined on a Turner Designs model 10 fluorometer.Discharge during the traces was determined directly withPrice meters, rather than by using the stage and rating curve.

Stream Gaging and Discharge DataNine gaging stations and four discharge observation

points, involving 17 channel cross sections, were establishedfor monitoring stream and spring flow. Three sites wereinstrumented with continuous-stage recorders:

Chenault Karst Window (CHEN recorder), Owens KarstWindow (OWEN recorder), and Sinking Creek at CherrywoodLane in the Tashamingo Subdivision (TASH recorder). TelogInstruments WLS-2109 single-channel digital recorders werecoupled with Druck PDCR 830 pressure transducers.Observations of the water depth, calibrated in feet, weremade every second, and the mean was recorded every 15minutes. Two of the Druck transducers have a 0 to 10 psirange (23.11 ft.; 7.04 m of water), and the third has a rangeof 0 to 20 psi (46.22 ft.; 14.08 m of water), with an accuracyof +/- 0.3 percent. The data were downloaded to a laptopcomputer, either directly from the recorder or via a TelogInstruments

Stream Gaging and Discharge

data transfer unit, then uploaded to a data base on theVAX 8550 at KGS. The pressure transducers weremounted in stilling wells constructed of PVC, whichwere imbedded in the channel bank. Elevations wereleveled to a datum scribed on each stilling well toobtain elevation head.

Discharges were measured using the partial-sectionsmethod (Buchanan and Somers, 1976) and Price flowmeters (Teledyne-Gurley models 622 and 625).Because the water-level recorders were installed inkarst valleys and windows, ponding influenced thehydrograph during high flow. Determination of theinception of ponding was based on topography and theelevation of the stage recorder relative to the sinkingpoint of the stream, changes in the hydrograph curve,and other onsite observations. Ponding at the swallowholes had two effects on discharge measurement. First,even at sites where flow was still confined to a distinctchannel, measurements were logistically more difficultto obtain because of deep water (6 to 15 ft.; 2 to 5 m).Second, accuracy of the Price flow meters was reducedbecause of sluggish velocities of the ponded flow.

Perhaps the most important water-level recorderwas the one in the Tashamingo Subdivision atCherrywood Lane because it provided the onlycontinuous record of runoff from the headwaters of thesoutheastern branch of the basin. Sinking Creek flowsunder Cherrywood Lane through three corrugated steelculverts penetrating an earth-fill causeway. Thecauseway creates an effective dam, and virtually allflow upstream of Cherrywood Lane is forced throughthe culverts until the roadway is overtopped at a stageof 11.2 feet (3.4 m). The station misses minordischarge from small springs and overland flowbetween the recorder and the swallow holes.

The Chenautt Karst Window and Owens KarstWindow sites did not have the benefit of a structure tocontrol discharge. The Chenault Karst Window stillingwell was a few meters upstream of the outcrop of theMacedonia Bed, which acted as a control section duringlow flow. The control shifted to the channel duringmoderate flow, and shifted again to pressure flow whenthe pool from the flooded swallow holes reached theelevation of the stilling well. At Owens Karst Windowthe stilling well was installed in the channel from themiddle spring. At Owens Karst Window four majoropenings discharge into three channels, and dischargesfrom the three channels were summed and applied tothe rating curve. control was by the channel from lowstage until backflooded. At stages over 12 feet (3 m)the banks of the channels from the springs wereovertopped and channel control was lost.

7

Discharge data at Garretts; Spring were obtainedfrom a flume operated by Dr. Gary Felton of theUniversity of Kentucky Department of AgriculturalEngineering. The spring is naturally impounded, but hasbeen further dammed for an irrigation supply.Conditions at the site imposed limits on the dimensionsof the flume, which had a maximum capacity of 30cubic feet per second (cfs) (0.85 cubic meter persecond [cms]). Unfortunately, the maximum dischargeat Garretts; Spring is known to exceed 60 cfs (1.7 cms)(Fig. 6). Furthermore, the dam supporting the flume hasleaked at various times since the flume's installation.Also, a secondary spring downstream is known to havereceived a small percentage of flow from the basin.When possible, discharges from Garretts Spring werealso measured with flow meters.

Determination ofSwallow-Hole Capacity

The suspected principal control on the flooding ofSinking Creek Karst Valley was the intake capacity of theswallow-hole zone at the western end of the valley.Access to the footwall area of Sinking Creek Karst Valleywas denied during the first year of the project, whichprecluded direct observation of both stage and inflow atthe swallow holes. Measurement of the intake capacity byindirect methods was tried until access was obtained inJanuary 1991. The methods used to measure intakeincluded determining head loss coupled with estimatedhydraulic characteristics of the conduits, budgetingmeasured outflows at Garretts Spring and Chenault KarstWindow, budgeting estimated storage and inflow,measuring inflow at critical points in the stage hydrograph,and directly observing inflow at the swallow holes andoutflow at Owens Karst Window. The two

, A

8 Flooding of the Sinking Creek Karst Area in Jessamine and Woodford Counties,

budgeting techniques proved too imprecise to develop arating curve, although the values set limits on realisticswallow-hole inflow rates. The most useful data wereeventually obtained by observing discharge at criticalstages and directly measuring inflow.

Rating-Curve DevelopmentRating curves were constructed using techniques

developed by the U.S. Geological Survey (Kennedy,1984). A gage-zero flow for each station was selectedby choosing the offset resulting in the largest Pearsoncorrelation coefficient for the regression of thedischarge on the stage. Data were insufficient to definehysteresis from overbank storage.

The stage-discharge relationships for all three stage-recorder sites and the Sinking Creek swallow holes arecomplex. The rating curve for the Sinking Creek andChenault Karst Window swallow holes consists of free-fall and confined-flow limbs (Figs. 7 and 8). The ratingcurve for Tashamingo is in two parts, free-fall andponded (Fig. 9). The rating curve for Owens KarstWindow is even more problematic: the flow datasuggest a

multiple-step curve, but data are insufficient to clearlydefine the curve. A single, straight line was regressed tothe available Owens Karst Window data (Fig. 10).Discharge data for the free-fall segment of the curve forSinking Creek swallow holes were measured at a crosssection upstream from the divergence of the firstdistributary. Discharge data for the pressure-flowsegment were derived from discharges measured atOwens Karst Window and were coupled with stageobservations at either the swallow holes orTashamingo. Only 85 percent of the flow throughOwens Karst Window was accounted for at the threedischarge measuring stations during moderate to highflow. The Owens Karst Window data were adjusted forthe unaccounted-for flow because the swallow holerating curve was to be used with the HEC-1 model.

Discharge Hydrograph Computationand Flow Budget Modeling

A computer program was written to calculatedischarge from stage data measured at Tashamingo,Owens Karst Window, and Chenault Karst Window us-

Discharge Hydrograph Computation and Flow Budget

ing algorithms developed from the rating curves forthese sites. The program was written in DigitalEquipment Corporation Datatrieve language to accessthe stage data stored in a data base and the dischargehydrographs for these sites were computed with thissoftware. The flow into Owens and Chenault KarstWindows was then summed to give an estimatedminimum discharge at Garrefts Spring.

Precipitation DataPrecipitation data were obtained from the National

Oceanic and Atmospheric Administration (NOAA)weather station at nearby Blue Grass Field, 4.6 miles(7.3 km) northwest of Garretts Spring, and a volunteerNOAA station at Keene, 3.8 miles (6.1 km) southeast ofthe spring. Additional precipitation data were obtainedfrom a weighing-bucket recording rain gage installed byFelton at Garretts Spring. These data were used whenavailable because the gage is within the study area.

Land-Use and SCSCurve-Number Estimation

After the basin boundary was mapped, land use wasdetermined by Felton (personal communication, 1992).These data were compiled into Soil ConservationService (SCS) cover type and hydrologic conditioncategories based on the percentage of each hydrologicsoil group in each sub-basin (McDonald and others,1983). A weighted-average runoff curve number wasthen determined for each sub-basin as defined by theU.S. Army Corps of Engineers (SCS, 1986). Thesevalues were used in modeling runoff in the basin.

Runoff ModelingThe Louisville office of the U.S. Army Corps of

Engineers was contracted by the Federal EmergencyManagement Agency to determine the 1 00-year floodplain for the Sinking Creek Basin (U.S. Army Corps ofEngineers, 1990). Their study was completed with theuse of the HEC-1 Flood Hydrograph Package(Computer Program 723-X6-L201 0) developed by theCorps to predict the impact of storm runoff . The Corpsmade the program and data files available to KGS. TheCorps data were coupled with the land-use andswallow-hole-capacity data gathered by this research tocompute the model flood hydrographs for Sinking Creek.

9

RESULTSGround-Water Dye Tracing

The qualitative ground-water traces resulted in severalsignificant findings. The Nicholasville 7.5-minutetopographic quadrangle map indicated that theheadwaters of the southeastern branch of the basinextended several hundred meters east of U.S. Highway68 (dashed area on Figure 5). However, tracing indicatedthat surface-water flow in this area had been divertedunderground to the south, out of the basin. All tracesconducted in this area were made during low flow, andwhether high-flow discharge into Sinking Creek occurs isunknown. Also, the watershed of the northwestembranch was mapped, and two sub-basins weredelineated (Spangler, 1989). The three main swallowholes of Sinking Creek were independently traced toOwens Karst Window; all principal springs at OwensKarst Window were found to receive flow from each ofthe three swallow holes. Finally, since it was thoughtpossible that the flow from Chenault and Owens KarstWindows did not join underground and that the GarrettsSpring rise pool was a double resurgence, dye detectorswere placed in the two obvious boils in the rise pool andindependent traces were run from both Chenault andOwens Karst Windows. Both detectors were positive forboth traces, but dye mixing in the rise pool may haveaffected both detectors. A physical examination of thespring revealed that the bottom of the rise pool wascompletely covered with talus. However, Garretts Springis a distributary resurgence. Positive traces to GarrettsSpring were also detected at Hoffmans Spring, a smallspring on the northern bank of Clear Creek,approximately 600 feet (200 m) downstream of GarrettsSpring. Discharge from Hoffmans Spring is small, evenduring high flow from Garretts Spring.

Quantitative dye traces were used to measure meanflow velocity and determine effective conduitcross-sectional area (Table 1). Two traces wereconducted from Sinking Creek to Owens KarstWindow. The centroid of the dye plume and dyerecovery (85.1 percent) were calculated for the firsttrace. Only velocity was calculated for the second traceto Owens Karst Window. Because of a higher thanexpected velocity, the leading edge of the dyebreakthrough curve was missed for a trace from OwensKarst Window to Garretts Spring. Its centroid wasestimated.

The straight-line distance that dye traces traveledwas measured from topographic maps. Previousresearchers have used a meander distance tostraightline distance ratio of 1.5:1 (Mull and others,1988; Thrailkill and others, 1990) for studies inKentucky. For this study the meander ratio wasdetermined by measuring

10 Flooding of the Sinking Creek Karst Area in Jessamine and Woodford Counties,K t k

the meander and straight-line distances of passages within the basin, and averaging the ratio. The ratio waswith flowing streams from maps of five Inner Blue Grass 1.11:1, and is reasonable in light of the linear nature ofcaves (O'Dell and O'Dell, 1992), including one cave many cave passages in the Inner Blue Grass.

Two quantitative traces were run by Baumgartner(1991) from Chenault Karst Window to Garretts Spring.Both traces were run under low-flow conditions andyielded a mean conduit cross-sectional area at GarrettsSpring, for the combined flow from both sub-basins, of72.4 feet2 (6.73 M2). This area is virtually the same asthe cross section calculated for the trace from OwensKarst Window to Garretts Spring, minus the ChenaultKarst Window flow. Baumgartner used the dischargedata from the flume, while discharge for the Owens KarstWindow trace was measured with Price meters.Baumgartner's dye recovery averaged 59.7 percent.Substantial flow was observed leaking through the damat the flume during the period Baumgartner ran histraces, which explains the poor dye recovery. Therefore,the discharge he used was likely to be too small,resulting in a smaller cross-sectional area. The similaritybetween the cross-sectional areas of Baumgartner andthis study may be coincidental. Alternatively, the missedflow may have roughly equaled the contribution fromChenault Karst Window. Data from the March 19 tracesuggest that the Owens Karst Window to Garretts Springconduit is twice the cross-sectional area of the ChenaultKarst Window to Garretts Spring conduit. Both

tributaries were likely to have been under pressure flowduring the trace.

Estimate of Peak DischargeDuring 1989 Flood

Calculations of the maximum discharge for theFebruary 1989 flood were made because it provided thehighest stage recorded to date for Sinking Creek KarstValley (Table 2). The cross-sectional area of the January7,1992, trace was used to represent the Sinking Creek-Owens Karst Window conduit under pressure-flowconditions. Data from the March 19, 1992, trace wereused for the Owens Karst Window-Garretts Springconduit area. The Darcy-Weisbach equation was chosento approximate the flow regime because it representsenergy loss from turbulent flow in pipes. The choice ofcross-sectional shape of the conduit is significant forestimating discharge because of the influence of conduitsurface area on head loss. However, because of thetributary-distributary and overflow-conduit planpostulated for the conduit system, any cross-sectionshape would be valid only for a short reach of conduitand is therefore generally arbitrary. A circular crosssection


was chosen. The Darcy-Weisbach friction factor wascalculated for the conduits between Sinking Creek andOwens Karst Window and from Owens Karst Windowto Garretts Spring using velocity and discharge datafrom the quantitative dye traces. The value used inthese calculations was the apparent friction factor, andrepresented total head loss in these conduits. Thevalues calculated fall well within ranges reported byother researchers for karst conduits (Ford and Williams,1989). The elevations of flotsam marks in SinkingCreek, Owens Karst Window, and Garretts Spring wererecorded soon after the 1989 flood.

These calculated discharges compared favorably withdischarges projected from lesser events. The flow fromOwens Karst Window to Garretts Spring includes anunknown contribution from several additional inputsdownstream of Owens Karst Window. Inspection of flowbudget data from three flood events suggests 25 percentof the flow from the southeastern branch may becontributed by this unaccounted-for area. Also, flow fromsprings in Owens Karst Window that is measurableduring low and moderate flow becomes inaccessibleduring high flow. Furthermore, the quantitative dye

trace data suggest only 85 percent of the flow isaccounted for during high flow. Seventy-five percent of70.6 ft.3/sec. is 53 ft. 3/sec. (1.5 m3/sec.) (Table 2),which compares favorably with the calculated flow intoOwens Karst Window from Sinking Creek of 47.3ft3/sec. (1.3 m3/sec.); if the unaccounted-for flow intoOwens Karst Window is considered, the comparison iseven better

Stage HydrographsStage data have contributed directly to understanding

the hydrology of the basin. A hydrograph for the ChenaultKarst Window, Tashamingo, and Owens Karst Windowrecorders for the largest flood event during the project ispresented in Figure 11. An important consideration ininterpreting the stage data is the position of the recordersrelative to the swallow holes. Both the Tashamingo andChenault Karst Window recorders are many metersupstream of the swallow holes, where backwater effectsare minimized. Under ideal circumstances two recordersshould be used, one upstream to record inflow and oneat the swallow holes to record head. However, only staffgages were installed at the swallow

Stage Hydrographs

holes, and the data obtained from them was unevenlydistributed through time.

The hydrograph from the Tashamingo recorder wastypical of unconfined flow until ponding of the swallowholes rose to its intake. After ponding reached therecorder, the hydrograph flattened out, and slowlydropped as storage was removed from the karst valley.The hydrograph closely paralleled stage data for OwensKarst Window when both features were flooded (Fig.11).

The recorder at Owens Karst Window was closer tothe swallow holes than the other two recorders and wasinfluenced by backwater effects earlier in a flood event.The stage recording at Owens Karst Window showed aconsistent, rapid rise and fall in stage from 2 to 4 feet (0.6to 1.2 m). Discharge data have been very difficult toobtain for this limb of the hydrograph because of its shortduration. During a series of observations on March 18,1992, to record this change, discharge increased from 5.7cfs (0.16 cms) to 10.2 cfs (0.29 cms), while stage rose1.4 feet (0.43 m) in 2.25 hours. The rapid increase instage and discharge at Owens Karst Window was causedby the onset of flow from its northern

13

springs. The northern springs at Owens Karst Windowbegin flowing at stage 1.6 feet (0.49 m) at the SinkingCreek swallow holes and discharged vigorously whenthe swallow holes were completely inundated. Thehigher northern spring is the outlet of a higher conduitnow acting as an overflow route.

The stage hydrograph for Chenault Karst Window isdistinctly different from Owens Karst Window orTashamingo, showing both a rapid rise and fall in stage(Fig. 11). While the recorder is nearly 900 feet (300 m)upstream of the swallow holes, the available staff gagereadings at the swallow holes suggest the fall in stagecontinues until free-fall flow is restored. The swallowholes at Chenault Karst Window are roughly 10 feet (3m) lower in elevation than the swallow holes at OwensKarst Window, but the conduit from Chenault KarstWindow to Garretts Spring is substantially shorter thanthe conduit from Owens Karst Window to GarrettsSpring. Also, the gradient from the Chenault KarstWindow swallow holes to Garretts Spring is slightlysteeper, 0.007 versus 0.006, than from the OwensKarst Window swallow holes, suggesting that theChenault Karst Window to Garretts Spring conduit canaccommodate greater discharge before the inception ofpressure flow. The rapid drop in stage at ChenaultKarst Window implies that it has an efficient flow routeto Garretts Spring.

Stage data for Garretts Spring are converted directlyto discharge by software in the monitoring equipmentand the stage data are not retained. In general terms,the stage hydrograph at Garretts Spring shows suddendrops when Chenault Karst Window and Owens KarstWindow empty.

Discharge HydrographsDischarge hydrographs for the December 1990

storm for Owens Karst Window, Chenault KarstWindow, and Garretts Spring are presented in Figure12. Two curves are shown for Garretts Spring, anestimate using the sum of the Chenault and OwensKarst Window discharges, and the recorded dischargefrom the flume. Four discharge measurements made bywading are also shown. The value for January 4 is lowbecause an inappropriate Price meter was used.

The difference between the hydrographs for ChenaultKarst Window and Owens Karst Window is striking. Once aflow rate of approximately 37 cfs (1.05 cms) was reachedat Owens Karst Window, it remained constant for anextended period of time before tapering off gradually atfirst, then dropping sharply to pre-flood levels. The suddendrop is due to the depletion of storage in Sinking CreekKarst Valley, and the sudden reduction in flow into Owens


Karst Window. In contrast, the Chenault Karst Windowhydrograph shows a series of broad, but steep-flankedpeaks. This pattern reflects the rapid flooding andemptying of the karst window.

The shape of the estimated hydrograph for GarrettsSpring is nearly the same as the Chenault Karst Windowhydrograph because of the relatively flat curve of theOwens Karst Window hydrograph (Fig. 12). However,except during the peak flows from Chenault KarstWindow, the overwhelming majority of the flow iscontributed by Owens Karst Window. The hydrograph forthe flume exhibits considerably more detail than theestimated hydrograph, but its magnitude is almost alwaysroughly 10 cfs (0.28 cms) less than the estimate. Thisdiscrepancy is due to the unaccounted-for flow discussedunder "Methodology." More important is the parallelism ofthe estimate and flume curves. Most of the major peaksmatch exactly. A peak in the flume data on December 25that is not shown by the estimate is probably due to thedischarge of storage in Chenault Karst

Window that is below the elevation of the stage recorder.The longer duration of flow recession recorded by theflume is caused by the flow contribution between OwensKarst Window and Garretts Spring that is not accountedfor by the estimate.

Although only a limited number of dischargemeasurements were made by wading at Garretts Spring,the Price meter measurements suggest the estimateddischarge is reasonably correct. Furthermore, high flowscertainly exceed the flume's capacity (Fig. 6). However,the flume shows greater detail and, by design, is moreprecise when flow is not bypassing it.

Both the estimated and measured hydrographs atGarretts Spring reveal the importance of the timing ofstorage depletion in the karst windows. Also, therelative magnitude of the contribution of flow from thetwo subbasins is clearly illustrated. The high flows fromOwens Karst Window in the eastern sub-basin do notsupport conduit blockage as the cause of flooding inSinking Creek Karst Valley.

Runoff ModelingThe HEC-1 Flood Hydrograph Package (Computer

Program 723-X6-L2010) developed by the U.S. ArmyCorps of Engineers was used to model the effects of

changing land use and moisture conditions in thesoutheastern branch of the drainage basin. The data filewas originally set up by the Corps using their surveydata for channel gradient, channel width, bridges, andculverts. The data file was modified for this study byapplying the rating curve for the inflow capacity of theSinking Creek swallow holes and by making newestimates of the SCS curve number from soil surveymaps and land-use data gathered for this study. Eachtime the model was run, hydrographs were calculatedfor the 1989 storm and 12-hour design storms of 10-,50-, 100-, and 500-year frequencies.

Although suburban development is accelerating in thebasin, agricultural land use still predominates.Approximately 79.5 percent of the Garretts Spring Basin isin pasture or row crop. The total area up-gradient of theSinking Creek swallow hole is 2,959.7 acres (1197.7 ha).Land use in this area is distributed as follows: 76 percentagriculture (pasture 63 percent, row crops 13 percent), 8.5percent golf course, 8 percent residential (farmsteads,individual lots, subdivisions), 6 percent woodland, 1 percentlakes, ponds, etc., and 0.5 percent roads and highways.For this study, future suburban

Swallow-Hole Inflow RatingThe stage-discharge relationship for the Sinking

Creek swallow hole is complex (Fig. 7), and overall istypical for a swallow hole (Bonacci, 1987). However, thepressure-flow segment has a subvertical slope, possiblybecause a component of the discharge continues asunconfined flow through normally abandoned conduits,joints, and bedding planes during high-flow events. Flowinto high-stage swallow holes at Sinking Creek and frommultiple high-stage springs at Owens Karst Window hasbeen observed during floods. The regression lines forthe free-fall and pressure-flow segments converge at astage of 1.9 feet (0.58 m) and a discharge of 25.5 cfs(0.72 cms). Field observations indicate that all majorswallow-hole openings are completely inundated at thisstage. The estimated discharge from Sinking Creek forthe February 1989 event has also been plotted onFigure 7.


development was defined as 1 -acre (0.4 ha) lots, andall development was assumed to occur in formerpasture. Furthermore, the increased impervious areacreated by access roads for the additional developmentwas not considered. The increase in development wasalso assumed to be evenly distributed between eachhydrologic soil type. However, most development willprobably be on group B soils, since they are the mostcommon in the basin, and type C and D soils primarilyoccur in poorly drained areas. The relative increase inSCS curve numbers for suburban development on typeB soil is twice that of the same development on type Dsoil, and would increase surface runoff relative to theabove assumptions.

The percentage of each soil in the basin wasdetermined from soil survey maps (McDonald andothers, 1983) and classified into hydrologic soil groups.The hydrologic groups were found to be distributed asfollows: 69 percent type B, 23 percent type C, and 8percent type D. No type A occurred in the basin, andtype D occurred predominantly along the course ofSinking Creek. Curve numbers for each land-usecategory were individually calculated according tosoil-type occurrence.

Field experience gained during the course of thestudy indicated that vegetative cover and antecedentsoil-moisture content are critical to both volume andrapidity of runoff in the Garretts Spring Basin.Antecedent moisture condition 11 represents "averagemoisture content," and antecedent moisture condition IIIrepresents "nearly saturated soil." Condition IIIfrequently occurred during the winter months when therewas little evaporation or plant uptake. The SCS runoffcurve numbers were calculated for both soil-moistureconditions. The weighted mean for the southeasternbranch for condition 11 was 70 and for condition III was85, under current land use. Frozen ground will produceeven higher curve numbers and greater runoff.

The HEC-1 computer model was first run using basinareas and SCS curve numbers determined by the Corpsand the initial estimate of swallow-hole capacity. Thesecond run used the new swallow-hole rating curve, newrunoff curve numbers (moisture condition 11), and thereduced basin area to reflect ground-water tracingresults. The third, and all subsequent runs, used thenew data and moisture condition Ill. For the fourth runthe land use data were modified to reflect futuresuburban development of 20 percent of thesoutheastern branch. Current suburban development is2.3 percent. Residential development of other size lotswas not considered. The fifth model run was for a totallyforested watershed.

The results for the current land-use model revealedthe importance of antecedent soil moisture. A12-hour/100-year-frequency storm of 5.3 inches (13.5cm) will produce a stage elevation at Tashamingo of920.5 feet (280.57 m) under antecedent soil-moisturecondition II, while the same storm will result in a stageelevation of 923.0 feet (281.33 m) under antecedentsoil-moisture condition Ill. This assumes little or noevaporation or transpiration. A more intense storm ofshorter duration but greater frequency will also flood thebasin. The total accumulated precipitation for the 1989storm from February 13-16 was 8.4 inches (21.3 cm) atKeene and 7.1 inches (18.0 cm) at Bluegrass Field. Therain on February 13, 0.64 inch (1.6 cm), was sufficientto saturate the soil in the prevailing conditions of nearfreezing temperatures and absence of transpiration. Forthe 1989 storm, using the Keene total, the modelcalculated a stage at Tashamingo of 924.1 feet (281.67m) for condition II and 927.2 feet (282.18 meters) forcondition Ill. For comparison, the maximum stageactually reached at Tashamingo in 1989 was 928.3 feet(282.95 m); the elevation of Cherrywood Lane is 925feet (281.94 m). Increasing development in the basin to20 percent with 1-acre (0.4 ha) lots resulted in anincrease in stage to 927.4 feet (282.67 m), 0.2 foot(0.06 m) higher than with current land use, which wouldcover Cherrywood Lane at the causeway with 2.4 feet ofwater. Treating the basin as virgin forest resulted in astage of 925.8 feet (282.18 m) at Tashamingo, 1.4 feetlower than current land use, which would still blockCherrywood Lane with nearly a foot of water.

Miscellaneous FindingsThree Jessamine County residents who had lived in

the Sinking Creek area for many years wereinterviewed concerning the history of flooding in thearea. The late Mr. Howard Owens, former owner ofOwens Karst Window, recounted repeated floodingssince his childhood in the early 1900's. He recalled hisfather deliberately waiting for floods to help him "raft biglogs out of the sinkhole." Another resident remembered1989 being the third time Delaneys Ferry Road wasblocked since the 1940's. He recalled having to haulfeed to cattle stranded by one flood. The third residentrecalled Delaneys Ferry Road being blocked four timessince 1957. He noted that he used to hunt for duckalong flooded Sinking Creek. All three peopleinterviewed had been in continuous residence in thearea since at least 1957. The Lexington andNicholasville newspapers from 1930 through 1989 werechecked for stories on floods, but other than the 1989flood no mention of Sinking Creek was found.

Miscellaneous Findings

A common cause of sinkhole flooding is theobstruction of an outlet by natural or man-made debris.Only occasional small pieces of trash and limitedquantities of natural debris have been seen in swallowholes in the Garretts Spring Basin. The natural debrisconsists of wood and leaves that either float during aflood or rot, break up, and are carded through theconduit (Figs. 13-14). Although there is a trash dump inOwens Karst Window, the trash has not moved into theswallow hole area where water pressure would hold it inplace. The trash is generally above flood level and is onthe spring side of the karst window. During floodsnumerous small springs discharge along the slopebelow the base of the trash pile and at its base,indicating that flow is not affected (Fig. 15). If thesediment load in a sinking stream is excessive it canaccumulate on wood and leaves and temporarily blocka swallow hole. Likely sources of high sediment runoffare tilled fields and construction sites. Water sampleshave not been collected for suspended sediment, butthe sediment mantle left by the 1989 flood in OwensKarst Window and Chenault Karst Window was verythin (less than 0.05 in. or 1 mm). The changes observedin stream channels since 1989 suggest a loss ofsediment from channels and swallow hole areas.Because the swallow holes are free of trash andgenerally clear of natural debris, the flooding is unlikelyto be related to limited capacity at the swallow hole.

The only significant, enterable cave known in theGarretts Spring Basin is Dry Ridge Cave. The cave hasbeen partially mapped by the Blue Grass Grotto of theNational Speleological Society (O'Dell and O'Dell,1992). Dye traces have shown that the cave is drainingan isolated sub-basin of the northwestern branch. Thesurveyed length of the cave is 274 feet (83.5 m). Typicalpassage dimensions are 2 to 6 feet wide and 9 to 12feet high (1 to 2 m wide, and 3 to 4 m high). A small pitor shaft cave is also known in the vicinity of ChenaultKarst Window and a small cave receiving drainage froma sinkhole on Dry Ridge Pike. Neither has beenmapped.

HYDROGEOLOGY ANDPALEOHYDROLOGY

Although Thrailkill and others (1982) found thatargillaceous units are not important inhibitors of groundwater in the Inner Blue Grass, field observations indicatethat the Brannon Member and Macedonia Bed arelocally aquitards within the Garretts Spring Basin. TheBrannon crops out near the crest of ridges. Groundwater perched on the top of the Brannon emerges atnumerous small springs rimming the basin. One outcropof the Macedonia Bed is known in the stream bed ofChenault Karst Window. Water from Chenault Springflows on the

17

Macedonia to within a hundred meters (few hundred feet)of the Chenault Karst Window swallow holes. Theprojected outcrop of the Macedonia at Garretts Spring isjust above the elevation of the spring, suggestingheadward erosion of the Macedonia toward ChenaultKarst Window. Owens Karst Window is less than 10 feet(3 m) below the elevation of the stilling well. Jointing andunmapped faulting play a major role in conduit locationand orientation in the Inner Blue Grass (Thrailkill andothers, 1983) and probably in the basin.

Garretts Spring has an annual median discharge of6.5 cfs (0.18 cms) (Felton, personal communication,1992). The maximum discharge measured to date is58.6 cfs (1.66 cms), although discharges exceeding thisamount are known to have occurred. The flow fromOwens and Chenault Karst Windows joins undergroundwithin 1,500 feet (457 m) of the resurgence at the spring.


The recession limb of discharge hydrographs forGarretts Spring is stepped (Felton, personalcommunication, 1992).

Floods in Chenault Karst Window recede morequickly than in Owens Karst Window or Sinking Creekbecause of the smaller size of the northwestern branchof the basin, and the relatively more efficient flow fromChenault Karst Window to Garretts Spring. Dischargefrom the Chenault Karst Window spring flows in a welldefined channel with minor tributaries along its course.As the flow approaches the swallow-hole end of thekarst window the channel bifurcates into a distributarysystem feeding over 17 swallow holes. No relationshiphas been observed between stage in Chenault KarstWindow and the rate of discharge from Owens KarstWindow.

Ground-water tracing shows that flows from the threeprincipal swallow holes in Sinking Creek are tributaries toa single conduit, which branches into a distributarysystem as it approaches Owens Karst Window.Nominally, three springs are active in Owens KarstWindow but a fourth is active during high flow. Themiddle spring

is lowest in elevation and the southern spring is 1.7 feet(0.5 m) higher. The two northern springs flow into asingle channel and their discharge is treated as onespring. The lower northern spring is 5.9 feet (1.8 m)higher than the middle spring and the higher northernspring is 14.2 feet (4.3 m) higher. Both the middle andsouthern springs remain active during low flow, themiddle spring persisting the longest. During extremelow flow all four springs stop flowing. However, a smallkarst-window-like opening near the middle spring isalways flowing, indicating that ground-water flow occursbelow the floor of the karst window. Further flowcontinues in Sinking Creek even when the Owens KarstWindow springs are not flowing. Several additionalopenings, from 1 to 10 feet (0.3 to 3 m) above the mainsprings, discharge during high flow. The northernsprings have significant impact on stage changes inOwens Karst Window (Fig. 16). When flow begins orends at the northern spring, its discharge, coupled withoverland and quick-return ground-water flow in theOwens Karst Window catchment, rapidly floods OwensKarst Window. During extreme high flow, hundreds ofsmall openings discharge into Owens Karst Window.

Hydrogeology and Paleohydrology 19

Sinking Creek Karst Valley differs in geomorphologyand hydrology from an open-upstream poIje only in size(White, 1988), although the structural controlsassociated with poIjes are absent. Downstreamreaches of the valley are characterized by steep,cliff-forming valley walls with local reliefs of 50 feet (15m). Flooding of the valley occurs frequently in winterand early spring, and persists for days, sometimesweeks. Springs rim the valley just above the grade ofSinking Creek. The swallow-hole zone at the footwall ofSinking Creek Karst Valley receives flow via adistributary system that has three branches. Eachbranch feeds a cluster of macro-swallow-hole openings,many of which only accept flow during high stage.Throughout the footwall area are hundreds ofswallow-hole openings, varying in size from 1 inch (2.5cm) to 16 inches (40 cm) in diameter.

The parallel between the Tashamingo and OwensKarst Window hydrographs reveals the close match ofoutflow from Sinking Creek and outflow from OwensKarst Window. Soon after the Tashamingo hydrographbegins to rise, flooding at the swallow holes reachessufficient depth to activate the northern springs inOwens Karst Window. Once Owens Karst Window isflooded, its outflow no longer increases rapidly with

stage because inflow into the downstream conduit iscontrolled by the pressure limb of the hydrograph.Inflow from Sinking Creek is then slowed by thehydrostatic pressure in flooded Owens Karst Window.The hydrographs remain parallel, with inflow intoOwens Karst Window nearly matching out flow, untilSinking Creek Karst Valley empties. The suddenstoppage of flow from the northern spring then allowsOwens Karst Window to empty rapidly.

A potentially important factor in flooding atTashamingo is the possible contribution of flow fromthe pirated eastern tip of the southeastern sub-basin.During extreme events flow may be diverted belowground, and perhaps on the surface, to the west.Unfortunately, an opportunity to trace or even observethe area during a major flood has not occurred since itspotential significance was recognized.

Spangler (personal communication, 1989) speculatedthat a wide valley that extends from the vicinity ofTashamingo Subdivision north to Shannon Run (atributary of South Elkhorn Creek) may be an abandonedchannel of a surface-flowing Sinking Creek. The piracy


of Sinking Creek, if it occurred, would have happenedafter the initial development of the swallow holes at thefootwall of present-day Sinking Creek Karst Valley. It ispossible that the conduits from the swallow holes toGarretts Spring may have had insufficient geologic timeto adjust to the higher inflows from the geologicallysudden increase in catchment area.

CONCLUSIONSThe discharge rate from Sinking Creek is controlled

by the stage in Owens Karst Window. The dischargefrom Owens Karst Window is controlled by the hydraulicparameters of the conduit system to Garretts Spring.The conduit is not blocked by any man-made debris, butdischarge is limited by the conduit diameter, gradient,length, and roughness. This conclusion is supported bythe absence of trash in the swallow holes, the largemeasured discharges and flow velocities, and cross

sectional areas of the conduits as determined byground-water dye traces. The intake capacity for theSinking Creek swallow holes at the moment floodingbegins is 25.4 cfs (0.72 cms). The maximum capacity isapproximately 47.3 cfs (1.34 cms); unfortunately, inflowsinto Sinking Creek Karst Valley can exceed hundreds ofcubic feet per second (tens of cubic meters per second).Hydrologic modeling of the basin suggests thatantecedent moisture conditions are critical to thepotential flooding from a given storm. A12-hour/100-year-frequency storm will flood the basinnearly to the elevation of Cherrywood Lane if it occurswhen soil moisture is high and there is little loss toevaporation or transpiration. Modeling also suggests thatthe February 1989 storm would have flooded the valleyto the elevation of Cherrywood Lane even if there wasno development in the basin; but further developmentwill cause a limited increase in depth of flooding.

POTENTIAL SOLUTIONSIt is not the intent of this report to recommend a

specific solution to the flooding at TashamingoSubdivision, but rather to outline some options available toplanners. There may be other solutions not mentionedhere.

The intuitively obvious course of action is to enlarge theSinking Creek swallow holes. However, this researchshows that efforts to improve the intake capacity of theswallow holes by excavation at Sinking Creek, ifsuccessful, will only increase the stage at Owens KarstWindow, which will then negate the increased flow fromSinking Creek. If the swallow holes at Owens KarstWindow are also cleaned, only a small increase incapacity would be gained because of the hydrauliclimitations of the conduit to Garretts Spring. The conduitwould have to be enlarged and smoothed from OwensKarst Window to Garretts Spring, a distance of 3,800 feet(1, 158 m) to the west, to improve its discharge capacity.

Although there is no evidence the Sinking Creekswallow holes are blocked at this time, they could becomeblocked in the future. This would be a likely consequenceif trash was dumped into the creek upstream of theswallow holes, or if the sediment load of Sinking Creekwere to increase. Regulations to control sediment runofffrom construction sites should be considered, as well ascontinued efforts by conservation agencies to control soilloss from farming. Furthermore, dumping of any kind intothe headwaters of Sinking Creek should be prohibited.Structures to prevent flood debris and sediment fromblocking swallow hole openings have been used in Europeto reduce peak flood stage in karst valleys. Although theconstruction of these structures would not substantiallyimprove the

Potential Solutions

outflow capacity of the swallow holes, they couldprevent a further reduction in capacity.

Pumping after flooding begins, while technicallyfeasible, would require that very high-capacity pumps beavailable year round on 24 hours notice. The six pumpsavailable to the Kentucky Disaster and EmergencyServices (DES) in February 1989 had a combinedcapacity of 9,400 gallons per minute (gpm) or 21 cfs(0.6 cms) (Patrick C. Conley, DES, personalcommunication, March 13,1989). At that pumping rate,in addition to the natural discharge, it would have takenabout 9 days to lower the water to the level ofCherrywood Lane. By allowing the water to drainnaturally, Cherrywood Lane was open to traffic 14 daysafter it was blocked. Siphoning water from the basin,while not requiring constant pumping, would be muchslower per pipe, and would require the construction ofstaging ponds for each 25 feet (7.6 m) of lift (typicalmaximum practical suction lift). At least one pumpwould be needed for priming the pipes.

To create a new, gravity-flow outlet for SinkingCreek, deep excavations would be needed. Theshortest route for a diversion ditch is to the head ofClear Creek, 2,200 feet (670 m) southwest. A cut with amaximum depth of 70 feet (22 m) would be required.An alternative route, north to the head of ShannonsRun, would have a maximum depth of 20 feet (6.3 m),but would be over 5,000 feet (1,524 m) long.Furthermore, water users downstream of Sinking Creekwould have their supply substantially reduced.

The HEC 1 modeling clearly shows that a significantincrease in suburban development in the basin willhave an adverse, but relatively small, impact on flooddepth in the Sinking Creek Karst Valley. Retentionbasins can be required in new developments to containincreased runoff, and lengthen the runoff travel time.Sinking Creek Karst Valley is itself acting as astorm-water retention basin, and has a long storagetime compared to a normal valley. Therefore, retentionbasins in the headwaters should be designed for aworst-case scenario and will have to retain their storagefor several days.

Land-use management could be used to mitigate theimpact of future floods. Many communities use flood-prone land for recreation areas. The U.S. Army Corps ofEngineers has prepared a study (1990) that legallydefines a 100-year flood plain for the basin. The areawithin the flood plain could be designated unsuitable forfurther development and the flood-prone property couldbe gradually acquired by local government. The naturalbeauty of Sinking Creek, with proper management, wouldmake it a very scenic recreation area.

21

ACKNOWLEDGMENTSThe successful conclusion of this research would not

have been possible without the cooperation of manyindividuals and organizations. The help and cooperationof Woodford and Jessamine County property owners issincerely appreciated, as is the help of County JudgeExecutives Sherman Dean and Neil Cassidy andemployees of Jessamine County. Mr. LawrenceSpangler shared results of ground-water traces heconducted during 1986 and 1987. The U.S. Army Corpsof Engineers, represented by Mr. George Herbig,graciously provided surveying resources for accurateelevations of the water-level recorders, and valuabledata files for the hydrologic modeling. Dr. Dan Carey ofthe Kentucky Geological Survey modified the HEC-1data files, and executed the hydrologic program. Withouthis help the process would have taken much longer.Special thanks is due Dr. Gary Felton, of the Universityof Kentucky College of Agriculture, who generouslyshared data from the flume installation at GarrettsSpring, provided additional help with the surveying, anddetermined land use in the basin. Finally, we thank thestaff of the Kentucky Geological Survey, who helped withequipment installation and field work; reviewed, edited,and drafted this report; and worked in the field for longhours in sometimes dangerous, and frequentlymiserable, conditions to gather the data essential to thisstudy.

REFERENCES CITEDAley, Tom, and Fletcher, M. W., 1976, The water tracer's

cookbook: Missouri Speleology, v. 16, no. 6, 36 p.

Baumgartner, R. M., 1991, Analysis of the spatialdistribution of sinkholes related to a ground waterbasin in the Inner Bluegrass Karst Region ofKentucky: Lexington, University of Kentucky, M.S.Thesis, 122 P.

Bonacci, Ognjen, 1987, Karst hydrology: New York,Springer-Verlag, 184 p.

Buchanan, T. J., andSomers, W. P., 1976, Dischargemeasurements at gaging stations, in Techniques ofwater-resources investigations of the United StatesGeological Survey: U.S. Geological Survey,Applications of Hydraulics, Book 3, Chapter A8, 65p.

Cressman, E. R., 1965, Geologic map of the KeeneQuadrangle, central Kentucky: U.S. GeologicalSurvey Geologic Quadrangle Map GO-440.

Davis, S. N., Campbell, D. J., Bentley, H. W., andFlynn, T. J., 1985, Groundwater tracers: U.S.Environmental Protection Agency Off ice ofResearch and Development and National WaterWell Association, 200 p.


Ford, D. C., and Williams, P. W., 1989, Karstgeomorphology and hydrology: London, UnwinHyman, 601 p.

Hamilton, D. K., 1950, Areas and principles ofgroundwater occurrence in the Inner Blue GrassRegion, Kentucky: Kentucky Geological Survey, ser.9, Bulletin 5, 68 p.

Jones, W. K., 1984, Dye tracer tests in karst areas: TheNSS Bulletin, Journal of Caves and Karst Studies, v.46, no. 2, p. 3-9.

Kennedy, E. J., 1984, Discharge ratings at gagingstations, in Techniques of water-resourcesinvestigations of the United States GeologicalSurvey: U.S. Geological Survey, Applications ofHydraulics, Book 3, Chapter Al 0, 59 p.

McDonald, H. P., Sims, R. P., Isgrig, Dan, and Blevins,R. L., 1983, Soil survey of Jessamine and WoodfordCounties, Kentucky: U.S. Soil Conservation Service,in cooperation with the Kentucky AgriculturalExperiment Station and the Kentucky Departmentfor Natural Resources and EnvironmentalProtection, 94 p.

Mull, D. S., Smoot, J. L., and Liebermann, T. D., 1988,Dye tracing techniques used to determinegroundwater flow in a carbonate aquifer system nearElizabethtown, Kentucky: U.S. Geological SurveyWater Resources Investigations Report 87-4174, 95p.

O'Dell, G. A., and O'Dell, C. S., 1992, Guide book,1992 Speleofest: Blue Grass Grotto, NationalSpeleological Society, 60 p.

Palmquist, W. N., and Hall, F. R., 1960, Availability ofground water in Bourbon, Fayette, Jessamine, andScott Counties, Kentucky: U.S. Geological SurveyHydrologic Investigations Atlas HA-25, 3 plates,scale 1: 125,000.

Palmquist, W. N., and Hall, F. R., 1961,Reconnaissance of ground-water resources in theBlue Grass Region, Kentucky: U.S. GeologicalSurvey Water-Supply Paper 1533, 39 p.

Quinlan, J. F., ed., 1987, Qualitative water-tracing withdyes in karst terranes: Practical karst hydrogeology,

with emphasis on groundwater monitoring: NationalWater Well Association, 26 p.

Smart, R L., 1984, A review of the toxicity of twelvefluorescent dyes used for water tracing: The NSSBulletin, Journal of Caves and Karst Studies, v. 46,no. 2, p. 21-33.

Soil Conservation Service, 1986, Urban hydrology forsmall watersheds (2d ed.): U.S. Soil ConservationService, Engineering Division, Technical Release 55,156 p.

Spangler, Larry, 1989, Karst hydrology of Steeles Run,Prestons Cave-McConnels Spring, and GarrettsSpring Basins, Lexington, Kentucky [abs.]: The NSSBulletin, Journal of Caves and Karst Studies, v. 53,no. 1, p. 40-41.

Thrailkill, John, Byrd, P. E., Sullivan, S. B., Spangler, L.E., Taylor, C. J., Nelson, G. K., and Pogue, K. R.,1983, Studies in dye-tracing techniques and karsthydrogeology: University of Kentucky WaterResources Research Institute, Research Report140, 89 p.

Thrailkill, John, Spangler, L. E., Hooper, W. M., Jr.,McCann, M. R., Troester, J. W., and Gouzie, D. R.,1982, Groundwater in the Inner Bluegrass KarstRegion, Kentucky: University of Kentucky WaterResources Research Institute, Research Report136, 108 p.

Thrailkill, John, Sullivan, S. B., and Gouzie, D. R., 1990,Flow parameters in a shallow conduit-flow carbonateaquifer, Inner Bluegrass Karst Region, Kentucky:University of Kentucky, Department of GeologicalSciences, unpublished manuscript, 31 p.

U.S. Army Corps of Engineers, 1990, Revision to floodinsurance study, Jessamine County, Kentucky(unincorporated areas), 1978: U.S. Department ofHousing and Urban Development, FederalInsurance Administration, 2 p.

White, W. B., 1988, Geomorphology and hydrology ofkarst terrains: Oxford, Oxford University Press, 464P.

Mission Statement

The Kentucky Geological Survey at the University of Kentucky is a Statemandated organization whose mission is the collection, preservation, and disseminationof information about mineral and water resources and the geology of theCommonwealth. KGS has conducted research on the geology and mineral resources ofKentucky for more than 150 years, and has developed extensive public data bases foroil and gas, coal, water, and industrial minerals that are used by thousands of citizenseach year. The Survey's efforts have resulted in topographic and geologic mapcoverage for Kentucky that has not been matched by any other state in the Nation.

One of the major goals of the Kentucky Geological Survey is to make the resultsof basic and applied research easily accessible to the public. This is accomplishedthrough the publication of both technical and non-technical reports and maps, as wellas providing information through open-file reports and public data bases.

Date post:	03-Oct-2020
Category:	Documents
Upload:	others
View:	2 times
Download:	0 times

FLOODING OF THE SINKING CREEK KARST AREA IN JESSAMINE … · UNIVERSITY OF KENTUCKY, LEXINGTON...

Documents