Belfry Bulletin

Search Our Site

Article Index

 

More Notes On Water Studies In Wookey Hole Cave, Somerset

T. Chapman, A. Gee. A. V. Knights, C. Stell and R.D. Stenner

The plan of the cave.

The plan shows the location of the sites where water samples were collected.  The plan was based on Trevor Hughes' 1982 drawing, which he compiled from unpublished surveys made by several members of the Cave Diving Group. The plan included here was made by scanning a large drawing in six sections.  A grid was drawn separately on one layer in an AutoCAD file.  The scanned sections were moved to their correct positions on the grid, trimmed, joined up and cleaned up until a print could be made at the final scale.  Using this print, a plan was traced in ink, and scanned to produce an Adobe PhotoShop 4 file.  This file was edited and text was added.  A draft was printed, and approved by the Wookey Hole management to be used to illustrate the present water studies.

The location of the sample sites is shown in the Figure 1.  A collection was started on 30.11.96, but after collecting samples from Chambers 3 and 9, the trip had to be abandoned.  Samples were collected on 14.12.96, and 25.01.97.  Water levels in November and December were extremely high. In January the water level had fallen, but it was still high.

The results from the samples collected in December 1996 and January 1997 were utterly unexpected, and the possible implications were intriguing.  A high priority was placed on collecting a sample from upstream of Wookey 23.  After many attempts between March 1997 and July 1997 failed, samples were collected on 20.07.97 from the same sites as on 14.12.96, with the addition of a sample from Wookey 25, immediately before the long descent into the 25th Sump.

The analytical techniques used will be available in detail elsewhere.  Ion balances are included in the tables of results.  They show where analyses have been satisfactory, and where analyses have been adversely affected by high levels of suspended calcium carbonate in the water.  These "colloids" made it impossible to analyse samples for total hardness calcium and bicarbonates with the usual accuracy.

Standard errors have been calculated (as 10 x 10-5 Molar, - ppm as CaC03):-  total hardness, calcium and aggressiveness to CaC03, 0.8; alkaline hardness 3.0; Non-alkaline hardness, 3.8; magnesium, 0.46; sodium, 3.0; potassium, 1.2; chloride, 2.6; sulphate, 0.3; nitrate, 0.23.  The reliability of the recent analyses is indicated by the cation/anion balances.

The units chosen in this study was 105 x Molar.  By making this choice, the data for the species associated with water hardness are numerically identical (to within analytical precision limits) to parts per million (ppm) as calcium carbonate, the unit widely used by limestone geomorphologists.  Using a unit based on Molarity was especially suitable when calculating ion balances.

 

RESULTS

Table 1. Data for samples collected 14th December 1996. Units: 105 x M (as specified; non-alkaline hardness di-valent, ion balance mono-valent).  Total, Mg, Ca, alkaline hardness and non-alkaline hardness figures are identical to concentrations in ppm as calcium carbonate. "Coll." represents the estimated concentration of "colloidal" calcium carbonate.  The accuracy of figures in italic script is seriously lower than usual because of analytical difficulties caused by high concentrations of colloidal calcium carbonate. The errors in the figures in parentheses are the result of titrating colloidal calcium carbonate as alkaline hardness, confirmed by large ion imbalances and impossible non-alkaline hardness data.

Site

Total

Mg

Ca

Alk

Non

Cl

K

Na

SO42

N03

Call.

Ion

 

Hard.

 

 

Hard

A

 

 

 

 

 

 

Bal.

 

 

 

 

 

Hard.

 

 

 

 

 

 

 

Road Bridge.

245.6

34.6

211

(259)

(-14)

45.5

3.9

30.3

14.8

45.0

>41

113

3rd Chamber

274.2

33.9

240

(261)

(14)

47.1

4.2

30.3

11.8

46.8

 

55

9th Chamber

239.7

33.5

206

(258)

(-19)

45.5

4.0

31.0

15.3

37.4

 

116

Sump 20

233.4

34.4

199

(257)

(-23)

43.9

4.1

30.9

14.5

43.1

 

128

Sump 22

267.9

32.3

236

(255)

(17)

45.5

4.0

30.2

14.1

45.0

 

72

S.Sump W 23

211.5

9.8

202

(235)

(-23)

42.3

4.2

34.2

14.1

40.8

 

119

 

Table 2. Data for samples collected 25th January 1997. Units and symbols as in Table 1

Site

Total

Mg

Ca

AIk.

Non

Cl

K

Na

SO42

N03

Call.

Ion

 

Hard.

 

 

Hard

A

 

 

 

 

 

 

Bal.

 

 

 

 

 

Hard.

 

 

 

 

 

 

 

Road Bridge.

279.7

42.2

238

253

27.0

47.9

4.0

34.5

12.4

39.3

1.5

19

3rd Chamber

274.7

42.2

233

255

20.0

46.3

4.0

33.8

17.3

37.6

1.5

40

9th Chamber

281.2

43.0

238

254

26.9

47.9

5.0

34.0

17.3

42.9

1.5

32

Sump 20

277.7

42.6

235

253

24.7

47.1

4.0

34.2

11.4

36.8

1.5

19

Sump 22B

281.2

42.0

239

252

29.4

46.3

4.0

34.4

11.0

36.0

1.5

8

Sump 22

276.1

42.4

234

249

26.8

45.5

4.1

34.1

 

 

3.1

 

S.Sump W23

251.8

13.6

238

231

20.4

40.6

3.8

36.2

13.6

42.7

3.8

30

 

Table 3. Data for samples collected 20th July 1997. Units and symbols as in Table 1.

Site

Total

Mg

Ca

AIk.

Non

Cl

K

Na

SO42

N03

Call.

Ion

 

Hard.

 

 

Hard

A

 

 

 

 

 

 

Bal.

 

 

 

 

 

Hard.

 

 

 

 

 

 

 

Road Bridge.

289.8

35.0

255

250

40.2

43.9

4.6

35.2

12.0

27.6

0.0

24

3rd Chamber

285.4

33.8

254

247

38.7

42.3

4.6

37.5

14.0

34.5

0.0

14

9th Chamber

Coli.

35.6

Coli.

Coli.

Coli.

44.7

4.6

37.8

13.9

30.6

 

 

Sump 20

289.4

34.6

255

250

39.8

42.3

4.6

36.5

13.8

41.4

0.0

9

Sump 22

287.0

34.8

252

249

38.2

42.3

4.8

33.4

14.6

33.2

0.0

10

S.Sump W23

270.6

11.9

259

239

31.9

39.8

9.7

28.8

11.0

8.6

0.0

32

Sump 25

286.7

35.4

251

251

35.9

44.7

4.6

32.8

10.0

32.3

0.0

5

On 30.11.96, the River Axe was very high and cloudy.  The analysis of the samples was very difficult because of high quantities of suspended calcium carbonate.  Filtering the samples did not remove the suspension.  The "colloidal" calcium carbonate interfered seriously with the total hardness and alkaline hardness titrations, making an alkaline hardness titration impossible, and seriously reducing the accuracy of the total hardness titrations.

The water in the Axe was completely clear by 12.12.96, but in spite of the complete absence of any visual warning of a likely problem, every sample still contained large concentrations of "colloidal" calcium carbonate.  This once again made reliable alkaline hardness determinations impossible, and reduced the accuracy of total hardness titrations. To make the nature of the problem caused by colloidal calcium carbonate clearer to the reader, Table 1 contains the alkaline hardness data which were obtained. Because negative values for non-alkaline hardness are impossible, the results are clearly grossly in error, the alkaline hardness figures being too high.  Titration of the sample from the Wookey Hole Road Bridge with HCI to a final stable end-point provided a minimum estimate of the concentration of "colloidal" calcium carbonate; 41 ppm CaCO3.  The initial value of the alkaline hardness was probably too high. The ion imbalance suggests that it was 60 ppm too high, but the non-alkaline hardness figure suggests a lower figure; approximately 49 ppm too high.  Therefore the true concentration of the colloidal calcium carbonate is likely to have been 89 or 101 ppm as CaCO3.

Samples from 25.01.97 contained small concentrations of "colloidal" calcium carbonate (1.5 to 3.8 ppm as CaCO3. The alkaline hardness, in which the dissolved and "colloidal" calcium carbonate species were titrated together, was corrected by subtracting the value obtained in the total hardness for the "colloidal" calcium carbonate.

Samples from 20.07.97 were free of "colloidal" calcium carbonate, with the single exception of the sample from Wookey 9, in which the "colloidal" calcium carbonate concentration was too high to permit the determination of total, alkaline or calcium hardness.

Comments on the results from 14.12.96

1.                  At every site, large concentrations of "colloidal" calcium carbonate were present.

2.                  At every site except the Static Sump in Wookey 23, magnesium concentrations were very similar (N.B. within the ranges of concentrations found in the present study, and within the range of pH in the water, although magnesium can be added, its natural removal is not possible).

3.                  Concentrations of chloride, sulphate, nitrate, potassium and sodium at all sites (including the Static Sump in Wookey 23) were similar.

4.                  At the Static Sump and chambers 20 and 9, calcium values were very similar.

Comment 2, above, will be examined in more detail.  The precipitation of magnesium as magnesium hydroxide will not take place from a solution containing 50 x 10-5 M Mg when the pH of the water is less than 11. This pH is far above the range which can exist naturally in the subterranean River Axe, so removal of magnesium by this means can be safely ruled out.  Indeed, laboratory measurements made after shaking natural waters with powdered limestones and dolomites have in every case led to an increase of magnesium in solution, rather than a reduction (Stenner, 1971, ibid, Table 1 p. 290, Table 2 p. 292).

There is evidence of the ability of magnesium carbonate to pass into solution by the incongruent solution of dolomite, but no evidence to suggest the removal of magnesium from solution.

Despite the analytical difficulties, the results show that the explanation of magnesium variations in the Axe which, prior to this study had been considered by Stenner to be most likely, was completely incorrect.  There was no magnesium gradient in the River Axe between Wookey 23 and the Entrance.

The outstanding feature of the result from 14.12.96 was the low magnesium level in the Static Sump in Wookey 23.  At the same time, the concentrations of many constituents in the sample were the same as in the other samples from the River Axe.  It seemed possible that water in the Static Sump had the same origin as water in the Axe, except that the latter water had dissolved a considerable quantity of magnesium.  Because there was a remote possibility that the similarities in the other hydro-chemical characteristics could have been a coincidence, another set of samples was needed.

The higher Ca contents in Wookey 22 and 3 samples were thought to have caused by inaccuracies in the total hardness titrations.  It is possible that the fraction of very small particles in the "colloidal" calcium carbonate was higher in these two samples, and a significant quantity of this fraction had been included in the rapid titration to the first unstable end-point.  A further set of samples was needed to resolve the uncertainties.

In conclusion, the very first set of samples had produced very exciting results.  There were uncertainties caused by the unfilterable suspended calcium carbonate.

Comments on the results from 26.01.97

1.                  Except at the Static Sump at Wookey 22, magnesium concentrations were very similar, but significantly higher than on 14.12.96, when the flow had been substantially greater.

2.                  Concentrations of calcium, sodium, potassium, sulphate, nitrate and chloride were very similar at all sites.  The data sets for sodium and sulphate on the two dates were different.  The data comprehensively supported the suggestion from the 14.12.96 results; that water in the Static Sump in Wookey 23 had the same origin as water in the Axe.  While there had been a remote possibility that the similarities on 14.12.96 could have been a coincidence, there is no possibility whatsoever that the different results from 26.01.97 could also have been a coincidence.

3.                  The sample collected underwater in Wookey 22, from where the river from Sting Corner enters the sump through boulders, gave results which were indistinguishable from those from the surface of the sump pool.

4.                  Because concentrations of suspended calcium carbonate were low, alkaline hardness data were more reliable, and total hardness and calcium data were much more reliable.  The increase of magnesium between the Static Sump and Wookey 22 was accompanied by an equal increment of alkaline hardness (within practical limits).  The conclusion is that magnesium from MgCO3 in dolomite or dolomitic limestone had dissolved as Mg(HCO3)2.  There was no change in calcium in true solution in the water.

Comments on the results from 20.07.97

The most important result was from the sample from Sump 25.  Results from this sample were similar to those from the main flow of the River Axe at all points downstream, while the magnesium, alkaline and total hardness results from the Static Sump were once again significantly different from those in all other samples.  This result had a considerable consequence on the understanding of the hydrology of the cave. Some of the minor abnormalities in the data from the Static Sump are likely to be the consequence of the static sump having been "stagnant" for several weeks, during which time several parties of divers had visited the site.

THE HYDROCHEMISTRY OF WOOKEY HOLE CAVE

The hypothesis is that in December 1996 and January 1997, water in the Static Sump in Wookey 23 did indeed have the same origin as water in the River Axe.

The following paragraph describes the position after examining the results from December 1996 and January 1997.  Following directly from the hypothesis of the origin of water in the Static Sump, there must be an "Unknown Junction" yet to be discovered, at some point upstream, where all the water in the Axe had a composition similar to that in the Static Sump.  From this "Unknown Junction", a small fraction of the Axe flowed into a route leading to the Static Sump.  The majority of the Axe flowed through a different route to Wookey 22, and in this route it entered a zone of dolomite or dolomitic limestone.  Here, the physical conditions (such as turbulent mixing) were such that the water dissolved the substantial concentration of magnesium carbonate seen in the results.

The results show that on 25.01.97, the water dissolved a higher concentration of magnesium between the "Unknown Junction" and Wookey 22 than on 14.12.96.  However, on both occasions, no change in magnesium was detectable in the considerable distance from Wookey 22 to the Entrance. These are very important observations, for which there are three possible explanations.

1.                  The "Unknown Junction" is a very large distance upstream of Wookey 22 (much farther than the distance from Wookey 22 to the Entrance).

2.                  The distance upstream is not crucially important, the most important factor being that the main body of the Axe flows through a zone where the physical conditions especially favour and maximise the solution of magnesium from the dolomite. This possibility presents a problem. When, in higher flow, water arrives at Wookey 22 with lower levels of magnesium, it follows that it must arrive there with a capacity to dissolve more magnesium.  Yet from Wookey 22 to the Entrance it fails to dissolve any more magnesium, in spite of contact with dolomitic conglomerate from Wookey 12 to the entrance.

3.                  (This is a modification of the second possibility).  Downstream of the "Unknown Junction", a part of the Axe flows through a dolomite zone where physical conditions encourage rapid reactions between water and rock, becoming saturated with magnesium to close to the low-flow value of approximately 50 x 10-5 M Mg.  As flow increases, this water is mixed with an increasing proportion of low magnesium water over-flowing from the route to the Static Sump.  This would explain the variable, flow-dependent concentration of magnesium in the Axe arriving at Wookey 22.

The results from the samples collected on 20.07.97 added nothing new to this particular aspect of the study.  Whichever explanation turns out to be the best explanation, results in the present studies have implications.

1.                  Water from the four major separate sources of the Axe (Swildon's Hole, Eastwater Cavern, St. Cuthbert's Swallet and - by far the biggest source - percolation water) must have coalesced upstream of the "Unknown Junction".

2.                  Important new information provided by samples collected on 20.07.97 concerned the location of the "Unknown Junction".  The results proved that the "Unknown Junction" lies beyond the present known limits of the cave; i.e. upstream of Wookey 25.

3.                  The possibility of making important discoveries in a route from the Static Sump to the "Unknown Junction" is very real.

4.                  There must be a zone of dolomite, dolomitic limestone or dolomitic conglomerate upstream of Sump 25, between Wookey 25 and the "Unknown Junction".

5.                  Where water from the "Unknown Junction" encounters the zone of dolomite, solutional activity will have caused considerable localised cavern enlargement (which could be masked by massive localised cavern breakdown). This is a direct consequence of the large quantity of magnesium carbonate being dissolved by the large river in a localised zone of the underground river system.

The water in the Static Sump will not always have the same chemical characteristics as water in the Axe, apart from elevated magnesium bicarbonate.  On the first two sample dates, the flow of the Axe was high, and it is possible that as flow falls, a level might be reached when water from the "Unknown Junction" ceases to flow to the static sump.  Water in the static sump will then reflect the levels of salts in the Axe the last time it flowed to the pool, and not the current levels in the river.  This will not negate the conclusions drawn from the results presented here.

The present article describes the present state of the understanding of the hydrology of the Wookey Hole system.  There are opportunities to refine this understanding before an attempt is made to explore this part of the cave.  The survey shows that there are four more several static sumps in Wookey 23, an intriguing one in Wookey 25 and another one in Wookey 20.  In the near future it is planned to analyse samples from these additional static sumps (together with a sample from Sting Comer).  It is also planned to analyse a selection of mud samples from Wookey 23 (because only those samples deposited in the last two thousand years by the St. Cuthbert's Swallet to the River Axe system will have a high lead content).

A summary of the earlier data is presented in Table Al in the Appendix, for comparison with the new data.

SUMMARY

1.                  On any occasion, there was no variation in any measured hydro-chemical component of the River Axe from Sump 25 to the Entrance.  In particular, there was no magnesium gradient in the River Axe.

2.                  The points of confluence of the tributaries which make up the River Axe are all upstream of Sump 25.

3.                  The Static Sump in the 23rd Chamber contained magnesium concentrations which were significantly lower than those in the River Axe.

4.                  Concentrations of magnesium and alkaline hardness were lower in the Static Sump, by an equal quantity, than those in the River Axe, while those of calcium, sodium, potassium, chloride, sulphate and nitrate were the same, proving that they shared the same origin.

5.                  There is a considerable possibility that exploration of the Static Sumps will result in the discovery of important extensions.

6.                  At an unknown distance upstream of Sump 25, the River Axe is a single unit with a composition similar to the Static Sump in the 23rd Chamber.  The physical conditions in which most of the River Axe acquires varying concentrations of magnesium bicarbonate are not known.

7.                  At times, analysis of samples for total, calcium and alkaline hardness have been made difficult by the presence of large concentrations of suspended calcium carbonate.

REFERENCES

1.                  Atkinson, T.C, Drew, D.P. with High, C. 1967 Mendip karst hydrology research project, phases one and two, Wessex Cave Club Occ. Pub. Ser 2 (1).

2.                  Gee, A. 1996 Recent exploration in "Wookey", Belfry Bull., 48(1), 7-10. 4.

1.                  3.   Hanwell, J.D. 1970         Digger meets diver, J Wessex Cave Club, 11(128) 34-9.

3.                  Hughes, 1982 1982 A sketch plan of Woo key Hole Cave. No grade, approximate scale, unpublished

4.                  Heathwaite, A.L., Knights, A.V. and Stenner R.D. 1998  In preparation.

5.                  Rose, L. 1983 Alkalinity, its meaning and measurement, Cave Science (Trans. B.C.RA.), 10(1),21-29.

6.                  Stenner, RD. 1969 The measurement of the aggressiveness of water towards calcium carbonate, Trans. C.RG. 11(3), 175200.

7.                  Stenner, RD. 1971 The measurement of the aggressiveness of water Parts II and ill, Trans. C.RG. 13(4),283-295.

APPENDIX

Table A1.  A summary of hydro-chemical characteristics of the River Axe at Wookey Hole between 1966 and 1978. Most of the samples were collected from the 3M Chamber.

 

Tern

Total

Mg

Ca

Alk.

NonA

Agg.

Cl

K

Na

SO42

T.An

D.O

Pres/

 

p'C

Hard.

 

 

Hard

Hard.

 

 

 

 

 

val=2

 

E.coli

 

 

 

 

 

 

 

 

 

 

 

 

 

%

 

No.

14

16

14

14

16

16

7

8

3

3

3

1

3

1

Mea

9.94

273

27.8

249

240

33.0

-0.2

58.5

4.3

27.5

20.5

37

89.8

90

n

 

 

 

 

 

 

 

 

 

 

 

 

 

 

S.D.

0.13

15.4

6.6

14.3

12.3

8.1

9.3

5.8

0.5

2.3

2.18

 

7.3

 

RSD

1.3

5.6

28.6

5.7

5.1

24.6

-5900

9.9

11.6

8.1

10.6

 

8.1

 

Min

9.7

244

18

219

212

22

-17

46

3.8

25.0

18

 

84

 

Max

10.1

295

45.5

267

260

50

11.2

64

4.8

29.2

28

 

98

 

 

Table A2.  Results of analyses of samples collected from the River Axe at Wookey Hole Bridge in 1996 and 1997.

 

Total

Agg

Mg

Ca

Alk.

NonA

Cl

K

Na

SO42

NO3

 

Hard.

 

 

 

Hard

Hard.

 

 

 

 

 

No.

6

4

6

6

4

4

6

6

6

4

4

Mean

273.6

-4.2

41.3

232

252

32.4

45.8

4.08

23.4

12.9

39.3

S.D.

15.8

1.44

6.0

14.0

1.6

4.9

2.13

0.39

3.2

1.07

7.2

RSD

5.8

34.8

14.4

6.0

0.6

15.0

4.66

9.5

9.5

8.3

18.2

Min

246

-6.5

34.6

211

250

27.0

42.3

3.7

29.5

12.2

27.6

Max

290

-2.8

51.3

255

254

40.2

47.9

4.6

38.8

14.8

45.3

While most of the hydro-chemical characteristics of the Axe have been stable over the period of more than thirty years, there is one exception.  A single sample, collected on 21.08.68, was analysed for sulphate, chloride and total anion content.  The results showed the water contained a nitrate content that was too low to be detected.  This conclusion was supported by a good anion/cation balance for the sample. Although this was the result of the analysis of a single sample, there is an identical situation at the Cheddar Spring.  A single sample, collected from the Cheddar Yeo spring on 8.10.68 was analysed for sulphate, chloride and total anion, and in this sample the concentration of nitrate was also too low to have been detected by this method.  At this spring, nitrate concentrations are also much higher now than the detection limit for nitrate in the methods used in 1968 to 1970. They are now in the range of 22 x 10-5 to 42 X 10-5 M.

Acknowledgement.

The authors wish to acknowledge the support given by the management of Wookey Hole Caves to members of the Cave Diving Group in their work in this cave, and to thank them for giving permission for this article to be published.

The following pictures are taken from The Great Cave of Wookey Hole - H.E. Balch, reprinted thanks to Robin Gray and Wells Museum.  Original photographs were taken by Harry Savory.



Plans of Wookey Hole from Balch's book