Belfry Bulletin

Search Our Site

Article Index

 

Streams Feeding St. Cuthbert’s Swallet

R.D. Stenner, Jingles, Estelle Sandford.

An article in the August 1994 B.B. explained how the latest stream studies started, and how the first analyses of the first set of samples had produced new information.  The article was low in actual data, and the aims of the studies were not explained.  This article aims to rectify these deficiencies.

Connections between surface sinks and inlets in the cave were discovered by a variety of methods between 1960 and 1972 (1). Until 1969, Plantation Stream sank spectacularly in Plantation Swallet (now filled in).  The streams were first studied intensively from 1965 to 1973.  From 1966 to 1968 a rectangular-notch weir was used in the old Plantation Stream (close to where the stream passed under the Ladywell Stream aqueduct).  Water samples were taken, and water temperatures measured.  The methods used have been discussed elsewhere (2).  A summary of the results is given in Table 1.

On 5 occasions, enough measurements were made between Traverse Chamber and the Entrance to work out the distribution of the streams in the cave, ignoring the small stream in Rocky Boulders series, which has no connection with the rest of the stream system (and which is thought to be joined in The Lake by minor seepage from the Main Stream at Mo's Dig).  The calculation of the distribution depended on two assumptions:

1.                    changes in the measured factors during the trip were negligible; and

2.                    that the discharge value of the inlet stream at P.1. was equal to that of Plantation Stream at the surface.

Evidence to support these two assumptions follows:

1.       On 26.11.67, the surface streams were sampled three times at four-hourly intervals. Temperature differences were 0.9T and O.TC, but salts showed no measurable changes.  This gave qualified support to the first assumption. In a 32-trip study in G.B. Cave in 1968, this assumption failed once, when there was a downpour during the trip.  Nevertheless, to maximise the chance of this important assumption being valid, it was recommended that stream surveys should be made in the shortest possible time, moving upstream (2).

2.       On 10.2.68, the size of Pulpit Passage East Inlet was measured directly, with a polythene sack, watch and measuring cylinder. Stream ratios along the stream passage to P.1. gave a value of 165 1/min for Plantation Inlet.  This was close to the value of 151 1/min given by the Plantation Swallet weir. Chemical changes between Plantation Swallet and P.1. were small (Table 1).

Chemical changes between the Pool outlet and Plantation Swallet were much greater (Table 1).  It was possible that somewhere in the marshy ground below the Pool, a separate stream was mixing with Plantation Stream. However, there was another possibility. The stream might leak into the marsh, become enriched with C02 from decaying vegetation, and seep back into the stream.

On 4.11.68, a change in the situation at Plantation Swallet was noticed.  Water was leaking to the Maypole Sink, which was unusual in relatively low flow.  But in addition, Plantation Stream was visibly shrinking in size between the Maypole Overflow corner and the weir.  Pulpit Passage East Inlet was again measured directly.  Stream ratios along the stream passage gave 140 1/min for Plantation Inlet at P.1., compared with 45 1/min at the weir, giving a measure of the leakage between the Overflow Corner and the weir; 95 1/min.

In 1969 and 1970, Dr. Tim Atkinson measured surface characteristics of Plantation Stream and St. Cuthbert's Stream weekly, part of a larger study of swallet and resurgence streams of Mendip.  His data (3 and 4) agreed well with those of Stenner.  During Atkinson's study, there was serious deliberate damage to the stream ways.  The leak through the stream bed increased, and much more of the water now sank close to the overflow corner.

In 1973, the bank at the corner was dug open, so most of Plantation Stream flowed down into the valley. The safety aspects of this action were discussed in an article in the B.B. of April 1974 (5), which contained the observation that the situation "need not be dangerous provided that it remains possible to put the stream rapidly back into Plantation if necessary."  But Plantation Swallet was filled in during the summer of that year.

For an uncertain number of years, Plantation Stream sank at the Maypole Overflow corner in dry weather, overflowing into the depression in normal conditions.  Dye tests proved that the stream still reached its old route through the cave.  At some date between 1978 and 1989, there was a major change.  At the site of old sluices which once fed water into the leadworks, the entire outflow of the Mineries Pool was diverted into the depression, joining the smaller St. Cuthbert's Stream to flow down the valley to the cave.

 

1966

 

 

 

 

 

 

1967

 

1968

 

 

1971

1973

DATE

26.03

23.04

1.06

22.06

27.07

20.08

23.10

28.01

29.11

28.01

10.02

9.11

11.04

19.08

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

DISCHARGE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SURFACE SINKS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Plantation Swallet

 

1200

 

280

5

 

900

415

300

 

165

140

 

 

Maypole Sink

 

90

 

 

 

 

38

44

 

8

52

 

 

 

Soak away Sink

 

90

 

 

 

 

56

38

 

17

9

 

 

 

Culvert

 

30

 

 

 

15

38

 

9

6

 

 

 

 

P.S. Weir

 

1200

303

284

5

 

940

415

303

140

151

45

 

 

CAVES SITES

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pulpit Pass. E. Inlet

 

50

 

 

 

 

38

19

 

11

9

 

 

 

Pulpit Pass. W. Inlets

 

30

 

 

 

 

15

19

 

5.6

0.0

 

 

 

Pulpit Pot

 

75

 

 

 

 

56

38

 

17

9

 

 

 

Disappointment Pot

 

 

 

 

 

 

 

 

 

0.6

0.0

 

 

 

Drinking Fountain

 

35

 

 

 

 

15

23

 

2.3

19

 

 

 

Old Route Stream

 

41

 

 

 

 

15

38

 

9

5.5

 

 

 

Maypole Stream

 

48

 

 

 

 

23

11

 

6

33

 

 

 

M.S. Traverse Ch. Choke

 

200

 

 

 

 

 

106

110

 

35

67

 

 

M.S. upstream of P.J.

 

200

 

 

 

 

155

106

110

 

35

67

 

 

Plant. S. upstream og P.J.

 

1200

 

 

 

 

940

415

300

 

165

140

 

 

M.S. d’stream of P.J.

 

1400

 

 

 

 

1100

520

410

 

200

160

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

DISCHARGE RATIO

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pulpit Pot : Old Route

 

1:8:1

 

 

 

 

2:4:1

3:5:1

 

0:49:1

0:27:1

 

 

 

M.S. :  Plant. S. at P.J.

 

1:6:0

 

1:10:0

1:20

1:13

1:6:1

1:3:0

1:3:0

 

1:4:8

1:2:1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

TOTAL HARDNESS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pool Exit

77

 

 

 

 

 

81

57

 

 

 

 

 

 

Plantation Swallet

108

77

105

114

151

116

87

90

140

108

88

108

125

142

Plant. S. upstream of P.J.

 

 

 

117

146

133

113

94

142

 

114

 

 

 

PERCENTAGE CHANGES

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pool Exit to P. Swallet

40.3

 

 

 

 

 

 

33.3

54.4

 

60.3

46.4

 

 

P. Swallet to P.J.

 

 

2.6

-4.0

14.7

29.9

4.4

1.4

 

29.5

 

0.7

 

 

Table 2. The distribution of the streams on the surface and in St. Cuthbert's Swallet in 1994 and 1995; estimates of discharge at each site (l/min), and changes in Total Hardness between some sites (ppm CaC03).

 

 

1994

1995

 

DATE

25.05

15.06

11.07

3.09

DISCHARGE

 

 

 

 

SURFACE STREAMS

 

 

 

 

Pool Exit

800

270

165

0

St. Cuthbert’s

2860

470

374

20

SURFACE SINKS

 

 

 

 

Maypole Sink

1520

310

171

3

Soak away Sink

2200

325

119

2

Culvert

660

145

255

19

CAVE SITES

 

 

 

 

Pulpit Pass. E.Inlet

 

23

42

0.18

Pulpit Pass. W.Inlets

 

297

77

0

Pulpit Pot

2200

320

119

0.18

Disappointment Pot

200

80

40

0

Drinking Fountain

220

80

32

0.3

Old Route Stream

660

145

255

19

Maypole Stream

1100

155

99

2.7

M.S. Traverse Ch. Choke

4400

780

545

23

M.S. upstream of P.J.

4400

780

400*

1.0

Plant. S. upstream of P.J.

2200

1300

400*

30

M.S. d'stream of P.J.

6600

2100

800*

31

DISCHARGE RATIOS

 

 

 

 

Pulpit Pot : Old Route

3.33:1

2.21:1

0.47:1

1: 1000

M.S : Plant. S. at P.I.

1:0.50

1: 1.67

1:1.00

30:1

TOTAL HARDNESS ppm Calcite

 

 

 

Pool Exit

71.8

97.9

103.1

 

St. Cuth's Stream Culvert

103.5

146.9

170.1

166.9

M.S. upstream of P.J.

128.6

154.9

175.6

190.6

Plant. S. upstream of P.J.

209.7

205.1

206.7

189.0

However, regular winter flooding (which had seriously hindered early attempts to enter the cave) has not returned as a result of this diversion, thanks to the effectiveness of the culvert.  Completed in 1965, the culvert vastly improved control over the drainage of the valley.  If it becomes choked, there will be a high risk of catastrophic flooding.

In spite of the complete diversion of the stream from its previous route, the stream inlet at P.J. continued to flow strongly.  There were two possible explanations:

1.       An unknown stream had indeed previously joined the old Plantation Stream, somewhere upstream of the Maypole Overflow comer. This unknown stream had in the past caused the majority of the chemical changes in the water between the Pool overflow and the Maypole Overflow pool, and was now the sole source of the P.J. Inlet.

2.       Under the new regime in the valley, Maypole Sink takes much more water than before.  Maypole Sink is known to be complex, feeding Maypole Series, Disappointment Pot, and Drinking Fountain inlets.  It is possible that the supply now swamps the previous routes, activating a new route intercepting the old Plantation Swallet to P.I. route, feeding water from the Maypole Sink to P.J.  This explanation looks attractive if the survey of the cave is looked at, noting the proximity of Disappointment Pot to Plantation Swallet.

The second phase of sampling started with the aim of describing the present hydrology of the major stream sinks, and to find out more about the present P.J. Inlet. Differences caused by the diversion would be quantified.  Sets of samples were taken from the cave, and from the surface streams.  The size of the surface streams was to be measured by salt dilution.  A standard KCI solution was added with a simple constant flow apparatus, and samples were analysed for potassium by Flame Emission Spectrophotometry.

Results from 1994 and 1995 are summarised in Table 2.  It was gratifying, and amazing, that the very first set of samples (25.5.94) gave results which were conclusive. The present P.J. inlet is indeed much smaller than in the first period of study. Its water is also much harder.  In its earlier state, the stream had been described as an over fit in the cave passages through which it flowed, and there was evidence of re-solution of stal. by the stream.   It was now often super-saturated with CaC03 at PJ. (capable of depositing stal).

The first set of results, shown in Table 2, eliminated the possibility that the present P.J. inlet might be flowing from the Maypole Sink.  The P.J. stream is much too hard to allow this possibility to be true.

However, the size of the inlet stream at P.J. is still a major stream.  Hardness measurements indicated a source which is much harder than the old Plantation Stream.  Furthermore, this previously unknown source had previously joined the Pool overflow stream ON THE SURFACE, somewhere upstream of the Maypole Overflow comer.

The old stream bed upstream of the Maypole Overflow comer was looked at regularly.  It remained suspiciously "soggy" in July 1994 (a hot dry month), and throughout the wet August and September of 1994. Then on 9.11.94, following a period of sustained heavy rain, a stream was seen flowing from a deep pool. Augmented by water overflowing from Ladywell Stream, the stream joined the former course of Plantation Stream a few metres upstream of the pool at Maypole Overflow comer.

On that first occasion, the flow was sufficient to overflow into the Maypole Sink, but it quickly shrank. On 12.11.94 it was still flowing, but no longer overflowed into Maypole Sink. By 28.11.94 the flow was reduced to isolated pools of open (but flowing) water, separated by stretches of marsh. After more very heavy rain, the stream was flowing strongly on 22.1.95, once again overflowing into the Maypole Sink. Still more heavy rain followed, and the stream was still flowing on 27.2.95 and 29.3.95.  By 30.04.95, surface water was no longer visible, and the stream bed reverted to a soggy marsh.  At the end of May 1995, more heavy rain increased the flow of the stream in the valley, but it was not enough to restart surface flow of the Plantation Swallet relic stream.

On 22.1.95, when the discharge value of St Cuthbert's Stream was the highest ever measured (it was certainly bigger after the great rainstorm of July 1968, but no measurements were made), two smaller intermittent stream were found, both of which flowed into the old Plantation Stream bed near the aqueduct.

Water temperatures on 22.1.95 confirmed that the newly found stream was spring fed ( Plantation stream 7.7oC, Pool Exit 5.00C).  This fact, together with the high Total Hardness of samples (145.7 ppm on 15.1.95), settled the matter beyond doubt. 

This stream was:

(a)     responsible for most of the changes in Plantation Stream between the Pool Exit and Plantation Swallet between 1966 and 1973, and

(b)     the main source of the inlet stream now entering the cave at P.J.

Stenner considers that water held in and underneath the marshy stream bed feeds water to the stream at P.J. when the stream has ceased to flow at the surface.  In this respect the situation is similar to that at Maypole Sink (where water held in an in filled depression under the sink continues to feed water to the Maypole Series many weeks after Maypole Sink becomes dry).

The suggested explanations are consistent with all the results of the previous water tracing experiments.

The results were examined for evidence of changes in temperature or solute concentrations during the flow of streams through the cave.  The conclusions were the same as those from G.B. Cave in 1968 (6):

Along Mendip-size streamways, significant changes take place in only four circumstances;

(1) In boulder chokes between the surface and inlets in the cave;

(2) At stream junctions;

(3) In percolation water trickling over flowstone formations;

(4) When the size of the stream is changing rapidly because of a rainstorm.

(Note; circumstances are different in the very long stream passages in other caving regions, where there is sufficient time for slow reactions involving humic acids to take place, as George Bray has demonstrated in O.F.D. (7)).

On 11.7.95, a sampling trip took place very soon after heavy rainfall.  Temperature and hardness changes were noted in the stream between successive sampling points, which were not caused by stream junctions.  They had been caused by the same exceptional circumstance that had been noted in G.B.  Cave, and warned about earlier.

For example, differences were found between the top of Pulpit Pitch and the bottom of Gour Passage Pitch.  However, the time taken to move between these two stations is much longer than the time taken to measure and sample at each stream junction.  For this reason, the stream discharge estimates for the trip are considered to be valid, except for the reservation about discharge values at PJ., for a different reason to be discussed later.

In a trip on 3.09.95, Estelle collected a set of samples when the cave was spectacularly dry.  The size of the surface stream at mid-day had been only 21 1/min, the lowest measurement of the summer.  Before the surface streams could be sampled by Stenner, there was a tremendous rainstorm, and by 7.30 that evening the stream size increased almost ten-fold to 204 1/lmin.  Fortunately, Estelle had made direct measurements of stream sizes, so the results of the sampling trip were not invalidated.  Instead, the results documented some of the effects of the storm. These will appear in a full report later.

Apart from the consequences of these rainstorms, no other instances of changes in chemical or physical changes along the stream courses were detected.  It is this stability which makes it possible to use solute or temperature changes at a stream junction to calculate stream size ratios. Beyond the present specialised application, the authors are confident that this discovery is of potential value in cave exploration, because:

If at any point in a streamway, a sudden change in temperature or water chemistry is measured, then that point marks a confluence of waters from different origins.

As the study progressed, it became clear that the rainfall pattern was very unusual.  In a secondary study, surface streams have been sampled regularly, giving data for comparison with earlier data, and with any future data. This study continued through the wet spring of 1995, through the unusually hot and dry summer which followed, and into the period of sustained wet weather that ended the drought.

In the summer of 1994, the bank of the stream entering the culvert was reinforced with mud-filled sand bags.  This prevented water sinking at the soak-away sink under the North bank, which was the main supply for the Pulpit Passage inlets and the N.E. Inlet in Arête Chamber (the last-named inlet sometimes split its flow between Pulpit Passage and the Showerbath at the Ledge Pitches).  Water from the Culvert flows via Arête Chamber to the Old Route Stream. Results from 11.7.95 confirm that the surface changes have changed the distribution of the stream between the Old Route and the New Route.  The effectiveness of the total drainage of the valley has been reduced, which must be weighed against any better control over the stream.

The prolonged period of unusually high flow in the winter and spring of 1994-5 had an unexpected consequence in the cave.  Water leaks from the Main Stream at Mo's Dig (just upstream of the Dining Room). During the earlier studies, the size of this leakage was assessed from the far side of the dig, under Cerberus Hall. It was (at that time) far too small to affect the stream distributions shown in Table 1.  Since the winter floods, a major part of the stream now leaves the Main Stream at this point.  By 23.08.95, the stream bed between Mo's Dig and P.J. was dry.  A small dam was built in September to encourage flow through Mo's Dig, which may assist future work at Sump 2.  There are three thoughts:-

(a)     it would be worth looking at the effects of the extra leakage from the other side;

(b)     it would be worth keeping an eye on the Lake - with a great deal more water going through, well you never know!;

(c)     although on 11.7.95, Estelle made an effort to divert all of the flow back into the Main Stream, the assumption that the discharges of the Main Stream at Traverse Chamber and at The Sewer are equal is no longer true.

Finally, on 3.09.95, Estelle noted that Pyrolusite Stream was flowing so strongly that it was making a significant contribution to the stream in Gour Passage.  Kanchenjunga Drip and Dining Room stream were also flowing well.  The three percolation inlets had continued to flow throughout an earlier period of major drought, from the summer of 1975 to the autumn of 1976.  Mineries Pool Outlet had been dry since 13.08.95, and on the same date flow from Ladywell had ceased.  Although the Pool Exit stream restarted on 8.09.95, Ladywell had still not restarted on 26.09.95, despite the heavy rainfall since the end of August.

References.

1.                    Irwin, D.J., 1991, St. Cuthbert's Swallet, B.E.C., pp 82.

2.                    Stenner, R.D., 1971, The measurement of the aggressiveness of water towards calcium carbonate Parts II and m., CRG Trans 13(4),283-296.

3.                    Atkinson, T.C., 1971, Hydrology and Erosion in a Limestone Terrain, University of Bristol, PhD Thesis (unpub).

4.                    Atkinson, T.C., 1995, Personal communication.

5.                    Collins, S.J., 1974, Water into Cuthbert’s, BEC Bel Bul(318), 70-72.

6.                    Stenner, R.D., 1973, A study of the hydrology of G.B Cave, Charterhouse-on-Mendip, Somerset. UBSS Proc 13(2), 171-226.

7.                    Bray, L.G., 1975, Recent chemical work in the Ogof Ffynnon Ddu system: further oxidation studies, Trans BCRA 2(3), 127-132.