The pond in the south-west Corner of Towerhouse Wood (ST473717), near Nailsea, North Somerset, owned by The Woodland Trust, produces bubbles of gas in large quantities (see photograph), though it appears that nobody has hitherto tried to find out the nature of the gas, or proposed any explanation for its formation here. This pond (about 5 metres in diameter) is one of several along a stretch of a stream fed principally by springs, eventually flowing into the river Land Yeo. Some gas bubbles are produced in the stream itself and at least one other pond in this stream shows gas production, though not with the same intensity as the area in question.
The wood is situated on a south-facing hillside, with Carboniferous Limestone (Clifton Down Limestone) above and Triassic Mercian Mudstone below. A fault line is marked on the geological survey map at a position very close to the pond. Coal deposits are confined to the south of this fault. Since the pond was in close proximity to the Nailsea coal measures, it seemed possible that the gas was derived from methane. Samples of the gas were collected on 16/3/1997 and sent for analysis by British Gas (GHG product code 97/453).
An inverted stainless steel tray (about 30 cm x 50cm) with a metal tube in one corner was placed in the pond. All of the gas collected under the tray then passed into the metal tube. A 500 ml glass bottle with screw-top stopper was washed several times with pond water, and while still full and with the open end under water, inverted and placed over the tube to collect the gas. Great care was taken to prevent atmospheric contamination. When the bottle was filled, the open end was covered with aluminium foil and the cap screwed on while still under the water, leaving a small amount of water in the bottle to secure the seal. A second bottle was filled in the same way. The gas was produced at about 200 ml per minute per square metre. The bottles were transported, still in the sealed and inverted position to the analytical laboratory. The samples were subjected to gas chromatography using a thermal conductivity detector with a Porapak column. Gas eluted from the column was oxidised using a copper oxide furnace, oxidising any potential methane to carbon dioxide that was trapped cryogenically. Carbon dioxide by-passing the furnace was also trapped. Methane could not be found, and this was confirmed using a capillary column with a flame ionisation detector; nor could any higher homologues of aliphatic hydrocarbons or hydrogen sulphide be detected.
Analysis of the gas showed nitrogen 73.48%; oxygen 18.37%; carbon dioxide 8.15%. The ratio of nitrogen to oxygen was very similar to that of air (78.08% nitrogen and 20.95% oxygen), but the concentration of carbon dioxide was significantly higher than that normally found in air (0.033%). The possibility of sample leakage was unlikely, and if it did occur, would presumably lead to the gas showing a greater similarity in composition to that of air, and would therefore not account for the enhanced CO2.
Carbon exists in Nature as two stable isotopes, carbon-12 and carbon-13. Biochemical processes that depend on carbon assimilation, for example photosynthesis, discriminate between these two forms and tend to utilize the lighter isotope, carbon-12, at the expense of the heavier isotope, carbon-13. The measure used to determine the depletion of carbon-13 depends on a technique known as mass spectrometry, which can very precisely separate the isotopes and measure the amounts present, expressing the result (delta13C) as parts per billion (parts in 1000,000,000). The indexation standard is set by the value found in the carbonate deposited in a Jurassic belemnite. Depletion is noted as a negative value, hence atmospheric air now has a delta13C value of –70/00. Correspondingly carbonate rocks tend to have enhanced amounts of the heavier isotope (delta13C +80/00), probably because the biological processes have utilized the carbon-12 isotope, leaving a larger proportion of the carbon-13. Biogenic methane and natural gas normally has delta13C of
–400/00 to –700/00. Bacterial oxidation would tend to enhance this depletion. Using mass spectrometry, the depletion of carbon-13 in the CO2 in the pond gas was found to be -23.150/00. This is within the range for bacterial samples (-20 to –30 0/00) (Fauré, 1986), indicating that the CO2 is not derived from carbonate or methane.
The absence of methane or its higher homologues does not necessarily indicate that these gases are not formed initially in small concentrations, since any of these hydrocarbons could be oxidised by bacteria in passage through the substrate.
There was no evidence of a major flow of water in the pond, though the possibility of some up welling was likely. No springs were found in the vicinity of the pond The bottom of the pond was of small rocks, and there was little evidence of organic matter. The bubbles appeared at irregular intervals from single origins with intervals of several minutes, over almost the entire surface of the pond. The pond has not been known to dry out in the six years during which it has been under observation, although the water level fluctuates about 25%. However, the rate of gas flow appears to be independent of the level of the water.
Ponds in other areas of the UK show similar phenomena (Gunn and Bottrell, 1997), though not with such high CO2 concentrations as in the present case. It is only possible to speculate on the reason for the flow of gas emanating from this pond and for its enriched CO2.
Perhaps the most likely hypothesis is that the gas is compressed during flow of water through one or more caverns in the hillside, and is released when the water rises to the surface. (Stanton, 1982). This should imply that the gas flow would depend on an erratic flow of water, and would show as diminished gas production when the water flow is decreased. This does not appear to happen in this pond, since the gas flow appears to be independent of the water level.
Another possible explanation might arise if water flowing by gravity in an underground passage were to become heated, and was unable to re-equilibrate on rising to the surface. The gas expelled would in theory then contain a greater proportion of carbon dioxide. However, the maximum proportion of CO2 that could be accounted for by this hypothesis would be 2.9% (Parkes and Mellor, 1939).
The enhanced CO2 might be explained by the bacterial decomposition of vegetation in the soil. The pond itself was almost devoid of growing plants, being shaded by trees. The rate of CO2 production amounted to about 17 kg per year per square metre. There did not appear to be sufficient organic matter in the basin of the pond to provide this; nor would this theory account for the nitrogen and oxygen in the gas. It is possible that two mechanisms are operating, one for the latter gases and another for the CO2 . However if the CO2 is produced by bacterial oxidation, it might seem strange that the oxygen / nitrogen ratio is not significantly lower than that of air, the calculated oxygen depletion being only 1.5%, which is insufficient to account for the enhanced CO2.
The author would welcome suggestions on possible mechanisms for the formation of these gas bubbles.
- Fauré, G., Principles of Isotope Geology, 2nd edition, John Wiley & Sons, New York pp. 491-512 (1986)
- Gunn, J. and Bottrell, S., Bubbling Springs: A Request for Information. British Cave Research Association, 24, 139 (1997).
- Parkes, G.D. and Mellor, J.W., Mellor’s Modern Inorganic Chemistry, Longmans, London p.163 (1939)
- Stanton, W.I., Mells River Sink; A Speleological Curiosity in East Mendip, Somerset. Proceedings of the University of Bristol Speleological Society 16, 93-104 (1982).
The author is very grateful to Dr Willie Stanton for his encouragement and for useful discussion. Thanks are also due to Dr Nigel Evans of British Gas, Loughborough for his interest and for providing the gas analysis. I am also grateful to Christopher M.G. Smith for his help in collecting the gas.
Published in Pennant 28, 20 – 24 (Jan 2001)