Sleets Gill

[Pitlamp]

Sleets Gill is one of a series of caves around the Dales which share similar characteristics:

* Fed mainly by percolation flow (autogenic).

* Act as a flood rising only.

* Normally drain to a resurgence lower than the cave entrance via immature underflow passage.

* Flood response highly unpredictable; may or may not be delayed (depending on various factors).

Some examples are:

Sleets Gill Cave
Hemplands Rising
Mammary Pot
Pool Bank Cave
Preacher's Cave

I've been associated with all of these over the years to a greater or lesser degree and have seen some alarming things. For example, when a bunch of us reopened Mammary Pot a few years ago we were chased out of what was then a dig by water rising from below during a heavy rain event. The depression in which this cave lies then became a resurgence about 15 minutes later. Normal water level is 8 m below this point. Assuming the water had been rising at a constant rate, it'd probably taken less than 60 minutes to rise by 8 m.

Another example was at Preacher's Cave. This starts as a 4.5 m shaft leading to a further descent of maybe 2.5 m to reach the main sump pool which is normally static. I kitted up at this point before heading in for a working dive at the (then) end of the sump. When I returned to dive base all my tacklebags were submerged so I scooped them up, swam up the slope and surfaced at the base of the entrance shaft. The water was rising noticeably so I decided not to dekit as normal and simply let the rising water lift me up the shaft to the surface. 5 minutes later the shaft was brim full and starting to resurge; it had been like riding up on a gentle elevator and (for once) I had no tank hauling to do.

On both these ocasions it was that day's rain after a long wet period (like now) which caused such fast water level rises. Sleets Gill Cave is analagous to these two sites - but much longer; this is why I wouldn't dream of going in there at the moment and why I think Langcliffe's 2 dry weeks rule of thumb is excellent advice.

 

[thehungrytroglobite]

Hemplands Rising
Mammary Pot
Pool Bank Cave
Preacher's Cave

I've never heard of these. Hopefully there is a warning for them on the CNCC guides (if they exist)?

 

[Pitlamp]

Ask me about these, next time paths cross. But in a nutshell Mammary Pot and Pool Bank Cave are flood rising caves at the foot of Whitbarrow. Hemplands Rising is a cave in Wharfedale (and potentially related to the Mossdale to Black Keld hydrological system) entered 3 or 4 years ago. Preacher's Cave is at the foot of the large outcrop of limestone on the west side of Wild Boar Fell.

I'm not sure warnings are as important as that for Sleets Gill because none of them would trap you if the water rose fast (unlike Sleets Gill). Worst case scenario is you'd be chased out by the water. I used these examples because they illustrate what can happen with the water in this kind of percolation-fed flood rising cave.

 

[aricooperdavis]

Just had a hunt through my old files - here's a nice clear rainfall event and cave water depth response from BH's data. Ignore the flood level line, that's a ridiculously conservative estimate!

From rising limb of hyetograph to rising limb of hydrograph is about 12 hours, but this was with relatively little antecedent rain.

 

[IanWalker]

I've never heard of these. Hopefully there is a warning for them on the CNCC guides (if they exist)?

This is easy to check by using the cave info search tool. Partial names match (eg Simp finds Simpsons Pot)

Preacher's Cave has an entry. No hazard but it does suggest contacting CDG before entry.

Preacher's Cave

Access information for Preacher's Cave (Mallerstang)

cncc.org.uk

 

[aricooperdavis]

Another interesting point is that if you take the rainfall over individual grid-cells and work out the correlation with the cave water depth for each one you can map this correlation in space. You get four distinct clusters - areas on the map where if rain falls it impacts the cave water depth quite differently.

Here are the clusters:

And here are the associated mean "delay" between rain and cave water depth change:

 

[IanWalker]

Just had a hunt through my old files - here's a nice clear rainfall event and cave water depth response from BH's data. Ignore the flood level line, that's a ridiculously conservative estimate!

From rising limb of hyetograph to rising limb of hydrograph is about 12 hours, but this was with relatively little antecedent rain.

3rd May 2021 looks like a bad day to go to Sleets Gill. I was curious what happened in the lead-up.

Rainfall low for about a month until 2nd/3rd May

The Wharfe shows generally flow levels had been low and falling for about a month until 2nd/3rd May.

Conclusion: 2 weeks of dry weather doesn't save you if it chucks it down!

 

[Pitlamp]

Yep; expect the unexpected.

 

[andrewmcleod]

What is


3rd May 2021 looks like a bad day to go to Sleets Gill. I was curious what happened in the lead-up.

Rainfall low for about a month until 2nd/3rd May

The Wharfe shows generally flow levels had been low and falling for about a month until 2nd/3rd May.

Conclusion: 2 weeks of dry weather doesn't save you if it chucks it down!

In this case (not necessarily to be extended to other cases) it would give you 12 hours, plenty of time for a normal caving trip even if the heavens burst immediately after you entered. If you went into Sleets Gill while it was raining, then you only have yourself to blame. But 12 hours even in the case of heavy unforecast rain is not too bad.

Whereas if it hasn't been dry, and then it only took 2 hours to flood (hypothetically) you are buggered.

 

[IanWalker]

Another interesting point is that if you take the rainfall over individual grid-cells and work out the correlation with the cave water depth for each one you can map this correlation in space. You get four distinct clusters - areas on the map where if rain falls it impacts the cave water depth quite differently.

Here are the clusters:

And here are the associated mean "delay" between rain and cave water depth change:

@aricooperdavis

I find this very interesting. Am I reading this correctly, that you can predict the extents of the rainfall catchment for a given cave? This would be very useful for - for example - caves that run beneath surface watersheds, and for finding hydrological connections between systems.

Example:- Northern Sump Index suggests Crescent Pot (East Kingsdale) could drain to Keld Head and Dale Barn Cave. Could your analysis show the liklihood of a Dale Barn connection, and rainfall events that cause it to flow?

 

[aricooperdavis]

@IanWalker with a lot of caveats, yes. The problem with this analysis is that, of course, it's correlational not causational. Rain doesn't tend to fall over a single grid-cell, and removing the conflating rainfall from adjacent grid cells requires an awful lot of rainfall and flow data.

 

[Pitlamp]

In this case (not necessarily to be extended to other cases) it would give you 12 hours, plenty of time for a normal caving trip even if the heavens burst immediately after you entered. If you went into Sleets Gill while it was raining, then you only have yourself to blame. But 12 hours even in the case of heavy unforecast rain is not too bad.

Whereas if it hasn't been dry, and then it only took 2 hours to flood (hypothetically) you are buggered.

Let's not forget the additional complication of lying snow which:

1. can melt, raising water levels even when there's been no rain for some time. (There is an account of this almost trapping a BPC party beyond The Wallows in Ingleborough Cave, for example.)

2. can be melted by rain thus producing a flood supplied by two or more precipitation events all coming at once.

I'm always a bit wary of trying to related the scale of a flood to weather data leading up to the flood. The problem is that some folk may rely on what they perceive to be a prediction, which may not take into account other variables and thus not be as reliable as might be thought.

What's that saying about guidance of the wise rather than obedience of fools?

 

[Steve Clark]

This is good stuff.

I'm interested to know how the rainfall data knows how much rain fell in each grid square in unit time? They look like 1x1km squares so presumably there's no physical on site monitoring at this scale. If it's radar, how much actual data is in it, or is it a smoothing exercise calibrated by actual local rain gauges? Also the monitored rain gauges being at a different altitude and temperature to the radar of actual storm cell dumps of rain up on the hill.

Is there a risk of the analysis being at a far higher resolution than the 'true' source data?

 

[IanWalker]

This is good stuff.

I'm interested to know how the rainfall data knows how much rain fell in each grid square in unit time? They look like 1x1km squares so presumably there's no physical on site monitoring at this scale. If it's radar, how much actual data is in it, or is it a smoothing exercise calibrated by actual local rain gauges? Also the monitored rain gauges being at a different altitude and temperature to the radar of actual storm cell dumps of rain up on the hill.

Is there a risk of the analysis being at a far higher resolution than the 'true' source data?

Looks to be estimates, interpolated from rain gauge data using a model

uk rainfall grid data at DuckDuckGo

DuckDuckGo. Privacy, Simplified.

duckduckgo.com

"The higher resolution therefore does not necessarily provide better representation of micro-climate effects."

 

[aricooperdavis]

Let's not forget the additional complication of lying snow

There are a couple of met office MIDAS weather stations nearby that allow you to guesstimate whether that could be an issue. It wasn't over the period of BHs data:

I would be surprised if the melting of surface snow by rain had a <12 hour response time, but it would certainly could melt without rain and cause potentially unexpected water level rises.

I'm interested to know how the rainfall data knows how much rain fell in each grid square in unit time?

The rainfall data is from NIMROD "nowcasting". It's generally pretty good, but there is a fair bit of interpolation going on there, so knowing whether a pattern you're observing is a physical one or a consequence of the interpolation is hard to say.

 

[JasonC]

Conclusion: 2 weeks of dry weather doesn't save you if it chucks it down!

... but conversely, if I understand all the foregoing comments (for which, many thanks) - if there have been (say) 5 or more dry days, and you are confident that today will also be dry, then the flood risk should be very low. At worst, you'd find the entrance unexpectedly sumped, but if it was clear, then you'd be ok to continue. Is that reasonable, or still too risky, in the opinion of experts?

 

[Pitlamp]

I don't in any way claim to be an expert - but my own choice would be to leave it until conditions are more certain. There's a lot of groundwater about currently.

 

[aricooperdavis]

I didn't see much sign of the cave water depth responding to rainfall after about 48 hours in the dataset that I was looking at, but it was very limited, so some things just might not have happened over that period (e.g. big snow melts, unusual hydrological phenomena). The question is, I suppose, how confident are you that it won't rain today? If the ground is already wet then you'd want to be confident beyond any doubt, which seems unfeasible.

Edit: and there are a number of reports of it responding long after this, so it must happen sometimes!

 

[langcliffe]

I don't in any way claim to be an expert - but my own choice would be to leave it until conditions are more certain. There's a lot of groundwater about currently.

I concur. There are three places where I don't want to be stuck on the far side of: the Drown and Glories in Mossdale Caverns, the Near Wallows in Ingleborough Cave, and Hydrophobia in Sleets Gill. The first two are reasonable predictable, but the general opinion seems to be that Sleets Gill ain't.

Having said that, those with whom I have been caving will know that I'm a wuss, and the place is probably readily accessible outside the periods I allow myself to venture in.

 

[Pitlamp]

You're not a wuss; you're a survivor.

Thing is, a caver might get away with playing water roulette 99 times out of 100 but, without such reasoned caution, eventually it'll all go wrong.