Forecasting river levels

IanWalker

Active member
The GOV.UK river level website shows current river levels for the Environment Agency's automatic monitoring stations around the country.

Now they have added an experimental tool to the Boroughbridge station that forecasts river levels 36 hours in advance.

https://flood-warning-information.service.gov.uk/station/8028

If expanded across the country this could be a very useful feature for anyone affected by river levels.

I understand the EA would like people to visit the page and use the feedback button so they can gauge interest for expanding the service to other locations.

Mention your feedback is on the river level FORECAST service.
 

Attachments

  • Capture.PNG
    Capture.PNG
    62.8 KB · Views: 279

Pitlamp

Well-known member
Am very grateful that you've flagged this up - which is potentially very useful.

Have just sent appropriate feedback, as requested.
 

kay

Well-known member
It could be worth including in your feedback the risk that flooding poses to cavers, as this is something they may not be aware of.
 

Simon Wilson

New member
This could be very good for cavers because what we often need to know is whether water is expected to be rising or falling. I've left feedback and asked for the service to be extended. I mentioned that I often drive through fords and have to plan my journeys according to my guess of river levels.
 

Pitlamp

Well-known member
The other obvious benefit of this in in rescue situations. Anyone remember the Sleets Gill job where the decision to dive out the two trapped cavers was mainly based on a consensus of what the water was likely to go on to do? All predictive tools which help arrive at such difficult decisions are of value.

I've already given my feedback but if someone else would like to mention this aspect, it would perhaps add a bit more weight.
 

mikem

Well-known member
Other rivers also have predictions:
https://www.ukriversguidebook.co.uk/forum/viewtopic.php?f=3&t=128160

Mike
 

Alex

Well-known member
Having compared the screenshot to the amount the river actually rose it does not seem very accurate. Yes there predicted rise did happen but to a much lower extent.
 

mikem

Well-known member
Funnily enough they can't currently predict how much rain will fall with any great accuracy, especially across an entire river catchment - if you check my previous link you'll see 3 of the rivers did follow the graphs, whilst the other 2 didn't, a pretty  good hit rate. Just treat it as another useful forecast & it will improve...

You also have to remember that most gauges are well downstream of the caves, so it's showing what has already happened.

Mike
 

IanWalker

Active member
Alex said:
Having compared the screenshot to the amount the river actually rose it does not seem very accurate. Yes there predicted rise did happen but to a much lower extent.
Taking this one example, it was forecast to:
rise through wednesday night, peaking at 11.5m on Thursday morning then dropping slowly.

What actually happened was :
rise through wednesday night, peaking at 10.5m on Thursday morning then dropping slowly.

General trend was spot on, and the minor difference of 10.5 versus 11.5 seems pretty good to me. In line with how i use weather forecasts for example. And definitely better than i would expect to have guessed 36 hours ahead.

If you feel able to guess better then feel free not to use the service! For the rest of us i expect this will be a great benefit in future years if developed further.
 

IanWalker

Active member
Points made about the location of river level gauges being too far downstream for our upland caves could be a useful thing to include in any feedback.

But there are already two on the Ribble at Langcliffe and Settle, many on the Wenning downstream of Clapham, two on Rawthey at Sedbergh, one on Wharfe at Kettlewell.

36 hour forecasts at these locations should more or less tell you what is expected in the caves, assuming it doesnt take too long for the cave water to reach the stations.

Should be enough to help judge general water levels in the caving region i think. And yes always with a bit of consideration and a pinch of salt and a look at several weather forecasts for example.

But its another bit of information that helps people make sensible plans and avoid driving up to a chosen trip only to find it impassable, dangerous, or not fun.
 

Alex

Well-known member
Taking this one example, it was forecast to:
rise through wednesday night, peaking at 11.5m on Thursday morning then dropping slowly.

What actually happened was :
rise through wednesday night, peaking at 10.5m on Thursday morning then dropping slowly.

Well your rounding, the difference is more like 1.5m. To me it is a good idea but it has to be more accurate for me to use it, otherwise I may as well just put a finger in the air and guess based on normal forecasts. Lets see what it does this weekend, as it is currently predicting almost flood levels by Sunday (12.5~m), yet the forecast is not showing that much rain at least less than what it has already been falling. But I know there are other factors, so lets see what happens.

What would be great is if the predicted dotted line stayed there for past dates, so we can compare predicted against actual and then we can tell for our selves when we look at these, how accurate the predictions generally are per river.
 

mikem

Well-known member
marysboy said:
General trend was spot on, and the minor difference of 10.5 versus 11.5 seems pretty good to me. In line with how i use weather forecasts for example. And definitely better than i would expect to have guessed 36 hours ahead.
Not really a minor difference when the river is only showing a less than 2 metre variation in levels, but as you state it's the general trend that is more important.

You do need to look at where the gauges are relative to your trip, as levels will often be rising there well before they are shown on the charts (or some caves well after, depending on the flow regimes) e.g. in South Wales the Pontneddfechan gauge is well downstream of Porth yr ogof, but before the little Neath river joins it, whilst the Craig y Nos one on Tawe is close to the outflow from Dan yr ogof & is usually a good indicator for ofd (although they source on opposite sides of the valley, so don't always correspond).

The most important thing is cavers don't take it as gospel - if the chart shows the river will rise at 4 pm that doesn't mean the cave will fill at that same time...

Mike
 

mikem

Well-known member
Alex said:
Well your rounding, the difference is more like 1.5m. To me it is a good idea but it has to be more accurate for me to use it, otherwise I may as well just put a finger in the air and guess based on normal forecasts. Lets see what it does this weekend, as it is currently predicting almost flood levels by Sunday (12.5~m), yet the forecast is not showing that much rain at least less than what it has already been falling. But I know there are other factors, so lets see what happens.

What would be great is if the predicted dotted line stayed there for past dates, so we can compare predicted against actual and then we can tell for our selves when we look at these, how accurate the predictions generally are.
The difference is closer to marysboy's statement - if you touch the chart it shows you the previous peak was 11.59, the forecast was slightly lower than that, the actual peak was 1O.46, so 1.1m at most...

At which point would you like them to freeze the forecast, as it is being regularly updated the chart would just be covered in possible outcomes?
 

Alex

Well-known member
I was thinking of a snap shot. i.e. Select a prediction that started two days ago and superimpose that on the graph. You can choose when the prediction was made by a drop down menu or something like that.
 

mikem

Well-known member
I'm sure they are doing that themselves, but think it unlikely they will make it public as it won't be a feature on the finished product. Just take your own image & compare for gauges you are interested in.
 

mikem

Well-known member
This website nicely combines rainfall & river levels in a single graph (although it is for fishermen & not a prediction):
https://www.farsondigitalwatercams.com/locations/stainforth

Mike
 
Top