In August this year, the Ministry of Housing, Communities and Local Government posted a blog titled Exploring the use of artificial intelligence to streamline the planning process:
Planning documents often contain vast amounts of valuable information, but accessing and using this data can be challenging. Artificial intelligence (AI) offers a potentially promising solution by efficiently processing and interrogating large volumes of data to extract key information and insights.
Draft minutes of the October 2024 meeting of the South West Aggregates Working Party – a group referred to in this post – also talked about the use of artificial intelligence in the planning process, particularly in the analysis of consultation responses:
EIW, LM and TB had recently attended a POS [Planning Officers Society] meeting. Various matters had been discussed including safeguarding, the Finch case (and adopting a cautious approach) and the use of artificial intelligence (AI) in the planning process, particularly in the analysis of consultation responses.
But artificial intelligence can work both ways, not only saving planning officers time in the analysis of hundreds or thousands of consultation responses, but also allowing Joe Public to sense-check consultants’ reports.
Aggregate Industries has this month submitted further information to Devon County Council in relation to its planning application for Hillhead Quarry near Uffculme – DCC/4399/2024:
Variation of conditions 2, 4, 6, 7, 12, 19, 22 and 25 of ROMP permission DCC/3655/2014... to vary the phasing; landform; drainage; and restoration of the site.
We had posted about this application earlier this year – here and here – which in part is to deal with the ponding problem at the quarry, and in part to deal with unsafe quarry faces – which would, fortuitously for the company, yield a further 460,000 tonnes of sand and gravel in the process.
In June, we wrote to Devon County Council saying:
We are concerned that the ES Chapter 9 Hydrological & Hydrogeological Impact Assessment Version 2 May 2024 for the above application is incorrect, more specifically that the calculated soakaway rate at Houndaller Plantation Pond of 5,000 m3/d is impossibly large.
The soakaway rate being pushed by Aggregate Industries and its consultants – five thousand cubic metres per day, equivalent to 1.8 million cubic metres per year – was used to conclude that:
5.8.8.7 ...Houndaller Plantation Pond has sufficient soakaway capacity for the design storm.
We pointed to a flaw in the author’s calculations, suggesting that the above assumption was incorrect. Devon County Council’s Flood Risk team objected to the application, and also asked for "Calculations for the current outflows from Houndaller Pond (abstraction / groundwater) to be submitted."
Aggregate Industries has now fundamentally redesigned its surface water management plans:
Following discussions at the site meeting held on 4th July 2024, an updated surface water management scheme has been prepared where, following mineral extraction operations, all surface water will be managed in the former extraction area, with Houndaller Pond only acting as an exceedance route.
Consultants have also had a stab at "Calculations for the Current Outflows from Houndaller Pond":
AIUK abstracts water from the groundwater-fed Houndaller Pond in accordance with Abstraction Licence No. SW/045/0002/055.The quarry operator has confirmed that they make full use of this licence. The rate of abstraction is set at 14 l/s, equating to 50 m3/hr. The annual rate of abstraction comes to 180,000 m3/year.This abstraction activity is a sustained process, occurring day after day, and the average groundwater level contours (Figure 5 in BCL’s report) are only c.1 m below the maximum levels (Figure 6), showing that abstraction (drawdown) impacts on water level in the pond are relatively limited.The fact that the pond can sustain long-term and regular abstraction at 14 l/s means that it would be similarly effective as a soakaway. It follows that occasional input to Houndaller Pond as a result of storm runoff would not be expected to raise the long-term water level in the pond.
So Aggregate Industries' consultants now appear to have moved from a soakaway rate of 1.8 million cubic metres per year to one approximating just 10% of that. However, even that is a leap.
Look at this claim again:
The fact that the pond can sustain long-term and regular abstraction… means that it would be similarly effective as a soakaway.
Is it correct? similarly effective?
We asked Google’s artificial intelligence chatbot Gemini to assess the above paragraph for accuracy. One of the Points Needing Clarification or Improvement was:
* "Similarly Effective as a Soakaway":While the pond can sustain the current abstraction rate, it's not necessarily directly comparable to a soakaway. A soakaway is designed to absorb and disperse water into the surrounding soil, while a pond primarily stores water.
Could Gemini be hallucinating?
We asked OpenAI’s ChatGPT instead: If a groundwater-fed pond can sustain long-term and regular abstraction would it be similarly effective as a soakaway?
If a groundwater-fed pond can sustain long-term, regular abstraction, it could theoretically serve some functions similar to a soakaway, but there are important differences in design, function, and environmental impact to consider.* Soakaway: A soakaway is designed specifically to manage surface runoff, allowing water to percolate directly into the ground and recharge the groundwater at a controlled rate. It's typically used in areas where rainwater needs to be quickly drained to avoid surface flooding.* Pond: A groundwater-fed pond typically has a different purpose, such as wildlife habitat, aesthetics, or water storage. While it may lose some water through infiltration, its primary function isn’t to allow for rapid or direct infiltration into the subsurface layers like a soakaway…. etc etc
It all sounds like common sense, but just to be sure, we also asked a human hydrogeology expert whether the pond would be similarly effective as a soakaway. He said:
Not true. No. No. No.
He explained that it is normally many times harder to get water into the ground than it is to take it out.
So what are Aggregate Industries’ hydrogeologists playing at? A simple error? A wild assumption? Or just trying to hoodwink the council?
The hydrogeologists working on the ponding problems at Hillhead, are the same ones working to implement Aggregate Industries’ permission to quarry Straitgate Farm, who will calculate infiltration rates and produce a surface water management plan for the site, a site that sits above flood-prone Ottery St Mary. It’s important they know their stuff, and worrying if they don’t.
