Licensed abstraction, scaled

Generated by build_demand_abstraction.py. Replaces the blanket 3 Mm3/yr with a figure that has a source - and a stated weakness.

Read this first. What follows is a licensed ceiling, scaled from nine licences in a single 100 km2 window. It is not demand and it is not use. It brackets the answer; it does not give it.

Why only 100 km2

The Environment Agency’s abstraction licence data is published under an Open Government Licence restricted to “a limited specified geographical area not exceeding 100 square kilometres and limited to internal use”. The study box is 1,333 km2. So core/abstraction.py filters server-side on a fixed 100 km2 window and no request ever asks for more; the records stay in raw/ and only derived statistics reach this chapter.

The window is E326,000-336,000 N122,000-132,000 - the Tone and Curry Moor ground that run 07’s intake would serve, and which the Curry Moor WLMP shows is already fed from the Tone.

What is in the window

The purpose split matters more than the scaling. 67% of licensed volume here is non-consumptive - one Transfer Between Sources licence on West Sedgemoor Main Drain plus a Lake & Pond Throughflow. Those move water rather than taking it out of the system. Scaling the unsplit total would have produced a number about four times the agricultural demand, and so been worse than the blanket figure it was meant to improve on.

Scaling to the study area

Not on area. The window is not a random sample of the box:

The window farms harder than the box, so raw-area scaling (13.29x) would inflate the answer. Scaling on agricultural area from CROME (12.29x) assumes instead that licensed volume per agricultural hectare is constant - still an assumption, but a weaker one.

The answer, and what it does to the model

Licensed agricultural abstraction over the study area is 2.33-3.19 Mm3/yr, the range being whether General Farming & Domestic is counted as agricultural.

The blanket 3.0 Mm3/yr the scheme runs have been using sits inside that range. So the assumption was not wrong - but it was unfounded, and it is now bounded instead. The scheme results to date do not need re-running on this account; they need re-labelling, from an assumed demand to a demand consistent with the licensed ceiling.

Why this is still a ceiling and not a demand

• Licences are held unused, at historic volumes, and sometimes defensively. A licensed m3 is permission, not need.
• Licences under 100 m3/d are excluded from publication entirely, which drops a long tail of small agricultural abstractions. This pushes the estimate DOWN.
• Nine licences. The sample is tiny and one of them carries 54% of the volume. A different 100 km2 window could give a materially different answer and there is no lawful way to check that from here.
• It is a ceiling on what may be TAKEN, not on what is NEEDED. The counterfactual the demand layer actually wants - what would have been used had water been available - is bounded below by use and above by neither of these. See the layer README.

What would improve it, in order of value

1. Abstraction returns 2019-24, in the EA’s Water Rights Trading spreadsheet under the same 100 km2 term. That gives actual use against licensed ceiling and closes the gap from the other side.
2. A wider extract by request to national.requests@environment-agency.gov.uk, which removes the nine-licence problem entirely.
3. Crop water requirement from CROME, which replaces scaling with modelling and is the only route to the counterfactual.

Environment Agency Water Abstraction Licences, England, 12 December 2023, OGL - use restricted to a geographical area not exceeding 100 km2 and to internal use. Licence records are not reproduced here; only derived totals. Crop Map of England 2024, Rural Payments Agency, OGL v3.

← All notes · More from Aquifer Storage