Data we cannot fetch ourselves — manual retrieval note

For papers/aquifers. Ordered by likely impact on the model, highest first. Everything here was attempted programmatically and failed, or is not served by any open endpoint.

Holes in the sources are expected; the point of this note is that they are named and addressable rather than discovered later.

1. BGS Aquifer Properties — the measured T and S values (max priority)

Why it is top. Transmissivity drives the leakage derivation, and leakage is what the aquifer case turns on. T is currently a published regional value (122.5 m2/d, WD/97/34 s.7.9) imported from the Otter Valley 30-50 km south. These sites would make it locally measured.

What we hold: the site INDEX - 131 locations in the box, 52 within 2 km of the Class 1A Triassic outcrop, nearest 2.3 km from the modelled store.

What is missing: the values. Probed, not assumed - the GeoIndex layer publishes these fields and no others:

AQPROP_LOC_ID, SITE_IDENT, EASTING, NORTHING, LOCATION

No transmissivity, storativity or hydraulic-conductivity column. It is a pointer to data held elsewhere.

Index endpoint (already fetched): https://map.bgs.ac.uk/arcgis/rest/services/GeoIndex_Onshore/boreholes/MapServer/9

How to get the values

1. Email ngdc@bgs.ac.uk (BGS National Geoscience Data Centre).
2. Ask for “pumping-test derived transmissivity and storage coefficient records from the Aquifer Properties database” for the sites in the table below.
3. State the use (research / feasibility) - Aquifer Properties Manual data is generally available for non-commercial use.
4. Ask for test duration with every value. WD/97/34 s.7.9 is explicit that storage coefficient in this aquifer rises with test length, and that short tests do not give a usable specific yield. A value without its duration cannot be interpreted.

Sites to quote - the 20 nearest the store that are also within 2 km of the target outcrop:

2. BGS geology at formation level — 250k or 50k (high)

Supersedes what were separate items 2 and 4: the same request settles both.

Status 2026-08-09: a DiGMapGB_250.zip was obtained but is a BGS sample tile — bounds E0–50000 N0–50000, parked at the coordinate origin, containing North Yorkshire. Zero polygons intersect the Somerset box. Quarantined at data/raw/geology/digmap250_SAMPLE_NOT_SOMERSET/. It loads cleanly as a valid EPSG:27700 shapefile, so it would have joined and clipped without error and contributed nothing.

Why this matters more than “better resolution”. The sample’s attribute table shows what the real dataset carries, and it is a different kind of information from what we have:

What we have now (625k hydrogeology) is ROCK_UNIT, CLASS, CHARACTER, FLOW_MECHA — productivity classes with no formation names and no ages. That is precisely why the polygons keep failing:

• ST12NW18 Milverton is mapped TRIASSIC ROCKS (UNDIFFERENTIATED) Class 1A and is actually Permian (Aylesbeare / Exeter Group). A MAX_PERIOD field would have caught that immediately.
• ST12NE6 Shepherd Water turns on whether “Keuper Sandstone” is the Sherwood. A LEX code on the polygon resolves straight to the Lexicon entry — so this request also closes the old item 4, and with it the risk that the layer model drops from two known tops to one.

Where:

• 1:250 000 — https://www.bgs.ac.uk/geological-data/ , ask the digital data team, digitaldata@bgs.ac.uk (the address in the sample’s own readme).
• 1:50 000 DiGMapGB-50 — https://www.bgs.ac.uk/geological-data/digmapgb/, free under the BGS non-commercial research licence. Better, if the licence is workable.

Ask for: bedrock theme, real coordinates (state plainly that the sample tile at 0,0 is not usable), covering E300000–375000, N100000–160000. Confirm the delivery includes LEX, FM_EQ, GP_EQ and MAX_PERIOD.

3. EA abstraction licences (high)

Why. WD/97/34 p.286 puts the whole Tone catchment resource at about 17 500 m3/d, “most of which currently supports environmental flows”, while every two-store configuration pumps 60 000 m3/d. That comparison is unresolvable without knowing what is already licensed and taken.

Where: Environment Agency abstraction licence data on data.gov.uk — search “water abstraction licences”; or an EA data request.

Ask for: licensed and actual quantities, split groundwater vs surface water, for the Tone catchment and the Sherwood / Otter Sandstone groundwater bodies.

4. BGS Lexicon — RESOLVED 2026-08-09 (was medium)

Closed by a click-through in the BGS viewer. No further action.

The Lexicon entry for HEY — Helsby Sandstone Formation gives:

• Parent Unit: Sherwood Sandstone Group (SSG)
• Previous names include “Keuper Sandstone Formation”, “Keuper Sandstone, Lower [Obsolete name and code, use HEY]” and “Otter Sandstone Formation [Obsolete Name And Code Use: HEY]”
• Age Anisian (Middle Triassic), 241.5–247.1 Ma
• “Pebbles may be common, particularly near the base of the formation, and thin units of hard intraformational conglomerate occur in the south-west”
• “The formation crops out almost continuously from the south Devon coast, northwards, to near Taunton, Somerset”
• Upper boundary transitional into the Sidmouth Mudstone Formation of the MMG (Mercia Mudstone Group)

Consequences

1. ST12NE6 Shepherd Water — the Survey’s “Keuper Sandstone” below 83 m is the Sherwood Sandstone Group. The second known top is confirmed, not assumed; its confidence is raised from medium to high. This was the single risk that would have collapsed the layer model to one known top and left no gradient at all.
2. The two gravel beds logged beneath the marl at Shepherd Water are diagnostic of the basal Helsby in the south-west, not incidental.
3. “Otter Sandstone” and “Helsby Sandstone” are the same unit under different names. WD/97/34 uses Otter Sandstone; the 50k map uses Helsby Sandstone. They must not be treated as two aquifers.
4. The outcrop ends near Taunton — independent support for the northward disappearance already read from the Cannington basement holes.

5. Scans BGS does not hold

ST65SW1 (390478), ST65SW2 (390479), ST65SW3 (390480) returned HTTP 404 - no scan held. All three are in the Bath/Bristol fringe, outside the study area. No action needed; listed only so the 253/256 download count is explained.

Not a fetch problem - a reading backlog

240 of 253 downloaded scans are unread, including about 19 deep holes within 5 km of the target outcrop. Nothing needs requesting; this is time. It is also the only item on this page already proven to move the model - two logs took the layer model from one known top to two.

← All notes · More from Aquifer Storage