Aquifer mapping - where we stand

Generated by aquifer_mapping_status.py. Three sources, kept separate on purpose: where they disagree, the borehole wins.

1. Every productive aquifer within 25 km of the store

Source: BGS 625k. Complete coverage, but classes only - no formation names, and it has been wrong at 8 of 13 borehole sites.

The finding that matters here: the target Sherwood is 224 km2 at 2.7 km, but the Upper Greensand is 402 km2 at 6.5 km, also intergranular, and has no representation in the model at all. The Great Oolite is bigger still but flows through fractures, which is poor for controlled storage and recovery.

2. What the 50k screenshots add

Formation names and ages, via click-through. Also the finding that one 625k polygon contains 6-9 distinct 50k units - so a 625k area is a mosaic, not a block of aquifer.

The rock column, youngest first - from core/stratigraphy.py, which is the single definition of order, age, thickness and role.

12 of 20 columned units have a colour keyed, so the rest cannot be drawn even though they are identified. A thickness of <= n is a Lexicon CEILING with no published floor rather than a range - that is DYS, CHAM, BCMU, whose low ends contribute nothing to any depth sum. Exactly one thickness in this column is measured in our own area: CDL. Every other figure is a distant basin, and the depth bounds below inherit that.

Mapped undifferentiated - polygons where BGS itself does not separate the units, so neither can we. These have no single place in the column and no single depth; what they have is a bracket.

The bracket runs from the least cover such a polygon can represent to the most. Its low end is the one to watch: it assumes the Penarth Group is at surface, which puts the aquifer a whole Blue Lias nearer than a Charmouth surface would.

Also confirmed by click, but not yet in the column (no Lexicon entry read):

How deep the Sherwood is beneath each surface unit - lower bounds, because units with no published thickness contribute nothing to them.

The ordering is certain even where the numbers are not: Sidmouth is always nearer the aquifer than Branscombe, and both are nearer than any of the Lias. That is a depth-to-aquifer constraint over the whole map, not at the 13 boreholes.

What is still unread on our own bedrock map. 171 polygons, 199 km2, carry a colour nobody has clicked (onshore only - 8 polygons and 41 km2 that the image draws over the sea are excluded, not counted as unread). The 625k is shown beside them as a PRIOR - it is wrong at 8 of 13 borehole sites, so it says only which click is worth making, never what the unit is.

75 km2 of unlabelled colour, across 15 distinct bins, sits inside 625k Lias polygons. Dyrham Formation is identified by click, entered in the column, and still not drawable for want of an RGB - so that is the ground it has to be in. Clicking one bin labels every polygon of its colour at once, which is the cheapest coverage left on the table; run georef_bedrock.py for the grid reference to click in each.

3. What the boreholes actually prove

13 logs read. This is the only direct evidence of what is underground.

Two known tops (Bishops Lydeard 28.6 m, Shepherd Water 83.0 m) give the only measured gradient: 11.5 m/km deepening, apparent dip 0.66 degrees.

4. What is still missing

The fourth row is the one that limits everything. Every area quoted from a bedrock map is the aquifer’s SURFACE expression. The scheme depends on the confined part beneath the Mercia Mudstone, and no map we have shows it - only boreholes can.

← All notes · More from Aquifer Storage