Can any basin be filled by gravity?
Generated by build_gravity_sites.py.
Generated by build_gravity_sites.py.
The operating concept turns on this. Divertible water arrives in ~9 events a year, median 2 days and 1.12 Mm3; taking a median event whole needs 560,000 m3/day for two days, and a 150,000 m3/day pump gets 27% of it. The events are too fast to pump. So the question is whether a basin can be filled by gravity instead.
That needs the basin’s top water level to sit below the river’s high-flow level at the diversion point.
The screen is the inverse of the old one
find_reservoir_sites.py looked for large sites and ranked them by storage per unit of embankment - which rewards deep cells on rising ground. Rising ground is exactly what a gravity diversion cannot use.
| old screen | this screen | |
|---|---|---|
| volume wanted | as large as possible | 1-2 Mm3 is enough |
| ground wanted | rising, for cheap embankment | low, below the river |
| ranked by | m3 stored per m3 of fill | head available |
So a site that scored badly before may score well here, and the best site from the old screen may be the worst.
The result
| site | name | river level m | TWL at 1 Mm3 | head | to river km |
|---|---|---|---|---|---|
| S3 | Pring Field | 5.3 | 3.7 | +1.6 m | 2.3 |
| S6 | Churley Stoke Gregory | 5.2 | 4.8 | +0.3 m | 1.7 |
| S8 | Pucklechurch Ilchester | 8.7 | 9.6 | -0.9 m | 1.3 |
| S4 | Coat Sewage Scheme | 7.9 | 9.4 | -1.5 m | 0.1 |
| S1 | Podimore | 12.9 | 14.5 | -1.6 m | 0.1 |
| S7 | Wood Stawell | 2.9 | 4.9 | -2.0 m | 1.3 |
| S5 | Draycott | 15.7 | 18.8 | -3.1 m | 2.1 |
| S2 | Meare | 7.5 | 16.7 | -9.2 m | 3.5 |
Head is river level minus the basin’s top water level at 1 Mm3. Positive means the river sits above the basin and can run in under its own weight.
2 of 8 sites have positive head at 1 Mm3, before any allowance for flood depth: S3 (+1.6 m), S6 (+0.3 m).
Gravity diversion is available at these sites. That is the single most useful result in the scheme layer, because it removes the constraint every earlier chapter was fighting - the intake pump - and replaces it with a structure that has no capacity limit worth speaking of at these volumes and no energy cost.
Sensitivity to flood depth
The DEM elevation at the channel is a proxy for the river’s level; a real diversion works off the flood level, which is higher. How many sites qualify as that allowance grows:
| flood depth above channel DEM | sites gravity-fillable at 1 Mm3 |
|---|---|
| 0 m | 2 of 8 |
| 1 m | 3 of 8 |
| 2 m | 6 of 8 |
| 3 m | 6 of 8 |
What this screen cannot settle
- There is no modelled river stage. The water layer says so plainly - level is not routed anywhere but a gauge - so the channel DEM elevation is standing in for a water surface. A LiDAR DTM over a watercourse returns something between the survey-day water surface and the bank top, and which varies. This is the weakest link in the chapter and the reason the head is reported rather than just a verdict.
- Only eight candidates were tested, and they were selected against the opposite criterion. This is a test of the existing shortlist, not a search.
- Nothing about the diversion structure. A gravity intake still needs a weir or sluice, a screened offtake, and a channel to the basin, none of which is sized or costed here.
- Nothing about flood risk. A cell that fills from the river by gravity also floods from the river by gravity. That is a consent question and a safety question, and on the Levels it is not a small one.