Quick answer: For the week of 28 August to 3 September 2026 England had its wettest week since early February — 26mm in the South East to 48mm in central England. Rivers responded at almost every monitored site and abstraction restrictions fell by more than half. Groundwater fell at every site, reservoir storage dropped to 58% (19.2% below average), 11 reservoirs are under half full and all ten areas in drought stayed in drought. A surface that is finally wet says nothing about where the water table is.
What the rain did — and didn’t do
The Environment Agency’s report for 28 August to 3 September 2026, published on 4 September, is the first this summer to open with good news. Rainfall improved river flows, soil moisture deficits and environmental conditions across much of England. The late-month rain took August to 78% of the long-term average nationally — although the South East managed only 54%.
Rivers responded. Flows increased at almost every site the agency reports on, though most are still normal or lower for the time of year: 25% below normal, 20% notably low and one site exceptionally low. A week earlier, 19% were exceptionally low.
Abstraction restrictions halved. There are now 608 restrictions on abstraction licences, down from 1,412 the week before, because flows in some catchments rose back above their triggers. The formal Section 57 restrictions on 302 licences in East Anglia remain until 31 October, and 1,123 voluntary restrictions remain across East Anglia, Thames, Wye and Severn.
Then the other half of the report.
Groundwater fell at every site in England. The seasonal recession continued regardless of the rain. Half of chalk sites are classed as normal for the time of year, including those in the South East and East, but sites in the Wessex Downs, Yorkshire Wolds and the North and South Downs are below normal or lower. Tilshead, in the Upper Hampshire Avon chalk, remains exceptionally low, as does Jackaments Bottom in the Cotswold oolite. Skirwith in the Eden Valley sandstone is below normal.
Reservoirs kept draining. Storage for England was 58% for the week ending 1 September, down another 1.8 percentage points and now 19.2% below average. Eleven reservoirs are less than half full. All but one are classed below normal or lower, and nine are classed exceptionally low — the report’s list runs Ardingly, Clatworthy, Clywedog, the Dove Group, Grafham and Rutland, Hanningfield, Haweswater and Thirlmere, and Wimbleball. Haweswater and Thirlmere are new to that list — both United Utilities reservoirs in the North West — while Blithfield has dropped off it.
Nothing changed on the map. Ten Environment Agency areas remain in drought and four in prolonged dry weather. Temporary use bans remain in force at ten water companies. Wessex Water moved from drought level 1 to level 1b. South East Water applied on 28 August for a drought permit on the River Ouse; Southern Water’s drought order application for the River Test had its public hearing on 25 August. The agency’s line is that “further significant rainfall is needed to return to normal water resource conditions for many water companies”.
Why a wet week is the wrong week to test a soakaway
We wrote two weeks ago about the risk of running BRE Digest 365 infiltration tests in the driest ground in living memory: the test result can flatter the design, and the winter water table — which the method assumes you know — is nowhere to be seen in an August pit.
This week’s report adds the opposite trap. After 40mm of rain, the top of the profile is wet, the test pit fills with less effort and the temptation is to treat the result as a wet-weather test. It isn’t. The rain has wet the soil; it has not touched the aquifer. The Environment Agency’s own numbers show groundwater still receding at every monitored site, which means a test carried out this week is still a test at or near the lowest water table of the year. Where BRE Digest 365 asks for evidence of the highest likely water table — lowland sites, land near a watercourse, anywhere on the chalk or limestone that the report lists as exceptionally low — that evidence still has to come from somewhere else: borehole records, published aquifer data or a monitoring period through the winter.
There is also a smaller, practical point. A single wet week on baked ground produces uneven wetting — the agency has noted all summer that rain on hard ground runs off rather than soaking in. A pit dug into partially re-wetted soil can give a less repeatable result than either a dry-season or a genuinely wet-season test. BRE Digest 365’s rule of three fills, taking the lowest, exists for exactly this reason. Record the antecedent rainfall with the test date, and if the three results diverge, treat that as information rather than noise.
And on the water supply side
For water efficiency and water neutrality evidence, the report changes nothing about the argument set out in our last drought update: reservoirs at 58% and falling, two North West reservoirs now exceptionally low, and a growing list of suppliers in formal drought response. Demand has eased with the cooler weather but the agency notes it remains high for some companies, particularly in the South East.
If you are preparing a Part G calculation or a water neutrality statement for a site in the Wessex, Solent, Kent, Thames, Anglian or — increasingly — United Utilities areas, the supplier’s drought status this autumn is worth a line in the report. It is context a planning officer will recognise.
Three things to do this month
Don’t rebrand a September test as a wet-season test. If the site needs a high-water-table check, it needs winter monitoring or third-party groundwater data, whatever the weather did last week.
Log antecedent conditions with every test. Rainfall in the preceding fortnight and the Environment Agency area status on the day. It costs nothing and it answers the question a reviewer will ask in March.
Size with the lowest of three. It is the method, and in a year with this much variability it is also the only defensible instinct. SoakMonkey runs BRE Digest 365 from your test results and gives you a printable design for the drainage strategy — it just needs the right test going in.
Sources: Dry weather and drought in England: 28 August to 3 September 2026 (Environment Agency, GOV.UK, 4 September 2026) · Dry weather and drought in England: 21 to 27 August 2026 (Environment Agency, GOV.UK) · BRE Digest 365 — Soakaway design (2016)