The EA's reply on the dispersion model: RP03 supplied, RP02 not held
The Environment Agency says it does not hold the dispersion model that WSP's soil reports cite (AQMAU-C2478-RP02) and sent a later issue, RP03, dated before Phase 1 soil sampling began. RP03 describes its own numerical predictions as assumption-based and not for quantitative comparison. The EA's plans comparing model and soil results were withheld under regulation 13 (personal data), and several questions about how the sampling was designed are answered only by reference to WSP's published reports.
- Environment Agency, response to EIR request EIR2026/46693, covering letter, 22 October 2026. Request and response on WhatDoTheyKnow: https://www.whatdotheyknow.com/request/aqmau_c2478_rp02_dispersion_mode
- Environment Agency Air Quality Modelling & Assessment Unit (AQMAU), ‘AGC Chemicals PFOA Dispersion. Air dispersion and deposition modelling study’, AQMAU-C2478-RP03, 12 pages, supplied with the response of 22 October 2026 (same WhatDoTheyKnow link).
- Plot ‘2.1.P1 2 and 3 AQMAU’ (one figure showing the three soil investigation phases over the AQMAU output), supplied with the response of 22 October 2026 (same WhatDoTheyKnow link).
- WSP for the Environment Agency, Phase 1 (December 2024), Phase 2 (July 2025) and Phase 3 (March 2026) factual reports, published by Wyre Council (see the Reports page on this site).
On 29 September 2026 we asked the Environment Agency (EA), under the Environmental Information Regulations 2004 (EIR), for the dispersion model that its contractor WSP cites in the Hillhouse soil reports, and for the records of how the sampling was designed. A dispersion model is a computer calculation of how a release into the air spreads and settles. The EA answered on 22 October 2026 (reference EIR2026/46693); the normal deadline was 27 October. The request and the response are public on WhatDoTheyKnow.
Everything quoted below is the statement of the EA, its Air Quality Modelling & Assessment Unit (AQMAU) or WSP, the EA’s contractor. These are their positions, not findings we have established. The model concerns PFOA, and this page says PFOA wherever the documents do.
What the EA has provided
Three documents: a covering letter with an appendix, an AQMAU report numbered AQMAU-C2478-RP03 (“RP03”), and one plot of the three soil investigation phases drawn over the model output. What the letter does with each of our seven questions:
| Our request | The EA’s answer, 22 October 2026 |
|---|---|
| 1. The model, RP02, with inputs and outputs | “We do not hold AQMAU-C2478-RP02 but have enclosed AQMAU-C2478-RP03. We do not hold any supporting information.” We are advised to contact AQMAU. |
| 2. Later versions, sensitivity analyses, reviews | “We do not hold any updates, sensitivity analyses or reviews.” |
| 3. Any comparison of model and soil results | Refused under regulation 13 (personal data). The EA says it holds “a limited number of plans” plotting average soil concentrations against the predicted deposition pattern. |
| 4. Sampling design records | Redirected to the Phase 1 and 2 factual and interpretative reports and the Phase 3 factual report on the Council’s website. |
| 5. Sample locations and results | Redirected to the Phase 2 and 3 reports; “Exact locations of samples are witheld as they were taken from private gardens”, with approximate locations shown in the reports’ figures. |
| 6. Sampling depth | Redirected to the reports; the letter gives the 600 mm target depth and a reason (below). |
| 7. Residents’ requests | No guidance or policy held; the letter describes how locations have been chosen (below). |
What RP03 says
- Issue history. Its front page lists AQMAU response dates, one per line: 23 March 2023 (WD01), 05 April 2023 (WD02), 23 May 2023 (WD03), 18 August 2023 (RP01), 22 August 2023 (RP02) and 05 April 2024 (RP03). It does not say what changed between issues. A footnote in RP03 cites an AGC document dated December 2023, which is after the date listed for RP02, so RP03 appears to include material added after RP02.
- Purpose. The study is to “advise on likely locations of peak off-site ground level concentrations or peak off-site concentrations of accumulated deposition” (paragraph 1.1), and the requesting team was “interested in gaining a rudimentary understanding of the potential area that may have been affected” (1.2).
- Release location and source terms were assumed. “The release location and source terms (including PFOA particle size, dimensions, and velocity of the emission) are largely unknown; therefore, assumptions were made regarding these parameters” (3.2). It assumed two possible release locations (the SLG vent and the ARK Plant stack), each modelled at stack heights of 10 m and 30 m, and “1000 kg of PFOA emissions per year” (3.3), in all 16 sensitivity cases listed in its Table 1. Beyond listing it among its assumptions, the text does not say where the 1,000 kg figure came from. RP03 describes using weather data from Blackpool Airport, about 13 km away, for 1980 to 1989 (3.9).
- Limits on use. “Due to the high levels of uncertainty in the source terms, we recommend that numerical predictions should not be used for quantitative comparative analysis” (3.14). The plots are “as a percentage of the maximum predictions”, and “our predictions do not infer any significant harm to surrounding residential areas as they are made to identify the potential locations to investigate land contamination” (2.4).
- Results, as stated. “Peak off-site PFOA ground level concentrations and deposition are predicted to be at the sensitive receptors which border the installation boundary to the south-east, south and west” (summary; paragraphs 3.11 and 3.12 give it for each particle-size scenario). For the largest particles, “peak deposition is expected to be on-site with the majority of the particles depositing within a few hundred metres” (2.2). The contour figures have a percentage-of-maximum legend and, in the copy supplied, no distance scale.
What this means
- RP03 carries a response date of 5 April 2024, before WSP’s soil sampling began. WSP’s reports give the sampling dates as 9 September to 4 October 2024 (Phase 1), February to March 2025 (Phase 2) and October to November 2025 (Phase 3). RP02 (22 August 2023) and RP03 (5 April 2024) both come before all three. RP03 therefore appears not to reflect WSP’s soil results, and it cites no soil measurements.
- WSP cites RP02, not RP03. The reference lists of the Phase 1, 2 and 3 reports each give “AQMAU-C2478-RP02”; the Phase 1 report dates the assessment to “August 2023”. RP03 was dated about five months before Phase 1 fieldwork began. The EA says it does not hold RP02. No document we hold says whether the EA or WSP considered RP03.
- AQMAU’s own account of its limits. AQMAU describes an assumption-based study meant to indicate where to look, with numerical predictions it says should not be compared quantitatively. It also says that “[i]n all modelled scenarios, we find that the spatial distributions of peak predicted concentrations are consistent” (3.14) and that the weather data are “reasonably representative of conditions experienced at the facility” (3.9). These are AQMAU’s statements about its own work; we have not tested the assumptions.
What it does not answer
- Whether WSP also considered RP03, and whether RP02 differed from RP03. The EA says it does not hold RP02, any supporting information, or any updates or reviews; the letter does not address a record of changes between issues. We cannot tell whether the conclusions, the 1,000 kg a year assumption or the sensitivity runs changed between issues.
- What “validate” means. The letter says the soil data were “primarily collected to validate the AQMAU model outputs”. RP03 says its numbers are not for quantitative comparison and are to identify locations to investigate. Both statements are the EA’s. “Validate” may mean a qualitative check of where deposition is predicted, as RP03 3.13 describes the plots as “indicative only of the likely locations”; the documents do not say which is meant. The records of what comparison was made, and against what criterion, were not provided. WSP’s Phase 3 report uses “validation” differently, for “data validation of Phase 2 findings”.
- Where 500 m, 1 km and 5 km came from. We searched RP03’s text for “km”, “radius”, “500”, “soil”, “sampl” and “grid”. It does not mention 500 m, 1 km or 5 km as sampling or zone distances, and does not mention soil sampling; its only reference to soil is “The potential for elevated soil concentration is the pressing issue” (1.2). The only kilometre figure in it is the weather station’s distance. The letter points to the reports for the “rationale”; it does not say whether separate design records are held. For the 1 km advisory zone, see our earlier post on its definitions.
- The model-versus-soil plans. They exist, and were withheld “in its entirety” under regulation 13, an absolute exception, because redaction “would not adequately prevent the identification of the private properties sampled”. Protecting residents’ addresses is not in dispute. The letter says redaction was considered and would not prevent identification. It does not say whether aggregation, for example by distance band or grid cell, was considered; our question 3 did not propose it, and our question 5 offered generalised locations.
- Sensitivity analysis. RP03 says “We have undertaken some analysis to understand the sensitivity of these assumptions” (3.2), sets out 16 sensitivity cases in Table 1 and reports their findings (3.14 to 3.17), but did not include their contour plots, saying “none of the sensitivity alters the conclusions regarding the locations of peak off-site PFOA ground level concentrations or deposition” (3.17). The letter says no “sensitivity analyses” are held. One reading is that it means none beyond those inside RP03. The additional sensitivity contour plots and the underlying model files were not supplied.
- Depth. The letter gives the 600 mm target as “generally considered to represent the principal zone of potential human exposure for shallow soil contamination”. That is the EA’s reason. It does not address sampling below 0.6 m, and WSP’s data do not show what happens below that depth. WSP’s Phase 3 report says average PFOA concentrations “appear to increase with depth across the dataset, with the biggest increase being seen between 0.4 and 0.6m depth” (section 4.3, Table 4-2; see The Numbers). Table 4-2 covers 0 to 0.6 m only, and the 0.4 to 0.6 m interval is within the target depth.
- Supporting information at AQMAU. The letter advises us to contact AQMAU, and WSP’s reports describe AQMAU as part of the EA (“EA Air Quality Modelling & Assessment Unit”). The letter does not describe where it searched.
- Residents’ land. The letter says no guidance is held, that sampling locations have followed “a regularly spaced grid pattern” with “practical factors such as access restrictions; agreement of landowners”, and that “[w]here residents have volunteered their land for inspection, we have considered this and, where appropriate to do so, we have incorporated such land”. Wyre Council’s own wording is in the FAQ answer.
What this does not show
- It does not show the sampling was wrong, or that it was well designed. It shows which documents the Agency holds and what one of them says about itself.
- It does not show where PFOA is or is not in the ground. Only measured results can do that, and the reports’ results are the measurements.
- It does not show that the model’s directions are right or wrong, or what was released. RP03 says the release location and source terms were largely unknown, and that it made assumptions about them.
- It does not show that RP02 and RP03 differ. We do not have RP02.
- It does not show anything about other PFAS. RP03 concerns PFOA and APFO (ammonium perfluorooctanoate, which it reports AGC saying was the main substance used and was used interchangeably with PFOA); it says its particulate-matter results are “not contaminant specific” and does not assess other PFAS individually.
- Our reading is limited to the three documents supplied and the WSP reports. We have not seen the plans withheld under regulation 13, any AQMAU records beyond RP03, or any separate sampling-design records beyond the plans and descriptions in WSP’s reports.
Questions this leaves open
Each can be answered with a document:
- Does the Agency’s record show whether RP02 was held on 29 September 2026, and by whom?
- Which AQMAU issue was current when each phase was designed, and which did the design rely on?
- What record shows the change from RP02 to RP03?
- What records set the 5 km, 500 m and 1 km limits and the south-east weighting of Phase 3, and what role did the model have?
- What comparison of model and soil results was made, and can a version of the plans with property detail removed or aggregated be provided?
- Are the sensitivity runs described in RP03 held, and if so, can they be provided?
On 22 October 2026, on the same WhatDoTheyKnow thread, we asked the Agency for an internal review under regulation 11 of four of its answers: that RP02 is not held, that no sensitivity analyses are held, the answer on the sampling design, and the withholding of the model-versus-soil plans in their entirety. We also made further requests for records: RP02 receipt and retention records, any change log between RP02 and RP03, per-phase records of which AQMAU output was used, how the 5 km, 500 m and 1 km limits were chosen, any existing coarse versions of the withheld plans, the model’s settings and input files, the representativeness of the weather data, and the building configurations modelled. The letter gives 40 working days for an internal review, which by our count ends on 17 December 2026. We will publish the answer, or the absence of one, here and on the Open questions page. The three documents are available from the WhatDoTheyKnow thread linked above.