Air Movement: Pooling, Odour and Ammonia Evidence
Updated: 1 day ago
Updated 18 November 2026
Summary of resident's findings
The applicant's odour and ammonia modelling does not represent night-time cold-air drainage, which LiDAR shows carries air from the sheds downhill directly past homes in Braycross, Higher Hall, Lower Hall and Brayford. Its ammonia "reduction" also rests on an incorrect and overstated baseline.
Drainage. The sheds sit at about 205-218 m on ground falling north. Both downhill routes from the sheds pass within 10-60 m of the applicant's own receptors: Braycross Cottages (5), Higher Hall Farm (1), Lower Hall (2) and Brayford (14, 15, 17), roughly 25-70 m lower than the sheds (Figures 1 and 3).
The applicant's own measurements. Its noise consultant's on-site weather station, about 100 m north of the sheds, recorded near-calm (mean 0.4 m/s), cooling, southerly night air on two of three March nights. That is air flowing from the sheds towards Higher Hall (Figure 6). The odour and ammonia models used past weather reconstructed by a global weather model on a 7 km grid instead.
Pooling. Brayford receptors 15 and 17 and Beara Hill (9) sit within 2 m of the valley floor, where drained cold air collects on still nights (Figure 2).
Odour. Still nights are unassessed. If the exhaust stays in the cold layer, a screening calculation using the applicant's own emission rates gives roughly 7-110 ouE/m³ at homes, against a 3 ouE/m³ benchmark. If it rises above the layer, the result is far lower. The model also assumes an 11 m/s chimney exit speed, which the specified fan cannot reach (10.4 m/s maximum) (Figure 4).
Ammonia baseline. The applicant applied the free-range "outdoor" factor (0.084 kg/bird/yr) to all 40,000 existing birds. GOV.UK says the combined free-range factor is 0.03. On that basis today's farm emits about 1,200 kg NH₃/yr; the proposal would emit about 4,224 kg, roughly 3.5 times more, not a reduction (Figure 5).
Independent review. On 16 November an independent review by Dr Michael Bull, a chartered engineer who chaired the Institute of Air Quality Management (IAQM) working group that wrote its odour guidance for planning, was placed on the planning file. It confirms the ammonia baseline is overstated "by a factor of approximately 3.6", and his own modelling shows ammonia rising by 31–107% at the three protected points he examined. On odour, he finds a key adjustment in the applicant's model "has not been adequately justified" and asks for it to be tested. So the central findings on this page are no longer the residents' alone.
The analysis is illustrative screening, not a replacement dispersion model. Its purpose is to show that the conditions most likely to cause odour and ammonia problems at Brayford are the ones the submitted modelling cannot represent, and to ask the council to require that they be assessed.
For transparency, the report by Michael Bull of Michael Bull & Associates, consultants in air quality and odour assesment, was commissioned by Communities Against Factory Farming in analysis of application 82141. His findings mirrored our own and have been included in support.
1. Cold-air drainage: where air from the sheds goes at night
On clear, still nights the ground cools the air above it, and that cold, heavy air slides downhill. Whatever is released into it (fan exhaust from 6.2 m chimneys, gable-fan air, odour, ammonia) travels with it. Air leaving the sheds splits into two routes, both passing homes.

Route | Passes Receptors | Distance along route | Ground level | Below sheds |
West | 5 Braycross Cottages | 296 m | 188 m | ~27 m |
West | 14, 13, 15 Brayford | 628-752 m | 153-171 m | ~45-60 m |
West | 17 Brayford (Bray valley floor) | 1,023 m | 144 m | ~70 m |
East | 1 Higher Hall Farm (applicant's farmhouse) | 266 m | 187 m | ~26 m |
East | 4 Catkins (100 m off route) | 399 m | 193 m | ~20 m |
East | 2 Lower Hall | 587 m | 168 m | ~45 m |
East | Higher Hall Row Wood stream (ancient woodland AW2, AW4, AW5) | 900-1,200 m | 161-186 m | ~30-50 m |

Why this matters for the applicant's modelling. The submitted reports describe the site as "on a col", use Ordnance Survey 50 m terrain resampled to 100 m, and take past weather reconstructed by the GFS global weather model at about 7 km resolution, adjusted by ADMS's FLOWSTAR module to an effective 100 m (odour) or about 230 m (ammonia) wind field (Odour report s.4.1, 4.6; Ammonia report s.4.6). At those resolutions the 300-700 m drainage routes in Figure 1 are barely resolved, and nothing in the reports shows that thermally driven downslope flow is represented. The applicant should be asked to show how it is.
The applicant's own site measurements support this. The applicant's noise consultant ran a weather station about 100 m north of the sheds, at monitor position 2, from 17 to 20 March 2026, recording every 10 minutes (Noise Impact Assessment, Appendix A, Tables A2.1-A2.5).
On two of the three nights (17-18 and 19-20 March), 66-76% of night readings came from SE-SW, at a mean of 0.4-0.45 m/s, while the air cooled to about 2 °C.
Air moving from the south at that point flows from the sheds towards Higher Hall Farm and Catkins, the east drainage route.
On the milder middle night (minimum 5.9 °C), winds were more variable.
The report itself records a highest wind of 2.5 m/s and a median of 0.6 m/s over the survey.
Winds were also often southerly by day that week, so not every southerly reading is drainage. But still, cooling, southerly night air at the site is exactly what the drainage analysis predicts. The applicant's 7 km GFS weather data was not checked against it.

2. Cold-air pooling: where it collects
Drained air collects on valley floors and lingers there. That is where odour and ammonia would build up and persist on still nights, rather than dispersing. Several of the applicant's receptors sit in those zones.

Receptor | Height above nearest valley floor | Pooling zone |
17 Brayford | 1 m | Valley floor, most frequent pooling |
9 Beara Hill | 1 m | Valley floor |
15 Brayford | 2 m | Valley floor |
14 Brayford | 7 m | Within 10 m |
3 The Bungalow | 9 m | Within 10 m |
12 Withygate | 11 m | Within 25 m |
2 Lower Hall | 12 m | Within 25 m |
16 Brayford | 13 m | Within 25 m |
AW4 / AW5 ancient woodland | 11 / 20 m | Within 25 m |
The west route ends in the Bray valley at Brayford, and the east route ends in the Higher Hall Row Wood stream valley among ancient woodland. Both are pooling zones.
3. Odour: still nights vs the applicant's annual figures
On still nights the result turns on one question the applicant never asks: does the warm fan exhaust stay in the cold drainage layer, or rise above it?
If the plume stays low (chimney height, or about 12 m after limited rise), the screening puts hourly odour well above the 3 ouE/m³ benchmark along the drainage routes: about 10-110 ouE/m³ at chimney height, and 7-28 ouE/m³ from Catkins outwards at 12 m.
If it rises to about 25 m before following the drainage, most homes fall below 3. The outer Brayford homes (16, 17) still reach about 2-5 ouE/m³.
If it stays level at its release altitude while the ground falls away, concentrations at homes are near zero.
Plume-rise estimates from the specified fan suggest a 10-40 m rise on still nights. Strong ground-level inversions, low winter ventilation and building wake effects push it lower. The applicant's predicted maximum annual 98th percentiles at these homes are 0.5-2.2 ouE/m³.
The honest conclusion is that the outcome ranges from negligible to well above the benchmark. The submitted model cannot tell which, because it does not represent the drainage layer at all.

Why one bad night type can decide the 98th percentile. The applicant's test is the 98th percentile of hourly concentrations: the level exceeded in 2% of hours, or about 175 hours a year (Odour report s.3.2). Drainage sends the plume down the same routes every time, unlike ordinary winds, which spread it around the compass. A receptor's 98th percentile therefore exceeds 3 ouE/m³ if drainage reaches it for about 175 hours or more a year, eg. 22 nights it smells for 8 hours.
Year | Night hours with wind ≤2 m/s and cloud ≤40% | Share of year |
2022 | 185 | 2.1% |
2023 | 147 | 1.7% |
2024 | 139 | 1.6% |
2025 | 354 | 4.0% |
Source: Open-Meteo historical reanalysis (ERA5-based), grid point at the applicant's own weather location (51.094 N, 3.862 W). A coarse grid tends to overstate wind in sheltered valleys, so these counts are probably low. Counting nights up to 3 m/s gives 354-643 hours a year.
Points specific to the submitted odour model
The modelled chimney cannot reach the modelled exit speed. App2 and App6 model every chimney at 11 m/s and 0.8 m diameter, and the Environmental Statement (5.1.13) promises "greater than 11m/s". The fan the noise report specifies is a Big Dutchman FF091-6ET in a 920 mm chimney. Its data sheet gives 10.4 m/s at 0 Pa, and about 9 m/s at typical working pressures of 20-30 Pa (Noise report, Appendix C). Its speed control is listed as "T, FU, Triac" (variable).
Night ventilation is a fraction of capacity. The noise report assumes only 25% of roof fans run at night (4 of 14 per shed). The odour model's fixed 11 m/s overstates night-time exit speed, and therefore plume rise, when drainage occurs.
The documents contradict each other on the equipment. The Plant Specifications describe Skov DA 820 stepless chimneys; the noise report specifies Big Dutchman fans. The council cannot tell which system the odour model represents.
Night emissions are not small. The applicant's Figure 2b shows ridge-fan odour of roughly 10,000-40,000 ouE/s per shed through most of each crop, with peaks around 140,000 ouE/s at clean-out. Four fans at about 6 m³/s each, at the report's internal concentrations of 300-2,300 ouE/m³, give about 7,500-57,000 ouE/s per shed, consistent with that figure.
Benchmark status. The report itself notes that H4, the source of the 3 ouE/m³ benchmark, was withdrawn in December 2025 (s.3.2).
Environmental Health accepted the report because it predicts a reduction compared with the existing farm. That comparison does not test still nights.
What an independent expert says about the odour model
Dr Michael Bull reviewed the applicant's odour assessment for Communities Against Factory Farming. His review does not produce alternative odour figures. It asks the applicant to justify and test the assumptions the prediction rests on.
The key adjustment is unjustified. The model lowers the smell inside the sheds when ventilation is high, using a square-root relationship. Dr Bull finds this "has not been adequately justified and could materially influence the predicted odour impacts", and recommends testing the model with it removed.
Borrowed figures must be shown to fit. IAQM guidance says that where standard data are used, "it should be demonstrated that the information is likely to be a reasonable representation of odour emission rates on the study site". The applicant has not done so.
Today's farm is probably overstated. The odour added for the free-range field "is likely to result in an overestimate of the odour emissions from the existing development", which flatters the predicted improvement.
The emission chart raises questions. Comparing the applicant's Figures 2a and 2b "suggests that odour emission rates at the end of the cycle are higher for the existing case compared with the proposed", although the new farm holds more than four times the birds. Figure 2b "also suggests that there is a thinning of the crop around day 28-30", which no part of the application assesses.
One model is not enough. The IAQM's September 2026 guidance "supports selecting complementary assessment tools where appropriate, including observational evidence". As Dr Bull quotes it, IAQM guidance also warns that model uncertainty "can be considerable" and recommends using at least two methods. The applicant used one model, measured nothing at the site, and stated no uncertainty.
There is no official pass mark. The Environment Agency's H4 benchmark was withdrawn in December 2025. IAQM leaves the criterion to the assessor, "somewhere in the range of 1 to 10 ouE/m3". So the pass mark of 3 is the consultant's choice, not an official standard.
So the prediction that odour will fall rests on one model, an untested adjustment and borrowed figures, with nothing measured at Higher Hall.
4. Ammonia: an overstated baseline, and drainage onto woodland
The ammonia report's comparison with the existing farm is built on an existing-farm emission about 3.6 times too high. Corrected, the proposal is roughly a 3.5-fold increase.
Applicant (Ammonia report Table 3) | GOV.UK method | |
Existing housing | 0.024 kg/bird/yr x 40,000 | Included in combined factor |
Existing range | 0.084 kg/bird/yr x all 40,000 birds | 0.084 applies only to the 10% of time on the range |
Existing total | 0.108 kg/bird/yr = 4,318 kg/yr | 0.03 kg/bird/yr = 1,200 kg/yr |
Proposed (176,000 indoor) | 0.024 = 4,224 kg/yr | 0.024 = 4,224 kg/yr |
GOV.UK states that for free-range broilers the housing factor is 0.024 and the outdoor component 0.084, giving "0.03" on a 90% housing / 10% ranging split (GOV.UK ammonia emission factors, updated 4 December 2024). The applicant applied the full outdoor component to every bird, all year. Its own report says the birds only range "during the second half of the flock cycle", so even 10% overstates it.
Independently confirmed. Dr Bull's review finds the same error. The applicant "has not accounted for the proportion of time the birds spend in the sheds and the ranging areas", so today's emissions are overestimated "by a factor of approximately 3.6". His corrected figure for today's farm, 0.038 g/s (his Table 4), matches ours: about 1,200 kg a year.
He then re-ran the model with the emissions correctly allocated. The proposal raises annual average ammonia at all three protected points he examined:
Protected point | Today, corrected (µg/m³) | Proposed (µg/m³) | Change |
Ancient woodland (AW1) | 1.052 | 1.832 | +74% |
North Exmoor SSSI | 0.00551 | 0.00720 | +31% |
South Exmoor SSSI / Exmoor Heaths SAC | 0.00267 | 0.00554 | +107% |
Modelled without terrain or calms, as in the applicant's own comparison run (its Table 6). Dr Bull notes these are not a precise reproduction of the applicant's figures, because he used a different weather dataset.
Why 3.5 times and 31–107% agree. The 3.5-fold figure is how much ammonia the whole farm gives off. The percentages are how much reaches each protected point. The new sheds would push air out of roof chimneys at speed, which spreads it further than today's ground-level range. Dr Bull says this "would improve the dispersion of emissions" but "does not fully compensate for the increased emissions". So 3.5 times more ammonia is released, and more of it reaches the woods and Exmoor, not less.
Dr Bull recommends that the council obtain a revised ammonia assessment before deciding, considered by a suitably qualified ecologist, with in-combination effects assessed where appropriate.
Consequences
The "existing" ammonia concentrations at Higher Hall Row Wood (e.g. AW1 3.04 µg/m³, 304% of the Critical Level) are inflated mainly by that ground-level range source. The apparent improvement to 0.77 µg/m³ is therefore not a like-for-like comparison.
Even on the applicant's figures, the proposal alone uses 77-81% of the 1.0 µg/m³ Critical Level at AW1 and AW2, leaving about 0.2 µg/m³ of headroom. The background is already 1.24 µg/m³ (APIS, per the report).
Calms were left out of the detailed run. The report says calms "cannot be used with terrain or spatially varying deposition", so the terrain/deposition run excluded them. The test that calms do not matter was run without terrain (s.5.1-5.2). Drainage nights are exactly calm-plus-terrain conditions, and the east route runs down the woodland stream (Figure 1).
Annualising the still-night screening concentrations over 140-350 hours gives an extra contribution at AW2/AW4/AW5 between about 0.02 and several µg/m³, depending on how high the plume starts. That range is too wide to settle the point. But it comfortably spans the 0.2 µg/m³ headroom, so the question cannot be dismissed without proper modelling.
5. Objection and requests
The odour (App2) and ammonia (App6) assessments do not represent night-time cold-air drainage. Environment Agency LiDAR shows the sheds on ground falling north, with both downhill routes passing within about 10-60 m of the applicant's own receptors at Braycross Cottages, Higher Hall, Lower Hall and Brayford, 25-70 m below the sheds. They then reach valley floors where cold air pools, at Brayford and in wooded valley.
The applicant's own noise survey weather station, about 100 m north of the sheds, recorded near-calm, cooling, southerly night air on two of three March nights: air moving from the sheds towards Higher Hall. Yet the assessments rely on 7 km GFS weather data and 50 m terrain resampled to 100 m, and the detailed ammonia run excluded calms.
The chimneys are modelled at a fixed 11 m/s exit velocity, but the fan specified in the noise report reaches only 10.4 m/s at zero pressure (about 9 m/s in use), and only a quarter of fans are assumed to run at night. Whether still-night odour reaches homes depends on whether the warm exhaust stays in the cold layer.
A screening calculation using the applicant's own emission rates gives roughly 7-110 ouE/m³ at these homes if it does, against a 3 ouE/m³ benchmark, and the submitted model cannot answer that question.
Separately, the ammonia baseline applies the free-range outdoor factor (0.084 kg/bird/yr) to all 40,000 existing birds, contrary to GOV.UK guidance (combined factor 0.03). That overstates the existing emission about 3.6-fold and presents a c.3.5-fold increase as a reduction.
An independent review by Dr Michael Bull confirms that error, finds ammonia rising by 31–107% at the protected points he modelled, and recommends that the council obtain a revised assessment before deciding.
The council are asked to reject the application, or in the least to not rely on these assessments until they are corrected.
Requests to the council
Require the ammonia baseline to be recalculated with the GOV.UK free-range factor and the comparison with existing re-presented.
Require the odour and ammonia modelling to be re-run with terrain and calms together, site-representative low-wind stable meteorology, and variable fan exit velocity at minimum ventilation.
Ask the applicant to demonstrate how downslope cold-air drainage towards Brayford and Higher Hall Row Wood is represented, using LiDAR-resolution terrain.
Ask Environmental Health to reconsider its acceptance, which rested on the predicted reduction compared with the existing farm.
Given the applicant's own 77-81% of the Critical Level at the ancient woodland, assess the corrected figures against Natural England's air pollution standing advice, as Natural England's 1 October response requires.
Require the applicant to confirm which ventilation system will be installed (Skov DA 820 or Big Dutchman FF091-6ET), and to re-run the models with that fan's real exit velocity and night-time ventilation, including plume behaviour in still, stable conditions.
Obtain the revised ammonia assessment the independent review recommends, considered by a suitably qualified ecologist, with in-combination effects assessed where appropriate.
Require the applicant to justify the square-root adjustment to odour at high ventilation, with sensitivity tests including its removal, and to state the uncertainty in its odour predictions.
6. Method, limitations and sources
Method
Terrain: Environment Agency LIDAR Composite DTM 2022, 2 m, nine 5 km tiles (SS62NW-SS73NW), mosaicked and averaged to 10 m.
Shed and chimney positions: georeferenced from the applicant's App2 Figure 4 (British National Grid axes; accuracy about ±3 m).
Drainage routes: steepest-descent (D8) paths from each pair of sheds on a depression-filled DEM. Drainage footprint: multiple-flow-direction spreading from the shed outlines.
Pooling zones: height above nearest stream (channels with more than 0.2 km² catchment); 10 m and 25 m bands.
Odour and ammonia screening: Gaussian plume, Pasquill-Gifford stability F (Briggs rural coefficients), 1 m/s, receptor directly downwind along the drainage route. Effective release heights 6.2 m, 12 m and 25 m; level-plume case for comparison. Emissions: odour 10,000-25,000 ouE/s per shed (read from App2 Figure 2b, cross-checked against fan flow); ammonia 0.1339 g/s total (App6 Table 4a). Each route takes half the sheds.
Plume rise: Briggs stable-atmosphere formulas, using the FF091-6ET data (0.92 m, 9.2 m/s at full speed, 4.6 m/s at half), exhaust 22 °C, air 3 °C, lapse rates of +0.035, +0.1 and +0.2 K/m, and single vs four merged chimneys. Result: 10-40 m at 1 m/s.
On-site weather: the applicant's noise report, Tables A2.1-A2.5 (459 ten-minute readings). Night = 18:30-06:00; station position read from the report's Figure 1.
Frequency of still, clear nights: Open-Meteo historical reanalysis, 2022-2025, hourly, at 51.094 N, 3.862 W.
Limitations
This is screening, not a full dispersion model. The Gaussian plume is not strictly valid below about 1 m/s, and it ignores valley confinement.
The largest uncertainty is plume rise. Whether warm exhaust stays in, or escapes, the drainage layer moves the odour results from near zero to well above the benchmark. The objection presents this as a question the applicant must answer, not as a predicted exceedance.
The odour values are hourly levels on still nights, not 98th percentiles. The percentile argument depends on how often such nights occur, estimated from coarse reanalysis.
The station data cover only three nights in March. Southerly winds also occurred by day, so not every southerly reading reflects drainage.
Ammonia annual-mean effects are highly sensitive to plume height. Only the baseline error is a firm numerical finding.
Sources
North Devon Council, application 82141: App2 Odour Report (AS Modelling & Data, 13 May 2026); App6 Ammonia Report; Plant Specifications; Environmental Health (18 Sep 2026), Environment Agency (18 Sep 2026) and Natural England (1 Oct 2026) responses
Michael Bull and Associates Ltd, Proposed Broiler Chicken Rearing Houses, Higher Hall Farm: Review of Ammonia and Odour Issues, version 1 (2 November 2026), for Communities Against Factory Farming; on the NDC planning file, 16 November 2026.
Institute of Air Quality Management, Guidance on the Assessment of Odour for Planning, version 2.0 (September 2026), as cited in the review above.