Brayford valleys - where does the farm air go?
Updated: 1 day ago
Revision 2 • 20 November 2026
Smell, ammonia and dust don’t stay at the farm gate. Where they go depends on the land, the weather and how the sheds are run. This article looks at what happens on still nights, when cold air drains into the valleys, and whether the applicant’s odour model can see it and therefore predict its effects.
Evidence of cold air pooling
On the morning of 14 November 2026, frost lay on the low ground around Brayford while the hilltops stayed green [1]. The Met Office station at Chivenor, about 20 km to the west near the coast, recorded −1.3°C at midnight under a clear sky, and it stayed around freezing until dawn [2].

Taken 08:54. Looking west-north-west across Brayford towards the Bray valley from an elevation of ~190m. Frost lies on the low ground along the western cold-air route traced in our analysis [30]; the hilltops beyond stay green.

Taken 08:59. Taken just east of Brayford Primary School from an elevation of ~175m, facing west-south-west across the village and the River Bray valley. Frost lies on the low ground at the end of the western cold-air route our ground analysis traces from the sheds [30].

Taken 09:00. From just east of Brayford Primary School, facing north-east up the fields. Frost on the low ground, green on the higher fields behind: cold air settles lower on still nights.
The pattern has a simple cause. Cold air is heavy. On a still, clear night the ground cools, the air touching it cools, and that air slides downhill and collects in the valleys. Frost marks where the coldest air sat overnight. Air from the chicken sheds that isn’t sent high into the sky would settle into the valleys in the same way.
The applicant’s weather data can’t show this. It comes from a model with a grid of about 7 km, which gives one temperature for the whole area, and even maps at 100 m detail show the area as a smooth wash, with no sign of the frost pockets [3][4].
The frost is only on the low ground. The hilltops stayed green.
That’s because cold air is heavy. On a still, clear night the ground cools, the air touching it cools, and it slides downhill and collects in the valleys.
The modelling won't see this. 7km weather patterns can't see this and 100m detail shows the area as a smooth wash, with no sign of the frost pockets where cold air sits.
Frost marks where the coldest air sat overnight. Air from the chicken sheds that isn't sent high into the air will settle down in the valleys in the same way, indicated here by the frost.
Two ways downhill
The sheds sit on high ground between valleys, at about 200 to 215 m above sea level. The applicant’s own consultant describes the site as “on a col amongst moorland edge slopes with deeply incised river valleys” [4].
Height data shows two routes for air draining off the site [5]. The western route runs past Braycross Cottages and through Brayford past the school to the floor of the River Bray valley, about 70 m below the sheds. The eastern route runs past Higher Hall Farm, Lower Hall and The Bungalow to the stream valley in the ancient woodland. The frost in the photos lies at the end of the western route. Either way, whatever drains off the site passes homes and the school before settling in a valley.
How air leaves the sheds
Air leaves through 12 chimneys on the roofs, three per shed, each 0.8 m across [4]. How fast it leaves decides how high it goes, and so whether it comes down on the village.
The applicant’s odour model has air leaving every chimney at 11 m/s, every hour of the year [4]. Fast air shoots up and carries the smell high. But the fans don’t have to run that fast: they are “stepless”, so they can run at any speed [6], and the applicant’s own noise report assumes only 25% of the fans run at night [7].
The chimneys are barely above the roof. The sheds are about 6 m to the ridge; the odour model puts the chimneys at 6.2 m, while the Environmental Statement says 6.9 m [4][8]. Ventilation also keeps running when the sheds are cleared out: the odour report says “it is normal to maintain ventilation” during clean-out [4].
By fixing every chimney at its fastest speed, every hour, the model wrongly assumes the air carrying the smell is always sent as high as it can go, and therefore miss the village when it comes down.
How often is the air still at night?
On the applicant’s own measurements, often. During its noise survey, from 17 to 20 March 2026, the wind on site was 1 m/s or less in 94% of night-time readings [7]. In other weather, the applicant’s wind rose shows the wind blowing most often from the west, carrying air east towards Lower Hall, The Bungalow and Withygate Barns [4].
We counted still, mostly clear night hours (wind 2 m/s or less, cloud cover 40% or less) in historical weather data for the area [9]:
Year | Still hours | Nights |
2022 | 185 | 66 |
2023 | 147 | 53 |
2024 | 139 | 54 |
2025 | 354 | 79 |
Still air matters because of how the model’s pass mark works. The model works out the smell at each home for every hour of the year, sets aside the smelliest 2% of hours (about 175 hours), and passes the home if every other hour is at 3 odour units or less [4]. That benchmark of 3 comes from Environment Agency guidance withdrawn in December 2025; the guidance that replaced it gives no number [10].
If the smell reached a home at above 3 on still nights, those hours alone would fill all or most of the 175 hours set aside. In 2022 and 2025 they would fill them completely, so a home would fail on still nights alone. In 2023 and 2024 they would leave only 28 to 36 hours spare, so a few smelly hours in any other weather would tip it over. The effect of this on residents is why the model has to get still nights right.

Above: another illustration of still, cool air. In the evening mist forms in the valleys and sits there until the morning sun warms the ground and disperses the mist into the atmosphere. This evening picture, taken on Exmoor to the east of the farm, shows its location on the hillside and cool air drifting down past it, where it would collect odour and deposit it in the village. Can the applicant's analysis see or predict this?
Can the applicant’s model see still nights properly?
The model itself, ADMS, is a recognised tool, and it was run with its settings for calm weather, hills and buildings switched on [4]. The problem is what it was fed. Its weather comes from a global forecasting model with a grid of about 7 km, in which valleys a few hundred metres wide aren't detailed.
It sees the land in 100 m squares, which barely picks up the routes the cold air takes. It fixes the fans at 11 m/s, so the smell is always launched as high as it can go. And it gives one number each for 17 chosen homes and sites [4]. A model that can’t see the valleys, and always runs the fans flat out, will show still nights at their best.
Our own check
We ran a simple screening calculation using the applicant’s own figures. It isn’t a substitute for full modelling, but it shows what happens when you change the things the model holds fixed.
We used 12 chimneys, 6.2 m high and 0.8 m across; a smell inside the sheds of about 2,300 odour units, typical of late in the crop; air leaving at 22°C; and a quarter of full airflow at night [4][7]. Unless a picture says otherwise, the weather is air moving at 1 m/s with a temperature inversion, where warmer air sits on top of cold air near the ground.
How to read the below graphics
Each one is a slice through the air. Pictures 1 to 4 and 6 to 7 run from the A399 junction in the west to Withygate Barns in the east, with the sheds in the middle.
Distances are from the centre of the four sheds to a door or window of each property, measured on Google Earth. Picture 5 is the exception: it measures along the downhill route.
The shading shows how strong the smell is. The darker the shade, the stronger the smell.
The white dashed lines mark 3, 5 and 10 odour units. Where a line touches the ground, people would smell that level.
The tag under the sheds is the smell inside them late in the crop: about 2,300.
Each picture shows both directions, but on any one night the air goes one way:
west, as cold air drains towards the village and school; or
east, on a light westerly, towards Lower Hall, The Bungalow and Withygate Barns.
What do 3, 5 and 10 mean?
3 is the benchmark for acceptable odour the applicant uses.
5 to 10 is the range in which, as the applicant’s report puts it, research finds a significant share of complaints.
These numbers are judged on the smelliest 2% of hours across a year. Each picture shows a single hour, so it isn’t a pass or fail on its own: it illustrates the level of smell people would experience during that hour.

Picture 1: What the applicant’s model assumes
Cold still night (0°C), all 12 chimneys at 11 m/s.
The air shoots up and settles at about 30 m.
None of the indicator lines reaches the ground.
The strongest smell at ground level, out to 1.3 km from source, is about 1.6.
This is the picture the applicant’s case relies on: air is pushed high in the air, taking the smell with it.

Picture 2: What happens if the fans are slowed?
Mild still night (15°C). The same amount of air, but all 12 chimneys at a quarter speed instead of a quarter of the fans flat out.
The air only rises to about 17 m.
As it spreads, its lower edge reaches the ground:
3 at about 480 m;
5 at about 555 m.
Between 3 and 5:
the village hall, about 4.3;
the playground, about 4.7;
Lower Hall, about 3.6.
Above 5:
the school playground, about 5.8;
The Bungalow, about 5.6;
the Old Baptist Chapel, about 9.3;
Withygate Barns and the A399 junction, about 9.1 each.
The smell passes over Braycross Cottages, the closest home.
Same smell, same amount of air, slower fans: the smell comes down onto the village.
Why doesn’t the smell come down in pictures 1 and 2?
A flat-ground calculation lifts the smell and keeps it at that height. It never brings it back down; the smell only spreads, slowly.
In real life it does come down, in three ways. Pictures 5 to 7 show each one.

Picture 3: What if the smell comes out at ground level?
For example, through open doors. We used 10,000 odour units a second, about a quarter of the night-time total. The application gives no figure for this.
The smell gets no lift, so it stays low.
Everything within about 800 m is above 10:
Braycross Cottages, about 75;
Lower Hall, about 23;
the school, playground and village hall, about 17–21;
The Bungalow, about 18.
Between 5 and 10:
the Old Baptist Chapel, about 7.4;
Withygate Barns and the A399 junction, about 5.9 each.
Smell released low down stays at the height people breathe.

Picture 4: The worst moment: the doors open at de-stocking
The birds have been caught, the doors are open and everything is being loaded onto lorries. The night is cold and clear (0°C), with a strong inversion and air drifting at 0.5 m/s.
4 chimneys are running, one per shed, at a quarter speed. They carry about a third of the smell.
The other two-thirds pours out of the doors at ground level.
Every home on the line is far above 10.
We kept the total smell the same as a normal night. We added nothing for the rush when the doors open, or for litter stirred up during catching. Both would add more.
The application doesn’t say whether the doors are open during loading. That part is our assumption.
At the moment the doors open on a still night, every home on this line is in strong smell.

Picture 5: Cold air draining downhill
On a still night, cold air slides downhill like water. It takes with it any smell that started low down: from open doors, slowed fans, or air caught behind the roofs.
The ground line here is the real ground along the two routes.
West: past Braycross Cottages and the school, through Brayford, to the Bray valley floor.
East: past Higher Hall Farm, Lower Hall and The Bungalow, to the woodland stream valley.
The cold air collects on the valley floors and sits there until morning. The frost in our photos marks exactly where.
The applicant’s weather data is on a 7 km grid. It can’t see a valley a few hundred metres wide.
Picture 5 shows where the air goes, not how strong the smell is. The depth of the cold layer is drawn to illustrate it, not measured, and heights are exaggerated about 4 times. See these articles for more detail: Air Movement and Full Report section 4. Following, read Environmental Health's acceptance response which states the expansion will likely reduce odour impacts compared to the current setup and requires no extra mitigation.

Picture 6: Dawn, when the night layer breaks up
This is the same night as picture 1, with the smell sitting about 30 m up.
After sunrise the ground warms, and a mixing layer grows upwards from it.
When that layer reaches the smell, it brings it down to the ground within a short time. This is known as “fumigation”.
With air moving at 1 m/s the smell level would be:
Place | Smell level |
Braycross Cottages | about 26 |
Lower Hall | about 14 |
Village hall | about 13 |
School playground and The Bungalow | about 12 |
Old Baptist Chapel | about 7 |
A399 junction and Withygate Barns | about 6 |
Night layer break-up typically lasts about half an hour to an hour. At 2 m/s the figures halve.
Even in the applicant’s best case, the smell can come down on the village in the early morning, a nice start to the day.

Picture 7: The sheds themselves, on a breezy night
A cloudy night with a 3 m/s breeze.
The chimneys sit less than 1 m above the roof ridge.
In a breeze, air tumbling over the roofs can catch the smell and pull it down before it rises.
The nearest homes get more smell. Braycross Cottages goes from about 7.6 without the sheds to about 11 with them.
Beyond about 400 m the sheds make little difference.
The applicant’s model does include the sheds, so this effect is in their numbers. We show it to explain why the nearest neighbours get more smell in a breeze.
(Method: Huber–Snyder building method, using the sheds’ 6 m ridge height.)
What can’t our check tell you?
This is a simple calculation by local residents, not a full odour model. Apart from Picture 5, it treats the ground as flat. It treats the four sheds as one point, which overstates the smell closest to them. It is less reliable at very low wind speeds, as in Picture 4. And Pictures 3 and 4 use figures we chose, because the application gives none for open doors or de-stocking.
What it does show is that the answer depends on how the fans are run, how the air gets out, the temperature of the air in the sheds compared with the air outside, the time of night and the shape of the land. The applicant’s model fixes the fans at full speed and sees the land in 100 m squares.
What it does show is that the answer depends on how the fans are run, how the air gets out, the temperature of the air in the barns versus the air outside, the time of night and the shape of the land. The applicant’s model fixes the fans at full speed and sees the land in 100 m squares - it simply isn't an accurate representation and so the results can not be relied upon.
What are we asking for?
We ask for the still nights to be modelled properly, with realistic fan speeds at night, cold air draining down the valleys, and the morning break-up. We ask the applicant to confirm the chimney height, 6.2 m or 6.9 m, and that the model used the right one. And we ask for the model to show what reaches the homes, school and woodland on both downhill routes: Braycross Cottages, the school and Brayford to the west, and Higher Hall Farm, Lower Hall and The Bungalow to the east.
We’re not against farming. We’re asking that the still nights, when the village would smell the farm most strongly, are properly looked at. Otherwise North Devon Council Planning will not have all the data they need to make an informed decission on our behalf.
Sources
[1] Author’s photographs, Brayford, 14 November 2026.
[2] Met Office, Chivenor observations, 14 November 2026.
[3] Windy.com, weather model maps for the Brayford area, 14 November 2026.
[4] AS Modelling & Data Ltd, Odour Report for Higher Hall Farm, 13 May 2026 (Environment Agency permit application and application 82141): sections 2, 3, 3.2 to 3.5 and 4, Table 1a and Figure 3b.
[5] Environment Agency LiDAR height data.
[6] Associated Plant Specifications, SKOV DA 820 fans (application 82141).
[7] Matrix Acoustics, Noise Impact Assessment, 8 June 2026 (application 82141), Table A2.
[8] Acorus Rural Property Services, Environmental Statement, August 2026 (application 82141), section 3.1.4.
[9] Open-Meteo, historical weather data for the Brayford area, 2022 to 2025.
[10] GOV.UK, Odour management: comply with your environmental permit, published 3 December 2025, replacing Environment Agency guidance H4.
[11] D. B. Turner, Workbook of Atmospheric Dispersion Estimates, US Environmental Protection Agency, 1970.
[12] Huber–Snyder building downwash method, using the sheds’ 6 m ridge height.