Weather & Air Pollution Forecasting

Weather, wind and waves decide when and how work can be done at sea, on the coast and on exposed sites. Air quality decides how much can be emitted, and where. We turn atmospheric and ocean data into site-specific forecasts and clear decision support.

Our work covers four connected areas: meteorology, wind, waves and air-pollution dispersion. They share one foundation, a good description of the atmosphere, and each is presented so that planners, operators and authorities can use it directly.

Weather forecasts start from numerical weather prediction: physical models of the atmosphere that are run on a grid and stepped forward in time from the observed state of the weather. Satellites, ground stations, weather balloons and ships supply that starting point, and the model returns pressure, temperature, humidity, cloud, precipitation, visibility and wind.

Global models are coarse. For a harbour, an industrial site or a stretch of coast, we refine them to the local scale using terrain, land use and the position of the coastline, and we check them against observations, so that their behaviour at your site is understood rather than assumed.

Wind is reported as a mean speed and a gust speed at the standard height of 10 metres above the surface. The mean describes the steady flow, and the gust is the strongest short burst within it. Direction completes the picture. Forecasts are usually given in knots (kt) or metres per second, in steps of one to three hours.

Wind matters for crane and lifting operations, work at height, vessel movements, drone flights, structural loads, ventilation and the spread of smoke and emissions.

• On 28 and 29 September the mean wind stays between about 6 and 11 kt, close to 10 kt for most of Saturday, with gusts of 13 to 14 kt.
• There are short lulls: the mean speed falls to about 6 kt on the evening of 27 September and to about 3 kt around midnight at the start of 1 October.
• The wind can also change quickly: on 2 October the mean speed rises from under 2 kt at midnight to about 14 kt by midday, with gusts close to 18 kt.

Waves are the sea's response to wind. Their size depends on how strong the wind is, how long it blows and over what distance of open water it acts (the fetch). Waves that are still driven by the local wind are called wind sea. Waves that have travelled out of the area where they were generated are called swell, and they are longer and more regular.

A forecast describes a sea state with three values:

• Significant wave height (Hs): the average height of the highest third of the waves. It is close to the wave height an observer would report, and it is the main value for operational limits.
• Maximum wave height (Hmax): the largest single wave expected in a period. It is typically about 1.6 to 2 times Hs, so a sea with an Hs of 1.5 m can contain a wave of close to 3 m.
• Swell height: the part of the sea state that comes from distant weather systems rather than from the local wind.

Wave period and wave direction complete the description and are part of the forecast too.

• Hmax follows Hs: the red line stays at about 1.8 times the blue one throughout.
• The largest seas come on 29 September: Hs reaches about 1.7 ft, and the largest single waves exceed 3 ft.
• Swell stays small, below about 0.7 ft, so almost all of this sea state is wind sea, which is why the waves follow the wind.
• The short peak on 1 October (Hs about 1.5 ft, Hmax about 2.7 ft) follows the rise in wind speed that morning, and when the wind drops below 2 kt on 2 October the waves die away almost completely.

• Marine and offshore operations: weather windows for lifting, vessel access and installation
• Ports, harbours and coastal works
• Wind-energy and other renewable-energy sites
• Construction sites with cranes, façade work and work at height
• Aviation, drone and event operations
• Environmental assessment and permitting

Where a weather forecast asks what the atmosphere will do, dispersion modelling asks where something released into it will go. Wind direction and speed carry a plume, atmospheric stability and mixing height decide how fast it is diluted, and rain removes part of it from the air.

We combine meteorological data with emission data to estimate concentrations at the places that matter: the site boundary, a neighbourhood, a school, a sensitive habitat.

• Air-quality and dispersion modelling
• Emission scenario studies: comparing process, stack-height or abatement options
• Short-term forecasts of episodes with poor air quality
• Analysis and reporting of monitoring data

Two typical programmes show how these disciplines come together in practice. Each is built up in stages that can be taken separately or together, so that observations, models and operations are put in place in a sensible order and every stage produces something usable.

A chemical transport model (CTM) simulates how pollutants are emitted, transported, transformed and removed in the atmosphere. Coupled to a weather forecast model, it turns a forecast of wind, temperature and mixing into a forecast of concentrations. Putting such a system into operation for a city involves three stages:

1. Emission measurement: mobile (traffic) and stationary (industrial, energy and other fixed) emission sources are measured, to establish how much is emitted, of which pollutants, and when.
2. Emission inventory: a spatially and temporally resolved inventory is compiled for the target city from the measurements and supporting activity data, in the form the model requires.
3. Operational forecasting: the CTM, coupled to a weather forecast model, is run routinely in production and delivers regular air-quality forecasts for the city.

The result is a repeatable forecasting chain, from the sources on the ground to a forecast that planners, operators and authorities can use.

Reliable warnings of extreme weather depend on three things: good observations, a numerical weather prediction (NWP) model that has been configured and tested for the region, and an operational system that uses the observations as they arrive. A national programme therefore typically proceeds in three stages:

1. Observation network: the national meteorological observation network is strengthened: synoptic (surface) stations, weather radar and radiosonde (upper-air) soundings.
2. Model analysis and sensitivity study: numerical weather prediction models are analysed, and the sensitivity of their configuration (for example physics schemes, resolution and domain) is studied to find the best model set-up for the country.
3. Operational forecasting with data assimilation: the NWP models are run routinely with data assimilation, so that observations are fed into the forecast and extreme events are better represented.

• Site-specific weather, wind and wave forecasting
• Weather-window and operability studies
• Air-quality and dispersion modelling
• Air-quality forecasting systems based on chemical transport models: emission measurement, inventories and operational runs
• Extreme-weather prediction: observation-network improvement, model analysis and sensitivity studies, and data-assimilation systems
• Monitoring-data analysis and reporting
• Decision support for weather-sensitive operations