Observations of boundary layer depth over an urban/rural transition

A field campaign has been undertaken in the UK to investigate the behaviour of the boundary layer over a rural/urban transition. The project was supported by HM Treasury under the Invest-To-Save Budget, and was a collaboration between the UK Meteorological Office, the University of Salford, the University of Essex and QinetiQ (formerly D.E.R.A.) in the UK. The project goal was to deploy two identical Doppler lidar systems near a rural/urban transition with the aim of measuring the behaviour of the boundary layer over the transition. The boundary layer parameters derived from the lidar data could be compared to parameters from the UK Met Office NWP Mesoscale model, the NAME dispersion model and the ADMS dispersion model with the aim of identifying potential improvements in the models.

The field campaign was carried out in July 2003 at RAF Northolt approximately 20 km west of central London. The conurbation of Greater London is made up of a conglomeration of numerous civic centres roughly aligned along the River Thames in an east-west direction. The urban/rural boundary is therefore not sudden change, but more of a gradual transition over many kilometres. In the area around the Aerodrome the land in the sector North-east turning anticlockwise to the South-east is on average more rural. The landscape from North turning clockwise to due South is more urban, and mostly residential 2 storey buildings. The topography to the southwest is gently rolling hills approximately +/- 10 m above the level of the Aerodrome. In the vicinity of the aerodrome the land is flat and approximately 50 m above sea level. Roughly 10 km to the North-west the land rises with the hills being approximately 100 m above sea level.

The field trial was carried out over a period of three weeks in July 2003. This period covered some of the hottest days ever recorded in North-western Europe. During the first week of the trial the UK was under anticyclonic high pressure conditions. There was little cloud cover. Both the day and night temperatures were extremely high as was the humidity. The winds were predominantly from the east. During the second week of the trial the there were several thunderstorms and heavy rainfall occurred. The third week the weather had settled down to more typical low pressure, cloudy conditions with south westerly winds and occasional showers.

The instruments present at the trial consisted of the two Doppler lidar systems, an automatic weather station and a sonic anemometer deployed on a 2m mast. The mast height was determined by the airport officials since the airport was in full operational use. Met data was also provided by the UK Met. Office in various forms including NWP Mesoscale model data (12 km resolution) and NAME dispersion model output. A third source of data was obtained from the UK Met. Office was AMDAR data. AMDAR is data recorded by commercial aircraft. The AMDAR was processed to retrieve data just over the London area in the time periods of interest. Standard surface synoptic data from our site and nearby London Heathrow were also available.

The primary variables measured using pulsed Doppler lidar are wind velocity along the direction of the lidar beam (radial velocity) and signal backscatter intensity. The measured wind velocities are volume averaged velocities, where the volume is defined by the range gate length (112 m) and the beam diameter (0.5 m at a range of 4 km). The maximum range of the velocity measurements is dependant on the signal-to-noise ratio (SNR). Where the SNR is high the maximum range is 9 km. The lidars also have a minimum range of approximately 700 m. The backscattered signal is retrieved at a rate of approximately 10 Hz, but the data is subsequently processed to retrieve the radial wind velocity at a rate of 0.2 Hz. The lidar beam can be scanned to produce 3-D wind fields.

The backscatter intensity of the signal is a function of aerosol concentration levels in the atmosphere, the water vapour and cloud droplet content and distance from the lidar. Since aerosols are well mixed within the boundary layer, the lidar backscatter intensity gives a good indication of the top of the mixing level in the atmospheric boundary layer. It also identifies the cloud base height.

Lidar data from the trial on the 9th July 2003 indicated a significant slope in the depth of the well mixed layer. Figure 1 shows a slope in the retrieved radial wind velocity along an east – west section, taken from 14:48 to 15:12 UTC. This cross section cuts through from the rural area on the left to the urban area on the right. The wind velocity is retrieved where the SNR is high enough. Where the back scattered signal is below the noise level a wind velocity is chosen randomly, hence where there is insufficient signal the plotted velocity becomes very noisy. The height at which the signal goes below the noise level does not actually define the top of the boundary layer. To measure the top of the boundary layer we look at the rate of change of SNR with height. In a boundary layer which is topped by a strong inversion layer we would expect to see a significant drop in the SNR at this level. Using the data from figure 1 we can calculate that in the rural sector that the boundary layer top is at a height of 1200 m. For the urban sector however the atmosphere is well mixed up to a height of 2000 m. This gives a difference in boundary layer height of 800 m over a distance of 8.8 km.

The lidar data from this day has been analysed and compared to AMDAR temperature profiles to examine how the structure of the boundary layer changes from the rural to the urban sectors.