Observing horizontal wind transects over an urban terrain with scintillometer and lidar technologies

Deeper understanding of urban roughness-sublayer flow is important for example in air quality, wind loading, pedestrian comfort, and numerical weather prediction. Unfortunately, point measurements in the roughness sublayer are notoriously specific and of limited wider use. Remote-sensing offers the opportunity to measure area- or line-averaged flow that ought to be more representative than point measurements. In this work we focus on an inter-comparison of wind measurements over Helsinki as measured from anemometers, lidar and scintillometer.

As part of an urban-climate research network, Helsinki UrBAN (http://urban.fmi.fi), there are multi-year installations of many research-grade instruments. Three key instruments are exploited here. [1] The scintillometer, Scintec BLS-900, path is 4.2 km long, and it has two transmitters (from the same unit) and one receiver which are respectively 52.4 and 23.9 m agl. The mean building height is 16 m and the beam's effective height is 33.6 m. The most standard estimate of wind from a scintillometer is the wind speed perpendicular to the path: it is most-commonly estimated based on either spectral techniques or the time-lag-correlation function of the two beams. [2] Sonic anemometers are located close to the each end of the scintillometer path. [3] The lidar (HALO Photonics Streamline) is dopplerized and scanning: this allows the estimation of transects of wind vectors along the scintillometer path (by employing a lidar scanning strategy that uses two near-by lidar beams and trigonometric assumptions; Wood et al. 2013, STOTEN 442: 527--533).

An intensive measurement period took place on 1–14 October 2013, when the lidar made more thorough horizontal scans along the scintillometer path. Wind speed rmse of sonic anemometer (regarded as truth) compared with nearby lidar data was 0.57 m/s. For an average along the whole 4.2 km path, lidar–scintillometer comparison gave perpendicular wind speed rmse of 0.8–1.3 m/s (dependent upon data-screening). Given that scintillometers are most commonly studied over more homogenous terrain, here the lidar scans along the scintillometer path enables us to study specific cases where the perpendicular wind speed along the path was heterogeneous, and see how this influences the scintillometer measurements.

We will show some limits of the techniques, caused by assumptions and accuracy limits of the instruments. Altogether, there is clearly potential in remote-sensing of the roughness sublayer, but challenges exist.