Temperature profiling is important for wind energy research given the well-known impact of atmospheric stability on power performance (Sumner and Masson 2006; Wharton and Lundquist 2012; Vanderwende and Lundquist 2012), turbine wake characteristics (Magnusson and Smedman 1994; Aitken et al. 2014; Dörenkämper et al. 2015), and on turbine loads (Sathe et al. 2013). Although profiles of temperature measurements along towers are often used to assess atmospheric stability, or temperature differences between two levels used to assess a bulk Richardson number, profiles of temperature from remote-sensing instrumentation can also be used to assess atmospheric stability. Some of these methods provide adequate vertical resolution across a turbine rotor disk to assess the stability of fine layers of the atmosphere.
A radio-acoustic sounding system (RASS) combines radio and acoustic techniques to sense profiles of virtual temperature. A high-power acoustic source emits a signal, and the radar senses the signal backscattered from variations in atmospheric refractive index to estimate the speed of sound (May et al. 1989). RASS is at a TRL > 8. Numerous studies have evaluated the accuracy of RASS measurements by comparing to Radiosonde and in situ tower observations. Root-mean-square differences are less than 1°C in virtual temperature (May et al. 1989; Lundquist et al. 2017; Bianco et al. 2016). Vertical resolution of a RASS is identical to that of the wind-profiling radar system with which the RASS is coupled: usually the wind-profiling radar/RASS system operates in wind-measurement mode for a large fraction of each hour and operates in temperature-measurement mode for the remaining small fraction of the hour. A RASS is associated with 915-MHz wind profilers, which usually have a nominal 60-m vertical resolution, with the lowest level approximately 100 m above the surface. RASSs associated with 449-MHz wind profilers usually have a nominal 100-m vertical resolution, with the lowest level approximately 200 m above the surface.
Microwave Radiometers provide regular and automated measurements of temperature and moisture profiles up to altitudes of approximately 10 km (Ware et al. 2003; Bianco et al. 2005; Cimini et al. 2011; Friedrich et al. 2012). By observing atmospheric brightness temperature in the K-band (22–30 GHz) and in the V-band (51–59 GHz) and applying radiative transfer equations and neural network retrievals, these instruments can provide estimates of profiles of temperature, liquid water, Humidity, integrated water vapor, and cloud liquid water path. The neural network retrievals rely on historical soundings from the vicinity (Ware et al. 2013) as a basis. Multiple commercial systems are available at a TRL of > 8 (Figure 15 shows one example). The Radiometrics microwave radiometer vertical retrieval intervals are 50 m between the surface and 500 m above ground level, 100-m intervals from 500 m to 2 km, and 250-m intervals from 2 km to 10 km, all with a temporal resolution of 1 minute. Microwave radiometers, as they measure profiles of temperature and moisture, should be able to provide an assessment of boundary layer height. Based on 1 year of comparisons between microwave radiometers and radiosonde profiles, Collaud Coen et al. (2014) found reasonable agreement between boundary-layer height estimates.
Finally, radiosonde launches can also provide profiles of temperature, pressure, humidity, as well as wind speed and direction. Radiosondes are small, expendable packages attached to a balloon that provides lift to carry the instrument aloft. By maintaining a radio connection to a ground-based receiver at the location of the launch, data from the ascent of the balloon are transmitted to the base. Additionally, these data are often used in evaluation studies of remote-sensing instrumentation (Lundquist et al. 2017; Bianco et al. 2016), although they are not ideal, considering that they drift with the winds and do not provide a strictly vertical profile. Studies of horizontal drift of radiosondes in the lowest 5 km of the atmosphere suggest median drift is less than 10 km below an altitudes of 5 km (Houchi et al. 2010). The ascent rate of radiosondes is typically between 4.5 and 6 m s-1; lower ascension rates can result in erroneous data high in the atmosphere as the temperature and relative humidity sensors are not aspirated properly. Radiosondes are generally considered a TRL of > 8.
Radiosondes, and their profiles of virtual potential temperature, are usually considered the most reliable approach for defining the height of the atmospheric boundary layer. Most investigations that assess whether remote-sensing platforms can quantify boundary-layer height evaluate the new approach by conducting a comparison of those platforms to soundings.