Accuracy of Sonic Anemometers: Laminar Wind-Tunnel Calibrations Compared to Atmospheric In Situ Calibrations Against a Reference Instrument

Abstract

Two Gill Solent Ultrasonic anemometers, models 1012R2 and 1210R3, weretested in field parallel measurements against a windvane based hot-film anemometerwith additional sensors for temperature and wet-bulb temperature, the MIUU (MIUU:Meteorology Institute, Uppsala University) instrument. This instrument was shown toretain its precision from laminar wind-tunnel tests when used in atmospheric turbulentflow. This contrasts strongly to the observed results for the two sonic anemometers,which were first calibrated in laminar wind-tunnel flow. Individual three-dimensionalcalibration matrices were constructed, and were shown to reduce the remaining calibration uncertainty for the wind speed to 0.4–0.8% for all azimuths and for angles of attack within ±40°. In the field intercomparison tests of the sonics against the MIUU instrument, it was found that the precision not only of the mean wind speed but of all second-order moments studied (variances and covariances, with and without temperature) deteriorated by a factor of typically three to four. Most of the scatter appears to be random, but in the case of the wind speed, a clear dependence on wind direction is found as well. It is concluded that the correction for the effect of the vertical supporting rods of the R2 and R3 instruments, which gives nearly perfect agreement for laminar flow, does not work entirely satisfactory in the natural turbulent flow. This, in turn, is likely to be so because of high sensitivity of the wake behind the cylindrical supporting rods to the character of the approach flow.