Intercomparison of wind speed, temperature, and humidity data between dropsondes and aircraft in situ measurements

Abstract. Airborne measurements of winds, temperature (T), and relative humidity (RH) are critical due to their importance for atmospheric processes. Data quality from these in situ measurements is difficult to assess and requires independent observations. This work intercompares for the first time in situ measurements from the Turbulent Air Motion Measurement System (TAMMS) of horizontal winds and T, and a diode laser hygrometer (RH) deployed on a HU-25 Falcon flying mostly within the marine boundary layer over the northwest Atlantic to an independent set of measurements from dropsondes launched from a higher-flying King Air. Leveraging data from 162 joint flights from these two spatially coordinated aircraft during the NASA ACTIVATE campaign in winter and summer seasons between 2020–2022, a total of 555 pairs of Falcon-dropsonde data points are identified within 30 km horizontal separation, minimal vertical separation (usually < 1 m), and within 15 minutes. This analysis is based on the following range of conditions experienced: altitude = ~0.1–5 km; T = -19 – 27 °C; RH = 1 – 100 %; wind speed = 0.2 – 42 m s^-1. Based on scatterplots, correlation coefficients, and mean (in situ – dropsonde) error (ME), intercomparisons reveal good agreement for wind speed (r = 0.95, ME = 0.21 ± 1.68 m s^-1), the u/v wind components (r ~ 0.96–0.97, ME ~ 0.03 – 0.16 (± 1.62 – 1.67) m s^-1), wind direction (r = 0.94, ME = 0.00 ± 0.22 based on cosine of direction angles), T (r = 0.99, ME = 0.00 ± 0.71 °C), and RH (r = 0.91, ME = -3.86 ± 10.74 %). Sensitivity analysis shows that binning data into categories of horizontal separation distance, clear versus cloud, winter versus summer, altitude range, and terciles of the values examined variables did not yield major changes except for RH where there was more deviation especially above 70 %. The effect of statistics was examined by relaxing the vertical separation distance criteria to expand the number of pairs to over 360,000, without much difference in intercomparison metrics. The effect of averaging more points for each instrument in the final 555 pairs was also shown to lead to minimal change in agreement. Overall, these results provide confidence in the performance of the various measurement techniques for airborne field campaigns.