Comparisons of temperature, pressure and humidity measurements by balloon-borne radiosondes and frost point hygrometers during MOHAVE-2009

We compare coincident, in situ, balloon-borne measurements of temperature (T ) and pressure (P ) by two radiosondes (Vaisala RS92, Intermet iMet-1-RSB) and similar measurements of relative humidity (RH) by RS92 sondes and frost point hygrometers. Data from a total of 28 balloon flights with at least one pair of radiosondes are analyzed in 1-km altitude bins to quantify measurement differences between the sonde sensors and how they vary with altitude. Each comparison (T , P , RH) exposes several profiles of anomalously large measurement differences. Measurement difference statistics, calculated with and without the anomalous profiles, are compared to uncertainties quoted by the radiosonde manufacturers. Excluding seven anomalous profiles, T differences between 19 pairs of RS92 and iMet sondes exceed their measurement uncertainty limits (2 σ ) 31 % of the time and reveal a statistically significant, altitude-independent bias of 0.5 ± 0.2 ◦ C. Similarly, RS92iMet P differences in 22 non-anomalous profiles exceed their uncertainty limits 23 % of the time, with a disproportionate 83 % of the excessive P differences at altitudes >16 km. The RS92-iMet pressure differences increase smoothly from −0.6 hPa near the surface to 0.8 hPa above 25 km. Temperature and P differences between all 14 pairs of RS92 sondes exceed manufacturer-quoted, reproducibility limits (σ ) 28 % and 11 % of the time, respectively. About 95 % of the excessive T differences are eliminated when 5 anomalous RS92-RS92 profiles are excluded. Only 5 % of RH measurement differences between 14 pairs of RS92 sondes exceed the manufacturer’s measurement reproducibility limit (σ ). RH measurements by RS92 sondes are also compared to RH values calculated from frost point hygrometer measurements and coincident T measurements by the radiosondes. The influences of RS92-iMet T and P differences on RH values and water vapor mixing ratios calculated from frost point hygrometer measurements are examined.