Bias quantification and correction for meteorological sensors of an air-mass following drifting balloon

The atmospheric temperature profile in Arctic winter is an essential driver for the observed Arctic amplification of global temperature changes. During the cold season, the atmospheric temperature and moisture profiles in the Arctic result from the advection and transformation of air masses from lower latitudes. While the air masses move polewards over multiple days, they lose much of their initial heat and moisture. Capturing the complete transition process is challenging with fixed-in-place (Eulerian) observations.

 

Altitude-controlled drifting (CMET) balloons enable vertical soundings of the lower boundary layer over periods of several days and distances on the order of 1000 kilometers from an air-mass-following (quasi-Lagrangian) perspective, which is considered necessary for understanding Arctic air-mass transformations. In data from previous deployments, the sensors have been found to be prone to radiative bias, lag, and hysteresis. Precise measurements require distinguishing between sensor-related errors, small-scale atmospheric variability between adjacent ascending/descending legs, and the observed processes.

 

We use experimental setups established for radiosonde calibration to quantify the radiative bias in temperature measurements, as well as the constant offsets across different reference humidities and the temperature-dependent time lag for the humidity sensor. While the measured parameters are comparable to those of commercial-grade radiosondes, the vertical speeds of CMET balloons are much lower, resulting in reduced sensor ventilation.  This and other Arctic in-flight conditions are reproduced in our calibration experiments.

 

The radiative bias depends on the solar irradiance at the balloon's position. We estimate the incident solar radiation using the output of the solar panels surrounding the balloon's payload.

 

Our findings from the calibration experiments and irradiance estimation are applied to flight measurements using a combined processing tool, thereby providing an improved understanding of the data.