Integrated water vapor over the Arctic: Comparison between radiosondes and sun photometer observations

Abstract

The amplification of global warming because of the feedbacks associated with the increase in atmospheric moisture and the decrease in sea ice and snow cover in the Arctic is currently the focus of scientists, policy makers and society. The amplification of global warming is the response to increases in precipitation originally caused by climate change. Arctic predominant increases in specific humidity and precipitation have been documented by observations. In comparison, evapotranspiration in the Arctic is poorly known, in part, because the spatial and temporal sparsity of accurate in situ and remote sensing observations. Although more than 20 observations sites in the Arctic are available, where AERONET sun photometer integrated water vapor (IWV) measurements have been conducted, that information have been barely used. Here, we present a comparison of IWV observations from radiosondes and AERONET sun photometers at ten sites located across the Arctic with the goal to document the feasibility of that set of observations to contribute to the ongoing and future research on polar regions. Sun photometer IWV observations are averaged for three-time windows; 30 min, 6 and 24 h. The predominant dry bias of AERONET IWV observations with respect to radiosondes, identified at tropical and midlatitudes, is also present in the Arctic. The statistics of the comparison show robust results at eight of the ten sites, with precision and accuracy magnitudes below 8 and 2% respectively. The possible causes of the less robust results at the other two sites are discussed. In addition, the impact of selecting other temporal coincidence windows in the average sun photometer IWV used in the comparison were tested. Auto-correlation in diurnal sun photometer IWV could produce appreciable bias in the statistics used for the comparison. We suggest using only one pair of values per day, consisting in the daily mean IWV sun photometer and the IWV radiosonde observation value. This feature should be valid also for comparison of IWV from sun photometer and other instruments. Maximum 10% error level of IWV from sun photometer observations, when compared with radiosondes, have been found for the Arctic. It is in the same order of magnitude than at tropical and middle latitudes locations. It has been demonstrated the feasibility of AERONET IWV observations in the Arctic for research on this variable. AERONET standard instruments and its centralized-standard processing algorithm allow its IWV observations to be considered a relative standard dataset for the re-calibration of other instrumental IWV observations assuming radiosondes as the absolute standard dataset.