Tracking ALMA System Temperature with Water Vapor Data at High Frequency

The ALMA observatory is now putting more focus on high-frequency observations (frequencies from 275-950 GHz). However, high-frequency observations often suffer from rapid variations in atmospheric opacity that directly affect the system temperature $T_{sys}$. Current observations perform discrete atmospheric calibrations (Atm-cals) every few minutes, with typically 10-20 occurring per hour for high frequency observation and each taking 30-40 seconds. In order to obtain more accurate flux measurements and reduce the number of atmospheric calibrations (Atm-cals), a new method to monitor $T_{sys}$ continuously is proposed using existing data in the measurement set. In this work, we demonstrate the viability of using water vapor radiometer (WVR) data to track the $T_{sys}$ continuously. We find a tight linear correlation between $T_{sys}$ measured using the traditional method and $T_{sys}$ extrapolated based on WVR data with scatter of 0.5-3%. Although the exact form of the linear relation varies among different data sets and spectral windows, we can use a small number of discrete $T_{sys}$ measurements to fit the linear relation and use this heuristic relationship to derive $T_{sys}$ every 10 seconds. Furthermore, we successfully reproduce the observed correlation using atmospheric transmission at microwave (ATM) modeling and demonstrate the viability of a more general method to directly derive the $T_{sys}$ from the modeling. We apply the semi-continuous $T_{sys}$ from heuristic fitting on a few data sets from Band 7 to Band 10 and compare the flux measured using these methods. We find the discrete and continuous $T_{sys}$ methods give us consistent flux measurements with differences up to 5%. Furthermore, this method has significantly reduced the flux uncertainty due to $T_{sys}$ variability for one dataset, which has large precipitable water vapor (PWV) fluctuation, from 10% to 0.7%.