摘要：High-frequency (diurnal) temperature variations occur simultaneously at multiple depths separated by meters of snow in at least several and probably many Arctic and Antarctic thermocouple datasets. These temperature variations cannot be caused by heat conduction from the surface because their amplitudes are too large and there is no phase lag with depth, and they cannot be caused by heat advection because the air flux required is greater than is available. Rather, the simultaneous temperature variations (STVs) appear to originate within the box that houses the data logger as thermocouple-like offset voltages, wire heating or thermistor error. The STVs can be corrected by requiring that the temperatures vary smoothly with time at the greatest depth at which temperature is measured. The correction voltage determined in this fashion, when applied to the thermocouples at other depths, corrects the entire dataset. The method successfully removes STVs with 24 hour period that are up to 3.8°C in amplitude, and is superior to the averaging techniques commonly used to correct thermocouple data because it introduces no spurious (non-physical) temperature variations. The correction method described can be applied to all thermocouple data where temperature measurements have been made at depths > ~0.5 m into the snowpack. The corrections should allow more physical process and parameter information to be extracted more confidently from existing firn temperature data. Identification of the STVs and their probable cause also suggests how better data might be collected in the future.
