Fast-Flux: Surface Heat Fluxes Measurements during a Total Solar Eclipse

Rapid changes in solar insolation during a total solar eclipse produce highly nonstationary surface-layer turbulence and turbulent fluxes over time scales much faster than those associated with the morning or evening transition periods. To address issues related to nonstationary time series, previous experimentalists reported surface heat fluxes during an eclipse using compressed block-averaging timescales that approached the lower theoretical limits for the eddy covariance technique (e.g., five minutes). However, these shorter averaging times span the order of the time of the event itself (time of totality) while sampling only a subset of turbulent eddies. Hence, questions associated with the timescale of atmospheric response to such rapid flux changes remain open. We present surface heat flux measurements from the Solar Eclipse Experiment for Boundary Layer Turbulence (SEE-BLT), for which a horizontally log-spaced array of nine eddy-covariance stations were deployed over homogenous terrain to circumvent the challenges inherent in assessing fluxes at sub-minute timescales. We find that fast-flux calculations are possible by treating the sensor array as a single instrument, and that the lower limit of the averaging time scale is determined by ensemble topology. We additionally find that the median biases in fast-flux calculations are associated with fixed scales of the array geometry and are vanishingly small, at only a few W-m^-2. This ensemble-based, fast-flux technique resolves (detects) an additional 30 W-m^-2 in in the peak heat flux magnitude during the period of eclipse totality compared to the heat flux magnitude determined by single-mast, traditional eddy-covariance techniques. Finally, we used the fast-flux technique to determine the atmospheric response timescale from totality was 4.75 minutes.