Novel mass-aggregation-based calibration of an acoustic method of monitoring bedload flux by infrequent desert flash floods

The monitoring of bedload flux under flash flood conditions has been successfully achieved since 1992 using slot samplers in the semiarid Nahal Eshtemoa. In the present study, a surrogate bedload monitoring technique - the Japanese plate microphone - has been deployed and calibrated against data from the slot samplers. Since a slot sampler has a sensitivity threshold that becomes especially important when transport rates are low, different averaging periods should be considered for high and low fluxes. In order to overcome the deficiencies of time-based aggregation used hitherto, we have developed a new method involving mass aggregation and commensurably variable intervals, thereby enabling a more accurate analysis and optimizing the bedload sampler's capabilities. The data derived with this new method has then been utilized to calibrate the Japanese plate microphone. The Eshtemoa is an ephemeral gravel bed channel with a high proportion of fine gravel (< 0.02 m); for these conditions, acoustic sensors have not been calibrated as yet. Two multiple linear regression models incorporating the effect of median bedload grain size on pulse rate have been established to predict bedload flux and cumulative transported bedload mass. The coefficients in these models are statistically significant. Good predictions are obtained for bedload flux (adj. r² = 0.83) and for cumulative bedload mass (adj. r² = 0.98) during flood recession. Overall, the multiple linear regression models, used in conjunction with the mass aggregation method of estimating bedload flux, suggest that field calibration of acoustic devices is feasible under these conditions for ca. 90% of the duration of bedload transport.