STATISTICAL STUDY OF BIVALVE HIGH FREQUENCY MICROCLOSING BEHAVIOR: SCALING PROPERTIES AND SHOT NOISE ANALYSIS

Abstract

We consider bivalve (oyster) behavior using high frequency valvometry data. These data are obtained in the field in the Arcachon bay, with light-weight electrodes of millimeters size, attached to the molluscs, and linked by flexible wires to an electronic control unit which measures valve activity. The latter are able to move freely and open their valve with an amplitude which is linked to environmental and physiological conditions, in order to filter surrounding waters. We consider here daily data recorded every 0.1 sec for 16 different animals, corresponding to a total of 1 728 000 data points. Valvometry data show important intermittent activity at high frequency, with frequent and sudden "microclosing" events (meaning partial closures), at apparently random times and with random amplitudes. The overall frequency spectrum of valve opening time series show power-law behavior with a slope β = 1.6 ± 0.1. The random succession of closing events with random amplitudes suggests to consider a shot-noise analysis, with two independent stochastic processes representing respectively the times of these discrete shot noise events and their amplitudes. This shows that methods and approaches devoted to extreme values, chaotic dynamics, recurrence time statistics in geosciences and climate studies can also be used to study living organisms and their interactions with complex environments such as their local ecosystem.