ALBERT

Notes: The thermal response of a marine heat flow probe has been modelled in detail by a semi-analytic solution for a 1-D multi-layered cylinder. the simulation was done for the probe used for measurements reported by Lister et al. (1990). These measurements were reduced by the conventional method of subtracting a time delay to maximize the linearity of temperature plotted against Bullard's (1954) F(α, τ) function. the frictional heating from entry into the sea-floor could be extrapolated to equilibrium by the use of small, but rather variable, time delay. However, the decays from a calibrated heat pulse, used to calculate thermal conductivity, required large time delays that increased up the probe. We have found that the heater-pulse decays are correctly reproduced by a multi-cylindrical model corresponding to the measured dimensions and materials of probe construction, except for some super heating at early times due to the non-cylindrical nature of the heater spiral. the increase in delay up the probe is modelled well by the increasing thermal resistance of incomplete oil fill between probe contents and metal probe, and not by any plausible disturbed sediment layer outside that tube. On the other hand, the only way to reduce the fitted time delays for the heat of probe entry was to inject that heat into a sheared layer of sediment around the probe about 1 mm thick.The extrapolations of dissipating frictional heat were all accurate to within 1 millikelvin; the poor fits at the top of the probe coincide with very little frictional heating, and so the errors remain small. the situation regarding conductivities is more complex, with the best fits by the delay-time method deviating by between -1 percent and +1.5 per cent, depending on details of the model and the time interval of synthetic data used for the reduction. This appears to be the limit of reliability for this method of reduction for a probe of conventional construction and the usual time range of data. Using data from large time is theoretically advantageous, but in reality would lose accuracy due to the limited remaining rise from a heat pulse of practical size, and to the onset of mechanical disturbance to the instrument on the sea-floor. Only a substantial shortening of the response time of probe temperature sensors to changes in the outer metal tube could improve the accuracy of conductivity measurement beyond this level.