Deep-Ocean Bottom Pressure Measurement: Drift Removal and Performance

Abstract

Sixteen records from seven Digiquartz deep-ocean bottom pressure sensors have been in deployments of 3?12 month duration under the Gulf Stream in depths of 3300 to 4400 m. Particular attention is given (i) to characterizing any observed drift in their calibration in relation to their construction (bellows or Bourdon-tube) and to their prior history of pressurization, and (ii) to estimating and removing this drift from the records. Bellows-type sensors exhibited significant drift (0.2 to 0.85 db) in all of their deployments. Bourdon-tube sensors had less drift in their first deployment (0 to 0.45 db), and in subsequent deployments had either no drift or a small drift with different shape that may have been due to clock-frequency drift. An exponential decay with time [?exp(?αt)] was found to best represent the drifts, such a curve was fit in a least-squares sense to each pressure record and then subtracted from it Careful attention is given to estimating the uncertainty of the residual ?dedrifted? records, which is 0.02 dh for records that are at least a year long; the stability over a few days and resolution of these measurements is better than 0.001 db. As a consistency cheer, neighboring pairs of bottom pressure records are used to calculate geostrophic currents from their differences, and the comparison with directly observed currents confirms that the error in drift removal may be less than 0.02 db. Typical amplitudes of the deep-ocean tidal- and derided signals are respectively 0.7 and 0.13 db in this region, so that we infer that this methodology is suitable for studies requiring knowledge of deep-ocean dynamic pressures even for subtidal mesoscale periodicites.