Deep Argo Development Cruise R/V Tangaroa June 2014 Stephen Piotrowicz 1, Philip Sutton 4, Denise Fernandez 4, Shelley Hersey 3, Carol Janzen 6, Norge Larson 6, David Murphy 6, Peter Sommerville 3, Esmee Vanwijk 2, Matt Walkington 4, Nathalie Zilberman 5 1 NOAA, U.S.A.; 2 CSIRO, AUS.; 3 LEARNZ, N.Z.; 4 NIWA, N.Z.; 5 Scripps Institution of Oceanography, U.S.A.; 6 Sea-Bird Electronics
Deep SOLO Float characteristics 6000 m depth # of dive cycles: 160 (3.5-day cycle for the prototypes at present) Telemetry: Iridium Surface time: 30 minutes Hydraulic pump runs at 50 m (guess bottom depth to 1000 m), at max depth, and during ascent to maintain vertical speed. 2-meter wire rope used to feel the ocean bottom without snagging on the bottom or entraining mud. Wire length and density chosen for damping and keeping the float from hitting the bottom. Mass: 25 kg (SOLO 2 weighs 18.6 kg ) Housing: 13-inch glass ball Float will not hit bottom unless ocean bottom depth is less than 600 m from target depth for a 21 g/m, 2-m wire chain.
Deep SOLO Float characteristics New CTD from Seabird = SBE61 Energy/cycle is 24.3 kj to 5500 m Buoyancy is 21.6 kj (89%), communication is 0.36 kj (1%), and CTD is 2.4 kj (10%). (SOLO 2 needs less than 10 kj/cycle to 2000 m) Speed of ascent & descent is ~ 6 cm/s (SOLO 2 speed is 12 cm/s) Continuous profile to 500 m, at 2 dbar averaging bins, and then switch to discrete-mode: 5 dbar to 1000 m, 10 dbar to 2000 m, 20 dbar to 3000 m, 50 dbar to 6000 m (adjustable). Other types of deep floats include deep Apex (6000 m), Ninja (4000 m), and Provor (4000 m)
Accuracy Goals for Deep Argo Pressure +/- 3 decibars Kistler specified accuracy is +/- 7 decibars out of 7000 Temperature +/- 0.001 C Salinity +/- 0.002 PSU
Temperature Comparison SBE 35 reference thermometer Calibrated over oceanographic range against standards grade platinum thermometer Calibrated at triple point of water and gallium melt point SBE 3 thermometer Calibrated against secondary reference SBE 3s which are in turn calibrated against standards grade platinum thermometer Calibrated in small baths (60 micro-degrees standard deviation over 10 minute calibration interval) SBE 61 Calibrated against secondary reference SBE 3s which are in turn calibrated against standards grade platinum thermometer Calibrated in large baths (600 micro-degrees standard deviation over 10 minute calibration interval)
Upper plot has all points, lower plot is zoomed in SBE 61 is cooler than SBE 9 due to viscous heating of SBE 3 probe in 30 ml/sec pumped flow (SBE 61 pumps at 8 10 ml/s) Agreement below 2000 decibars is +/- 0.001 with the exception of 5 points out of 185 Temperature Comparison SBE 61 SBE 9 Primary Pressure in Decibars Pressure in Decibars 0 1000 2000 3000 4000 5000 6000 Temperature Difference (C) SBE 61 - SBE 9 0 1000 2000 3000 4000 5000 6000 5578 5581 5579 Temperature Difference (C) SBE 61 - SBE 9 5578 5581 5579
Upper plot has all points, lower plot is zoomed in Data is noisier than previous slide due to physical location of probes Agreement below 4500 decibars is +/- 0.001 with the exception of 3 points out of 142 Temperature Comparison SBE 61 SBE 35 Pressure in Decibars Pressure in Decibars 0 1000 2000 3000 4000 5000 6000 0 1000 2000 3000 4000 5000 6000 Temperature Diff (C) SBE 61 - SBE 35 5578 5581 5579 Temperature Diff (C) SBE 61 - SBE 35 5578 5581 5579
Error Caused by Thermal Disequilibrium of Strain Gauge Pressure Sensor Heat flow between the environment and the strain gauge causes an error due to temperature disequilibrium across the bulk of the strain gauge Thermal disequilibria distorts the strain gauge, causing the error Calibrations done with a 1.5 2 hour soak at temperature Exacerbated by 1 m/s profiling rate of this test (high dt/dt) Expect the error to be lower at Argo profiling rate of 10 20 cm/s (low dt/dt)
Improving Pressure Accuracy Dynamic error can be improved by mechanical design May impact length and mass of SBE 61 Mathematical correction seems intuitively possible but there are concerns about adding complexity to the calculation and noise to the signal
Salinity Comparison SBE 61 - Discrete Agreement between SBE 61 and discrete samples below 2000 decibars SBE 61 5578 and 5581 are fresh 0.0 to -0.004 SBE 61 5579 is salty to within 0.001 to 0.005 Comparison with SBE 9 is offset due to temperature difference (SBE 61 SBE 9) Pressure in Decibars Pressure in Decibars 0 1000 2000 3000 4000 5000 6000 0 1000 2000 3000 4000 5000 6000 Salinity Difference (psu) SBE 61 - Discrete Salinity Difference (psu) SBE 61 - SBE 9 5578 5581 5579 5578 5581 5579
Status and Work In Progress Temperature Meeting the accuracy goals Pressure Close to desired accuracy but can see some room for improvement Improve transient temperature error through mechanical design work Further work on accuracy via calibration process Conductivity Continue experiments on cell drift Salinity Close to desired accuracy, refinements in pressure and conductivity may yield improvements
That system, known as Argo, is one of the scientific triumphs of the age, http://www.nytimes.com/2014/08/12/science/in-the-ocean-clues-to-change.html