This function:

1. Divides the time series into down and upcasts
2. Applies T/S response matching code at the profile level according to several user-specified options
3. Computes chi from microconductivity using the full time series or large chunks of data in between gaps.

T/C response matching code

• Response matching code is not necessary for processing microconductivity.
• We want to do the response matching for SBE temperature and salinity. It’s still unclear whether we should do that profile-by-profile or with the whole time series.

apply_response_matching_code

= 0

• No response matching code is applied

= 1

• Response matching code is applied profile as long as it is at least 10 m deep
  • That should be plenty because you do it on the time series, not the profiles

use_old_code

= 1

• “Use old code” applies San’s legacy response matching code

= 0

• New code applies Matthew’s new response matching code, which applies a low pass filter

use_phys

= 1

• Uses a physical model with default parameters

= 0

• Uses polyfit

compute_spectra

= 0

• uses default parameters

= 1

• Computes spectra and polyval models on the fly

do_thermal_mass_corrections

= 0

• Does not correct for conductivity thermal mass error

= 1

• Adds adjustment for the conductivity thermal mass errr

Microconductivity processing

add_chi_micro_MHA_v3

• Microconductivity processing is done on long chunks of data (multiple profiles) without gaps in data longer than 1 second. It requires at least 2000 samples (at 16 Hz, that’s 2 minutes of data)

• Assumes microconductivity sampling rate is 320 Hz. For older, pre-SOM versions of the epsi/fctd boards, use 160 Hz instead of 320.

  remove_sbe_preemphasisMHA

  • corrects the microconductivity time series using gain = 21 and offset = 2.72

  DoOneChi_MHA_v3

  • Computes chi for the time period specified by tlim using the parameters in chi_param following steps in the appendix of Alford Pinkel 2000

    

  • FCTD_Default_ChiParam:

    fslow: 16            %Not used
       fs: 160           %Used in chi_from_variance but default should be 320
     gain: 22
   offset: 2.7200
     dTdC: 10
     nsec: 0.5000
   dt_sec: 2
     kmin: 2
     kmax: 200
  min_spd: 0.5000
        D: 1.4000e-07
  version: '0.2'
   plotit: 0
  

  • Integrates the filtered d(ucond)/dz spectrum from kmin = 1 cpm to kmax = 12.5 cpm

    • The spectrum is filtered with an anti-aliasing sinc2 filter

    

    Calculate chi from the P/sinc2 spectrum, between kmin and kmax:

    chi1=6*D*dTdC^2*sum(pcorr(ichi),'omitmissing')*dk;