Matlab MEDD spike test

Purpose

The MEDD test is a spike detection test. Its aim is to detect simple spikes and vertically short burst of spiky points. MEDD stands for MEDian with a Distance.

Short description

The test is based on three main steps:

• first the computation of a sliding median with some customizations
• then limits are computed that are at relative 2-dimensional distance d from the median
• finally this limits are also computed for the density profile There is a spike if both the density profile and the (temperature or salinity) profile are out of limits. If there is no conductivity sensors, then the spikes in temperature are evaluated using a bigger d value.

Scripts

This is composed of three routines:

• The main call QTRT_spike_check_medd_main.m that includes the configuration, launches the other routines, and computes final compound spike detection (density + salinity/temperature parameters)
• QTRT_spike_check_medd.m that computes the median, call the relative_2D_distance routine and computes unitary spikes suspicions
• relative_2D_distance.m that computes the limits located at relative 2-dimensional distance d from a curve

Documentation

See the specification in the doc folder for a more detailed explanation of the algorithms.

Inputs (in the call of routine QTRT_spike_check_medd_main.m) :

• PRES: pressure values
• TEMP: temperature values
• PSAL: salinity values. This array should be set to NaN(size(PRES)) if not available.
• DENS: potental density values referenced to 0 dbar and computed as per TEOS-10. Density values in validation material are computed using Gibbs SeaWater library - see details in last section. This array should be set to NaN(size(PRES)) if not available.
• LAT: latitude of the profiles in degrees

Outputs:

• SPIKE_T: array of size(PRES) with a 1 on temperature spikes, It is set to NaN if the computation could not be made
• SPIKE_S: array of size(PRES) with a 1 on salinity spikes, It is set to NaN if the computation could not be made

Potential density computation:

Here is the piece of Matlab code executed to retrieve the potential density referenced to 0 dbar needed as input for the algorithm and using the Gibbs Sea Water library:

• PRES_no_neg=PRES;
• PRES_no_neg(find( PRES<0 | PRES > 11000 ))=nan;
• [SA,in_ocean]=gsw_SA_from_SP(PSAL,PRES_no_neg,LON,LAT);
• CT=gsw_CT_from_t(SA,TEMP,PRES);
• DENS=gsw_rho(SA,CT,0);