demfish

Process biologging data from tagged demersal fish

This set of Matlab scripts is prepared for analysing output from a logger attached to demersal fish. A logger outputs time series of temperature and pressure. Amplitude and phase of diurnal and semidiurnal astronomical tides are used to estimate the location of the fish.

Requirements

We do not provide the following data with this software. You need to modify the code to read these data. Look for comments %%%IO in tseries_to_points.m. Example scripts are found in the example/ directory.

Logger output

Time series of temperature and pressure.

Tide model output

Time series of surface height hout(i,j,k) is necessary. hout(i,j,k) is the predicted surface height in meters, at longitudinal grid i and latitudinal grid j. The following variables are defined and used to specify the grid in tseries_to_points.m.

• LONGITUDE is a vector with a length of NX.
• LATITUDE is a vector with a length of NY
• DT is the time step in minute. Length of the time series is NT.

We used NAO.99b tidal prediction system.

Bathymetry data

The ocean depth data are necessary. The grid can be different from the grid used in the tide model output. If so, linear interpolation is used. The following variables should be defined in tseries_to_points.m.

• TOPOLON is a vector for longitude.
• TOPOLAT is a vector for latitude.
• TOPOZ is a matrix where TOPO(i,j) is the depth in meters at TOPOLON(i), TOPOLAT(j).

Numerical simulation

A novelty in the present approach is the use of temperature. Data of temperature near/on seabed is not usually available and we need to turn to simulation. We used output from JCOPE model. The output should be readable from points_with_temps.m.

How to run

Given all required data described above and relevant reading functions are implemented;

1. tseries_to_points.m to estimate candidate points.
2. points_with_temps.m to add temperature from simulation to the candidate points.
3. temps_to_track.m to examine candidate points and find most likely fish track.

Reference

Kawabe, R., Katsumata, K., in preparation

History

0.1 -- Repository created on GitHub

TODO

• Need to despike time series before analysis?
• Better gridding of topography near the coast -- probably all NaN's (v0.1).