README for cBathy
When you type 'help {toolboxName}' where you insert your own toolbox name, will give the Contents of the toolbox.
The outline below shows the order of how analysis routines are called.
This toolbox has been extacted from the full CIL version and SHOULD work on a standalone basis. In doing this, a number of checks have been removed since these depend on CIL databases and standards.
Two demo files are provided as well as example data. 'democBathy' runs the main cBathy processing to create phase 1 and 2 results (fDependent and fCombined) and plots the results. This is currently set to analyze a test data set, testStack102210Duck.mat, one used in the original cBathy paper in 2013. 'demoKalmanFiltering' takes an example day of cBathy results and performs Kalman filtering, then plots the results. The set cBathy input data is constained the folder 102210cBathys and is again from an example from the 2013 paper. NOTE that the results for the Kalman demo are stored in a different location, simply for demo purposes (since many people may end up using the toolbox). Normally you would save the results back to the input file location but now with the runningAverage field filled in. For comparison, the CRAB survey from three days prior is included in the file 19-Oct-2010FRFGridded.mat.
The first part (democBathy) needs only the stack information and the tide, although if the tide is not found the routine should continue but leave nans in the tide fields. The Kalman stage needs a process error routine specific to your site (findProcessError) and this usually needs the wave height, H. Adequate results can be found by guessing H, for instance H = 1 m in this demo. It makes no sense to Kalman filter if you have not removed the tide (averaging a non-stationary signal). The toolbox contains a routine doKal.m which is a CIL version, included only to illustrate the steps that we normally take. Note the use of fixBathyTide in Kalman filtering to make sure the tide has been compensated for. If tide has been fixed in a previous step, this routine will do no harm.
Note that the supplied example tide routine, fixBathyTide, is a CIL routine that normally looks for two input arguments, the first of which is the station name that we extract from the stack name (to ensure that you can't possibly call the tide routine for an incorrect station). If the stack name is not standard, the routine will assume that this is not a CIL station and will fall back to a single input argument version (time), leaving the user responsible for chosing the correct tide location.
Please let us know if problems arise HERE
Logicial Outline for analyzeBathyCollect version 2
- no tiles, each x,y done independently
- ie., csmInvertKAlpha does one x,y at a time, returns only fDependent
analyzeBathyCollect
input: xyz, epoch, data (data from stacks), bathy (with params)
output: bathy (final results)
prepBathyInput( xyz, epoch, data, bathy )
output: f, G, bathy with empty data, allx, ally
parfor loop on allx, ally
subBathyProcess( f, G, xyz, ax, ay, params )
output: fDependent
spatialLimitBathy( f, G, xyz, params, ax, ay )
output: subG, subxyz
csmInvertKAlpha( f, subG, subxyz, ax, ay, params )
output: fDependent
stuff fDependent back into main bathy
end parfor
convert bathy fDependent to final output