OSU Tidal predictions in python

Gary Egbert's group at Oregon State University produce inversions of open-ocean tidal signals to give predictions of ocean tides. There are a variety of solutions available at http://volkov.oce.orst.edu/tides/otps.html For this package, download the binary, non-netcdf, version of the predictions.

This software is based on a chain of software to decode these files. The files themselves are straightforward, but to get accurate tidal predictions the amplitudes and phases they provide must be corrected by astronomical tidal constants that depend on time. Hence, there is a bit of extra processing that is necessary (and somewhat arcane), provided by this package.

This package is extracted from https://github.com/ofringer/suntanspy and work that Matt Rayson (@mrayson) did to code up the older matlab code. The only changes are that we dropped suntanspy-specific libraries for interpolation in favour of scipy.interpolate, and use standard datetime functions to handle the dates.

Installing

This doesn't have a PyPI package yet, but its pretty easy to install into a pip or conda environment (if you aren't using pip, I'm not sure how to easily install, but putting otis_tide_pred.py in your local directory will work)

git clone https://github.com/jklymak/OtisTidalPrediction.git
cd OtisTidalPrediction
pip install -e .

Using

Once installed, you should be able to import otis_tide_pred as otp.

An example usage is in testOTIS.py.

import otis_tide_pred as otp
import numpy as np
import matplotlib.pyplot as plt

modfile = './DATA/Model_haw'

dates = np.arange(np.datetime64('2001-04-03'),
                  np.datetime64('2001-05-03'), dtype='datetime64[h]' )

lon = np.array([198, 199, ])
lat = np.array([21, 19])

h, u, v = otp.tide_pred(modfile, lon, lat, dates, z=None,conlist=None)

fig, ax = plt.subplots()
ax.plot(dates, u, dates, v)
plt.show()

Note that we have had to modify the paths in the file ./Data/Model_haw. This is an odd annoyance of the OSU data packaging.