APRP fix

This version fixes a slight inaccuracy in the APRP code.

Extend functionality of check_and_download

• Now enabled to specify a directory to download to, rather than having to give the full target filepath.
• Will automatically create the target directory if it doesn't exist, rather than raise an error.

Humidity and UTCI again

In this release the humidity relationship is rolled back to Alduchov and Eskridge (1996) as I didn't have a literature reference for the one I was using.

UTCI has the ability to take a surface pressure variable, which affects the humidity calculation and UTCI if specific humidity is provided to be converted to relative.

On the repo side, I think the leading "v" from the version tag no longer appears on conda releases, which is ideal.

Humidity fixes for UTCI

One main fix is the specific humidity calculation in the UTCI code, which was previously not covered by the tests and showed odd results. The humidity and UTCI code calculation of saturation vapour pressure is now the same.

The output from cos_mean_zenith_angle is transposed and is now on the same grid as the input data (which makes more sense).

The climate inputs to the utci module are now dicts of CMIP-style variables for consistency with the aprp module. This avoids the user having to remember which order to specify the variables in.

Addition of mean solar zenith angle

Code added to calculate the mean solar zenith angle and lit fraction over a time period, for example a 3-hourly climate model output. Support exists for all calendars available in cftime. This is useful for solar energy and UTCI applications as it enables us to estimate what the time-averaged direct normal irradiance from a climate model might be given the total horizontal (rsds) and diffuse horizontal (rsdsdiff).

One other change: twolayermodel has default year length set to 365.24219 days, with option to specify it.

Some repository setup and refactoring

Time for a release, as other projects will want to use utils.

• added: utils module
• added: geometry module
• removed: masked array keyword in humidity
• fixed: wrong global mean latitude weighting in aprp

Addition of Geoffroy/Held two layer model

This minor version adds in the Geoffroy/Held two layer model. Thanks to Glen Harris.

from climateforcing.twolayermodel import TwoLayerModel
scm = TwoLayerModel()
out = scm.run()

See the code in src/climateforcing/twolayermodel for details.

Fix netCDF4 dependency in setup.py

This could be a long afternoon

Fix README and setup.py for PyPI

v0.0.1

one more setup.py error

First release of UTCI and APRP code

Working functions and tests inplemented for utci and aprp. Lots to do on repository setup still.