invest-ucm-calibration 0.4.0

Automated calibration of the InVEST urban cooling model with simulated annealing

InVEST urban cooling model calibration

Overview

Automated calibration of the InVEST urban cooling model with simulated annealing

Citation: Bosch, M., Locatelli, M., Hamel, P., Remme, R. P., Chenal, J., and Joost, S. 2020. "A spatially-explicit approach to simulate urban heat islands in complex urban landscapes". Under review in Geoscientific Model Development. 10.5194/gmd-2020-174

See the user guide for more information, or the lausanne-heat-islands repository for an example use of this library in an academic article.

Installation

This library requires specific versions of the gdal and rtree libraries, which can easily be installed with conda as in:

$ conda install -c conda-forge 'gdal<3.0' rtree 'shapely<1.7.0'

Then, this library can be installed as in:

$ pip install invest-ucm-calibration

An alternative for the last step is to clone the repository and install it as in:

$ git clone https://github.com/martibosch/invest-ucm-calibration.git
$ python setup.py install

TODO

• Allow a sequence of LULC rasters (although this would require an explicit mapping of each LULC/evapotranspiration/temperature raster or station measurement to a specific date)
• Test calibration based on cc_method='intensity'
• Support spatio-temporal datasets with xarray to avoid passing many separate rasters (and map each raster to a date more consistently)
• Read both station measurements and station locations as a single geo-data frame

Acknowledgments

• The calibration procedure is based simulated annealing implementation of perrygeo/simanneal
• With the support of the École Polytechnique Fédérale de Lausanne (EPFL)