New implementation of Knutson scaling incl. fix to negative freqs and double counting (#734)

* avoid negative frequencies

* remove frequency correction

* remove intensity correction factors

* remove use of "variable" key

* add non-signficant changes plus correct error in increase percentage change

* fix freq correction

* fix bug in test: assert true should be assert equal

* update test due to application of only freq changes - and sometime new ones

* adapt tests in TropCyclone

* fix name and adapt climate scenario test

* reinsert check that no neg freq are modeled

* test_trop_cyclone: adaptation to tag refactoring

* pylint

* throw warning when negative freq

* handle neg freq_scaling; refactor code

* new tc scaling

* (partial) tests update

* changes in docstrings, gsmt output as dict, percentiles as string

* fix typo in dict assignment

* add info on available percentiles

* add lost function

* coherent baseline argument

* adapt tests to new implementation

* update tutorial

* update tests

* change docstring order, remove default values, change ref_year to target_year

* update tests

* add correct license statement

* spell out dict

* changes to docstring

* update percentiles docstring

* update RCP values for better consistency and get rid of renaming dictionary

* adjust file description

* spell out variables for operation

* spell out GMST

* spell out number of rcps

* make windows_props as default params

* simplify code

* apply no changes for basins we have no info for

* allow retrieving yearly projections and simplify use of constants

* beautification

* add docstring for new variable

* update license statement and better error raise

* fix typo

* update gmst info dictionary docstring

* explicit basin short names

* remove use of baseline argument

* remove tab, add default, shorten lines

* update test

* refactor gmst_info dictionary

* add ref for gmst data

* remove trailing whitespaces

* change Knutson data structure

* bring back new functions lost after merging develop

* update tutorial

* update unit tests tc_cc

* simply var name and code

* improve unit tests for cc scaling factors

* update tests for rescaled future tropical cyclones

* update var names

* update docstring

* update docstrng

* add warning when no categories

* Update CHANGELOG.md

* bring back original notebook

* update notebook

---------

Co-authored-by: emanuel-schmid <[email protected]>
Co-authored-by: Benoit P. Guillod <[email protected]>

Author: 3 people

CHANGELOG.md

@@ -16,6 +16,7 @@ Code freeze date: YYYY-MM-DD
 
 ### Changed
 
+- Improved scaling factors implemented in `climada.hazard.trop_cyclone.apply_climate_scenario_knu` to model the impact of climate changes to tropical cyclones [#734](https://github.com/CLIMADA-project/climada_python/pull/734)
 - In `climada.util.plot.geo_im_from_array`, NaNs are plotted in gray while cells with no centroid are not plotted [#929](https://github.com/CLIMADA-project/climada_python/pull/929)
 - Renamed `climada.util.plot.subplots_from_gdf` to `climada.util.plot.plot_from_gdf` [#929](https://github.com/CLIMADA-project/climada_python/pull/929)