<a id="recipe-climate-change-hotspot-rst"></a>

# Climate Change Hotspot

## Overview

In the context of a changing climate, it is found that
not all regions change at the same pace and the same
way. The regions that change at a faster rate than the rest
of the globe are labelled as climate change hotspots. Estimating
the location and magnitude of the hotspots is important
for climate change adaptation, and it is usually computed using
the projected climate variables’ differences between the regional
and larger scales.

One issue when trying to evaluate projections of climate change
is the vast amount of information available from the Coupled
Model Intercomparison Project (CMIP) exercises. Additionally,
results from the CMIP phases 5 and 6 can be quite different,
therefore a comparison between the two multi-model ensembles
can be made to evaluate their differences and similarities. To
account for the projections scenario uncertainty, data from
three different end-of-the-century radiative forcings is given
in the recipe.

This recipe compares regional surface temperature and precipitation
against larger scale means to obtain the hotspot magnitudes for both
CMIP5 and CMIP6 in the 2.6, 4.5 and 8.5 $Wm^-2$ radiative forcings
by the year 2100 against the preindustrial Era
(RCP2.6, RCP4.5, RCP8.5 for CMIP5 and SSP1-2.6, SSP2-4.5, SSP5-8.5 for CMIP6).
Recipe based on the work by [Cos et al. (2022)](https://doi.org/10.5194/esd-13-321-2022).

Note: This recipe is currently set to evaluate the Mediterranean
hotspot (with bounds start_longitude: -10, end_longitude: 40, start_latitude: 30, end_latitude: 45) but it can be set to any other rectangular region.

## Available recipes and diagnostics

Recipes are stored in esmvaltool/recipes/

> * `recipe_climate_change_hotspot.yml`: Loads and ensembles the data,
>   : computes the necessary climate change hotspot diagnostics
>     and plots the results figures.

Diagnostics are stored in esmvaltool/diag_scripts/cos22esd/

> * `climate_change_hotspot.py`: Calculates the regional field hotspot
>   for temperature and precipitation and the 10-year rolling mean
>   timeseries for regional and large-scale temperature and precipitation.
> * `hotspot_plotter.py`: Gathers the data output from the
>   `climate_change_hotspot.py` script and plots the hotspot fields and
>   the rolling mean timeseries [Figures 2, 3, S2 and S4 by
>   [Cos et al. (2022)](https://doi.org/10.5194/esd-13-321-2022).].

## User settings in the recipe

1. Script `climate_change_hotspot.py`

   *Required settings for script*
   * `baseline_period`: Historical period that serves as a reference to compute the time anomalies.
   * `future_periods`: List of the two future periods given in years (“YYYY-YYYY”) where the hotspot will be computed.
     Following the format [future period #1, future period #2].
   * `region`: list of longitudes and latitudes that enclose a rectangular region.
     In the form of [start_longitude, end_longitude, start_latitude, end_latitude].
   * `region_name`: Name of the region to be included in the provenance record.
2. Script `hotspot_plotter.py`

   *Required settings for script*
   * `baseline_period`: Historical period displayed in the figures’ titles.
   * `future_periods`: List of the two future periods given in years (“YYYY-YYYY”), following
     the format [future period #1, future period #2], used to identify the
     ancestor files and in the figure titles.
   * `region`: List of longitudes and latitudes that enclose a region.
     In the form of [start_longitude, end_longitude, start_latitude, end_latitude].
     Used in the title to identify the precipitation large-scale region.
   * `region_name`: Name of the region used in the plot titles.

## Modifying the datasets and scenarios used

`recipe_climate_change_hotspot.yml` can be modified to use different scenario combinations.
The standard recipe uses data from scenarios with the radiative forcings 2.6, 4.5 and 8.5 Wm^{-2} (referred to as 26, 45 and 85),
but any combination of three scenarios from the following list can be used:

```yaml
26: "RCP2.6/SSP1-2.6"
45: "RCP4.5/SSP2-4.5"
60: "RCP6.0/SSP4-6.0"
85: "RCP8.5/SSP5-8.5"
```

To specify which datasets are available for each scenario, lists of datasets can be attributed to a
specific CMIP project and scenario between the `documentation`
and `preprocessor` sections of the recipe as follows:

```yaml
cmip6_85: &cmip6_85
   - {...dataset keys...}
   - {...dataset keys...}
cmip5_85: &cmip5_85
   - {...dataset keys...}
   - {...dataset keys...}
cmip6_45: &cmip6_45
   - {...dataset keys...}
   - {...dataset keys...}
cmip5_45: &cmip5_45
   - {...dataset keys...}
   - {...dataset keys...}
```

These different dataset sections will be called at each diagnostic as `additional_datasets` using the
anchors `*cmip6_85`, `*cmip5_85`, etc. as in the example:

```yaml
pr_cmip6_85:
 variables:
   pr:
     mip: Amon
     short_name: pr
     preprocessor: ensemble_members
     additional_datasets: *cmip6_85
 scripts:
   pr_cmip6_85:
     <<: *script_input
```

In case of wanting to use other scenarios, the datasets and diagnostics must be
changed maintaining the format `cmip{phase}_{scenario}` and `{variable}_cmip{phase}_{scenario}`.
For example, if we want scenario 60 instead of scenario 85, we would need to include the
files available for `cmip6_60` and `cmip5_60`, and the previous diagnostic would change to:

```yaml
pr_cmip6_60:
 variables:
   pr:
     mip: Amon
     short_name: pr
     preprocessor: ensemble_members
     additional_datasets: *cmip6_60
 scripts:
   pr_cmip6_60:
     <<: *script_input
```

Finally, if the datasets that need to be included in the multi-model means are common for all scenarios,
the datasets could be simplified to:

```yaml
cmip6: &cmip6
   - {...dataset keys...}
   - {...dataset keys...}
cmip5: &cmip5
   - {...dataset keys...}
   - {...dataset keys...}
```

Note that the diagnostics’ `additional_datasets` will need to be modified accordingly.

## Variables

* tas (atmos, monthly mean, longitude latitude time)
* pr (atmos, monthly mean, longitude latitude time)

## References

* [Cos et al. 2022](https://doi.org/10.5194/esd-13-321-2022), Earth Syst. Dynam., 13, 321–340

## Example plots

<a id="fig-climate-change-hotspot-1"></a>
![image](recipes/figures/cos22esd/tas_45.png)![image](recipes/figures/cos22esd/pr_45.png)

<a id="fig-climate-change-hotspot-2"></a>
![image](recipes/figures/cos22esd/scenario_combination_tas-tas_jja.png)
