Last modified on Jun 20, 2024 14:59

Contributors: Marianna Benassi (CMCC), Silvio Gualdi (CMCC), Bas Amelung (TEC)

1. Short description of the dataset

With "Climate Suitability for Tourism indicators" we refer to a set of bioclimatic indicators designed to assess whether weather or climate conditions are suitable for touristic activities. These indicators are tailored for different kind of touristic activities. In particular, the Holiday Climate Index (HCI) is focused on urban tourism, while the Climate Index for Tourism (CIT) is focused on beach tourism. Both these indicators take into account different climate fields, in order to evaluate climate conditions from a touristic point of view. Different climate facets (precipitation, wind, cloud cover, and temperature) are considered, and to each of them a score is assigned. These scores are then aggregated, resulting in a total rating scale ranging from 0 to 100 for HCI, and from 0 to 7 for CIT (for further details on HCI see Scott et al., 2016; while on TCI see de Freitas et al., 2008). In general, the higher the value of the indicators is, the more suitable for touristic activities climate conditions are. In the framework of C3S European Tourism project, climate projections information is available. Climate projection values, based on Euro-CORDEX projections, are computed both in the historical period and for different future emission scenarios, in order to allow an assessment on how climate change will possibly shape the suitability of climate conditions from a touristic perspective.

2. Input data and pre-processing

The input data for the computation of HCI and CIT are the same: daily precipitation, daily total cloud cover, daily wind speed, daily maximum temperature and daily minimum relative humidity. In particular, relative humidity is needed for HCI scoring, where the effective temperature is taken into account. Effective temperature definition is given by the Missenard's formula (e.g.Gregorczuk and Cena, 1967):

where T is surface temperature in degree Celsius and RH is relative humidity. In this framework, maximum value of surface temperature and minimum value of relative humidity are used as proxy of daytime condition.

2.1. Climate projection data

Climate projections are the results of numerical simulations of the Earth climate over long time scales, typically until the end of the 21st century. These simulations may take into account the entire globe or a specific region. In the former case, Global Climate Models (GCMs) are adopted, while in the latter Regional Climate Models (RCMs) are used. In a climate change perspective, the aim is to provide a consistent representation of the climate system in the past and to assess the response of climate system to different possible future scenarios of greenhouse gas concentrations: the so-called Representative Concentration Pathways (RCPs).

In this framework, we use climate projection data from the Euro-CORDEX initiative - the European branch of the CORDEX program. The CORDEX program is sponsored by the World Climate Research Program and is focused on organizing an integrated international framework to produce regional climate change projections for all the land region world-wide.

Global climate models simulate weather and climate conditions over the large scale. The equations describing the behavior and the interactions of the different components of the Earth system (e.g. atmosphere, ocean, land surface etc.) are solved on set of discrete points – i.e. the model grid. The horizontal resolution of GCMs (i.e. the distance between two adjacent points of the model grid) is usually around 100 km. However, at the local scale climate signal is affected not only by the large- scale circulation, but also by several factors and processes (e.g. the effect of topography or of land- sea contrast) which cannot be properly represented with the typical resolution of a GCM. To solve this issue, regional climate models are used. RCMs basically increase the resolution of a GCM, but over a limited region. RCMs can reach a horizontal resolution of 10 Km, and hence resolve processes and phenomena which can be of particular interest for regional impact assessment. Regional models cannot be adopted alone, since at the boundary of the region of interest inputs from GCMs are needed. In this way, a consistent representation of climate conditions is guaranteed. This is the reason why we will identify each input model for the climate projection dataset as a RCM/GCM couple.

In the case of Euro-CORDEX data, the horizontal resolution is nominally 0.11deg, i.e. around 12.5 Km. The CORDEX data are available from the ESGF catalogue, and in this framework, the choice of the GCM/RCMs to be adopted has been based on the availability of the needed inputs. The limiting factor has been the availability of high frequency (i.e. sub-daily) outputs for relative humidity, necessary to derive the minimum relative humidity value to be included in the definition of effective temperature. The final set of GCM/RCM adopted is listed below together with the available climate change scenarios for each model. In general, the historical simulations span the period 1970-2005, while the scenario period span the period 2006-2100. The historical simulations are included to provide a reference for the future projections and are necessary to interpret trends in these projections.

3. Impact model

For the computation of both HCI and CIT, some ad hoc python scripts have been coded in the framework of the C3S European Tourism project. Even if the computation of the two raw indicators is common in both seasonal forecast and climate projection scripts, in order to better manage the I/O for the different kind of data, two separate scripts have been produced.

HCI is a rating varying from 0 to 100, where 0 defines condition potentially dangerous for tourists, while 100 indicates condition ideal for tourism. This score is defined as:

Here ET is the score from the effective temperature, CD is the score from the total cloud cover, PR the one from precipitation, and WN the one linked to wind speed condition. Given that for very strong wind and strong precipitation conditions negative values of the rating are foreseen, it is nominally possible to achieve slightly negative HCI values. The complete rating system for each facet is defined in Scott et al., 2016.

CIT score integrates the thermal, aesthetic and physical facets, taking into account the overriding capabilities of wind, precipitation, and lack of sunshine. Thermal conditions are expressed as scores on the nine point ASHRAE scale of thermal sensation, ranging from "very cold" to "very hot". The aesthetic component is represented by the percentage of cloud cover, whereas the physical aspect is addressed by including rain and wind. On days with less than 50% cloud cover, less than 3 mm rain and wind speeds of less than 6 m/s, CIT scores are fully determined by thermal conditions. In the remaining cases, CIT scores are co-determined by the limiting factor, of which rain is the most prominent one (resulting in the lowest ratings), followed by wind and cloud cover. Here, as in Morgan et al. (2000), the ASHRAE scale is based on the values of the skin temperature Ts.

Ts is determined following Green (1967) definition:

where TM is the maximum daily temperature (as before, proxy for daytime condition), s is the percentage of sunshine (here derived as reciprocal of total cloud cover), and W is the wind speed. The remaining parameters in the equation are representative of the thickness of clothing (h), the metabolic rate (M), and the albedo of clothing and skin (A). Here the standardized values of 0.008 (cm), 25 (cal/s), and 0.45 respectively are used, as in Morgan et al. (2000). In the following tables the link between skin temperature and ASHRAE scale (adapted from Morgan et al., 2000) and the CIT rating scale (adapted from De Freitas et al., 2008) are shown.

For both the indicators, and for both seasonal forecast and climate projection data, together with the raw daily value of the indicators, the number of days per month with good/fair/unfavorable conditions are computed. For HCI we define:

• good conditions: HCI > 70
• fair conditions: 50 < HCI < 70
• unfavorable conditions HCI < 50

For CIT, we maintain the same categories defined as:

• good conditions: CIT=5,6,7
• fair conditions: CIT=4
• unfavorable conditions: CIT=0,1,2,3

The code workflow followed for both HCI and CIT in both the seasonal forecast and climate projection cases is the following:

1. Reading of input ECVs from Euro-CORDEX data.
2. Pre-processing of the input variables: aggregation of sub-daily ECVs to daily values, homogenization of units, computation of effective temperature from maximum surface temperature and minimum relative humidity (for HCI)/ computation of skin temperature (for CIT).
3. Computation of the daily values of the indicators: definition of the proper score for each climate facet and combination to derive the daily value of the indicator. Multi-model statistics for these daily index values are additionally served at 10-day, monthly and seasonal time aggregations.
4. Computation of the derived indicators: for each month, computation of the number of days with good/fair/unfavorable conditions.
5. Writing of output files: the daily values of the indicator and each derived indicator are saved in netcdf files (following the common data model convention).

4. Sectoral Impact Indicators

The outputs of the computations described in the previous sections may be downloaded from the dedicated page of the Climate Data Store catalogue. In the following, we will show some example of the analysis which can be performed with the data available on the CDS. A complete set of analysis will be included in the dedicated app, as soon as it will become publicly available.

4.1. Climate projection data

To assess the potential role of climate change in shaping the suitability for tourism of different locations in the Euro-Mediterranean sector, climatologies of both HCI and CIT are computed starting from daily values. A climatology is a long-term average of a given variable, typically computed over a time period of 20 or 30 years. A monthly climatology will produce a mean value for each month over the period of interest (i.e. the average of all the January mean values for that 20 years, the average of all the February mean values for that 20 years and so on). The mean monthly values obtained in this way represent the "typical" monthly conditions in that climate state (e.g. in that period/under that emission scenario). The same approach can be applied also for example for seasonal or sub- monthly values.

In figure 1 we show a simple metric allowing to emphasize the effects of climate change under RCP8.5 scenario. In the panel the differences of winter (DJF) HCI conditions in the future (from left to right: near-term future, mid-term future, long-term future) compared to present day reference values are reported. The average values of HCI are foreseen to increase, especially over the Mediterranean sector.

Figure 1: DJF HCI climatologies under RCP8.5 scenario: from left to right differences between near-term future (2021-2040) DJF climatology and present day (1986-2005) DJF climatology, mid-term future (2041-2060) DJF climatology and present day DJF climatology, and long-term future (2081-2100) DJF climatology and present day DJF climatology. Thesesclimatologies have been computed from the HCI projection ensemble average.

A similar approach may be adopted also for the pre-calculated indicators (i.e. the number of days per month with good/fair/unfavorable conditions). For instance, in figure 2 the number of days per month with good HCI conditions over London under different emission scenarios and for different future time slices is shown.

Figure 2: Number of days per month with good HCI conditions over London: from left to right results from RCP2.6, RCP4.5, and RCP8.5 scenario. In each plot, the black bars represent the present-day reference values, the light-colored bars represent the near-term (2021-2040) projected values, the medium-colored bars represent the mid-term (2041-2060) projected values, and the dark-colored bars represent the long-term (2081-2100) projected values. These values have been computed from the HCI projection ensemble average.

5. Uncertainty or skill assessment

5.1. Climate projection data

For climate projection data, using an ensemble of different climate models allows to give an assessment on the uncertainty in climate projection data.

Different metrics may be adopted to show the spread across the different model projections (e.g. the variance/standard deviation across the ensemble members, the computation of different percentiles etc.)

In general, regions where the ensemble spread is higher may be considered as affected by more uncertainty, vice versa for regions where the ensemble spread is lower. In figure 4 we show the difference between the 90th and the 10th percentile for winter HCI conditions under the RCP8.5 scenario for the long-term future period (defined as in figure 1). In this case, it is clear how projections for Western Europe are affected by much more uncertainty than the ones for Eastern Europe. Analogues metrics may be computed for different time slices and/or emission scenarios.

Figure 3: Difference between the 90th and the 10th percentiles of the ensemble distribution of HCI long-term future (2081-2100) DJF values under RCP 8.5 scenario. Over Eastern Europe the HCI projections are affected by less uncertainty than the projections over Western Europe.

6. References

de Freitaset al., 2008: A second generation climate index for tourism (CIT): specification and verification. International Journal of biometeorology 52(5), 399-407

Gregorczuk and Cena, 1967:Distribution of Effective Temperature over the surface of the Earth.Journal International Journal of Biometeorology38 (6), 1179–1196.

Green, J. S. A., 1967: Holiday meteorology: reflections on weather and outdoor comfort. Weather22(4), 128-131

Morgan et al., 2000: An improved user-based beach climate index. Journal of Coastal Conservation6(1), 41-50.

Scott et al., 2016: An Inter-Comparison of the Holiday Climate Index (HCI) and the Tourism Climate Index (TCI) in Europe. Atmosphere7(6), 80

7. Related articles

• Page:
  Climate variables for cities in Europe from 2008 to 2017 documentation
• Page:
  Climate Suitability for Tourism Indicators (CST): Dataset description
• Page:
  Fisheries and marine properties
• Page:
  Fish Abundance and Catch Dataset: Products User Guide
• Page:
  Ocean Fronts Dataset: User Guide for Products