Last modified on Oct 09, 2023 16:14

Contributors: Carley Iles (CICERO), Marit Sandstad (CICERO), Clemens Schwingshackl (CICERO), Jana Sillmann (CICERO)

Issued by: Carley Iles (CICERO), Marit Sandstad (CICERO), Clemens Schwingshackl (CICERO), Jana Sillmann (CICERO)

Issued Date: 07/05/2021

Ref:  C3S_D2.1_202103_Product_User_Guide

Official reference number service contract: 2020/C3S_COP_69_2020

Acronyms

1. Introduction

1.1. Executive Summary

This dataset includes the 27 ETCCDI indices (Sillmann et al. 2013a, Kim et al. 2020) as well as a non-exhaustive set of five heat stress indicators (Schwingshackl et al. 2021).

The various indicator values (ETCCDI and selected heat stress indicators) are provided for historical and four future simulations (SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5) included in the Coupled Model Intercomparison Project Phase 6 (CMIP6) and used in the 6th Assessment Report of the Intergovernmental Panel on Climate Change (IPCC) Working Group 1.

The indices are calculated for one ensemble member for all models that had the necessary daily resolved data to calculate the indices (total daily precipitation and mean, minimum, and maximum daily near-surface air temperature for the 27 ETCCDI indicators; daily mean sea level pressure, daily mean near-surface specific humidity, and daily maximum near-surface air temperature for the heat stress indicators) for both historical and at least two of the future projection scenarios.

In addition, the dataset includes ETCCDI indices calculated from four models that have a large number of ensemble members in order for the user to estimate the associated uncertainty in the spread of model outcomes. The selection criteria for the four models (CanESM5, EC-Earth3, MIROC6 and MPI- ESM1-2-LR ) was the relatively large number of members in the ensemble for the historical and SSP experiments included in the dataset (SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5), as well as having all the variables required to compute the indices.

Both the ETCCDI indices and heat stress indicators are commonly used by the climate science and impacts communities and thus a precalculated set of indices will be highly useful (Sillmann et al. 2013b). 

The CMIP6 dataset is an extremely important source of climate information, hence having indicators calculated for this dataset is all the more important.

The processing and calculations performed to compute the 27 ETCCDI indices and the five heat stress indicators are computationally expensive. Hence, if researchers can download them from the Climate Data Store (CDS) instead of calculating them for themselves, this adds clear value for the user community.

In addition, calculating the ETCCDI indices consistently with the appropriate statistical methodology for percentile-based indices is important (Zhang et al. 2005), and inconsistencies due to the use of improper algorithms employed for their calculation can be avoided if the setup used here is applied and referred to. The ETCCDI indices here are calculated using the climdex.pcic.R package, which was developed, evaluated, and approved by the ETCCDI (Sillmann et al. 2013a).

The heat stress indicators are calculated based on near-surface air temperature, humidity, and surface pressure. Effects from radiation and wind are not considered. The heat stress indicators thus represent indoor conditions or calm conditions in the shade but are not representative for sunlit outdoor spaces. This also applies to other effects from wind or radiation that can affect heat stress, such as windchill or altered indoor conditions due to radiation through windows or other openings. Regarding radiation, the indicators thus provide a conservative estimate of heat stress, as including radiation would generally lead to an increase in heat stress. Considering wind could lead both to decreased and increased heat stress estimates, depending on the prevailing ambient conditions (e.g., whether near-surface air temperatures are higher or lower than the human body temperature).

To facilitate the usage of heat stress indicators in combination with absolute thresholds, this dataset additionally provides bias-adjusted heat stress indicators. Bias adjustment is carried out using the ISIMIP3b bias adjustment method (Lange 2019) and employing WFDE5 v1.0 (Cucchi et al. 2020), as a reference dataset. Providing both bias-adjusted data and data without bias adjustment is of great value for climate and impact studies since the calculation of heat stress indicators as well as bias adjustment are computationally very expensive.

This dataset is related to the WFDE5 dataset (Cucchi et al. 2020), also provided through C3S, which was used for bias adjustment of the heat stress indicators.

This dataset has been produced by CICERO with funding provided through the ERA4C project ClimInvest (grant agreement No 274250), the ERA-NET project SUSCAP (grant agreement No 299600), the H2020 project EXHAUSTION (grant agreement No 820655), the Belmont forum project HEATCOST (contract No 310672) and the Norwegian Research Council project ClimateXL (grant agreement No 244551) and compounded on behalf of C3S.

1.2. Scope of Documentation

This documentation describes the calculation of the indices (as also documented in Sillmann et al. 2013a and Schwingshackl et al. 2021), the data sources and metadata information. Details on the datasets and metadata can be found in section 2.2 and appendix 1 of this document, and in metadata contained in the netcdf files. Section 2.3 describes the input data used with complete references to peer reviewed literature in which the input data are presented. Section 2.4 and references therein describe the methodology and definitions and algorithms employed. Sections 2.3 and 2.4 and references therein together should provide the reader with all necessary information to reproduce the dataset.

1.3. Version History

Version 1: First publication of the ETCCDI indices and heat stress indicators.

Version 2: The equation used in Version 1 to calculate WBT was erroneous. We updated the calculation method (see section 2.4. Method for more details) and recalculated WBT and WBGT. Version 2 contains these updated WBT and WBGT data.

1.4. Known issues

Version 1 of this dataset contained an inadequate equation for calculating WBT (using the WBT calculation used by Dunne et al., 2013), which led to unreliable WBT and WBGT estimates.

In the updated Version 2, WBT is calculated based on the methodology presented by Buzan et al. (2015), which combines equations from Davies-Jones (2008, 2009) to iteratively calculate WBT from near-surface air temperature, near-surface relative humidity, and surface pressure. We used a Matlab implementation of the WBT formula (https://github.com/jrbuzan/WetBulb.m/blob/master/WetBulb.m) and translated it into python. We used the new WBT estimates to calculate WBGT (see section 2.4 Method for more details).

2. Product Description

2.1. Product Target Requirements

The complete set of 27 ETCCDI indices is provided within the CDS catalogue based on the CMIP6 simulations for a number of experiments/MIPs with the aspiration that those simulations are available at the CDS. In addition, a set of heat stress indicators (HSIs) based on CMIP6 simulations is provided that can be used to measure heat-induced impacts on human health. Compliance of the datasets with minimum requirements to be used with the CDS Toolbox (netcdf CF compliance) is overall ensured. A full documentation of the new datasets is made available, including (1) guidance on the use of the indicator datasets, including relevant scientific documentation and validation of the methodology allowing a full traceability of the product generation: from the model input data to the final indicator; (2) completion of the required Quality Assessment reports which are being implemented within the C3S_513 Evaluation and Quality Control of the SIS contract.

2.2. Product Overview

2.2.1. Data Description

The data consists of ETCCDI indices and selected heat stress indicators calculated for CMIP6. The ETCCDI indices have been defined by the Expert Team on Climate Change Detection and Indices (ETCCDI) to be a representative and descriptive set of indicators of temperature and precipitation extremes. The heat stress indicators combine near-surface air temperature, near-surface specific humidity and surface air pressure to give indications of adverse effects of heat on human health. As we do not include other variables like wind or solar radiation, the selected heat stress indicators represent indoor conditions or calm conditions in the shade.

Figure 1: A selection of CMIP6 based ETCCDI indices available from this dataset. Projected multi-model median changes for the end of century (2071-2100) relative to the recent past (1981-2010) are shown for four scenarios for five indices: TXx (annual maximum of daily maximum temperature), WSDI (warm spell duration index), FD (frost days- annual count of days when daily minimum temperature is less than 0°C), R95p (total annual precipitation from very wet days – i.e. days above the 95th percentile of daily precipitation from wet days) and SDII (simple daily intensity index of precipitation).

Table 1: Overview of key characteristics of the dataset.

2.2.2. Variable Description

The variables contained in the dataset are ETCCDI indices on monthly and yearly resolution, and daily resolved heat stress indicators HI, Humidex, UTCI, WBT and WBGT.

Table 2: Overview and description of variables.

2.3. Input Data

The main input data for this dataset is CMIP6 data (Eyring et al 2016) from experiments historical, ssp126, ssp245, ssp370 and ssp585. In addition, the WFDE5 v1.0 (Cucchi et al. 2020) dataset is used for bias adjustment of heat stress indicators. Mean, maximum and minimum daily temperature, and total daily precipitation is used for calculation of ETCCDI indices, whereas maximum daily temperature, daily mean surface pressure and daily mean surface specific humidity is used to calculate the heat stress indicators.

Most of the CMIP6 models used in this catalogue entry are now available in the CDS catalogue but indices offered in this dataset are calculated for more ensemble members than currently offered for some models, and therefore the CMIP6 data have had to be taken from the Earth System Grid Federation (https://esgf.llnl.gov).

Table 3: Overview of climate model CMIP6 data, summarising the model properties and scenario simulations used.

2.3.1. Input Data 1 – CMIP6 model output

CMIP6 (Eyring et al 2016) model data are available from the Earth System Grid Federation ESGF (https://esgf.llnl.gov/). Variables are daily resolution total precipitation (pr), daily maximum temperature (tasmax), daily average temperature (tas), daily minimum temperature (tasmin), daily mean pressure at sea level (psl) and daily mean near-surface specific humidity (huss). All data stem from the DECK experiment historical, and the ScenarioMIP experiments ssp126, ssp245, ssp370 and ssp585 (O'Neill et al 2016; https://www.wcrp-climate.org/modelling-wgcm-mip-catalogue/cmip6-endorsed-mips-article/1070-modelling-cmip6-scenariomip).

CMIP6 is a widely used comprehensive model intercomparison project. The historical experiment is the standard protocol for simulating the historical climate, and the four other experiments used are the most standard future pathway scenarios and span a wide range of possible futures.

The CMIP6 dataset is nicely streamlined, producing highly consistent outputs for the various models. However, some model inconsistencies remain. For instance, different models employ different calendar types, with years of slightly different length ranging from 360 days a year, to years with no leap years, to a calendar with leap years. In addition, some models have delivered data for a few more years than the prescribed standard. CanESM5 ran the future projection experiment SSP1-2.6 up to 2300. EC-Earth3 delivered data for the year 1849 for some members in the historical simulation in addition to 1850- 2014. FGOALS-G3 let their historical simulation run up to 2016, creating a two-year overlap between historical data and future projections. We have chosen to keep ETCCDI indices also for these extra years in the data set to keep consistency with the original CMIP6 data.

The models BCC-CSM2-MR, MIROC6, and NESM3 are not considered in the calculation of heat stress indicators as they do not provide all necessary input variables for calculating heat stress indicators. Temperature-related ETCCDI indices for the model NorESM2-LM are only provided from 1951 onwards, as minimum and maximum temperatures are not available for NorESM2-LM for the time before 1950. Precipitation-related ETCCDI indices for NorESM2-LM are provided for the full time period.

2.3.2. Input Data 2 - WFDE5 (bias-adjusted ERA5 reanalysis data)

The heat stress indicators have been bias corrected using the WFDE5 v1.0 reanalysis. WFDE5 is a bias-adjusted version of the ERA5 reanalysis, using the WATCH Forcing Data method and is available at 0.5 degrees spatial resolution from 1979-2018 (Cucchi et al. 2020). The WFDE5 dataset (https://doi.org/10.24381/cds.20d54e34) is available from the Copernicus Data Store.

The heat stress indicators were calculated from WFDE5 using the daily maximum of near-surface air temperature, and the daily mean of near-surface specific humidity and surface air pressure. Bias correction of the CMIP6 heat stress indicators was then performed using the WFDE5 derived heat stress indicators as the reference with the ISIMIP3b method (Lange 2019; 2020). Heat stress indicators calculated with WFDE5 were regridded by conservative remapping to the respective grid of each CMIP6 model.

It should be noted that using ERA5 as the reference dataset instead of WFDE5 gave lower values for the heat stress indicators (see Figure 2). This was due to higher daily maximum temperatures and higher humidity in WFDE5 relative to ERA5 (see Figure 3).

Figure 2: The effect of reference dataset choice for bias correcting the heat stress indicators. Time series of the annual maximum of heat index (HI). Raw values for ACCESS-CM2 are shown in black, bias-adjusted values using WFDE5 and ISIMIP3b are shown in solid red, and dashed red when ERA5 with ISIMIP3b is used. For comparison, a quantile delta mapping (QDM) technique using ERA5 is also shown in dashed purple. Region acronyms refer to SREX regions and are SAS - South Asia, EAS - East Asia, SEA - Southeast Asia, WAF - West Africa, EAF - East Africa, MED - South Europe/Mediterranean, CEU - Central Europe, CAM - Central America Mexico, ENA - East North America.

Figure 3: Differences between WFDE5 and ERA5 climatological means (1981-2010) for daily maximum temperature (top), specific humidity (middle) and surface pressure (bottom).

2.4. Method

2.4.1. Background

The Expert Team on Climate Change Detection and Indices (ETCCDI) has defined a set of climate indices that provide a comprehensive overview of temperature and precipitation statistics focusing particularly on extreme aspects (Zhang et al. 2011). The indices are used in several applications in climate research due to their robustness and straightforward calculation and interpretation.

Calculating the ETCCDI indices consistently with the appropriate statistical methodology for percentile-based indices is important (Zhang et al. 2005), and inconsistencies due to the use of improper algorithms employed for their calculation can be avoided if the setup used here is applied and referred to. The ETCCDI indices here are calculated using the climdex.pcic.R package, which was developed, evaluated, and approved by the ETCCDI (Sillmann et al. 2013a).

Various heat stress indicators have been developed to quantify the impact of heat stress on humans (de Freitas and Grigorieva, 2017). A subset of these indicators is applied in climate science, restricted to those heat stress indicators that can be calculated from climate model output and, additionally, restricted by computation time constraints. Here, we include heat stress indicators based on the selection of Schwingshackl et al. (2021). We include heat stress indicators that are used by meteorological organizations (HI, Humidex) and that are based on human heat balance models or theoretical considerations on the effect of temperature and humidity on the human body (UTCI, WBT, WBGT). Since heat stress indicators are often used to estimate the exceedance of absolute thresholds, we provide a bias-adjusted version of the heat stress indicators in addition to the version calculated from raw CMIP6 output.

The CMIP6 ScenarioMIP experiments are widely used to assess and understand future climate changes impacts. Having a comprehensive set of indicators calculated from these and the CMIP6 historical experiment is hence of great value in many applications. However, calculating the ETCCDI and heat stress indicators consistently across these large data sets is very time consuming, both in terms of computing time and person hours. Hence, having a publicly available shared dataset including these is very valuable. This is the purpose of this dataset.

2.4.2. Model / Algorithm

ETCCDI indicators are calculated using the R package provided by the Pacific Climate Impacts consortium. The code follows the official definitions identified by the ETCCDI and is publicly available from github at https://github.com/pacificclimate/climdex.pcic

The indices csdi, wsdi, r95p, r99p, tn10p, tn90p, tx10p and tx90p are calculated based on comparison with a baseline climatological period. We have chosen to calculate these using two separate base periods, 1961-1990 and 1981-2010. The first period is the base period used in the previous AR5 climate assessment from IPCC and is hence useful for comparison with older datasets, or previous literature. The other period is the updated climatological base period and is hence more useful if no such comparison is intended. Some reanalysis datasets of use for comparison also do not go all the way back to 1961, hence the newer base period may be more useful in such cases.

The heat stress indicators are calculated for CMIP6 output using daily mean near-surface specific humidity, daily mean surface pressure, and daily maximum near-surface air temperature. Daily mean surface pressure is calculated from daily mean sea level pressure by applying a height adjustment (see Schwingshackl et al., 2021 for details). Additionally, the heat stress indicators are calculated for WFDE5, for which daily maximum temperature is calculated as maximum, daily mean pressure and humidity as average of the hourly WFDE5 variables near-surface air temperature, near-surface specific humidity, and surface air pressure. The equations used to calculate heat stress indicators can be found under https://github.com/schwings-clemens/CDS_heat_stress_indicators .

In particular, HI is calculated based on the equation indicated on the NOAA official web page https://www.wpc.ncep.noaa.gov/html/heatindex_equation.shtml, Humidex based on the equation as reported by Blazejczyk et al. (2012), and UTCI according to the multiple linear regression developed by Bröde et al. (2013). WBT is calculated based on Buzan et al. (2015), which combines equations from Davies-Jones (2008, 2009) to iteratively calculate WBT from near-surface air temperature, relative humidity, and pressure. We used a Matlab implementation of the WBT formula (https://github.com/jrbuzan/WetBulb.m/blob/master/WetBulb.m) and translated it into python. WBGT is calculated as weighted mean of the WBT estimates and daily maximum near-surface air temperature (WBGT = 0.7*WBT + 0.3*T). For a more general review on heat stress indicators see de Freitas and Grigorieva (2017). Descriptions and review of the heat stress indicators used here is also included in Schwingshackl et al. (2021).

Bias adjustment was carried out using the ISIMIP3b method (Lange 2019; 2020), with WFDE5 v1.0 as the reference dataset (Cucchi et al. 2020). This is a trend-preserving quantile mapping method. We used it in its non-parametric version (i.e. no particular distribution was assumed). It was applied to the heat stress indicators themselves (rather than the input variables used to calculate them). The ISIMIP3b bias adjustment method adjusts the various percentiles of the model data distribution to match the reference dataset distribution. This is done for each grid cell and each month separately using 1981-2010 as the reference period. The application periods are 1951-1980, 1981-2010, 2011-2040, 2041- 2070, and 2071-2100. Before applying quantile mapping, the ISIMIP3b bias adjustment method detrends the data for each period by removing the linear trend in the variable to which it is applied, to avoid biased estimates of variability in the quantiles from the warming trend. This trend is then added back onto the bias-adjusted data at the end. To bias-adjust future periods, pseudo-future observations are generated by adding the change in each percentile between the model simulated reference and future period to the observational percentiles. The quantile mapping is then applied to the future simulations using the pseudo-future observations. The bias adjustment is performed on the original grid of each model to avoid effectively downscaling the models in the process, as would occur if using the high-resolution WFDE5 grid. In practice, this means that the WFDE5 reference heat stress indicator data were interpolated by conservative remapping to the grid of each CMIP6 model before bias adjustment. Python code for implementing the ISIMIP3b bias adjustment is available from the Zenodo repository https://zenodo.org/record/3898426, version 2.4.1. The code takes proleptic Gregorian calendars only. For 365 day calendars we inserted extra days containing missing values for leap days, and for 360 day calendars we additionally inserted 5 extra days of missing values evenly spaced throughout the year (i.e. every 72 days). These extra days were removed after bias adjustment.

2.4.3. Validation

ETCCDI indices have previously been presented, studied and validated for the CMIP5 dataset in Sillmann et al. (2013a) and Sillmann et al. (2013b). Parts of the ETCCDI dataset presented here has already been documented and studied and compared to observational and reanalysis datasets in (Kim et al. 2020, Thorarinsdottir et al. 2020).

Heat index (HI) is operationally used by the US weather service (https://www.weather.gov/safety/heat-index) and Humidex by the Canadian weather service (https://www.canada.ca/en/environment-climate-change/services/seasonal-weather-hazards/warm-season-weather-hazards.html#toc7, accessed on 04.05.2022). For UTCI, the Fiala model was selected after extensive validation (Psikuta et al., 2012). WBT was applied in several studies to estimate adaptation limits to future heat stress (e.g., Pal and Eltahir, 2016; Im et al., 2017), following a study that pointed to potential adaptability limits to climate change due to heat stress based on WBT (Sherwood and Huber, 2010). Wet-bulb globe temperature is used as a heat stress indicator in industry and sport events (Garzon-Villalba et al., 2016; Kjellstrom et al., 2013; Orlov et al., 2019). Several epidemiological studies were carried out considering various heat stress indicators, but these studies do generally not find a single indicator being superior to the others (see references in Schwingshackl et al. 2021). Moreover, epidemiological and validation studies sometimes use different equations for calculating an indicator (especially for WBGT), limiting the comparability across studies. We thus encourage the realization of further validation studies based on the heat stress indicators published in this dataset.

Figure 2 shows that the ISIMIP3b method performs similarly to the quantile delta mapping (QDM) documented in Cannon et al. (2015).

3. Acknowledgments

Dataset authors are Carley Iles, Marit Sandstad, Clemens Schwingshackl, and Jana Sillmann.

The calculation of the indicators in this dataset has received funding from the European Union's Horizon 2020 research and innovation program under grant agreement No 820655 (EXHAUSTION), from the Belmont Forum Collaborative Research Action on Climate, Environment, and Health, supported by the Norwegian Research Council (contract No 310672, HEATCOST), the ERA4C project ClimInvest (grant agreement No 274250), the ERA-NET project SUSCAP (grant agreement No 299600), and the Norwegian Research Council project ClimateXL (grant agreement No 244551). The compounding of the indicators was done on behalf of C3S Copernicus data store under service contract 2020/C3S_COP_69_2020.

We acknowledge the use of the WATCH Forcing Data methodology applied to ERA5 (WFDE5) dataset, provided by the Copernicus Data Store. Neither the European Commission nor ECMWF is responsible for any use that may be made of the Copernicus Information or Data it contains. We acknowledge the World Climate Research Program, which, through its Working Group on Coupled Modelling, coordinated and promoted CMIP6. We thank the climate modeling groups for producing and making available their model output, the Earth System Grid Federation (ESGF) for archiving the data and providing access, and the multiple funding agencies who support CMIP6 and ESGF.

The ETCCDI indices contained in this dataset has previously been presented or partially presented in Kim, Y.-H., S. Min, X. Zhang, J. Sillmann, and M. Sandstad, 2020: Evaluation of the CMIP6 multi-model ensemble for climate extreme indices, Weather and Climate Extremes, 29, doi: 10.1016/j.wace.2020.100269.
Thorarinsdottir, T. L., Sillmann, J., Haugen, M., Gisslib, N, Sandstad, M. (2020). Evaluation of CMIP5 and CMIP6 simulations of historical surface air temperature extremes using proper evaluation methods. Environ. Res. Lett. 15 124041

The heat stress indicators in this dataset are partially presented in:
Schwingshackl, C., Sillmann, J., Vicedo‐Cabrera, A. M., Sandstad, M., & Aunan, K. (2021). Heat Stress Indicators in CMIP6: Estimating Future Trends and Exceedances of Impact‐Relevant Thresholds. Earths Future, 9, e2020EF001885. https://doi.org/10.1029/2020EF001885 "

4. Concluding Remarks

The dataset consists of the 27 ETCCDI indices identified by the Expert Team on Climate Change Detection and Indices (ETCCDI; Zhang et al. 2011). These indices are used in several applications in climate research due to their robustness and straightforward calculation and interpretation. The indices are on yearly or monthly resolution and describe various aspects of temperature and precipitation extremes. The particular calculation methodology used to derive this dataset is contained in the climdex.pcic R package described in Sillmann et al. (2013a).

In addition, the dataset contains a selection of five daily resolved heat stress indicators; Heat index – (HI), Humidex, indoor Universal Thermal Climate Index (UTCI), wet-bulb temperature (WBT) and indoor wet-bulb globe temperature (WBGT). The various heat stress indicators are based on human heat balance models or theoretical considerations on the effect of temperature and humidity on the human body. HI and Humidex are used by meteorological organizations. Background on these indices and their calculation for this dataset is given in Schwingshackl et al. (2021).

All the indicators in this dataset have been calculated for a wide array of climate model data included in the CMIP6 climate dataset (Eyring et al. 2016), both for historical runs and various future projection scenarios ssp126, ssp245, ssp370 and ssp585 (O'Neill et al. 2016). The heat stress indicators are provided both with and without bias correction using the ISIMIP3b method (Lange 2019; 2020), with WFDE5 v1.0 as the reference dataset (Cucchi et al. 2020).

There are always limitations to the reliability of climate model results and their interpretation, particularly in future scenarios. The historical results can to a certain degree be used to assess the accuracy of the data for the various models by comparing to observations, although it is not entirely clear that this will fully validate the accuracy in a changing future climate. Limitations of the interpretation of the indicators themselves have been discussed in the various works cited throughout this document for the ETCCDIs (i.e. Zhang 2011, Sillmann et al. 2013a, Sillmann et al. 2013b, Kim et al. 2020, Thorarinsdottir et al. 2020) and for the heat stress indicators (Rothfusz, 1990; Steadman, 1979, Masterson & Richardson, 1979; as cited in Blazejczyk et al., 2012, de Freitas and Grigorieva, 2017, Fiala et al. 2012, Bröde et al. 2013, Davies-Jones 2008, Schwingshackl et al. 2021).

Appendix I

Variable Description
All ETCCDI variables were calculated using the climdex.pcic package https://github.com/pacificclimate/climdex.pcic .
Heat stress indicators were calculated according to https://github.com/schwings-clemens/CDS_heat_stress_indicators.
The design ranges indicated for the five heat stress indicators are obtained from de Freitas and Grigorieva (2017), except for Humidex where the range is defined by the Canadian weather service (https://www.canada.ca/en/environment-climate-change/services/seasonal-weather-hazards/warm-season-weather-hazards.html#toc7, accessed on 04.05.2022).

Consecutive dry days - cddETCCDI

Maximum number of days in a row with precipitation below 1 mm in a year. If a dry spell spans across multiple calendar years (as may happen in very dry regions), then CDD is not reported for that year and the accumulated dry days are carried forward to the year when the spell ends. Unit: day.

Cold spell duration index - csdiETCCDI

Annual count of days with at least 6 consecutive days when daily minimum temperature is below the calendar day 10th percentile of minimum temperature centered on a 5-day sliding window during the base period. Unit: day.

Consecutive wet days - cwdETCCDI

Maximum number of days in a row with 1 mm or more precipitation in a year. If a wet spell spans across multiple calendar years (as may happen in very wet regions), then CWD is not reported for that year and the accumulated wet days are carried forward to the year when the spell ends. Unit: day.

Diurnal temperature range - dtrETCCDI

Annual mean daily difference between minimum and maximum temperature. Unit: °C.

Frost days - fdETCCDI

Annual count of days when daily minimum temperature < 0°C. Unit: day.

Growing season length - gslETCCDI

Annual (January 1st to December 31st in the Northern hemisphere and July 1st to June 30th in the Southern hemisphere) count of days between first period of at least 6 days with daily mean temperature above 5 °C and first span after July 1st (NH) or January 1st (SH) of at least 6 days with daily mean temperature below 5 °C. Unit: day.

Ice days - idETCCDI

Annual count of days when daily maximum temperature < 0°C. Unit: day.

Total wet day precipitation - prcptotETCCDI

Total annually summed precipitation on days with precipitation ≥ 1mm. Precipitation is deposition of water on the Earth's surface, either rain, snow, ice or hail. Unit: mm.

Number of wet days – r1mmETCCDI

Number of days per year with 1 mm or more precipitation. Precipitation is deposition of water on the Earth's surface, either rain, snow, ice or hail. Unit: day.

Heavy precipitation days – r10mmETCCDI

Number of days per year with 10 mm or more precipitation. Precipitation is deposition of water on the Earth's surface, either rain, snow, ice or hail. Unit: day.

Very heavy precipitation days – r20mmETCCDI

Number of days per year with 20 mm or more precipitation. Precipitation is deposition of water on the Earth's surface, either rain, snow, ice or hail. Unit: day.

Very wet day precipitation – r95pETCCDI

Total annually summed precipitation on days with more daily precipitation than the 95th daily precipitation percentile on wet days in the base period. Precipitation is deposition of water on the Earth's surface, either rain, snow, ice or hail. Unit: mm.

Extremely wet day precipitation – r99pETCCDI

Total annually summed precipitation on days with more daily precipitation than the 99th daily precipitation percentile on wet days in the base period. Precipitation is deposition of water on the Earth's surface, either rain, snow, ice or hail. Unit: mm.

Maximum 1-day precipitation – rx1dayETCCDI

Maximum precipitation on a single day in period (year or month). Precipitation is deposition of water on the Earth's surface, either rain, snow, ice or hail. Unit: mm.

Maximum 5-day precipitation – rx5dayETCCDI

Maximum precipitation in five consecutive days in period (year or month). Precipitation is deposition of water on the Earth's surface, either rain, snow, ice or hail. Unit: mm.

Simple daily intensity index – sdiiETCCDI

Total annually summed precipitation on wet days (days with ≥ 1mm), divided by total number of wet days. Precipitation is deposition of water on the Earth's surface, either rain, snow, ice or hail. Unit: mm/day.

Summer days - sdETCCDI

Annual count of days when daily maximum temperature > 25 °C. Unit: day.

Cold nights – tn10pETCCDI

Percentage of days with minimum temperature below the corresponding calendar day 10th percentile of minimum temperature for a 5-day moving window in the reference period. The percentile is calculated using a bootstrap technique to increase robustness of results. Unit: %.

Warm nights – tn90pETCCDI

Percentage of days with minimum temperature above the corresponding calendar day 90th percentile of minimum temperature for a 5-day moving window in the reference period. The percentile is calculated using a bootstrap technique to increase robustness of results. Unit: %.

Minimum value of daily minimum temperature - tnnETCCDI

Minimum of daily minimum temperature in period (year or month). Unit: °C.

Maximum value of daily minimum temperature - tnxETCCDI

Maximum of daily minimum temperature in period (year or month). Unit: °C.

Tropical nights – trETCCDI

Annual count of days when daily minimum of temperature > 20 °C. Unit: day.

Cold days – tx10pETCCDI

Percentage of days with maximum temperature below the corresponding calendar day 10th percentile of maximum temperature for a 5-day moving window in the reference period. The percentile is calculated using a bootstrap technique to increase robustness of results. Unit: %.

Warm days – tx90pETCCDI

Percentage of days with maximum temperature above the corresponding calendar day 90th percentile of maximum temperature for a 5-day moving window in the reference period. The percentile is calculated using a bootstrap technique to increase robustness of results. Unit: %.

Minimum value of daily maximum temperature – txnETCCDI

Minimum of daily maximum temperature in period (year or month). Unit: °C.

Maximum value of daily maximum temperature - txxETCCDI

Maximum of daily maximum temperature in period (year or month). Unit: °C.

Warm spell duration index – wsdiETCCDI

Annual count of days with at least 6 consecutive days when daily maximum temperature is above the calendar day 90th percentile of maximum temperature centered on a 5-day sliding window during the base period. Unit: day.

Heat index – HI

Heat index is a heat stress indicator used by the US National Oceanic and Atmospheric Administration (NOAA) National Weather Service for issuing heat warnings. It is calculated using multiple linear regression based on near-surface air temperature and relative humidity (https://www.wpc.ncep.noaa.gov/html/heatindex_equation.shtml). Unit: °C. Design range: 20 °C to 60 °C.

Humidex – Humidex

Humidex is a heat stress indicator used by Canadian meteorological services. It is calculated as linear combination of near-surface air temperature and vapour pressure (Masterson & Richardson, 1979; as cited in Blazejczyk et al., 2012). Unit: °C. Design range: >20 °C.

Indoor Universal Thermal Climate Index - UTCI

UTCI is a multinode model of human heat transfer (Fiala et al., 2012), which can be approximated by a polynomial equation based on near-surface air temperature, solar radiation, vapour pressure, and wind speed (Bröde et al., 2013). In the present dataset, the influence of solar radiation and wind speed is not considered, and UTCI is calculated from temperature and vapour pressure solely, thus representing indoor conditions. Unit: °C. Design range: -90 °C to 60 °C.

Wet-bulb temperature – WBT

Wet-bulb temperature considers the cooling capacity of a human body through sweating by indicating the temperature an air parcel would have in case of complete water evaporation. We employ the WBT formula derived by Buzan et al. (2015), which combines equations from Davies-Jones (2008, 2009) to iteratively calculate WBT from near-surface air temperature, relative humidity, and pressure. We used a Matlab implementation of the WBT formula (https://github.com/jrbuzan/WetBulb.m/blob/master/WetBulb.m) and translated it into python. Unit: °C. Design range: 10 °C to 50 °C.

Indoor Wet-bulb globe temperature – WBGT

Indoor wet-bulb globe temperature is defined as weighted mean of WBT and near-surface air temperature (WBGT = 0.7*WBT + 0.3*T). Unit: °C. Design range: 10 °C to 50 °C.

Appendix II

Input Data Description

CMIP6 data

The Coupled Model Intercomparison Project Phase 6 (Eyring et al. 2016) is the largest and most up to date climate model intercomparison project, providing invaluable data used throughout the climate research community.

Table 4: Overview of key characteristics of CMIP6 data.

WDEF5 data

WDEF5 (Cucchi et al. 2020) is a bias-adjusted version of meteorological surface data from the widely used reanalysis dataset ERA5.

Table 5. Overview of key characteristics of the WFDE5 data.

References

Blazejczyk, K., Epstein, Y., Jendritzky, G., Staiger, H., & Tinz, B. (2012). Comparison of UTCI to selected thermal indices. International Journal of Biometeorology, 56(3), 515–535. https://doi.org/10.1007/s00484-011-0453-2

Bröde P, Błazejczyk K, Fiala D, Havenith G, Holmér I, Jendritzky G, Kuklane K, Kampmann B. (2012) The Universal Thermal Climate Index UTCI compared to ergonomics standards for assessing the thermal environment. Ind Health. 2013;51(1):16-24. https://doi.org/10.2486/indhealth.2012-0098.

Buzan, J. R., Oleson, K., and Huber, M.: Implementation and comparison of a suite of heat stress metrics within the Community Land Model version 4.5, Geosci. Model Dev., 8, 151–170, https://doi.org/10.5194/gmd-8-151-2015, 2015.

Cannon, A. J., Sobie, S. R., and Murdock, T. Q. (2015). Bias Correction of GCM Precipitation by Quantile Mapping: How Well Do Methods Preserve Changes in Quantiles and Extremes? J. Clim. 28, 6938– 6959. doi:10.1175/JCLI-D-14-00754.1.

Cucchi, M., Weedon, G. P., Amici, A., Bellouin, N., Lange, S., Müller Schmied, H., Hersbach, H., and Buontempo, C. (2020). WFDE5: bias-adjusted ERA5 reanalysis data for impact studies, Earth Syst. Sci. Data, 12, 2097–2120, https://doi.org/10.5194/essd-12-2097-2020

Davies-Jones, R. (2008). An efficient and accurate method for computing the wet‐bulb temperature along pseudoadiabats. Monthly Weather Review, 136(7), 2764 – 2785. https://doi.org/10.1175/2007MWR2224.1

Davies-Jones, R. (2009). On Formulas for Equivalent Potential Temperature, Monthly Weather Review, 137(9), 3137-3148. https://doi.org/10.1175/2009MWR2774.1

de Freitas, C. R., & Grigorieva, E. A. (2017). A comparison and appraisal of a comprehensive range of human thermal climate indices. International Journal of Biometeorology, 61(3), 487– 512. https://doi.org/10.1007/s00484‐016‐1228‐6

Dunne, J. P., Stouffer, R. J., & John, J. G. (2013). Reductions in labor capacity from heat stress under climate warming. Nature Climate Change, 3(6), 563. https://doi.org/10.1038/nclimate1827

Eyring, V., Bony, S., Meehl, G. A., Senior, C. A., Stevens, B., Stouffer, R. J., and Taylor, K. E. (2016) Overview of the Coupled Model Intercomparison Project Phase 6 (CMIP6) experimental design and organization, Geosci. Model Dev., 9, 1937-1958, https://doi.org/10.5194/gmd-9-1937-2016

Fiala, D., Havenith, G., Bröde, P. et al. (2012) UTCI-Fiala multi-node model of human heat transfer and temperature regulation. Int J Biometeorol 56, 429–441. https://doi.org/10.1007/s00484-011-0424-7

Garzon-Villalba, X. P., Mbah, A., Wu, Y., Hiles, M., Moore, H., Schwartz, S. W., et al. (2016). Exertional heat illness and acute injury related to ambient wet bulb globe temperature. American Journal of Industrial Medicine, 59(12), 1169–1176. https://doi.org/10.1002/ajim.22650

Im, E.-S., Pal, J. S., & Eltahir, E. A. B. (2017). Deadly heat waves projected in the densely populated agricultural regions of South Asia. Science Advances, 3(8). https://doi.org/10.1126/sciadv.1603322

Kim, Y.-H., S. Min, X. Zhang, J. Sillmann, and M. Sandstad, (2020) Evaluation of the CMIP6 multi-model ensemble for climate extreme indices, Weather and Climate Extremes, 29, doi: 10.1016/j.wace.2020.100269

Kjellstrom, T., Lemke, B., & Otto, M. (2013). Mapping occupational heat exposure and effects in South-East Asia: Ongoing time trends 1980–2011 and future estimates to 2050. Industrial Health, 51(1), 56–67. https://doi.org/10.2486/indhealth.2012-0174

Lange, S. (2019) Trend-preserving bias adjustment and statistical downscaling with ISIMIP3BASD (v1.0), Geosci. Model Dev., 12, 3055–3070, https://doi.org/10.5194/gmd-12-3055-2019

Lange, S. (2020) ISIMIP3b bias adjustment fact sheet, available from https:// www.isimip.org/gettingstarted/isimip3b-bias-correction/, accessed 06/05/2021

Masterson, J., & Richardson, F. (1979). Humidex, a method of quantifying human discomfort due to excessive heat and humidity, Environment Canada (Vol. 151, pp. 1–79). Downsview, Ontario: Atmospheric Environment Service.

Napoli, C. D., Pappenberger, F., & Cloke, H. L. (2018). Assessing heat-related health risk in Europe via the Universal Thermal Climate Index (UTCI). International Journal of Biometeorology, 62(7), 1155– 1165. https://doi.org/10.1007/s00484-018-1518-2

O'Neill, B. C., Tebaldi, C., van Vuuren, D. P., Eyring, V., Friedlingstein, P., Hurtt, G., Knutti, R., Kriegler, E., Lamarque, J.-F., Lowe, J., Meehl, G. A., Moss, R., Riahi, K., and Sanderson, B. M. (2016) The Scenario Model Intercomparison Project (ScenarioMIP) for CMIP6, Geosci. Model Dev., 9, 3461-3482, doi:10.5194/gmd-9-3461-2016

Orlov, A., Sillmann, J., Aaheim, A., Aunan, K., & de Bruin, K. (2019). Economic losses of heat-induced reductions in outdoor worker productivity: A case study of Europe. Economics of Disasters and Climate Change, 3(3), 191–211. https://doi.org/10.1007/s41885-019-00044-0

Pal, J. S., & Eltahir, E. A. (2016). Future temperature in southwest Asia projected to exceed a threshold for human adaptability. Nature Climate Change, 6(2), 197–200. https://doi.org/10.1038/nclimate2833

Psikuta, A., Fiala, D., Laschewski, G. et al. Validation of the Fiala multi-node thermophysiological model for UTCI application. Int J Biometeorol 56, 443–460 (2012). https://doi.org/10.1007/s00484-011-0450-5

Rothfusz, L. P. (1990). The heat index equation. Fort Worth, TX: National Weather Service Technical Attachment (SR 90–23).

Schwingshackl, C., Sillmann, J., Vicedo-Cabrera, A. M., Sandstad, M., & Aunan, K. (2021). Heat Stress Indicators in CMIP6: Estimating Future Trends and Exceedances of Impact-Relevant Thresholds. Earths Future, 9, e2020EF001885. https://doi.org/10.1029/2020EF001885

Sherwood, S. C., & Huber, M. (2010). An adaptability limit to climate change due to heat stress. Proceedings of the National Academy of Sciences, 107(21), 9552–9555. https://doi.org/10.1073/pnas.0913352107

Sillmann, J., V. V. Kharin, X. Zhang, F. W. Zwiers and D. Bronaugh, (2013a) Climate extremes indices in the CMIP5 multi-model ensemble. Part 1: Model evaluation in the present climate. J. Geophys. Res. Atmos., 118, 1716-1733, doi: 10.1002/jgrd.50203.

Sillmann, J., V. V. Kharin, F. W. Zwiers, X. Zhang and D. Bronaugh, (2013b) Climate extremes indices in the CMIP5 multi-model ensemble. Part 2: Future climate projections. J. Geophys. Res. Atmos., 118, 2473-2493, doi: 10.1002/jgrd.50188

Steadman, R. G. (1979). The assessment of sultriness. Part I: A temperature‐humidity index based on human physiology and clothing science. Journal of Applied Meteorology, 18(7), 861–873. https://doi.org/10.1175/1520‐0450(1979)018⟨0861:TAOSPI⟩2.0.CO;2

Thorarinsdottir, T. L., Sillmann, J., Haugen, M., Gisslib, N, Sandstad, M. (2020). Evaluation of CMIP5 and CMIP6 simulations of historical surface air temperature extremes using proper evaluation methods. Environ. Res. Lett. 15 124041

Urban, A., Hondula, D. M., Hanzlíková, H., & Kyselý, J. (2019). The predictability of heat-related mortality in Prague, Czech Republic, during summer 2015—A comparison of selected thermal indices. International Journal of Biometeorology, 63(4), 535– 548. https://doi.org/10.1007/s00484-019-01684-3

Zhang, X., G. Hegerl, F. Zwiers, and J. Kenyon (2005), Avoiding inhomogeneity in percentile-based indices of temperature extremes, J. Climate, 1641–1651, doi:10.1175/JCLI3366.1.

Zhang, X., L. Alexander, G. C. Hegerl, P. Jones, A. K. Tank, T. C. Peterson, B. Trewin, and F. W. Zwiers (2011), Indices for monitoring changes in extremes based on daily temperature and precipitation data, WIREs Clim. Chang., 2, 851–870, doi:10.1002/wcc.147.

Related articles

• Page:
  Data Stream 2: AgERA5 historic and near real time forcing data: Product User Guide and Specification (PUGS)
• Page:
  Windstorm footprints: Product User Guide
• Page:
  NEMO-ERSEM and POLCOMS-ERSEM: User Guides for Products (v1.1)
• Page:
  Sea level and ocean wave time series and indicators
• Page:
  Product user guide for sea level and ocean wave products - time series and indicators