Contributors: Christiana Photiadou (SMHI), Peter Berg (SMHI)
Acronyms
1. Scope of the document
This document presents the data set of climate impact indicators (CIIs) and essential climate variables (ECVs) from EURO-CORDEX regional climate model (RCM) projections. The data set is described in a concise manner with focus on: space and time extent and resolution, data formats, metadata, description of variables, quality of the data set and limitations.
2. Executive Summary
The specific data set provides CII for the period of 1971-2100 at grid resolutions of 5 km and 0.11 degrees (about 12.5 km). The indicators are calculated using a selected ensemble of EURO-CORDEX RCMs with both bias and non-bias adjusted variables. The indicators provided here are derived from the Operational water service; however, they are intended for users active not only in the water sector but also interdisciplinary sectors such as agriculture and energy.
The indicators are calculated as annual or seasonal means over the reference period, as changes over three future periods for three RCPs (2.6, 4.5, 8.5), and for three degree scenarios for global mean temperature increase of 1.5, 2.0 and 3.0 °C above pre-industrial conditions. The indicators cover mean temperature, precipitation, consecutive dry days, etc.
3. Product description
3.1. Introduction
Using Climate Impact Indicators (CIIs) is an efficient way to make climate information accessible to users within a sector, as the specific indicator concentrates the outcome of the climate models into a quantity that is sector relevant.
The CII should contain the condensed climate information needed, which can make the subsequent analysis relatively quick and efficient, in comparison to going through a full climate modelling chain.
The Operational Water Service consists of pan-European hydrological and climate indicators on catchment and grid resolution (0.11° and 5km) based on the EURO-CORDEX 0.11° (about 12.5 km) resolution. The produced indicators follow user requirements from two previous proof-of-concept contracts (SWICCA and EDgE).
This data set consists of a set of CIIs listed in Table 1. The CIIs are currently calculated from a sub- ensemble of EURO-CORDEX EUR11 including daily mean temperature and precipitation. The indicators are calculated at different grid resolutions and from bias adjusted and non-bias adjusted variables, and are mainly calculated as mean and/or seasonality; where "mean" indicators are calculated as the mean annual values over a 30-year period, while by "seasonality" the indicators are calculated as the mean monthly values of an indicator (averaged over each month over a 30- year period). Further documentation on the bias adjustment methodology can be found in the Algorithm Theoretical Basis Document (ATBD) document "Bias adjustment specification" in the Documentation tab.
The indicators are provided for different time ranges; absolute values are given for a reference period (e.g. 1971-2000) and the future changes for different 30-year time-slices that are defined as time periods or as degree scenarios. The time periods are: early century (2011-2040), mid-century (2041-2070) and end-century (2071-2100), while the degree scenarios are defined according to model specific periods when the global mean temperature has increased by 1.5, 2.0 and 3.0 °C
above pre-industrial conditions (see definition in section 3.4). CIIs for the future time-slices are presented as either relative (100*(future - historical)/historical) or absolute (future – historical) changes depending on the variable.
In particular, an ensemble of EURO-CORDEX (daily mean temperature and precipitation) were bias adjusted using EFAS-Meteo and a new bias adjustment method developed by SMHI. Further documentation on the bias adjustment method and reference data set can be found in the related document. The ensemble was then used as forcing to the multi-model E-HYPEcatch, E-HYPEgrid, and VIC-WUR hydrological models to calculate the water indicators. The water indicators can be found in data set entitled "Water Indicators for the European water sector using EURO-CORDEX EUR11".
3.2. Geophysical product description
Table 1: List of variables in this data set.
Short name | Long name | Unit | Aggregation | Definition or URL? |
2m air temperature | Mean air temperature | K (ECV) | daily | The ambient air temperature close to the surface. |
Precipitation | precipitation | kg m-2 s-1 (ECV) mm day-1 (CII) | daily | Precipitation is defined as the deposition of water to the Earth"s surface in the form of rain, snow, ice or hail. The essential climate variable (ECV) data is given as the mass of water per unit area and time. The data originate from EURO-CORDEX RCM simulations and are bias adjusted using the EFAS- Meteo reference dataset The climate impact indicator (CII) of precipitation is defined as the monthly/annual mean of the liquid water equivalent daily precipitation, averaged over a 30 year period. For future periods the indicator is given as a relative change against the reference period (1971-2000). |
Longest dry spells | Consecutive dry days | days for reference period | Mean | Longest dry spell is defined as the maximum number of consecutive dry days (dry day: daily precipitation < 1mm) over a 30 year period. For |
Number of dry spells | Number of dry spells | no. of spells for reference period | mean | Number of dry spells is defined as the number of dry periods (dry day: daily precipitation < 1mm) for more than 5 days for a 30 year period. For |
Highest five-day precipitation amount | Maximum 5-day precipitation | mm/5day | Max, mean | Highest five-day precipitation amount is defined as the maximum of 5-day precipitation totals. The value is given as a maximum over a 30 year period. For future periods the indicator is given as a relative change against the reference period |
3.3. Product target requirements
DATA DESCRIPTION | |
Horizontal coverage | Europe (EFAS-Meteo domain) |
| 5km x 5 km (ECV, daily; CII monthly and annual mean and future changes) |
| Spatial gaps over non-land grid points in bias adjusted data, and for regions outside the CORDEX simulation domain in the interpolated 5 km |
Vertical coverage | Single level |
Vertical resolution | Surface |
Temporal coverage | 01/1971-12/2100 |
Temporal resolution | Daily, 30 year annual and monthly means and changes |
Temporal gaps | The HadGEM based ECVs sometimes lack the last month of 2099. |
Update frequency | No further planned updates |
File format | NetCDF 4 |
Conventions | Climate and Forecast (CF) Metadata Convention v1.6, Attribute |
Available versions | 1 |
Projection | lambert_azimuthal_equal_area (5km); and rotated grid (0.11°) |
Data type | Grid |
3.4. Input data
The climate projections are taken from the EURO-CORDEX ensemble of regional climate models (downloaded from ESGF, but now potentially available in the CDS-entry: https://cds.climate.copernicus.eu/datasets/projections-cordex-domains-single-levels?tab=overview). The ensemble consists of three different global climate models (GCM), where one of them (MPI-ESM-LR) comes with two different realizations (marked by the RIP-Realisation- Initialization-Physics code), as presented in Table 2. Different realisations of the same model essentially mean that the GCM scenarios are starting from a different Earth system state such that the natural climate oscillations affect the model at short to multi-decadal time scales throughout the simulation. A set of four different RCMs have downscaled the GCM simulations to the EURO- CORDEX 0.11-degree grid (about 12.5 km), and here we use outputs of daily precipitation and daily mean temperature (2m height). When selecting models from the download form, it is recommended to include all GCMs and all RCMs to sample uncertainty related to model definitions. Further, the two realisations of MPI-ESM-LR with REMO2009 can be used to estimate the influence of natural variability on the projection results.
The ensemble consists of the complete set of EURO-CORDEX ensemble members (in May 2019) that include the time period 1971-2100, RCPs 2.6, 4.5, and 8.5, and includes the ECVs required for simulations with the hydrological models E-HYPE and VIC-WUR. VIC-WUR also uses solar radiation, thermal radiation, wind speed, surface pressure and humidity.
Table 2: List of EURO-CORDEX EUR-11 members used in this data set.
GCM | RCM | RCP | RIP |
EC-EARTH | CCLM4-8-17 | 2.6, 4.5, 8.5 | r12i1p1 |
EC-EARTH | RACMO22E | 2.6, 4.5, 8.5 | r12i1p1 |
EC-EARTH | RCA4 | 2.6, 4.5, 8.5 | r12i1p1 |
HadGEM2-ES | RCA4 | 2.6, 4.5, 8.5 | r1i1p1 |
HadGEM2-ES | RACMO22E | 2.6, 4.5, 8.5 | r1i1p1 |
MPI-ESM-LR | RCA4 | 2.6, 4.5, 8.5 | r1i1p1 |
MPI-ESM-LR | REMO2009 | 2.6, 4.5, 8.5 | r2i1p1 |
MPI-ESM-LR | REMO2009 | 2.6, 4.5, 8.5 | r1i1p1 |
The degree scenarios are defined as the 30-year time period when the driving GCM reaches a global mean temperature increase of 1.5, 2.0 or 3.0 degrees above pre-industrial conditions (1861–1890; Joshi et al., 2011), following the method of Nikulin et al., (2018). Table 3 presents the extent of each time period of the degree scenarios for each of the driving GCMs.
Table 3: Degree scenario time periods for RCP8.5 with the different driving GCMs.
Degree scenario: | 1.5 | 2.0 | 3.0 | ||||
GCM | RIP | First | Last | First | Last | First | Last |
ICHEC-EC-EARTH | r12i1p1 | 2005 | 2034 | 2021 | 2050 | 2047 | 2076 |
MOHC-HadGEM2-ES | r1i1p1 | 2010 | 2039 | 2023 | 2052 | 2042 | 2071 |
MPI-M-MPI-ESM-LR | r1i1p1 | 2004 | 2033 | 2021 | 2050 | 2046 | 2075 |
MPI-M-MPI-ESM-LR | r2i1p1 | 2002 | 2031 | 2018 | 2047 | 2044 | 2073 |
4. Workflow and Quality assurance
4.1. Workflow
The workflow followed in this data set is presented in Figure 1. It contains the production of an additional data set entitled Water indicators for the European water sector using EURO-CORDEX EUR11, which is also available in the CDS catalogue. In this document we focus on the steps for the production of the climate indicators.
Figure 1: Workflow overview
The data set workflow includes (see Figure 1):
- Step 1: Retrieving daily data from EURO-CORDEX EUR 11
- Step 2: CII calculation using EUR11
- Step 3: Bias adjustment
- Step 4: CII calculation using bias adjusted EUR11
4.1.1. Step 1: Retrieving data
Data were retrieved from ESGF nodes as an interim solution since the CDS catalogue did not include the EURO-CORDEX EUR11 ensemble. Data were retrieved May 2019. Daily mean temperature and daily precipitation were extracted for 8 members of the EURO-CORDEX EUR-11 ensemble (Table 2). QA consisted of checks and pre-processing routines for completeness of the data extraction; checking that the temporal and spatial scale and metadata were correct and complete.
4.1.2. Step 2: CII calculation using EUR11
The CII in Table 1 are calculated using raw output of daily mean temperature and precipitation from EUR11. The indicators are first calculated at the original resolution, 0.11°. Then the EUR11 variables are remapped at 5km resolution and the indicators are calculated once more. Both calculation sets are available in the CDS catalogue.
4.1.3. Step 3: Bias adjustment
The two variables were bias adjusted using a newly developed method based on time separation scaling (based on Haerter et al., 2011; Berg et al. 2012). The data were adjusted to the EFAS-Meteo reference dataset (Ntegeka et al. 2013). A complete documentation on the method, the reference period used and the QA evaluation are available in document C3S_424.SMHI_Biasadjustment_v1.
4.1.4. Step 4: CII calculation using bias adjusted EUR11
The CII (Table 1) are calculated using the bias adjusted EUR11 variables at 5km resolution. The indicators are usually calculated as mean and/or seasonality; where "mean" indicators are calculated as the mean annual values over a 30-year period, while by "seasonality" the indicators are calculated as the mean monthly values of an indicator averaged over each month over a 30-year period.
The indicators are calculated for different time ranges; absolute values are calculated for a reference period (e.g. 1971-2000), while future changes are calculated for three 30-year time-slices; early century (2011-2040), mid-century (2041-2070) and end-century (2071-2100). The indicators for the future periods are presented as either relative (100*(future - historical)/historical) or absolute (future – historical) changes depending on the variable. These CII are also available in CDS. The calculations are performed mainly with the climate data operators tool (CDO; Schulzweida, 2019). The functions used are: time average – timmean, seasonality – ymonmean or monsum & ymonmean, number of dryspells and longest dry spells– eca_cdd,1,5, highest five-day precipitation amount – timselsum,5 & eca_rx5day.
4.1.5. Quality Assurance
The QA procedure is ongoing for this data set as the production is continuing. Checks are conducted for each step in the production of the indicators, i.e. selection of the CIIs, selection of the Regional Climate Models (RCM), calculations and definitions, and additional checks on the ranges and outliers of the meteorological and hydrological indicators and metadata. We follow the guidelines developed in C3S_422_Lot1_SMHI, Quality Assurance Checklist (QUACK) and the quality assurance indicators are linked to the quality checks implemented in the production chain. The following checks were agreed between SMHI and WUR for producing Meteorological and Hydrological Indicators (Table 4).
Table 4: Checks performed on the extracted and bias adjusted ECVs and the calculated Meteorological and Hydrological Indicators, with the corresponding QUACK indicator.
Dimension | Criterion | Indicator | Short description of pre and post processing |
Input/output data (Meteo and Hydro ECV) | Scientific & methodological quality | Appropriateness | Check realizations for both historical and future periods |
Completeness | Check units | ||
Check calendar | |||
Check and count missing value (single and | |||
Check values below zero for precipitation. If negative values after interpolation then these are | |||
Reliability | Check value ranges for daily mean, min, max | ||
Completeness, | Check for variables the dimensions, shape, and | ||
Processing | Scientific & methodological quality | Validation, Appropriateness | The CDO commands were tested against other software (icclim python module, climdex R library) in previous projects and produced the same results. |
Input data (Meteo Indicators) | Scientific & methodological quality | Appropriateness | Check file ending (.nc) |
Completeness | Check file size EURO-CORDEX derived indicators | ||
Check units for the derived indicators | |||
Check and count missing value CIIs (set flag) | |||
Check time steps for derived indicators | |||
Check value ranges for each indicator | |||
Remove catchments outside the EFAS Meteo | |||
Output (Meteo and Hydro Indicators) | Scientific & methodological quality | Reliability | The spread of the whole ensemble is presented and no further sub-selection is planned to reduce |
Ensembles are created from all scenarios to find outliers with possible errors, i.e. ensemble members with CII values which deviate strongly from CII values of other ensemble members. | |||
CII values under climate scenario conditions are also compared to values of CIIs under reference conditions with EFAS-Meteo forcing, in order to qualitatively assess the range of projected change in CIIs. | |||
Absolute bias for temperature indicators and relative bias for precipitation indicators are estimated. | |||
Qualitative assessment of validity of E-HYPE and VIC-WUR output data is performed through diagnostic map plots. The validity of spatial patterns in hydrological variables is assessed through visual inspection of mapped aggregates (averages, sums) of output variables such as discharge and soil moisture. Values of HYPE variables at selected spatial points, e.g. large river outlets, are semi-quantitatively assessed through comparison with expected ranges based on external data, e.g. observations or previous HYPE model results. | |||
Input data (ECV and CII) | Scientific & methodological quality | Transparency, Appropriateness | Metadata are defined for each indicator considering the user friendliness and |
Completeness | Metadata follow the CF conventions and international standards set previous projects. Also follow the Common Data Model and respective | ||
Processing | Scientific & | Transparency | The scripts are reproducible as much as possible.
|
For the comparison of the CII with other data sources, we specifically compared with the IMPACT2c and ECLISE runs (Van Vliet et al. 2015, Roudier et al. 2016; Donelly et al. 2017). However, from our previous experience (Merks et al. 2020) the comparison with other open source climate services and data, has shown that differences are bound to arise from a number of parameters:
- Different model ensemble (and size) used between this contract and the other sources; As there is an uncertainty and variety between the GCM-RCM a different selection of models translates to varying results when computing CIIs. The larger the variability in the outcomes of the models the larger the ensemble size should be to minimise the risk that the selected ensemble has on the different mean and variance compared to the full ensemble.
- Different bias adjustment methods and reference datasets. It is extremely difficult to find openly available CII calculated with the same bias adjustment method and reference dataset.
- Different hydrological models (water balance based hydrological model (E-HYPE) vs a land surface model (VIC)): different ways of simulating the same processes, for example differences in modelled runoff and evapotranspiration due to different snow schemes give rise to the uncertainties present in assessments (Haddeland et.al. 2011). Van Vliet et. al., (2015) also showed in a comparison between VIC and HYPE over Europe that for most hydrological indicators the uncertainties originating from the climate models were larger compared to the uncertainties from the hydrological models. Only for evapotranspiration there were important differences between the two hydrological models.
References
Berg, P., Feldmann, H., & Panitz, H. J. (2012). Bias correction of high resolution regional climate model data. Journal of Hydrology, 448, 80-92.
Donnelly, C., W. Greuell, J. Andersson, D. Gerten, G. Pisacane, P. Roudier, and F. Ludwig. 2017. Impacts of climate change on European hydrology at 1.5, 2 and 3 degrees mean global warming above preindustrial level. Climatic Change 143:13-26.
Greuell W et al. (2015) Evaluation of five hydrological models across Europe and their suitability for making projections underof climate change. Hydrol Earth Syst Sci Discuss 12:10289–10330
Haddeland, I., et al., 2011. Multimodel Estimate of the Global Terrestrial Water Balance: Setup and First Results. Journal of Hydrometeorology, 12 (5), 869–884.
Haerter, J., Hagemann, S., Moseley, C., & Piani, C. (2011). Climate model bias correction and the role of timescales. Hydrology and Earth System Sciences, 15, 1065-1073.
Joshi, M., Hawkins, E., Sutton, R., Lowe, J., & Frame, D. (2011). Projections of when temperature change will exceed 2 C above pre-industrial levels. Nature Climate Change, 1(8), 407-412.
Merks, J., Photiadou, C., Ludwig, F., Arheimer, B. (2020) Comparison of open access global climate services for hydrological data. Hydrological Sciences Journal, in publication.
Nikulin, G., Lennard, C., Dosio, A., Kjellström, E., Chen, Y., Hänsler, A., ... & van Meijgaard, E. (2018). The effects of 1.5 and 2 degrees of global warming on Africa in the CORDEX ensemble. Environmental Research Letters, 13(6), 065003.
Ntegeka, V., P. Salamon, G. Gomes, H. Sint, V. Lorini, M. Zambrano-Bigiarini, and J. Thielen (2013) EFAS-Meteo: A European daily high-resolution gridded meteorological data set for 1990 – 2011, JRC Tech. Report, doi: 10.2788/51262.
Roudier, R., J.C.M. Andersson, C. Donnelly, L. Feyen, W. Greuell & F. Ludwig 2016. Projections of future floods and hydrological droughts in Europe under a +2°C global warming, Climatic Change 135, 341-355
Schulzweida, Uwe. (2019, October 31). CDO User Guide (Version 1.9.8). http://doi.org/10.5281/zenodo.3539275.
Van Vliet, M.T.H., C. Donnelly, L. Strömbäck, R. Capell & Ludwig, F. (2015). European scale climate information services for water use sectors. Journal of Hydrology 528: 503-513
