Input Data & Bias Correction


The ISIMIP framework includes bias-corrected climate-input data sets on a 0.5°x0.5°C global grid and at daily time steps. Additional socio-economic and human-influence data are also available, and detailed in the table below.

The methods used for bias correction and statistical downscaling of climate model output have been updated and improved in every ISIMIP phase. See ISIMIP3b Bias Correction, ISIMIP2b Bias Correction and Fast-Track Bias Correction for information about the methods used in the different ISIMIP phases.

For the ISIMIP3 simulation round, the protocol newly includes a counterfactual “no-climate change baseline” (detrended climate + observed socio-economic forcing) to allow for the detection and attribution of observed impacts according to the WGII-AR5 Chapter 18 definition.

If published input data sets for any reason need to be replaced or updated, you will find details of these changes in the input data change log below.

Information about accessing ISIMIP input data.

Availability of ISIMIP2b input data.

Use the table's search box to limit input data sets to a chosen simulation round, data type, or to search for a specific data set.

Data set Protocol relation Data type Simulation round Description
Protocol
None
Protocol
None
[3]
Protocol
None
Protocol
Simulated atmospheric climate
ISIMIP3b
Pre-industrial (1601-1849), historical (1850-2014) and future (2015-2100) conditions provided based on the bias adjustment and statistical downscaling of CMIP6 output from MRI-ESM2-0, using ISIMIP3BASD v2.4.1 and W5E5 v1.0. CMIP6 experiments used in ISIMIP3b are piControl, historical, ssp126 (SSP1-RCP2.6), ssp370 (SSP3-RCP7) and ssp585 (SSP5-RCP8.5).
Protocol
Simulated atmospheric climate
ISIMIP3b
Pre-industrial (1601-1849), historical (1850-2014) and future (2015-2100) conditions provided based on the bias adjustment and statistical downscaling of CMIP6 output from IPSL-CM6A-LR, using ISIMIP3BASD v2.4.1 and W5E5 v1.0. CMIP6 experiments used in ISIMIP3b are piControl, historical, ssp126 (SSP1-RCP2.6), ssp370 (SSP3-RCP7) and ssp585 (SSP5-RCP8.5).
Protocol
Simulated atmospheric climate
ISIMIP3b
Pre-industrial (1601-1849), historical (1850-2014) and future (2015-2100) conditions provided based on the bias adjustment and statistical downscaling of CMIP6 output from MPI-ESM1-2-HR, using ISIMIP3BASD v2.4.1 and W5E5 v1.0. CMIP6 experiments used in ISIMIP3b are piControl, historical, ssp126 (SSP1-RCP2.6), ssp370 (SSP3-RCP7) and ssp585 (SSP5-RCP8.5).
Protocol
Simulated atmospheric climate
ISIMIP3b
Pre-industrial (1601-1849), historical (1850-2014) and future (2015-2100) conditions provided based on the bias adjustment and statistical downscaling of CMIP6 output from UKESM1-0-LL, using ISIMIP3BASD v2.4.1 and W5E5 v1.0. CMIP6 experiments used in ISIMIP3b are piControl, historical, ssp126 (SSP1-RCP2.6), ssp370 (SSP3-RCP7) and ssp585 (SSP5-RCP8.5).
Protocol
Simulated atmospheric climate
ISIMIP3b
Pre-industrial (1601-1849), historical (1850-2014) and future (2015-2100) conditions provided based on the bias adjustment and statistical downscaling of CMIP6 output from GFDL-ESM4, using ISIMIP3BASD v2.4.1 and W5E5 v1.0. CMIP6 experiments used in ISIMIP3b are piControl, historical, ssp126 (SSP1-RCP2.6), ssp370 (SSP3-RCP7) and ssp585 (SSP5-RCP8.5).
Protocol
Other human influences
ISIMIP3a
Historical annual country-level wood harvesting data, covering the period 1850-2017.
Protocol
Land use
ISIMIP3a
Historical land-use (LU) changes for the time period 1860-2018.
Protocol
Socio-economic
ISIMIP2a
Annual gridded GDP data for 1861-2016.
Protocol
Observed atmospheric climate
ISIMIP3a
ISIMIP2a
W5E5 global meteorological forcing data processed for ISIMIP2a/ISIMIP3a.
Protocol
Socio-economic
ISIMIP2a
Population (1860 - 2017), country level and on 0.5°x0.5° grid level
Protocol
Observed atmospheric climate
ISIMIP3b
The W5E5 dataset was compiled to support the bias correction of climate input data for ISIMIP3b. For a full technical description, see "Data source" below. Data is available for the time period 1979-2016.
Protocol
Observed atmospheric climate
ISIMIP2a
EWEMBI global meteorological forcing data processed for ISIMIP2a.
Protocol
Other
ISIMIP2b
ISIMIP2a
ISIMIP Fast Track
Extension of GRanD reservoirs & dams data. For some dams for which no construction/commissioning years were given in GRanDv1.1, they were added by PIK according to various sources (see column “SOURCE”). The dam location data previously to 2010 is based on GRanDv1.1 (Lehner et al., 2011, DOI: 10.1890/100125), but adjusted to match the river-routing network in DDM30. After 2010, the data has been provided by Dr. Jida Wang, from Kansas State University (KSU). Mapping algorithm developed by Andreas Dobler. Manual mapping done by M.Rocío Rivas, PIK.
Secondary
Other
ISIMIP Fast Track
Increased human and economic losses from river flooding with anthropogenic warming. The data set includes: - Static maps (land use, economic variables, population) used to compute impacts - For each model realization, the merged global maps of economic damage for the historical (hist) and future (rcp8p5) periods, as well as statistic tables of impacts aggregated per continent and per country - For each model realization, the merged global maps of population affected computed with current socioeconomic conditions and with SSP3 and SSP5 population projections, and the statistic tables of impacts aggregated per continent and per country - Tables of variables used as input in the impact assessment procedure - Tables describing aggregated impact results
Secondary
Socio-economic
ISIMIP2b
Annual global population data at 0.5° and 2.5' resolutions for the years 2006-2100 based on the population projections described in Jones & O’Neill (2016)
Secondary
Simulated atmospheric climate
ISIMIP2b
Future (2006-2099) conditions provided based on CMIP5 output from MIROC5. These data are analogous to the data described here: https://www.isimip.org/gettingstarted/details/21/ but for this RCP not included in the ISIMIP2b simulation protocol.
Secondary
Simulated atmospheric climate
ISIMIP2b
Future (2006-2099) conditions provided based on CMIP5 output from HadGEM2-ES. These data are analogous to the data described here: https://www.isimip.org/gettingstarted/details/52/ but for the this RCP not included in the ISIMIP2b simulation protocol.
Secondary
Simulated atmospheric climate
ISIMIP2b
Future (2006-2299) conditions provided based on CMIP5 output from IPSL-CM5A-LR. These data are analogous to the data described here: https://www.isimip.org/gettingstarted/details/53/ but for this RCP not included in the ISIMIP2b simulation protocol.
Secondary
Simulated atmospheric climate
ISIMIP2b
Future (2006-2099) conditions provided based on CMIP5 output from GFDL-ESM2M. These data are analogous to the data described here: https://www.isimip.org/gettingstarted/details/51/ but for this RCP not included in the ISIMIP2b simulation protocol.
Protocol
Land use
ISIMIP2b
Time-varying historical crop-land areas for 6 land-use types: rain-fed crop land, irrigated crop-land, rain-fed bioenergy, irrigated bio-energy, pastures, natural vegetation and urban areas. For more specific information, thd data are provided in 5 different resolutions of land-use types: landuse-totals: the land surface is broadly categorized into rainfed (cropland_rainfed) and irrigated cropland (cropland_irrigated), and into pastures (pastures) landuse-5crops: the distribution of the five predominant crop types (c3ann, c3per, c4ann, c4per and c3nfx), delineating both occurences of rainfed/irrigated crops and occurences of nutrition/bioenergy crops landuse-15crops: provides areas specific to either important species, e.g. wheat, or species groups, such as pulses landuse-pastures: pastures areas are split into managed pastures (managed_pastures) and rangeland (rangeland) landuse-urbanareas: provides the per-cell coverage of urban areas Data is available for the time period 2006-2099.
Protocol
Socio-economic
ISIMIP2b
Country-level population data for the years 1861-2009.
Protocol
Socio-economic
ISIMIP2b
Annual GDP data on a 0.5°x0.5° grid for 2006-2099.
Protocol
Socio-economic
ISIMIP2b
Annual country-level GDP data for the SSP2 scenario for 2006-2100 based on a linear temporal interpolation of the OECD data.
Protocol
Simulated atmospheric climate
ISIMIP2b
Pre-industrial (1661-1860), historical (1861-2005), future (RCP2.6 (2006-2100), RCP6.0 and RCP8.5 (2006-2099)) and extended-future (RCP2.6 and RCP8.5 (2100-2299)) conditions provided based on CMIP5 output from IPSL-CM5A-LR.
Protocol
Simulated atmospheric climate
ISIMIP2b
Pre-industrial (1661-1860), historical (1861-2005), future (RCP2.6 (2006-2100), RCP6.0 and RCP8.5 (2006-2099)) and extended-future (RCP2.6, 2100-2299) conditions provided based on CMIP5 output from HadGEM2-ES.
Protocol
Simulated atmospheric climate
ISIMIP2b
Pre-industrial (1661-1860), historical (1861-2005) and future (RCP2.6 (2006-2099), RCP6.0 (2006-2099) and RCP8.5 (2006-2099)) conditions provided based on CMIP5 output from GFDL-ESM2M.
Protocol
Other human influences
ISIMIP2b
ISIMIP2a
Catch data by country or EEZ or LME or globally. Covers 1950-2015.
Protocol
Land use
ISIMIP2a
Time-varying historical crop-land areas for 12 crop-functional types and pastures for rainfed and irrigated land use for the years 1700-2012. Based on MIRCA 2000 and historical trends from HYDE 3.0.
Protocol
Other
ISIMIP3b
ISIMIP3a
ISIMIP2b
ISIMIP2a
ISIMIP Fast Track
Land-sea masks, on a global 0.5°x0.5° lat-lon grid, for the different ISIMIP simulation rounds.
Protocol
Other human influences
ISIMIP2b
Water abstraction for domestic and industrial uses consistent with SSP2. Histsoc (1901-2005); rcp26soc (2006-2099); rcp60soc (2006-2099)
Protocol
Simulated ocean climate
ISIMIP2b
Non bias-corrected 3D oceanic temperature fields on original GCM grid with monthly resolution from the GCMs: IPSL-CM5A-LR, GFDL-ESM2M, MIROC5, HadGEM2-ES.
Protocol
Simulated atmospheric climate
ISIMIP2b
Non bias-corrected 3D humidity and temperature fields on original GCM grid with monthly resolution from the GCMs: IPSL-CM5A-LR, GFDL-ESM2M, MIROC5, HadGEM2-ES.
Protocol
Emissions
ISIMIP2b
ISIMIP2a
ISIMIP Fast Track
Atmospheric CO2 concentration for 1661-2299. Values are constant for 1661-1860, following observations from 1861-2005, and correspond to rcp26, rcp60 and rcp85 from 2006-2299.
Protocol
Simulated atmospheric climate
ISIMIP2b
Uncorrected daily wind components at 250hPa and 850hPa on original GCM grid for the GCMs: GFDL-ESM2M (1861-2100), MIROC5 (1850-2300), IPSL-CM5A-LR (1850-2300) & HadGEM-ES (1959-2299).
Protocol
Observed atmospheric climate
ISIMIP2b
The EWEMBI dataset was compiled to support the bias correction of climate input data for ISIMIP2b and the ISIMIP2a protocol extension. For a full technical description, see "Data source" below. Data is available for the time period 1979-2016.
Protocol
Socio-economic
ISIMIP2b
Annual country-level population, urban population and age structure (in five-year age bands) based on SSP2 from IIASA (population & age structure) and NCAR (urbanshare). Data is available for the following time periods: Annual country-level population (2006-2099), urban population (2006-2099), age structure (2010-2100).
Protocol
Socio-economic
ISIMIP2b
Annual global population data at 0.5° and 5' resolutions for the years 2006-2099 based on the national SSP2 population projections as described in Samir and Lutz, (2014).
Protocol
Simulated atmospheric climate
ISIMIP Fast Track
Daily-resolution, bias-corrected climate data from the global climate models MIROC-ESM-CHEM, NorES1-M, IPSL-CM5A-LR, GFDL-ESM2M, HadGEM2-ES covering the period 1950-2099 (historical run up to 2004, then split into 4 RCPs), downscaled to a 0.5°x0.5° lat-lon grid. For a full technical description, see the DOI link below (Hempel, et al., 2013a).
Protocol
Simulated ocean climate
ISIMIP2a
The ESM2M run is a fully coupled free running climate model forced by historical and future scenario GHGaerosol, solar boundary conditions without flux adjustment or other restoring to observations using TOPAZv2 as in Dunne et al 2013. DOI: 10.1175/JCLI-D-12-00150.1 Data is available for the following time periods: historical (1950-2005), rcp26 (2006-2100), rcp45 (2006-2100), rcp60 (2006-2100), rcp85 (2006-2100).
Protocol
Simulated ocean climate
ISIMIP2a
The COBALT run is as described in: http://www.sciencedirect.com/science/article/pii/S0079661113001079. It uses the same basic ocean-ice setup as Dunne et al., 2013 but is forced with reanalysis. Data is available for the following time period: historical (1959-2004).
Protocol
Other
ISIMIP2b
ISIMIP2a
ISIMIP Fast Track
** This data set has been expanded with missing construction/commissioning years, see https://www.isimip.org/gettingstarted/details/75/ ** The dam data is based on GRanDv1.1 (Lehner et al., 2011, DOI: 10.1890/100125), but adjusted to match the river-routing in DDM30. After 2010, the data has been provided by Dr. Jida Wang, from Kansas State University (KSU). Use of these data sets is recommended but, if unfeasible, groups may use their default dataset. For a comparison of this dataset with the default WaterGAP and LPJmL datasets, see https://www.arcgis.com/home/item.html?id=d966db9c7b2949ac8380458d7020adf9
Protocol
Observed atmospheric climate
ISIMIP3a
ISIMIP2a
Daily-resolution observed climate data on a global (land and ocean) 0.5°x0.5° lat-lon grid from the Global Soil Wetness Project Phase 3 (GSWP3), based on the reanalysis data set 20CR and using the bias targets GPCC, GPCP, CPC-Unified, CRU and SRB. The data set covers the period 1901‐2010. For more information about GSWP3's scope, experiment design, framework, etc. see https://www.isimip.org/documents/405/GSWP3.forcing.HyungjunKim20140306.HJKIM.pdf
Protocol
Socio-economic
ISIMIP2b
ISIMIP2a
Annual country-level GDP data for 1850-2016.
Protocol
Observed atmospheric climate
ISIMIP2a
Daily-resolution observed climate data on a global (land and ocean) 0.5°x0.5° lat-lon grid. PGMFD v.2 (Princeton) is from he Terrestrial Hydrology Group at Princeton University, based on the reanalysis data set NCEP/NCAR Reanalysis I and using the bias targets CRU, SRB, TRMM, GPCP & WMO and validated against GSWP2. The PGMFD v.2 data set covers the period 1901‐2012. WATCH (WFD) Derived from the WATCH Project, based on the reanalysis data set ERA-40 and using the bias target GPCC. The data set covers the period 1901‐2001. WATCH+WFDEI Based on the reanalysis data set ERA-Interim and using the bias target GPCC. The data set covers the period 1901‐2012, where the data for 1901-1978 are taken from WFD, and from 1979 onwards from WFDEI.GPCC
Protocol
Land use
ISIMIP2b
Historical land-use (LU) changes from the HYDE3.2 data (Klein Goldewijk, 2016) (see Figure 3). Data for the time period 1861-2005.
Protocol
Land use
ISIMIP Fast Track
Harvested areas based on the MIRCA data set, covering irrigated and rainfed landuse.
Protocol
Socio-economic
ISIMIP2b
Annual global population data at 0.5° and 5' resolutions for the years 1861-2005 on a 0.5° and 5' grid.
Protocol
Simulated ocean climate
ISIMIP2a
Ocean data from the GCM IPSL-CM5A-LR. Data is available for the following time periods: historical (1950-2005), rcp26 (2006-2100), rcp45 (2006-2100), rcp60 (2006-2100), rcp85 (2006-2100).
Protocol
Simulated atmospheric climate
ISIMIP2b
Pre-industrial (1661-1860), historical (1861-2005), future (RCP2.6 (2006-2100), RCP6.0 and RCP8.5 (2006-2099)) and extended-future (RCP2.6, 2100-2299) conditions provided based on CMIP5 output from MIROC5.
Protocol
Other human influences
ISIMIP3a
ISIMIP2b
ISIMIP2a
We provide the fertilization rate per our five canonical crop types, i.e. for both annual and perennial C3 and C4 crops, as well as for nitrogen fixing crops. The variables’ names are fertl_c3ann, fertl_c3per, fertl_c4ann, fertl_c4per and fert_c3nfx. The data are provided on an annual times step on the land surface of a global 0.5° × 0.5° grid. The units are kilograms per year and per hectare. ISIMIP2a/ISIMIP2b: Data is available for the following time periods: historical (1861-2005), rcp26soc+rcp60soc (2006-2099) (see below for detailed information). ISIMIP3a: Data is available for the time period: historical (1860-2018).
Protocol
Other human influences
ISIMIP3a
ISIMIP2b
ISIMIP2a
ISIMIP2a/ISIMIP3a: Monthly, 0.5° gridded data for 1860-2016. ISIMIP2b: Annual and monthly, 0.5° gridded data for 1850-2099 derived from the average of three atmospheric chemistry models (i.e., GISS-E2-R, CCSM-CAM3.5, and GFDL-AM3) in the Atmospheric Chemistry and Climate Model Intercomparison Project (ACCMIP) (0.5° x 0.5°).
Protocol
Observed atmospheric climate
ISIMIP2a
Daily-resolution observed climate data on a global (land and ocean) 0.5°x0.5° lat-lon grid from the Terrestrial Hydrology Group at Princeton University, based on the reanalysis data set NCEP/NCAR Reanalysis I and using the bias targets CRU, SRB, TRMM, GPCP & WMO and validated against GSWP2. The data set covers the period 1901‐2012
Protocol
Socio-economic
ISIMIP Fast Track
Historical (1950-2009) and future projections (2010-2100) of country-level and gridded population data based on the Shared Socio-economic Pathways (SSPs) for SSP1-SSP5.
Protocol
Simulated atmospheric climate
ISIMIP2b
Uncorrected pressure and wind fields at 10m above the land surface with 3-hrly resolution for the GCMs IPSL-CM5A-LR (1861-2299), MIROC5 (1861-2099) and GFDL-ESM2M (1861-2099).
Protocol
Other
ISIMIP2b
ISIMIP2a
ISIMIP Fast Track
DDM30 River-routing network data.
Protocol
Other
ISIMIP2b
Total climate-driven regional sea level projections on a global 0.5x0.5° grid. Projection is a combination of ocean dynamic sea level fields and steric global contribution from the ISIMIP GCMs, and the fingerprints from glaciers and the Greenland and Antarctic ice sheets. Tthe time-dependent fingerpints were generated by scaling a time-independent fingerprint (see Bamber and Riva 2010) with the respective global timeseries for each contribution. Uncertainty for fingerprints is only from the global timeseries. Data is available for following time periods: historical (1861-2005), rcp26 (2006-2099; HadGEM2-ES and IPSL-CM5A-LR extended to 2299), rcp60 (2006-2099) (See below for detailed information).
Protocol
Observed atmospheric climate
ISIMIP2a
Daily resolution climate data on a global 0.5°x0.5° lat-lon grid: Derived from the WATCH Project, based on the reanalysis data set ERA-40 and using the bias target GPCC. The data set covers the period 1901‐2001. Related website: http://www.waterandclimatechange.eu/about/watch-forcing-data-20th-century When referring to the WATCH data set, the following paper should be cited: Weedon, G.P., S.S. Gomes, P.P. Viterbo, W.J. Shuttleworth, E.E. Blyth, H.H. Österle, J.C. Adam, N.N. Bellouin, O.O. Boucher, and M.M. Best, 2011: Creation of the WATCH Forcing Data and Its Use to Assess Global and Regional Reference Crop Evaporation over Land during the Twentieth Century. J. Hydrometeor., 12, 823–848, doi: 10.1175/2011JHM1369.1
Protocol
Observed atmospheric climate
ISIMIP2a
Daily-resolution observed climate data on a global (land only) 0.5°x0.5° lat-lon grid, based on the reanalysis data set ERA-Interim and using the bias target GPCC. The data set covers the period 1901‐2012, where the data for 1901-1978 are taken from WFD, and from 1979 onwards from WFDEI.GPCC

Input data changelog

July 29, 2020

We have deleted the entries from GRanD without the year of construction, except for 29 large dams (capacity >0.5 km3), for which we have searched and filled out the year of construction. This change means that the reservoirs/lakes without dams are removed from the ISIMIP3 databases, since for these ones there is no year of construction ...

July 23, 2020

In addition to the dimensions that are regularly to be reported for ISIMIP output variables (time,lat,lon), an alternative set of dimensions is now given in the comment column of the respective table for variables with a number of levels (e.g. layers or depth). More information about level dimensions can be found here and here ...

July 6, 2020

Thanks to the initiative of Jonas, Christoph and Sam the ISIMIP3 protocol now provides an additional soil data set that meets criteria specific to the Agriculture Sector. The data are based on HWSD soil data and were aggregated to the ISIMIP 0.5° grid. The aggregation has been performed using the dominant soil class on cropland (using the 5 arc- ...

June 19, 2020

A flaw in ISIMIP3BASD v2.3.1 caused bad bias adjustments of all variables except ps, rlds, and tas. The problem is particularly severe for precipitation. The flaw was corrected in ISIMIP3BASD v2.4.1, which was used to re-do the bias adjustment of climate model output for the affected variables. All climate models and climate scenarios are affected ...