CMIP6 statistically downscaled climate datasets

Access statistically downscaled climate datasets based on model output from the Coupled Model Intercomparison Project Phase 6 (CMIP6). Statistical downscaling predictors are also available.

What is downscaling? Downscaling is a method of providing global climate model (GCM) output at a finer spatial resolution. Consider a low resolution video where you cannot discern details because each image in the video is composed of a small number of large pixels. In contrast, in a high definition video, you can see even the minute details of the people and background because each image is composed of many small pixels. Downscaling is a similar process, where a low resolution grid of few large pixels is converted to a high resolution one made up of many small pixels. Converting to a higher resolution allows us to discern more details within the image or, in this case, make more spatially detailed future projections for climate. However, it is important to note that a downscaled projection is not necessarily better or more accurate; rather, it provides information at a scale that is more convenient for many applications.

There are two approaches to downscale climate model output: statistical downscaling and dynamical downscaling. In dynamical downscaling, a regional climate model (RCM) provides projections at a higher resolution for a particular region with a global climate model providing the input and drivers from the outside world to the borders of that region. With statistical downscaling, as provided here, a statistical relationship between coarser-scale global climate model output and local historic data from weather stations is derived to produce higher resolution projections.

What datasets are available? The first drop-down menu below provides a link to access statistically downscaled projections of minimum and maximum temperature and total precipitation. The projections follow the scenarios used in CMIP6 (Coupled Model Intercomparison Project Phase 6), the Shared Socio-economic Pathways (SSPs), which are a set of alternative storylines about how the world might develop, and greenhouse gas emissions might change, over the coming century.

The second drop-down menu provides a link to download statistically downscaled climate indices. An index is a quantitative value that is calculated from climate datasets and provides more targeted information about related to temperature or precipitation. Some examples of climate indices are, the hottest day of each year and driest day of each year. A full list of the indices provided can be found here.

There are also links at the bottom of the page to CMIP6 daily predictor variables. A predictor variable is output from a global climate model intended for use in the kind of statistical downscaling described above. As well, observational datasets are provided, also for use in statistical downscaling. A total of 26 variables are provided for users with the capability to perform their own statistical downscaling.

Data are available for download at the following resolutions:

  • Temporal resolution: available at a daily scale
  • Spatial resolution: approximately 10km
  • Download temperature and precipitation data

    Individual statistically downscaled global climate model (GCM) temperature and precipitation output, named the Canadian Downscaled Climate Scenarios–Univariate method from CMIP6 (CanDCS-U6), are available for download.

    Please see the technical notes for more details

    In addition, please note that the CanDCS-U6 dataset is the same dataset made available by Pacific Climate Impacts Consortium (PCIC).

  • Download climate extreme indices

    Statistically downscaled climate indices

    Download statistically downscaled multi-model output of climate indices in NetCDF format.

    Please see the technical notes and the definitions table for more details.

  • Technical notes on statistical downscaling method

    The CMIP6 statistically downscaled climate scenarios, named Canadian Downscaled Climate Scenarios–Univariate method from CMIP6 (CanDCS-U6), are based on the Bias Correction/Constructed Analogues with Quantile mapping version 2 (BCCAQv2) procedure. Please see the technical notes for more details.

    Please visit the Pacific Climate Impacts Consortium (PCIC) for statistically downscaled datasets based on other downscaling methods, such as the Bias-Correction Spatial Disaggregation (BCSD) and Bias Correction/Constructed Analogues with Quantile mapping reordering (BCCAQ) procedures.

    Please note that the CanDCS-U6 dataset is the same dataset made available by PCIC.

Date modified: