Results from these simulations inform the Intergovernmental Panel on Climate Change Assessment Reports. We are currently in the sixth phase of CMIP (i.e., CMIP6), with climate model results from this phase informing the recent IPCC Sixth Assessment Report (AR6; IPCC, 2021). The CMIP5 climate model results were reported in AR5 (IPCC, 2013a) and have been widely used to assess climate change impacts and in adaptation studies. The Canadian Centre for Climate Modelling and Analysis has been involved in climate model development for several decades, with versions of the Canadian ESM taking part in both CMIP5 (CanESM2) and CMIP6 (CanESM5). Details for the CMIP5 and CMIP6 ensembles are provided in Table 4.1.
1. Pre-industrial simulations (corresponding to year 1750 or 1850)
The pre-industrial simulation is designed to represent climate conditions which occurred prior to industrialization. The concentration of greenhouse gases (GHGs), land cover, emissions of atmospheric aerosols and other forcings are all set to their pre-industrial levels and the simulations then run for several hundreds of years. Even though there is no imposed change in external forcing (e.g., an increase in atmospheric CO2) in these simulations, there is year-to-year variability in the simulated climate simply because of the natural variability of the climate system. The long-term climate, however, remains stable, i.e., there is no systematic increase or decrease (‘drift’) in temperature or other climate variables. The pre-industrial simulation serves as a control for comparison with the historical and future simulations both at the global and regional scale.
2. Historical simulations (generally from 1851 to ‘present’)
The pre-industrial simulations are also used to launch the historical simulations - any year in the pre-industrial simulation can be used as the starting point for the historical simulations since any year is representative of the pre-industrial climate. Each climate modelling centre will typically launch, or initialize, between 5 and 10 historical simulations[1] from different years of the pre-industrial simulation. Unlike the pre-industrial simulation, however, the historical simulation includes external forcings from increasing concentrations of GHGs in the atmosphere and increasing emissions of aerosols (i.e., soot, also associated with increasing use of fossil fuels), and from changes in land cover (associated with deforestation related to agriculture). In addition to these forcings, information about changes in solar activity and volcanic forcing are also included. While all these forcings affect the climate in different ways, the primary response of the climate system over the historical period to the combined effect of all these forcings has been warming.
While each historical simulation is distinct in terms of the year-to-year climate, over the long historical period they all show similar increases in temperature associated with increasing GHG concentrations in the atmosphere. Note that none of the historical simulations can be expected to correspond to the actual climate observed at corresponding dates in the past – the historical climate that we have observed is an individual realization of the highly chaotic climate system. To be able to reproduce the exact observed year-to-year variability, we need to know the exact state of all components of the climate system in 1850 (or any other year) to initialize a perfect climate model – neither nor which are possible. However, models simulate the overall statistical characteristics of the historical climate well (e.g., climate trends, time averages, amplitude of variability).
For CMIP5, the historical simulation was from 1850 to 2005; for CMIP6 it was from 1850 to 2014. Users should be aware that there are many other research-related experiments undertaken as part of the model intercomparison projects (MIPs; e.g., the atmospheric model intercomparison project - ‘AMIP’; ‘land-hist*’ labels). If downloading raw climate data, users should select the ‘historical’ experiment label for climate adaptation analyses over the historical period.
The statistical results from historical simulations are usually compared with observations (e.g., for temperature and other climate variables) at the global scale, as well as at different geographical scales and at several time frequencies, e.g., annual, seasonal and monthly. Comparison with observations helps to evaluate model results and assess model limitations.
3. Future scenario simulations (e.g., from the ‘present’ to 2100)
The climate model simulations for the future are initialized from the historical simulations, in exactly the same way that the historical simulations are initialized from the pre-industrial simulations. For CMIP5, the future simulation period is 2006-2100; for CMIP6 it is 2015 to 2100. The main difference between the historical and future simulations is that the forcings (e.g., concentrations of GHG, of aerosols, changes in land cover) are known during the historical period but not for future.
The change in forcings for the future is dependent on human activities, which are uncertain. Prediction of future human activities comes under the purview of the environmental and energy policy analysis community, which uses integrated assessment models (IAMs) to consider the social and economic factors driving the emissions of GHGs. This community develops coherent and consistent descriptions of how future society may operate, particularly with regards to fossil fuel consumption, and uses IAMs to calculate anthropogenic emissions and concentrations of CO2, methane (CH4), nitrous oxide (N2O) and other GHGs, atmospheric aerosols, as well as estimates of future agricultural area, for each emissions pathway they develop. These future emissions scenarios typically span a range of conditions, from low to moderate to high emissions, and some include mitigation actions. Each emissions scenario provides the necessary forcing information to drive climate models. The resulting simulations of future climate, known as climate scenarios, are all equally plausible descriptions of possible future states of the climate system in response to the prescribed forcing. None of the climate scenarios should be considered as a forecast or prediction, but rather as an example of a possible future climate. When using climate scenarios in adaptation planning, therefore, it is advisable to use multiple scenarios to capture the range of possible future climates.
There are many possible emission scenarios and it is not possible to run climate models for all of them. The CMIP experimental protocol recommends a core set of climate model experiments which consider emissions ranging from low, to moderate, to high.
[1] The set of climate model results stemming from simulation experiments using the same forcing is generally known as an ensemble.