Numerical climate simulations have been through many different crucial stages before the appearance of the CRCM. In a general context, this section recalls the main goal and realization of this new science.
Earth past climate studies give important knowledge on the natural variability of the global climate system and on ecosystems responses to climate change. However, they cannot answers many questions regarding processes and feedbacks occurring inside the climate system and between its various components. Experimentation with the climate system is inconceivable, both because of the scale of the experiment as well as the risks involved. Involuntarily, Humankind started a planetary scale climate experiment, but it would be unwise to wait decades or more to find out the results and learn from our mistakes. In order to develop a strategy and give society a chance to adapt to upcoming climate change, it is important to develop sophisticated climate system models. Even though these simulators do not reproduce perfectly all the real climate system characteristics, they are the only tools at our disposal to test the behavior of this complex system in a perturbed environment.
The advent of computers in the 1950's and the exponential growth of computing power in the following decades have allowed the development of numerical simulation of climate processes. Climate processes are based on fundamental physical laws (e.g. energy, mass, momentum conservation), which can be stated mathematically. Writing the mathematical equation describing climate processes is equivalent to creating a model. By modifying certain quantities in the equations, we learn how the system would react to certain constraints. By adding more equations, in order to describe a greater amount of processes and interactions, we create a model of the whole climate system, including the atmosphere, ocean and biosphere. The theoretical limit of this complexity is the computing power and computing time required to solve the set of equations.
Atmospheric Global Climate Models (GCM) are complex models based on the physical principles representing atmospheric motion in three dimensions. GCMs reproduce the energy cycle in the climate system, from its entrance as shortwave radiation (solar energy) to its departure as longwave radiation (infrared). They also simulate the effects of this energy on the different elements of the climate systems, calculating the results in terms of temperature, precipitation, humidity and other important climate variables. The GCMs are the most complete models but also the most difficult to use. In these GCMs, the climate system is represented by a series of thousands of equations, each of which is solved at each iteration. For each timestep (~20 min), and for each grid point (between 350 and 500 km), they execute over 100 000 instructions. Computing power is by far the factor limiting spatial resolution.
Even with their high complexity and the use of the most powerful supercomputer, the GCMs are only an approximation of reality. The coarse resolution of GCMs limits the representation of topography, the different types of soil and vegetation, the lakes distribution, etc, which influence the climate. This limited spatial resolution also forces the parameterization of subgrid climate processes. Phenomena such as cloud formation, precipitation, evaporation and sol humidity are introduced in the GCMs as fixed values or through parameterization which link the processes physically and statistically to large scale variables (temperature, pressure, humidity).
In the 1960's, only a few research centers were working at developping GCMs. Quickly, with the increase in computing power, simulation of large-scale atmospheric circulation started to attract attention from a number of research centers around the world. In Canada, the developpement of a GCM began in the mid 70's at the Atmospheric Environment Service (AES) in Toronto (Boer and McFarlane, 1979). The model's origins come from a spectral model used for weather forecast (Daley et al. 1976) to which were added the necessary parameterization to simulate large-scale circulation. This model became the first Canadian GCM (GCMi). It had a horizontal resolution of 625 km, used 10 vertical levels and, except for one innovation, used the same general characteristics than first generation GCMs (Boer et al. 1984). The innovation that set him apart was the introduction of a parameterization scheme for momentum transport by subgrid vertically propagating gravity wave. The parameterization of gravity wave originating from atmospheric flow above topography allowed to significantly improve jet stream representation.
At the end of the 80's, the second generation of GCM was designed, significantly increasing resolution and improving characteristics, processes and parameterizations. In the second version of the Canadian GCM (CGCMii), values prescribed for sea surface temperatures, ice and cloud covers were replaced by schemes which evaluated these fields in an interactive way (McFarlane et al., 1992; Boer et al., 1992). Spatial resolution of the second version was increase to ~450 km.
In the 90's, a third version (GCMiii) was started in Victoria (BC) at the Canadian Centre for Climate Modelling and Analysis (CCCma). This new version uses a sophisticated surface scheme (CLASS, Verseghy, 1991; Verseghy et al. 1993) and a better parameterization of cloud cover and convective processes.
The scientific community recognizes the major role played by Coupled GCM (CGCM), global climate models coupled to surface models, ocean models and sea ice models, with respect to our understanding of physical processes responsible for the equilibrium, evolution and natural variability of the climate system. CGCM are the most sophisticated tools to make climate projections when atmospheric and surface conditions are altered. However, because of their complexity and the time required to produce simulations, these models are quite demanding in terms of computer resources. For this reason, the CGCM have to use large grids (order of hundreds of km) than the weather forecasting models (order of tens of km). But, the impact of climate change on the environment, society and economy requires a finer scale than what is currently available from CGCMs. High-resolution regional climate models are then an essential tool to study the climate system and future climate changes. Even if they are equally computationally expensive as the CGCM, they can reach much finer scales.
Because it was impossible to regularly perform high-resolution integrations of CGCMs, Giorgi and other scientists from the Center for Atmospheric Research (NCAR) showed it was possible to nest a high resolution model inside a low-resolution CGCM (Dickinson et al., 1989; Giorgi 1990; Giorgi and Marinucci 1991). Their strategy consisted in interpolating low-resolution atmospherical fields of CGCMs onto a regional grid such that the high-resolution regional model could use them as lateral boundary conditions. This pioneer work paved the way to the development of regional climate models (RCM) which have been developed since in a dozen research centers around the world, including in Canada.
The Canadian Regional Climate Modeling and Diagnostics Network (CRCMD) is located at UQAM and is responsible for the development of the CRCM for the past fifteen years. The goal is to develop a Canadian expertise in regional climate modeling and to develop, maintain, validate and use this most sophisticated regional climate simulator. At the beginning of the 90's, the Canadian expertise in regional climate modeling was literally non-existent. At that time, René Laprise thought of using a new dynamical core, based on Euler's equations (Tanguay, Robert et Laprise, 1990), as a foundation for a new RCM. A new CRCM was then developed between 1991 and 1994 as part of Daniel Caya's Ph.D. work (Caya and Laprise, 1999). The CRCM is characterized by three main components: a dynamical core, an ensemble of physical parameterization and a series of driving and diagnostics scripts. The dynamical core is based on Euler's equations (perfectly elastic) and can be used at all spatial scales. The model's equations are solved by top of the line semi-implicit and semi-Lagrangian numerical algorithms, which makes it about five times faster than other RCMs. The physical parameterizations are based on the Canadian GCM of second generation (CGCMii; McFarlane et al., 1992). The series of scripts used to manage the large quantities of simulated or observed data were specifically designed to handle regional limited area simulations and their subsequent analysis.
As part of the first phase (1991-1996), the series of diagnostics has been improved to make data handling operations automatics and the dynamic of the model was optimized (Caya et al. 1998). Two simulations of five years were made over the Canadian West. They corresponded to present-day and double CO2 concentration and ended in 1996, after which they were analyzed and published (Laprise et al., 1998). These relatively long simulations were useful to detect certain weaknesses in the first version of the CRCM, most notably the empirical formulation of cloud cover and convective processes.
During phase two (1996-2001), two versions of cumulus parameterizations were implemented in the CRCM. The first one was developed for meso-scale models by Kain and Fritsch (1990) while the second is the adapted version of Bechtold et al. (2001) and is used by the French meteorological community. The diagnostics were increased to include higher statistical moments as well as some diagnostics budgets. Three simulations of 10 years with varying aerosols and CO2 concentrations were completed and analyzed. With time, the group efforts were dedicated at comparing CRCM simulations with available observations, while special attention was given to fields that beneficiated the most from the high resolution (e.g. hydrological processes and surface fluxes). The group is also studying nesting strategies while trying to define the limits in the use of RCMs. To address this issue, two ten years simulations were done with different nesting strategies and a comparison allowed to analyze their influence on the simulated climate (Caya et al., en préparation).
During the last 10 years, the Network has developed from non-existant to thriving and has reached an expertise such that it can keep up with international work in the field of regional climate modeling. The training of highly qualified personnel in the use and development of RCM is also part of the activities of the CRCMD Network. Many graduate students and research assistants have worked on different aspects of the model, from developing specific modules, to using the model and its output. The Network currently employs ~38 master's students, ~13 Ph.D. students, ~9 post-doctoral fellow and ~22 research assistants.
Boer, G.J., et N.A. McFarlane (1979):
The AES atmospheric general
circulation model. Report of the JOC Study Conference on Climate
Models: Performance, Intercomparison and Sensitivity Studies, Vol. I,
GARP Publ. Ser. No. 22, pp. 409-460.
Boer, G.J., N.A. McFarlane et M. Lazare (1992) Greenhouse gas- induced climate change simulated with the CCC second- generation general circulation model. J. Climate 5:1045-1077.
Boer, G. J., N. A. McFarlane, R. Laprise, J. D. Henderson et J.-P. Blanchet, 1984a: The Canadian Climate Centre Spectral Atmospheric General Circulation Model. Atmos.-Ocean, 22, 397-429.
Boer, G. J., N. A. McFarlane et R. Laprise, 1984b: The Climatology of the Canadian Climate Centre General Circulation Model as Obtained from a Five-year Simulation. Atmos- Ocean, 22, 430-473.
Caya, D., S. Goyette et R. Laprise, 1994: La modélisation des climats: du global au régional. Interface: la revue de la recherche, ACFAS, 15 (1), 14-25.
Daley, R., C. Girard, J. Henderson et I. Simmonds, 1976: Short-Term Forecasting with a Multi-Level Spectral Primitive Equation Model. Part I - Model Formulation. Atmosphere, 14, 98-116.
McFarlane, N. A., G. J. Boer, J.-P. Blanchet et M. Lazare, 1992: The Canadian Climate Centre Second Generation General Circulation Model and its Equilibrium Climate. J. Climate, 5, 1013-1044.
Verseghy, D. L., 1991: CLASS - A Canadian Land Surface Scheme for GCMs, I. Soil Model. Int. J. of Climatology, 11, 111-133.
Verseghy, D. L., N. A. McFarlane et M. Lazarre, 1993: CLASS - A Canadian Land Surface Scheme for GCMs, II. Vegetation Model and Coupled Runs. Int. J. of Climatology, 13, 347-370.
Dickinson, R. E., R. M. Errico, F. Giorgi et G. T. Bates, 1989: A Regional Climate Model for the Western United States. Clim. Change, 15, 383-422.
Giorgi, F., 1990: Simulation of Regional Climate using a Limited Area Model Nested in a General Circulation Model. J. Clim., 3, 941-963.
Giorgi, F., et M. R. Marinucci, 1991: Validation of a Regional Amospheric Model Over Europe: Sensitivity of Wintertime and Summertime Simulations to Selected Physics Parameterizations and Lower Boundary Conditions. Quart. J. Roy. Meteor. Soc., 117, 1171-1206.
Tanguay, M., A. Robert et R. Laprise, 1990: A semi-implicit Semi-Lagrangian Fully Compressible Regional Forecast Model. Mon. Wea. Rev., 118, 1970-1980.
Caya, D. et R. Laprise, 1999: A semi-implicit semi-Lagrangian regional climate model: The Canadian RCM. Mon. Wea. Rev., 127(3), 341-362.
Caya, A., R. Laprise et P. Zwack, 1998: On the effect of using process splitting for implementing physical forcings in a semi-implicit semi-Lagrangian model. Mon. Wea. Rev., 126(6), 1707-1713.
Laprise, R., D. Caya, M. Giguère, G. Bergeron, H. Côté, J.-P. Blanchet, G. J. Boer et N. McFarlane, 1998: Climate and Climate Change in Western Canada as Simulated by the Canadian Regional Climate Model. Atmos.-Ocean, 36(2), 119-167.
Laprise, R., D. Caya, A. Frigon and D. Paquin, 2003: Current and perturbed climate as simulated by the second-generation Canadian Regional Climate Model (CRCM-II) over northwestern North America. Clim. Dynamics, 21, 405-421. (DOI:10.1007/s00382-003-0342-4).
Riette, S. and D. Caya, 2002: Sensitivity of short simulations to the various parameters in the new CRCM spectral nudging. Research activities in Atmospheric and Oceanic Modelling, edited by H. Ritchie, WMO/TD - No 1105, Report No. 32: 7.39-7.40groupe de recherche.