MEC-Surface & Hydrology (MESH)

This system is a stand-alone Land-Surface-Hydrology scheme designed for both forecasting and open loop simulations. The CLASS land-surface scheme (LSS) is currently operational in the Canadian GCM and RCM frameworks, and MESH contains a standalone version of CLASS 3.7 that has been modified to capture important hydrological processes. MESH can also host other LSS. The SVS LSS is a modification of the current GEM land-surface model operationalized in ECCC (ISBA) that also has been modified to incorporate important hydrological components. A version of MESH with SVS is linked to GEM in a two-way coupled mode often called GEM-Hydro. In short, MESH is comprised of two main surface water and energy budget simulation models. The CLASS Canadian Land surface scheme (MESH-C) and the SVS land-surface scheme (MESH-V) with the former more readily adaptable to open-loop simulations, and the latter (SVS) designed to work in conjunction with land-data assimilation systems in a short-term forecasting and possibly re-analysis mode, providing the assimilation variables are available (GEM-Hydro). In either case, the framework for the two MESH systems is essentially the same and differs in the model vertical energy and water budget components and the ability to assimilate land-surface information. The stand-alone version is designed to run in an offline mode however, and since it is based on operational code, can be run experimentally in a coupled mode with the GEM atmospheric system, WRF, reanalysis or station data or from an RCM framework. It is important to note that other land-surface schemes/hydrological models could potentially be incorporated to the MESH system.

Underlying these schemes are a within grid routing model (WATROF) that allows for the timing of water transfer to the local stream within the model landscape unit.  At this point, it is assumed that all flow from the GRU is summed up at the grid level and then routed downstream using the numeric flow routing techniques between basin grid elements (WATROUTE).  A variation of WATROF that attempts to conceptualize the non-contributing areas surface routing phenomenon that is particularly acute on the Prairies is PDMROF, which is also available for routing at the grid scale in within the system.  As a hydrology modelling system, MESH captures many of the important land-surface processes necessary for cold-regions simulation and Canadian Hydrological realities; however it does have some limitations and makes important simplifying assumptions.  These are necessary for two main reasons, maintain computational efficiency in large basin simulation and maintain model parsimony by limiting the degrees of freedom and model parameters.    MESH uses a GRU approach to parameter identification in the 1-D surface and water budget model.  All these 1-D model parameters can be prescribed or obtained through optimization and there are many examples of using a priori parameter estimates and also of calibrating important parameters to enhance model performance.  The Ostrich platform is coded in to the MESH system to allow for parameter optimization and this can be done using many different objective functions, including of course, minimize observed and simulated errors in streamflow. 

  • See the CCRN MESH Update for details on the advancements made by CCRN over the course of the network

 

 

Cold Regions Hydrological Modelling (CRHM)

The Cold Regions Hydrological Modelling (CRHM) platform is developed at the Centre for Hydrology at the University of Saskatchewan (Pomeroy et al., 2007; 2016). As its name implies, CRHM has been developed explicitly for modelling cold regions, being based on over 50 years of research on Canadian hydrology. CRHM is semi-distributed, based on HRUs which can also be grouped to permit application to larger areas. Unlike other models, CRHM does not require there to be a stream in a modelled basin, making it uniquely suited for modelling Prairie and Northern basins. CRHM is modular, in that it contains different modules for each hydrological process. These modules are selected and combined to form a model for a given location.
CRHM has modules to compute the following processes:

  • Blowing and avalanching snow, including the erosion, deposition, redistribution and sublimation of snow
  • Interception of rain and snow by plant canopies, including forests
  • Energy balance snowmelt for open snowfields, slopes and under plant canopies
  • Infiltration to frozen and unfrozen soils including macropore flow
  • Energy budget evaporation and evapotranspiration
  • Soil moisture storage and drainage to groundwater and phase change from freezing and thawing soils including permafrost
  • Depressional storage and variable contributing areas for runoff generation
  • Runoff from rainfall and snowmelt as saturation overland flow, infiltration excess overland flow, detention flow through snowpacks and organic layers, shallow subsurface flow through unsaturated and saturated porous media and saturated groundwater flow from field to river basin scales using linear and spatially distributed routing functions
  • Groundwater flow in directions that differ from surface water flow
  • Glacier firnification, firn melt, ice melt, mass balance, ice flow and routing of surface and sub-glacial storage and runoff.

The hysteretic change in connected fractions of Prairie basins due to the “fill and spill” of depressional storage has been added to CRHM by incorporating the Pothole Cascade Model (PCM), which is described in (Shook and Pomeroy, 2011). Use of PCM in was shown to dramatically improve the performance of CRHM in modelling Prairie basins (Pomeroy et al., 2012, 2014).

CRHM is has strongly physically based modules that describe its surface processes, meaning that its parameters are values that can be measured or estimated using remote sensing/GIS. CRHM does not allow for any form of calibration of model parameters. The exceptions to this rule are the groundwater parameters, which cannot be measured directly, as groundwater is modelled conceptually. Fortunately, the groundwater parameters only change over geological time scales. Therefore, hydrological models developed using CRHM can be used under nonstationary conditions, such as produced by changes in climate, agriculture and drainage.

 

Canadian Hydrological Model (CHM)

The Canadian Hydrological Model (CHM) is a modular, multi-physics, spatially distributed modelling framework designed for representing hydrological processes, including those that operate in cold-regions (Marsh et al., submitted). Unstructured meshes permit variable spatial resolution, allowing coarse resolutions at low spatial variability and fine resolutions as required. Model uncertainty is reduced by lessening the necessary computational elements relative to high-resolution rasters.

CHM is in the later stages of development at the UofS Centre for Hydrology by Chris Marsh and provides a novel tool for examining cold region hydrological systems. Key features include efficient terrain representation, allowing simulations at various spatial scales, reduced computational overhead, and a modular process representation allowing for an alternative-hypothesis framework. Using both physics-based and conceptual process representations sourced from the current cold regions literature allows for comparison of process representations and importantly, their ability to produce emergent behaviours.

CHM uses a novel multi-objective approach for unstructured triangular mesh generation that fulfills hydrologically important constraints (e.g., basin boundaries, water bodies, soil classification, land cover, elevation, and slope/aspect). This provides an efficient spatial representation of parameters and initial conditions, as well as well-formed and well-graded triangles (TINs) that are suitable for numerical discretization. CHM uses high-quality open source libraries and high performance computing paradigms to provide a framework that allows for integrating current state-of-the-art process algorithms. The impact of changes to model structure, including individual algorithms, parameters, initial conditions, driving meteorology, and spatial/temporal discretization can be easily evaluated.

The CHM meshing algorithm reduces the total number of computational elements and preserves the spatial heterogeneity of predictions.  It is envisioned that advances in CHM will make it attractive as an alternative hydrological land surface scheme core for future developments of MESH where explicit representations of surface and subsurface hydrology warrant the additional complexity.