About MESH

MESH (Modélisation Environnementale communautaire - Surface Hydrology) is the hydrology land-surface scheme (HLSS) of Environment and Climate Change Canada's (ECCC's) community environmental modelling system (Pietroniro et al., 2007; Wheater et al., 2022), and is complimentary to ECCC's GEM-Hydro modelling platform. MESH is purposefully a bilingual acronym, demonstrating its Canadian origin.

MESH allows different hydrological land-surface and water routing and management models to co-exist within the same modelling framework so that they can easily be compared for the same experiment using exactly the same forcings, interpolation procedures, grid, time period, time-stepping and output specifications.

While containing physics from Numerical Weather Prediction (NWP) systems, MESH is able to reads atmospheric forcings from file instead of obtaining them from an atmospheric model, and supports many types of file structures and formats. This makes it possible to test incremental changes to NWP systems in an offline mode, and to drive the HLSS with forcing data from other sources such as direct observations or reanalysis products.

MESH has been used for advancing the representation of hydrological processes (notably for cold regions), for multi-decadal climate analyses, for flow forecasting, both in research and operationally, and as a testing platform to distribute some of ECCC's numerical models to its partners. Early stages and recent updates to the system are described in Pietroniro et al., 2007 and Wheater et al., 2022. As a conceptual framework for model development, it was initiated using different degrees of model coupling that ranged from loosely linked models, requiring separate calibration of the land-surface and hydrological components, to one with complete linkages between the atmospheric model, land-surface scheme and river routing model, including two-way coupling between the land-surface and atmosphere (Pietroniro, 2001; Voisin et al., 2002; Fassnacht and Soulis, 2002; Snelgrove et al., 2005; Seglenieks & Soulis, 2008; Yirdaw et al., 2009).

MESH was initially developed and demonstrated as “WATCLASS” at the University of Waterloo, coupling the Canadian LAnd Surface Scheme (CLASS, now CLASSIC) and the two-dimensional riverine routing scheme of WATFLOOD, using the latter’s approach for semi-distributed parameterization (i.e., the grouped response unit, or GRU approach). MESH has since evolved into a robust framework that supports many hydrological modelling applications.

The modelling framework

MESH was designed to be a flexible, accessible and portable modelling framework:

• Regularly tested for compatibility with open-source compilers

• Fully scalable from netbooks and laptops to multi-node workstations and high-performance supercomputers / clusters (HPC)

• Cross-compatible with Linux, Windows and MacOS

It also supports a number of options to test with ECCC’s land-surface systems, including:

• Support for deep soil profiles (e.g., permafrost under changing climate with CLASS/CLASSIC)

• Glacier dynamics

• Various options for reservoirs, including simple or complex fitted parameterization, multi-zone managed models and volume-stage-discharge tables

• Various options for water management, including district- or point-based irrigation, abstraction and diversion

• Cold regions hydrology, including snow redistribution (spatial)

• Statistics-based prairie pothole models for variably-contributing areas (i.e., prairie hydrology)

Open

Simple diagram of process discretization in MESH with an example of stacked options

How MESH compares with ECCC’s operation model

GEM-Hydro is ECCC’s operational flow forecasting model. Around 2007, it was named MEC (Modélisation Environnementale communautaire, not to be confused with ECCC’s Modèle Environmental Couplé), from which the name of MESH is also derived. In literature, it has also been described as GEM-Surf and the Surface Prediction System (SPS).

The primary distinction between the two systems is that GEM-Hydro maintains the ability to couple directly to the GEM (Global Environmental Multi-scale) model, ECCC’s 3D atmospheric modelling system which drives its NWP systems. In contrast, MESH is provided static atmospheric information from input files. However, compared to GEM-Hydro, MESH includes many experimental features, developed by the community, to develop and explore new representation for physical processes as well as feedbacks between various elements of the land-surface.

Open

Comparing ECCC's GEM-Hydro and MESH systems at the high level

Process representation and data flow

As a framework, MESH manages the translation of data and the exchange of information between activated options, including coordinating time-stepping between them.

Open

Options, models or algorithms are generally grouped into one of three categories: within-tile, between-tile and between grid. Tiles are sub-grid elements parameterized using grouped response units (GRUs). Not all HLSSs within MESH recognize this level of sub-grid discretization. Grids could also be considered subbasins (i.e., hydrological response units, HRUs, used in other models), and do not necessarily represent rectangular elements typical of many NWPs. Aside from the options for process representation, MESH also contains pre-processors and post-processors.

The most recent official version of MESH contains the following core routines:

Open

Core routines in the most recent official version of MESH

Parameter flexibility

MESH supports a variety of file formats and data structures, which permit the user to parameterize options using a variety of approaches.

Open

Parameterization options for the HLSS and routing components in MESH

Open

Core types of files supported by the most recent official version of MESH

MESH also supports mixing parameterization approaches and the use of different types of files in the same setup.

Modes for data delineation

MESH supports different modes for data delineation (i.e., basin or domain representation).

Open

Modes for data delineation supported by MESH

• Multi-point (1D) – A succession of single point runs, useful for the simulation of vertical fluxes for multiple spatially-disjointed Fluxnet sites

• Point (1D) – A single point run, useful for the simulation of vertical fluxes at a Fluxnet site

• Uniform Distributed Grid (2D) – Running over a domain of uniformly-sized grids, useful for emulating WATFLOOD domains and the mode WATCLASS and earlier versions of MESH used

• Hybrid Distributed Grid (2D) – Running over a multi-resolution domain where the HLSS runs at one resolution of uniformly-sized grids and the routing runs at another, useful for emulating GEM-Hydro

• Uniform Distributed HRU – Running over a domain of HRUs or subbasins derived from vector-based geospatial fabric, useful for emulating HRU-based models or for focusing on setups derived from hydrological datasets than using NWP-based information

Contact

For general inquiries or to contact MESH Support, please email MESH@ec.gc.ca.

At a glance

Quick links

Recently updated

Popular labels