A new Pan-Canadian Sub-km external surface prediction system to improve fluxes over ecosystems, cities, and water surfaces in numerical weather prediction systems
The HRDSPS consists of an external land SURFace model, called GEM-Surf, and a land-surface-characteristics generator, called PreX . The GEM-Surf model is based on the operational EC model (the Global Environment Multi-scale (GEM) model) with the important distinction that GEM-Surf only integrates land and near surface processes, with all atmospheric forcing provided by observations, atmospheric analyses and/or forecasts.
The low computational cost of GEM-Surf compared to that of a fully three-dimensional Numerical Weather Prediction (NWP), allows Gem-Surf to be run at the resolution of the most detailed surface characteristics databases available. PreX is a state-of-the-art surface processor developed specifically to provide a better estimate of urban and non-urban land surface characteristics needed for the model, based on high-resolution datasets available over Canada.
In this study, we present the methodology behind the state-of-the-art surface processor, the improvements done to the land surface schemes integrated in the HRDSPS as well as preliminary results of the HRDSPS over an experimental grid over Southern Ontario, Canada. In the experimental set-up, which mimics the planned operational set -up, the HRDSPS was run in a continuous cycle mode, forced with best estimates of atmospheric forcing (precipitation from the Canadian Precipitation Analysis CaPA, near-surface air conditions from EC's screen-level analyses, and radiative forcing from short-range GEM forecasts) and surface conditions (PreX).