Adaptation Of The Canadian Land Data Assimilation System to Urban Areas

Recent advances in surface and hydrological forecasting at Environment and Climate Change Canada (ECCC) were achieved through the more precise initialization of the surface state with the Canadian Land Data Assimilation System (CaLDAS). More recently, a satellite-based approach has been developed (Carrera et al. 2015, 2019; Bélair et al. 2023) and already used for daily operation of the ECCC’s National Surface and River Prediction System. A limitation of this system is that it doesn’t consider adequate representation of urban areas, albeit urban areas represent a tiny portion of the Canadian landscape. 

Such satellite-based surface data assimilation system is likely to be used also in the next generation of short-term weather prediction for a domain covering Canada and the northern part of USA. In this NWP system, detailed physical processes in the urban canopy are represented with the Town Energy Balance TEB urban scheme and the Soil, Vegetation and Snow scheme SVS. In this context, it is important to adapt the land surface data assimilation system to improve coherency and accuracy in urban areas.  

This study aims to show the impact of the different steps required towards a more realistic representation of urban areas in CaLDAS. The TEB scheme is added in background state in addition to SVS. Then, the surface canopy temperatures are added in the analyzed variables. Preliminary results for 2022 summer show a global improvement of 3-h forecasts of 2-m air temperature, reducing bias up to 0.5°C over large cities. Impacts of the new method on the soil moisture and snow prescription are also investigated. Implementation of such system could greatly improve the initial conditions used by surface-atmosphere coupled systems for weather and environmental forecasts and analysis in urban areas.