In our personal and professional lives, we are constantly battling the lack of two key resources: time and money. NHMSs around the world are no exception — the prohibitively high costs of deploying and maintaining the on-premises HPC platforms required for running local Numerical Weather Prediction (NWP) Models, and the timeframes needed to deploy and maintain such platforms even if/when budgets are obtained, are an exercise in meteorological frustration. What’s more, achieving high SLAs for any NWP platform requires highly-specialized engineering capabilities often not available to NHMSs, especially in emerging economies. The result? Significant compromises on the accuracy and timeliness of forecasts in regions that are often particularly weather sensitive.
CBAM, Tomorrow.io’s recently-launched NWP platform, provides a new way to look at — and solve — this problem.
CBAM, Tomorrow.io’s Bespoke Atmospheric Model, enables NHMSs to deploy customized, country-specific Operational NWP on the cloud. With CBAM, NHMSs can customize their own forecast models to fulfill their missions, and provide highly accurate forecasts to their governments and citizens. This new level of accuracy and reliability is achieved on the cloud, circumventing the costs and bureaucracy of deploying on-premises platforms.
CBAM can be configured to run at any temporal and spatial resolution desired. When needed, it can be nested within the best available global or regional models (e.g. ECMWF) and assimilate all available observations, including those deployed by the NHMSs themselves, as well as private and public observation networks. CBAM also enables NHMSs to expand their observation network by ingesting Tomorrow.io’s proprietary observations, derived from its “virtual sensing” technologies, such as microwave signal attenuation, connected vehicles, mobile devices and more. In addition, Tomorrow.io’s proprietary “dynamic nudging” technique eliminates the drifting of the forecast when running only several times per day.
CBAM is completely customizable. Tomorrow.io’s team of world-leading scientists and engineers works with NHMSs to design the ultimate configuration for each — geographical domain, forecasting horizon, temporal and spatial resolution, weather parameters of interest — and allows for quick iterations without paying for unnecessary compute power or being constrained by local hardware capabilities.
With the paradigm shift to a cloud-based NWP platform, meteorological services can change their own focus: Rather than seeking CapEx budgets, which are hard to come by, allocating manpower over several years, and locking the organization into a single type of deployment for multiple years, NWP-as-a-Services is an OpEx expenditure, and benefits from the rapid deployment and ongoing upgrades and updates made possible by the cloud. As a result, NHMSs are freed to do their actual job: providing better weather forecasting for their communities. With NWP-as-a-Service, NHMSs are able to not only maximize those scarce resources — time and money — but to do so while getting cutting-edge technologies that are typically not available to the public sector.
It’s the rare win-win: technology that reduces costs, vastly improves performance, and frees specialists to do what they do best, with huge ramifications for the safety, financial success, and well being of the regions that are dependent on their forecasts.