The forecasting, storage and delivery architecture at the core of the OpenWeather Platform. Numerical annotations are reproduced from the source brochure.
Global data collection
The Platform brings together complementary sources of weather information. Observations describe conditions as they develop, while forecasting models provide a broader view of how the atmosphere may change.
Weather observations
Weather stations and connected sensors provide local measurements of conditions such as temperature, humidity, pressure and wind. Radar helps track precipitation, while meteorological satellites provide a wider view of clouds and atmospheric conditions. Together, these sources support current weather information and forecasts for the coming hours.
Global and regional models
Outputs from global and regional forecasting systems contribute to forecasts over longer periods. OpenWeather combines these inputs with observations and its own processing methods. Historical and reanalysis data provide additional information for understanding past conditions and developing the forecasting system.
Explore our data sources
Data processing and OWHL forecasting
Data arrives with different formats, resolutions and update cycles. Our processing systems prepare and combine these inputs so that information from different sources can be used together. OpenWeather algorithms and machine learning help correct biases, incorporate new observations and refine predictions for particular locations.
Alongside data prepared in advance, the Platform performs calculations on demand to provide the weather parameters required for particular locations.
The OWHL™ OpenWeather Hyper-Local Forecasting Model brings these capabilities together through complementary forecasting components.
Nowcasting and short range forecasting
The nowcasting component uses recent station, radar and satellite information to follow rapidly changing conditions and produce forecasts for the coming hours. Frequent updates allow it to incorporate new observations as weather systems develop.
Medium and long range forecasting
For longer forecast periods, OpenWeather combines outputs from global forecasting systems with its own processing and machine learning. These components support forecasts over the coming days and longer outlooks for planning and analysis.
Bringing forecasts together
The fusion stage combines information from the forecasting components to provide a unified view of the weather. Forecast data also feeds the historical archive, supporting later analysis and further development of the models.
Explore the OWHL forecasting model
Quality control and verification
Quality control forms part of data preparation and processing. We check incoming information for anomalies and inconsistencies and apply methods to improve the data used in forecasting.
We also evaluate forecast performance against reference observations. Meteorological station measurements support checks of weather parameters, while radar provides a reference for precipitation. Statistical and diagnostic measures help compare source data, assess processing methods and identify systematic errors.
The results inform continuing improvements to our models and algorithms. Verification connects the forecasts we produce with the conditions subsequently observed.
How we assess data quality
Petabyte scale storage with Deker
We developed Deker™ to store and manage extensive weather archives at petabyte scale. Weather information spans many locations, time periods and parameters. Deker organises these dimensions so that the Platform can retrieve the specific data required for a calculation, an analysis or a customer request.
Deker supports parallel processing and simultaneous reading and writing of data. This allows the Platform to work with large volumes of information while continuing to incorporate new data and maintain its archives.
Stored historical weather and archived forecasts support retrospective analysis, forecast evaluation and model development. They also provide the foundation for datasets that help customers investigate past conditions and assess weather patterns over time.
Global data delivery
Our delivery infrastructure makes weather data available to millions of users worldwide. Distributed data centres, API servers and caching systems support responsive access and allow the service to scale with demand.
Customers can request data through APIs, download prepared bulk datasets or arrange tailored data feeds. These delivery methods bring information from the Platform into operational systems, analytical tools and customer applications.
Datasets from the OpenWeather Platform
The Platform supports a broad range of datasets. Some contain information generated through OpenWeather modelling and processing; others incorporate information from external sources, such as official weather warnings.
- Current and forecast weather: current conditions and forecasts across different time horizons.
- Historical weather: past weather data and archived forecasts for analysis and evaluation.
- Weather maps: geographical layers showing weather conditions, forecasts and historical information.
- Alerts and risk data: official warnings, weather alerts and specialist indicators such as fire weather risk.
- Energy and renewables: solar irradiance, wind information and renewable generation datasets.
- Road weather: weather information and risks relevant to routes and transport operations.
- Environmental and climate data: air quality information, climate indices and related datasets.
- Bulk weather datasets: prepared files for standard or agreed locations and data requirements.
Each dataset has a defined scope. Coverage, parameters, resolution, time intervals, update frequency and delivery options depend on the dataset selected.
Explore datasets
Choose data for your organisation
Choose the datasets that meet your requirements and combine them through OpenWeather Enterprise. The Base Licence provides a defined package, with options to add datasets, increase capacity and customise formats and delivery.
Your agreement sets out the included data, usage allowances and service arrangements, giving your organisation a clear basis for integration and ongoing use.
Explore Enterprise
