Meteomatics weather API delivers fast, direct, simple access to an extensive range of global weather and environmental data.
Our weather API was originally created for our own internal use to improve weather data access and build solutions for industry.
We soon recognised the power of the capability we had created and decided to make our weather API available to external users. Enabling industry, academia, and research, to innovate and apply weather data to their particular use case.
1) Underpinned by solid foundations
The Meteomatics API is underpinned by our Meteocache that aligns data in time and space to ensure that the weather API can efficiently and very quickly, return the requested data query in spite of the size and format of the original data set.
We also hold a significant amount of data in memory which allows for a highly robust API that gives industry-leading performance.
2) Single source of weather data
Users are offered a single API endpoint to access weather and earth data covering the globe, from weather forecasts, including; nowcast, forecast, seasonal, climate and space! Historical data (back to 1979 dependent on location). Observational data sets such as satellite, station data, weather drones, and observations from the Internet of Things. Plus, all our web feature services (WFS’s) and web mapping services (WMS’s) can be accessed using the same weather API.
Simplifying access to weather data reduces the complexity of access commonly associated with meteorological information. Plus, it makes it possible for businesses to easily integrate our single weather API right across their workforce management, planning, business insight, trading, and decision-making systems.
3) On the fly calculation for most up-to-date forecasts
The Meteocache holds lots of its data in memory, which allows the API to calculate a user's request using the latest available observations, what we call calculating ‘on the fly’.
Calculating on the fly helps mitigate the problem of querying the last available run of a numerical weather prediction model.
Numerical weather models typically take a long time to produce their results (the fastest NWP models are run every hour). Sometimes models may not capture a particular event and the model cannot correct or adjust its output until the next model run.
This ‘on the fly’ approach gives users confidence that they are receiving the most accurate and up to date weather data possible.
Meteomatics continuously ingests new observational data, with some datasets updating at 1-minute intervals.
Calculating on the fly can also minimize the need for manual intervention and allow customers to automate decision making, achieve greater consistency across their products and services.
4) Downscaling weather forecasts to 90 metres.
We improve the accuracy of weather data by taking into account the detail of the topography.
As numerical weather prediction models typically use a coarse model resolution to ensure it can create a weather forecast on a global scale within the compute capability available, which has a few disadvantages.
Firstly, the model grid points are for a fixed area, (for example ECMWF project their model onto 8-10km grid points - at the equator), which can result in atmospheric events not being included in the model if they occur over areas smaller than the grid points.
Also, coarser model resolutions do not sufficiently capture the detail of the underlying topography which makes it difficult to forecast in areas that are not flat and become even more problematic in mountainous terrain.
Plus, physical parameterizations are based on assumptions for horizontally homogeneous and flat terrain, by assuming that the surface is flat and not taking into account the friction created by a landmass as air passes over (impacting precipitation forecasts for example).
So we decided to incorporate NASA’s highest-resolution topographic data set, which is at a resolution of 90m and mapped it to the neighbouring cells in the model data we ingest. Allowing the API to accurately calculate a weather forecast by taking into account the topography that represents ground truth.
Plus, we created an interpolation scheme that allows for the forecast of the actual height of the query specified in the lat / long, giving us the ability to generate the forecast for the height a user is interested in.