Meteorological forecasting

Meteorological forecasting

Customised meteorological forecast service

We offer high-resolution (up to 366 m), customisable meteorological forecasts, that offer reliable forecasts up to 7 days in advance. This service is based on the Weather Research and Forecasting (WRF) model, a dynamical model that constitutes the state-of-the-art in regional modelling of the atmosphere.

The complexity of the model allows it to take into account local topography and land use, to reproduce and predict localised phenomena, like urban heat islands. It also allows to analyse intra-urban differences and trends across many variables, both general (temperature, precipitation or humidity) and sector-specific (solar radiation levels, apparent temperature and many more).

We work hand in hand with the Santander Meteorology Group to develop the model further, fine-tuning its multiple applications to achieve an optimal model configuration. This research group has extensive experience with the model, having participated in multiple projects of national and international scope.

Key features

High resolution (~366m)
Ensemble forecasting
Uncertainty quantification
Statistical post-processing with machine learning techniques for increased accuracy

Detailed description

Our forecast service covers any location worldwide and provides outputs for one or more domains or regions. To optimise the results and ensure a higher accuracy, we offer a tailored configuration of the model (i.e. parameterisation choice) that depends on the problem addressed.

We offer forecasts up to 7 days in advance, although the forecasting time-window depends highly on the resolution. For the higher resolutions, we provide reliable forecasts up to 48 hours in advance, due to their higher computational cost.

Based on a dynamical model that reproduces the physics of the atmosphere, our forecast service is able to provide predictions for a myriad of variables: temperature at surface, precipitation, wind, humidity, insolation and sector-specific climate indexes.

As the variables vary wildly across sectors, this is not an exhaustive list. If you don’t find the index you are looking for, drop us an e-mail!

Depending on your needs, the outputs can be offered in several formats:

  • Ready-to-use dynamic meteograms, integrable in websites
  • Gridded forecast variables in the native grid or interpolated to a customized geographical projection, in NetCDF format.
  • Text files in several formats (JSON, XML) with forecasts in a set of locations or regions of interest.
  • Standard mapping service (WMS) using our visualization system datlas.
  • Smartphone app.
  • Issuing of alerts (as Early Warning System) through different channels (e.g. mailing lists, SMS), based on customized thresholds relevant for the user.

Swift & seamless set-up

To set up the product, we can use two different approaches:

  • Use tailored datasets: if the client has historical observations of the region of interest, we can adapt those data to feed the model and set it up, providing an increased accuracy.
  • Use historical data: we have access to climate datasets of historical observations coming from trusted sources such as Copernicus (the European Union's Earth observation programme) or the Earth System Grid Federation. They enable us to manage large datasets such as ERA5, that provides hourly estimates of a large number of atmospheric, land and oceanic climate variables, covering from 1979 to within 5 days of real time.

Added Value Services

The WRF model can be used to complement the forecasts, to provide additional features:

  • Retrospective forecasts: reproduce past past periods, to assess the model skill in a given domain.
  • Explore the local climate: to search for favourable locations or compute return periods.
  • Multi-physics forecast ensembles: by providing an range of plausible forecasts instead of a single forecast, we are able to quantify the uncertainty of the model.
  • Calibration system: based on Kalman Filters using real time observations.
  • Use of past observations: we can assimilate the observed data into the model to increase its accuracy.


The following is our pricing scheme (taxes not included):

Resolution 1 900€ / month
Size: 240Km x 240Km
Resolution: 3.3Km
Nested to one of the daily GFS outputs
72 forecast hours for up to 8 model variables
Historical year simulation for 600€/simulated month
Resolution 2 1500 € / month
Size: 80Km x 80Km
Resolution: 1.1Km
Nested to one of the daily GFS outputs
72 forecast hours for up to 8 model variables
Historical year simulation for 1000€/simulated month
After the second year a 20% discount is applicable. Customized services such as data assimilation, calibration and multi-physics are budgeted case by case.

Count with us for your next project! Contact us to get more information: