modelizacion_2_web.png Modeling 

 

Having good observations is crucial for understanding the atmospheric phenomena. But very often, is needed to generalize the point measurements to a wider temporal or spatial domain using numerical modeling.

interMET has experience on both the two main families of simulation models:

This models build mathematical expressions that try to simulate the observed relationships between the predictors and the predictand using a long enough time series. The main advantage of these techniques is that they can simulate very complex physical phenomena implicitly, are relatively easy to develop and have short execution times, what make ideal for simulating climatologies. On the other hand, they are sometimes black boxes that predict phenomena but do not help on understanding them, and a long time series is needed for training the model and validating.
These model simulate through mathematics the most important physical laws forcing the phenomena to be simulated. They do it in a discrete mathematical world, so there are aspects such domain and resolution that need to be carefully selected. They do not require very long time series for acceptable results but are complex to run and tune properly. They take more time to run than empirical models but give results in a spatial domain, usually in three dimensions.

Numerical simulation has experienced in the last years an spectacular evolution thanks to the revolution experienced by computers. Nowadays medium size computers are able to run complex simulation tools, some of them open source and with a community that make the use of these tools not very difficult for a company like interMET.

Our personnel have actively participated in numerical simulation projects giving support on many aspects like pre-procesing of input data, scenario design, comparison with observations, algorithms development, statistical analysis, running code and helping in the interpretation of the results.

Do you have a numerical simulation project?

If you have a numerical simulation project we would be willing to help you in:

  • Homogenization and validation of temporal series for model development.
  • Processing of high volume of data coming from other models (MERRA, ECMWF, GFS).
  • Development of forecasting tools based on statistical or physical models.
  • Development of very specific models for your application like renewable energies, pollution dispersion, weather prediction, hydrology, etc (XOQDOQ, CALMET, WindSim, WAsP, WRF, LOPM).
  • Skill Analysis of predictions, model intercomparisons, ensembles, etc.

If you have a numerical simulation project we would be willing to help you.

interMET Services

Ask Us