Meteorology and Data Mining

Modern meteorology deals not only with the study of the physics and nonlinear dynamics of the atmosphere, but also with the efficient statistical analysis of the huge amounts of information resulting from the numerical simulations performed both in operational and hindcast (retrospective) modes. Although meteorology was originally concerned with short-range prediction (1-3 days), this discipline has greatly evolved and nowadays the forecast horizons expand to seasonal time scales (predicting climate anomalies months in advance) and climate change scenarios (decades); in this case, the atmospheric models are coupled with oceanic ones, with slower dynamics, including also the effects of the cryosphere, biosphere (mainly land use), etc.

Our research is focused both in theoretical and applied topics of this discipline, with special emphasis in those aspects involved in seasonal forecasting and climate change scenarios. Our interests include the study of simplified atmospheric models (Lorenz-like models, barotropic vorticity model, etc.) to analyze theoretical aspects of predictability and ensemble forecasting in nonlinear spatiotemporal systems, and the numerical simulation of weather and climate at regional scale, using regional atmospheric models such as the open-source WRF model (Weather Research and Forecasting, www.wrf-model.org). We are also involved in the statistical analysis and postprocessing of the operational prediction systems, including studies of predictability/skill assessment and downscaling (local prediction). These activities include the development of efficient data mining techniques and their implementation in web portals as applications and services, so they can be used to provide callibrated/local observations and predictions to different users from impact sectors (agriculture, hydrology, energy, etc.).

http://www.meteo.unican.es

Group Projects