Very often observations hide their relation with the phenomena under investigation very well. In order to extract and detect this relationships is necessary to perform statistical analysis of the data. Correct analysis is necessary to get significant and robust conclusions with statistical significance.
Before any data analysis on observations it is necessary to perform some kind of quality check, often called "data validation". This pre-processing of data is crucial even for networks with expected very high reliability. This is due because there are many factors that can interfere the measuring process leading, if not detected, to wrong conclusions. Also sensors and accessories drift their response to the observed phenomena after many years in the field.
interMET personnel have many years of experience validating meteorological data. At the beginning this was mainly done through visual inspection, but now we have tools that help the meteorologist to pay attention to potentially wrong data because they are incoherent with climatology, in relation with other variables or close stations.interMET has expertise in many statistical techniques applied to environmental data such:
- Validation: climatological coherence, internal coherence tests, spatial coherence tests, spatial coherency tests, etc.
- Temporal series analysis: regression analysis, univariate and multivariate model fitting, autocorrelation analysis, spectral analysis, stochastic models, ARMA and ARIMA models, Markov models, bayesian models, etc.
- Statistical significance tests.
- Analysis of not linear processes: neural networks, fuzzy logic, etc.
- Principal component analysis, clustering, k-means.
We have a solid background in statistics and many years of experience with meteorological data analysis. We have some scientific publications that make use of these and other techniques and we offer you our skills to analyze your data.
With the first station taking first measurements on 1998 at 2080 m height, this network was during a long time a pioneer in alpine automatic measurments. This project has been for interMET not just a project, it has been our workbench that made our human and technical skills to be taken almost to the limit.
Met masts with mechanical anemometers and wind vanes have being the standard for wind power assessment during many years. This measuring principle relies on the conversion of the wind energy into mechanical energy that anemometers and wind vanes convert to an electric signal proportional to wind speed or wind direction. Like any other measuring technique, anemometers and wind vanes have their advantages and disadvantages. One of the limitations of anemometers and wind vanes is that that they need a solid structure to be held at a certain height, the other is that they take samples of wind almost in a single point. In the last decades, size of wind turbines have grown considerably. With this increment on rotor height, a more precise assessment of the wind profile is then necessary in order to make a better production forecasting. It is possible to use tall towers with sensors all their way up to 100 meters or more, but the increase of the costs of such towers increase considerably. This fact along with others like reliability, installation and maintenance costs and environmental impact make necessary to consider other measurement techniques that do not need towers. One option is to use a SODAR (SOund Detection And Ranging). They use the heterogeneities found on air to reflect ultrasonic sound pulses and retrieve their velocity using the Doppler effect theory. SODARs are at ground level but can reach very high levels of the atmosphere, depending on their configuration. SODARs specially focused on the first hundreds of meters of the atmosphere have become in the last years very helpful on the wind energy resource area.