Browsing by Author "Rodrigues, Carlos Veiga"
Now showing 1 - 9 of 9
Results Per Page
Sort Options
- An exploratory study on hindcasting with analogue ensembles of principal componentsPublication . Balsa, Carlos; Breve, Murilo Montanini; Rodrigues, Carlos Veiga; Costa, Luís S.; Rufino, JoséThe aim of this study is the reconstruction of meteorological data that are missing in a given station by means of the data from neighbouring stations. To achieve this, the Analogue Ensemble (AnEn) method was applied to the Principal Components (PCs) of the time series dataset, computed via Principal Component Analysis. This combination allows exploring the possibility of reducing the number of meteorological variables used in the reconstruction. The proposed technique is greatly influenced by the choice of the number of PCs used in the data reconstruction. The number of favorable PC varies according to the predicted variable and weather station. This choice is directly linked to the variables correlation. The application of AnEn using PCs leads to improvements of 8% to 21% in the RMSE of wind speed.
- Cluster-based analogue ensembles for hindcasting with multistationsPublication . Balsa, Carlos; Rodrigues, Carlos Veiga; Araújo, Leonardo Oliveira; Rufino, JoséThe Analogue Ensemble (AnEn) method enables the reconstruction of meteorological observations or deterministic predictions for a certain variable and station by using data from the same station or from other nearby stations. However, depending on the dimension and granularity of the historical datasets used for the reconstruction, this method may be computationally very demanding even if parallelization is used. In this work, the classical AnEn method is modified so that analogues are determined using K-means clustering. The proposed combined approach allows the use of several predictors in a dependent or independent way. As a result of the flexibility and adaptability of this new approach, it is necessary to define several parameters and algorithmic options. The effects of the critical parameters and main options were tested on a large dataset from real-world meteorological stations. The results show that adequate monitoring and tuning of the new method allows for a considerable improvement of the computational performance of the reconstruction task while keeping the accuracy of the results. Compared to the classical AnEn method, the proposed variant is at least 15-times faster when processing is serial. Both approaches benefit from parallel processing, with the K-means variant also being always faster than the classic method under that execution regime (albeit its performance advantage diminishes as more CPU threads are used).
- Hindcasting with cluster-based analoguesPublication . Balsa, Carlos; Rodrigues, Carlos Veiga; Araújo, Leonardo Oliveira; Rufino, JoséThe reconstruction of meteorological observations or deterministic predictions for a certain variable and station may be performed with data from other variables at that station, or from other nearby stations. This is a hindcasting problem, known from some time to be solvable using the Analogues Ensemble (AnEn) method. However, depending on the dimension and granularity of the datasets used for the reconstruction, this method may be computationally very demanding, even if parallelization is used. In this paper, the AnEn method is combined with K-means clustering, allowing for a considerable acceleration of the reconstruction task, while keeping the accuracy of the results.
- Hindcasting with multistations using analog ensemblesPublication . Chesneau, Alexandre; Balsa, Carlos; Rodrigues, Carlos Veiga; Lopes, Isabel MariaA hindcast with multiple stations was performed with vari- ous Analog Ensembles (AnEn) algorithms. The different strategies were analyzed and benchmarked in order to improve the prediction. The un- derlying problem consists in making weather predictions for a location where no data is available, using meteorological time series from nearby stations. Various methods are explored, from the basic one, originally de-scribed by Monache and co-workers, to methods using cosine similarity, normalization, and K-means clustering. Best results were obtained with the K-means metric, wielding between 3% and 30% of lower quadratic error when compared against the Monache metric. Increasing the predictors to two stations improved the performance of the hindcast, leading up to 16% of lower error, depending on the correlation between the predictor stations.
- Parametric study of the analog ensembles algorithm with clustering methods for hindcasting with multistationsPublication . Araújo, Leonardo Oliveira; Balsa, Carlos; Rodrigues, Carlos Veiga; Rufino, JoséWeather prediction for locations without or scarce meteorological data available can be attempted by taking meteorological datasets from nearby stations. This hindcasting problem can be successfully solved using the Analog Ensemble (AnEn) method. This paper presents a parametric analysis of the AnEn method, and two variations (based on K-means and fuzzy C-means clustering methods), when used to search for analog ensembles in a historical dataset. The study allowed to identify the parameter combinations that yield the best prediction accuracy, improving 13% on the systematic error and 5% on the random error of the previous results obtained with the same dataset. In addition, important performance gains were achieved at the computational level.
- PCAnEn - hindcasting with analogue ensembles of principal componentsPublication . Balsa, Carlos; Breve, Murilo Montanini; André, Baptiste; Rodrigues, Carlos Veiga; Rufino, JoséThe focus of this study is the reconstruction of missing meteorological data at a station based on data from neighboring stations. To that end, the Principal Components Analysis (PCA) method was applied to the Analogue Ensemble (AnEn) method to reduce the data dimensionality. The proposed technique is greatly influenced by the choice of stations according to proximity and correlation to the predicted one. PCA associated with AnEn decreased the errors in the prediction of some meteorological variables by 30% and, at the same time, decreased the prediction time by 48%. It was also verified that our implementation of this methodology in MATLAB is around two times faster than in R.
- Reconstruction of meteorological records by methods based on dimension reduction of the predictor datasetPublication . Balsa, Carlos; Breve, Murilo Montanini; Rodrigues, Carlos Veiga; Rufino, JoséThe reconstruction or prediction of meteorological records through the Analog Ensemble (AnEn) method is very efficient when the number of predictor time series is small. Thus, in order to take advantage of the richness and diversity of information contained in a large number of predictors, it is necessary to reduce their dimensions. This study presents methods to accomplish such reduction, allowing the use of a high number of predictor variables. In particular, the techniques of Principal Component Analysis (PCA) and Partial Least Squares (PLS) are used to reduce the dimension of the predictor dataset without loss of essential information. The combination of the AnEn and PLS techniques results in a very efficient hybrid method (PLSAnEn) for reconstructing or forecasting unstable meteorological variables, such as wind speed. This hybrid method is computationally demanding but its performance can be improved via parallelization or the introduction of variants in which all possible analogs are previously clustered. The multivariate linear regression methods used on the new variables resulting from the PCA or PLS techniques also proved to be efficient, especially for the prediction of meteorological variables without local oscillations, such as the pressure.
- Simulação do escoamento atmosférico sobre topografia urbana idealizada através do OpenFOAMPublication . Bittencourt, Bruno; Balsa, Carlos; Rodrigues, Carlos VeigaOs modelos de Dinâmica dos Fluidos Computacional (CFD) em larga escala não conseguem representar em detalhe o escoamento atmosférico em uma cidade ou ambiente urbano. Quando a resolução numa malha computacional é maior ou equivalente à escala característica dos obstáculos urbanos, estes constituem perturbações de submalha. Uma forma de contornar este problema consiste em parameterizar o efeito do tecido urbano no escoamento através da inclusão do arrasto adicional na simulação de larga escala. Tal pode ser feito aumentando artificialmente a rugosidade aerodinâmica do terreno (z0) e deslocando a altura do nível do solo através de uma altura de deslocamento (d). Este trabalho simula, com o software de código aberto OpenFOAM, o escoamento da camada limite atmosférica sobre um domínio computacional que consiste numa matriz quadrada de cubos, adotando o modelo de turbulência k-ω SST. A independência da malha foi verificada. O arrasto de pressão e de fricção foram quantificados e, então, definido o coeficiente de arrasto urbano (Cd). Foram verificados os efeitos de um baixo número de Reynolds nos perfis de velocidade e no valor do coeficiente de atrito, que se mostraram irrelevantes, também foram calculados a rugosidade aerodinâmica do terreno e a altura de deslocamento. Os resultados foram obtidos para uma camada limite atmosférica neutra e hidrodinâmicamente estável e foram comparados com dados da bibliografia e com dados experimentais em túnel de vento, mostrando boa concordância.
- Using analog ensembles with alternative metrics for hindcasting with multistationsPublication . Balsa, Carlos; Rodrigues, Carlos Veiga; Lopes, Isabel Maria; Rufino, JoséThis study concerns making weather predictions for a location where no data is available, using meteorological datasets from nearby stations. The hindcast with multiple stations is performed with different variants of the Analog Ensemble (AnEn) method. In addition to the traditional Monache metric used to identify analogs in datasets from one or two stations, several new metrics are explored, namely cosine similarity, normalization, and k-means clustering. These were analyzed and benchmarked to find the ones that bring improvements. The best results were obtained with the k-means metric, yielding between 3% and 30% of lower quadratic error when compared against the Monache metric. Also, by making the predictors to include two stations, the performance of the hindcast improved, decreasing the error up to 16%, depending on the correlation between the predictor stations.