Papers by Rosangela Cintra
arXiv (Cornell University), Jul 16, 2014
This paper presents an approach for employing an artificial neural network (NN) to emulate an ens... more This paper presents an approach for employing an artificial neural network (NN) to emulate an ensemble Kalman filter (EnKF) as a method of data assimilation. The assimilation methods are tested in the Simplified Parameterizations PrimitivE-Equation Dynamics (SPEEDY) model, an atmospheric general circulation model (AGCM), using synthetic observational data simulating localization of balloon soundings. For the data assimilation scheme, a supervised NN, the multilayer perceptron (MLP-NN), is applied. The MLP-NN is able to emulate the analysis from the local ensemble transform Kalman filter (LETKF). After the training process, the method using the MLP-NN is seen as a function of data assimilation. The NN was trained with data from first three months of 1982, 1983, and 1984. A hind-casting experiment for the 1985 data assimilation cycle using MLP-NN was performed with synthetic observations for January 1985. The numerical results demonstrate the effectiveness of the NN technique for atmospheric data assimilation. The results of the NN analyses are very close to the results from the LETKF analyses, the differences of the monthly average of absolute temperature analyses is of order 10 −2. The simulations show that the major advantage of using the MLP-NN is better computational performance, since the analyses have similar quality. The CPU-time cycle assimilation with MLP-NN is 90 times faster than cycle assimilation with LETKF for the numerical experiment.
2015 AGU Fall Meeting, Dec 18, 2015
InTech eBooks, Feb 28, 2018
Numerical weather prediction (NWP) uses atmospheric general circulation models (AGCMs) to predict... more Numerical weather prediction (NWP) uses atmospheric general circulation models (AGCMs) to predict weather based on current weather conditions. The process of entering observation data into mathematical model to generate the accurate initial conditions is called data assimilation (DA). It combines observations, forecasting, and filtering step. This paper presents an approach for employing artificial neural networks (NNs) to emulate the local ensemble transform Kalman filter (LETKF) as a method of data assimilation. This assimilation experiment tests the Simplified Parameterizations PrimitivE-Equation Dynamics (SPEEDY) model, an atmospheric general circulation model (AGCM), using synthetic observational data simulating localizations of meteorological balloons. For the data assimilation scheme, the supervised NN, the multilayer perceptrons (MLPs) networks are applied. After the training process, the method, forehead-calling MLP-DA, is seen as a function of data assimilation. The NNs were trained with data from first 3 months of 1982, 1983, and 1984. The experiment is performed for January 1985, one data assimilation cycle using MLP-DA with synthetic observations. The numerical results demonstrate the effectiveness of the NN technique for atmospheric data assimilation. The results of the NN analyses are very close to the results from the LETKF analyses, the differences of the monthly average of absolute temperature analyses are of order 10-2. The simulations show that the major advantage of using the MLP-DA is better computational performance, since the analyses have similar quality. The CPU-time cycle assimilation with MLP-DA analyses is 90 times faster than LETKF cycle assimilation with the mean analyses used to run the forecast experiment.
An Artificial Neural Network (ANN) is designed to investigate a application for data assimilation... more An Artificial Neural Network (ANN) is designed to investigate a application for data assimilation. This procedure provides an appropriated initial condition to the atmosphere to weather forecasting. Data assimilation is a method to insert observational information into a physical-mathematical model. The use of observations from the earth-orbiting satellites in operational numerical prediction models provides large data volumes and increases the computational effort. The goal here is to simulate the process for assimilating temperature data computed from satellite radiances. The numerical experiment is carried out with global model: the "Simplified Parameterizations, primitivE-Equation DYnamics"(SPEEDY). For the data assimilation scheme was applied an Multilayer Perceptron(MLP) with supervised training. The MLP-ANN is able to emulate the analysis from the Local Ensemble Transform Kalman Filter(LETKF). The ANN was trained with first three months for years 1982, 1983, and 1984 from LETKF. A hindcasting experiment for data assimilation cycle was for January 1985, with a MLP-NN performed with the SPEEDY model. The results for analysis with ANN are very close with the results obtained from LETKF. The simulations show that the major advantage of using MLP-NN is the better computational performance, with similar quality of analysis.
Proceedings of the 6th International Conference on Nonlinear Science and Complexity, 2016
Anais do 10. Congresso Brasileiro de Inteligência Computacional, 2016
An Artificial Neural Network (ANN) is designed to investigate a application for data assimilation... more An Artificial Neural Network (ANN) is designed to investigate a application for data assimilation. This procedure provides an appropriated initial condition to the atmosphere to weather forecasting. Data assimilation is a method to insert observational information into a physical-mathematical model. The use of observations from the earth-orbiting satellites in operational numerical prediction models provides large data volumes and increases the computational effort. The goal here is to simulate the process for assimilating temperature data computed from satellite radiances. The numerical experiment is carried out with global model: the "Simplified Parameterizations, primitivE-Equation DYnamics"(SPEEDY). For the data assimilation scheme was applied an Multilayer Perceptron(MLP) with supervised training. The MLP-ANN is able to emulate the analysis from the Local Ensemble Transform Kalman Filter(LETKF). The ANN was trained with first three months for years 1982, 1983, and 1984 from LETKF. A hindcasting experiment for data assimilation cycle was for January 1985, with a MLP-NN performed with the SPEEDY model. The results for analysis with ANN are very close with the results obtained from LETKF. The simulations show that the major advantage of using MLP-NN is the better computational performance, with similar quality of analysis.
Learning and Nonlinear Models, 2009
The goal of the present work is to employ artificial neural networks as a data assimilation metho... more The goal of the present work is to employ artificial neural networks as a data assimilation method applied to shallow water equation. This model is used to represent ocean dynamics. Data assimilation is a computational procedure to combine observation data with model data for identifying the best initial condition (analysis) to an operational prediction system. Here we compare two techniques: representer method and artificial neural network.
Ciência e Natura, Dec 1, 2007
This paper presents the capabiliity of HSB (Humidy Sensor Brazil) channels data from AQUA satelli... more This paper presents the capabiliity of HSB (Humidy Sensor Brazil) channels data from AQUA satellite,on retrieving the Integrated Water Vapor (IWV) of atmosphere, using Artificial Neural Network(ANN) with simulations of the brightness temperatures from RTTOV-7 radiative model. The results shows ANN as a new method to estimate IWV, with supervised training of observations data from the "RACCI/LBA" experiment in Rondônia, during period of September and October 2002. The Total IWV is also compared against radiosonde data, where all of the results are in good agreement with rms differences less than 4 mm and biases less than 1 mm. This method can also used to estimate the variability of distribution of water vapor in atmosphere through the on-line update training process. The knowledge of the vertical and horizontal distribution of the water vapor in global scale is applied for applications of numerical prediction, climatic modeling and climate global changes studies.
Incêndios florestais causam muitas alterações no clima e no meio ambiente, sendo uma das grandes ... more Incêndios florestais causam muitas alterações no clima e no meio ambiente, sendo uma das grandes preocupações relacionadas ao meio ambiente, sua prevenção e controle. Assim, para auxiliar no planejamento de atividades para sua prevenção, o cálculo do risco de incêndios se faz uma importante ferramenta, determinando a probabilidade das ocorrências destes em determinado local. Este trabalho tem como objetivo fazer o mapeamento das regiões de risco de incêndio no Município de Belo Horizonte. A modelagem proposta será realizada através de Redes Neurais Artificiais (RNA) com treinamento supervisionado. Espera-se obter uma rede neural para fazer a previsão de áreas propícias aos incêndios, apresentando as variáveis de entrada de qualquer período que se deseja determinar. Esta estimativa dará o delineamento de áreas prioritárias através de mapas que auxiliarão em atividades de prevenção e alocação de equipes brigadistas, buscando minimizar possíveis danos causados pelos incêndios. O que se...
All observational data received at "Centro de Previsão de Tempo e Estudos Climáticos" (... more All observational data received at "Centro de Previsão de Tempo e Estudos Climáticos" (CPTEC) from the "Global telecommunications Systems"(GTS) are archived with a consistent format BUFR (Binary Universal Form for the Representation of meteorologiacal data). The BUFR data are pre-processed, selecting data types and performing a few simple quality check and corretions. Useful by_products of the assimilation system are the innovation files (Observation Data Stream -ODS) which contain the differences between the observations and the first guess of ETA model, interpolated to the observation location. This paper presents the structure of ODS file and how the GTS data are archieved in it.Pages:
2015 Latin America Congress on Computational Intelligence (LA-CCI), 2015
Predicting the weather is dependent on the initial states specified in the computer model used to... more Predicting the weather is dependent on the initial states specified in the computer model used to make the prediction. The data assimilation (DA) schemes are state-estimation techniques to generate an appropriated initial states for numerical models. DA deals with observations and data from the nonlinear dynamical models, both data set are very large in use on operational weather centers. The output from the DA procedure is called analysis. Some DA techniques become computationally intensive. The artificial neural networks (NN) can be employed to improve the computational performance. Two DA schemes are analized here: the Local Ensemble Transform Kalman Filter [17], and a version of a variational assimilation method [2] named the representer method. The EnKF was applied to a 3D atmospheric global spectral model (SPEEDY model), while the representer scheme was applied to the 2D shallow-water modelfor simulating the ocean circulation. These DA techniques were emulated by multilayer perpectron neural network (MLP-NN). The goal of this paper is to show the speed up for the DA computer performance in comparison to the methods emulated. The data assimilation process by NN preserves the analysis quality of the former DA techniques. In our experiments, the NN applied to DA on the SPEEDY model was 75 times faster than EnKF.
ABSTRACT The description of the most physical phenomena is based on differential equations. But, ... more ABSTRACT The description of the most physical phenomena is based on differential equations. But, the modeling error is a permanent feature. One basic uncertainty is to identify the initial condition. For the operational prediction systems, a strategy to deal with such uncertainty is to add some information from the real world into the mathematical model. This additional information consists of observations (measurement values). However, the observed data might be carefully inserted, in order to avoid negative impacto on the prediction. Techniques for data assimilation are tools to produce an effective combination of two sources of data: observation and model, for computing the analysis. The analysis is the initial condition used in the prediction computer model. The goal of the present work is to employ artificialneural networks as a data assimilation method applied to shallow water equation used to represent ocean dynamics.
Studies in Fuzziness and Soft Computing, 2014
The predictability of the behavior of chaotic systems is of great importance because many real-wo... more The predictability of the behavior of chaotic systems is of great importance because many real-world phenomena have some type of chaotic regime. In chaotic systems, small changes in the initial conditions can lead to very different results from the original system trajectory. The prediction of chaotic systems behavior is usually very difficult, particularly in practical applications in which initial conditions are obtained by measurement instruments, very often subject to acquisition errors. Here we use “bred vectors” methodology to generate pairs of input/output that are then used to train Neural Networks and Neuro-Fuzzy Systems. We apply the approach to predict regime change for Lorenz strange attractors and the nonlinear coupled three-waves problem from solar physics.
Russian Journal of Numerical Analysis and Mathematical Modelling, 2013
Recently, ensemble Kalman filters have come into practical data assimilation for numerical weathe... more Recently, ensemble Kalman filters have come into practical data assimilation for numerical weather prediction models. We give an overview of ensemble Kalman filters and problems that arise with practical implementation of ensemble methods. We present our implementation of the local ensemble transform Kalman filter, one of ensemble square root filters using observation localization. Multiplicative and additive inflations are used to prevent filter divergence and to account for the model error. The implemented assimilation system is tested with the global semi-Lagrangian atmospheric model SL-AV using real observations for 2 months of cyclic assimilation (August and September 2012). The system works stably. Application of the ensemble filter significantly reduces first guess (background) errors and corrects the forecast biases. Numerical weather prediction models require estimates of the initial state of the atmosphere as input data. Data assimilation is a cyclic process to obtain such estimates (analyses) by improving first guess (also called background) estimates with available atmospheric observations. Data assimilation methods use the first guess and observations together with the information on their error distributions to seek for the best possible estimate in some sense (for example, maximum a posteriori estimate, or minimum variance estimate). Data assimilation consists of two steps, i.e., the analysis step: obtaining
... Rosângela Cintra, Dirceu Herdies ,José A. Aravéquia, Julio Tóta, Jose P. Bonatti Centro de Pr... more ... Rosângela Cintra, Dirceu Herdies ,José A. Aravéquia, Julio Tóta, Jose P. Bonatti Centro de Previsão de Tempo e Estudos Climáticos - CPTEC/INPE ... REFERENCES Cohn, SE, A. da Silva, J. Guo, M. Sienkiewicz, D. Lamich, 1998: Assessing the Effects of the Data Selection with ...
... MCG-CPTEC Conventional GTS and LBA DSA Satéllites Retrievals (NOAA) ... DAO Physical-space St... more ... MCG-CPTEC Conventional GTS and LBA DSA Satéllites Retrievals (NOAA) ... DAO Physical-space Statistical Analysis System. Mon. Wea. Rev., 126, 2913-2926. da Da Silva, Arlindo , M. Tippet, J. Guo, 1998: PSAS User's Manual. DAO Offoce Note 96-02, NASA/ Goddard Space ...
Uploads
Papers by Rosangela Cintra