%0 Audiovisual Material
%4 sid.inpe.br/mtc-m16c/2025/10.12.17.58
%2 sid.inpe.br/mtc-m16c/2025/10.12.17.58.32
%T Estimating Atmospheric Turbulence (Cn²) at Astronomical Sites Using Meteorological Data and Machine Learning
%D 2025
%A Cuadros, Edith Tueros,
%A Almeida, A.,
%A Cornejo, D.,
%A Carlesso, Franciele,
%A Guarnieri, Fernando Luis,
%A Dal Lago, Alisson,
%A Vieira, Luis Eduardo Antunes,
%@affiliation Instituto Nacional de Pesquisas Espaciais (INPE)
%@affiliation
%@affiliation
%@affiliation Instituto Nacional de Pesquisas Espaciais (INPE)
%@affiliation Instituto Nacional de Pesquisas Espaciais (INPE)
%@affiliation Instituto Nacional de Pesquisas Espaciais (INPE)
%@affiliation Instituto Nacional de Pesquisas Espaciais (INPE)
%@electronicmailaddress edith.cuadros@inpe.br
%@electronicmailaddress
%@electronicmailaddress
%@electronicmailaddress franciele.carlesso@inpe.br
%@electronicmailaddress fernando.guarnieri@inpe.br
%@electronicmailaddress alisson.dallago@inpe.br
%@electronicmailaddress luis.vieira@inpe.br
%B Semana Acadêmica da Geofísica Espacial (SAGE)
%C São José dos Campos
%8 22 a 26 set. 2025
%I Instituto Nacional de Pesquisas Espaciais (INPE)
%J São José dos Campos
%X Accurate estimation of the refractive index structure parameter (Cₙ²) is vital for optimizing astronomical observations and adaptive optics. This study explores machine learning models to predict Cₙ² from Differential Image Motion Monitor (DIMM) data using ERA5 reanalysis from ECMWF. Three models were tested: Multilayer Perceptron (MLP), Long Short-Term Memory (LSTM), and Extreme Gradient Boosting (XGBoost). Data preprocessing aligned time series and addressed missing values. Training and validation were performed with data from Paranal Observatory, Chile. The models will later be applied at the Laboratório Nacional de Astrofísica (LNA), Brazil. ERA5-derived features included temperature, pressure, humidity, and wind profiles; the target was DIMM-measured Cₙ². Results demonstrate the potential of combining reanalysis data and machine learning for atmospheric turbulence estimation, offering a cost-effective alternative to continuous in-situ monitoring.
%9 Poster
%@language en
%3 SAGE-Poster2025.pdf