Enhancing the Forecast of Ocean Physical Variables through Physics Informed Machine Learning in the Santos Estuary, Brazil

Felipe M. Moreno; Luiz A. Schiaveto Neto; Fabio G. Cozman; Marcelo Dottori; Eduardo A. Tannuri

doi:10.1109/OCEANSChennai45887.2022.9775449

Enhancing the Forecast of Ocean Physical Variables through Physics Informed Machine Learning in the Santos Estuary, Brazil

Felipe M. Moreno, Luiz A. Schiaveto Neto, Fabio G. Cozman, Marcelo Dottori, Eduardo A. Tannuri

Resultado de la investigación: Contribución a una revista › Artículo de la conferencia › revisión exhaustiva

Resumen

This work aims to improve the forecast of surface currents in the entrance of the Santos estuary in Brazil by applying Quantile Regression Forests (QRF) to estimate the error of the Santos Operational Forecasting System (SOFS), a physics-based numerical model for the region. This was achieved by using in-situ data, measured between 2019 and 2021, associated with historical forecasted data from the SOFS. The use of QRF to correct the SOFS forecasts led to a increase in skill of 0.332 in Mean Absolute Error (MAE) and almost eliminated the bias error of the predicted currents.

Idioma original	Inglés
Publicación	Oceans Conference Record (IEEE)
DOI	https://doi.org/10.1109/OCEANSChennai45887.2022.9775449
Estado	Publicada - 2022
Publicado de forma externa	Sí
Evento	OCEANS 2022 - Chennai - Chennai, India Duración: 21 feb. 2022 → 24 feb. 2022

Acceder al documento

10.1109/OCEANSChennai45887.2022.9775449

Otros archivos y enlaces

Enlace a la publicación en Scopus

Citar esto

@article{d04e2b43abe348269958eee3cba4915b,

title = "Enhancing the Forecast of Ocean Physical Variables through Physics Informed Machine Learning in the Santos Estuary, Brazil",

abstract = "This work aims to improve the forecast of surface currents in the entrance of the Santos estuary in Brazil by applying Quantile Regression Forests (QRF) to estimate the error of the Santos Operational Forecasting System (SOFS), a physics-based numerical model for the region. This was achieved by using in-situ data, measured between 2019 and 2021, associated with historical forecasted data from the SOFS. The use of QRF to correct the SOFS forecasts led to a increase in skill of 0.332 in Mean Absolute Error (MAE) and almost eliminated the bias error of the predicted currents.",

keywords = "Current Forecasting, Physics-Informed Machine Learning, Quantile Regression Forest, Santos Estuary",

author = "Moreno, {Felipe M.} and {Schiaveto Neto}, {Luiz A.} and Cozman, {Fabio G.} and Marcelo Dottori and Tannuri, {Eduardo A.}",

note = "Publisher Copyright: {\textcopyright} 2022 IEEE.; null ; Conference date: 21-02-2022 Through 24-02-2022",

year = "2022",

doi = "10.1109/OCEANSChennai45887.2022.9775449",

language = "Ingl{\'e}s",

journal = "Oceans Conference Record (IEEE)",

issn = "0197-7385",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

Enhancing the Forecast of Ocean Physical Variables through Physics Informed Machine Learning in the Santos Estuary, Brazil. / Moreno, Felipe M.; Schiaveto Neto, Luiz A.; Cozman, Fabio G.; Dottori, Marcelo; Tannuri, Eduardo A.

En: Oceans Conference Record (IEEE), 2022.

Resultado de la investigación: Contribución a una revista › Artículo de la conferencia › revisión exhaustiva

TY - JOUR

T1 - Enhancing the Forecast of Ocean Physical Variables through Physics Informed Machine Learning in the Santos Estuary, Brazil

AU - Moreno, Felipe M.

AU - Schiaveto Neto, Luiz A.

AU - Cozman, Fabio G.

AU - Dottori, Marcelo

AU - Tannuri, Eduardo A.

PY - 2022

Y1 - 2022

N2 - This work aims to improve the forecast of surface currents in the entrance of the Santos estuary in Brazil by applying Quantile Regression Forests (QRF) to estimate the error of the Santos Operational Forecasting System (SOFS), a physics-based numerical model for the region. This was achieved by using in-situ data, measured between 2019 and 2021, associated with historical forecasted data from the SOFS. The use of QRF to correct the SOFS forecasts led to a increase in skill of 0.332 in Mean Absolute Error (MAE) and almost eliminated the bias error of the predicted currents.

AB - This work aims to improve the forecast of surface currents in the entrance of the Santos estuary in Brazil by applying Quantile Regression Forests (QRF) to estimate the error of the Santos Operational Forecasting System (SOFS), a physics-based numerical model for the region. This was achieved by using in-situ data, measured between 2019 and 2021, associated with historical forecasted data from the SOFS. The use of QRF to correct the SOFS forecasts led to a increase in skill of 0.332 in Mean Absolute Error (MAE) and almost eliminated the bias error of the predicted currents.

KW - Current Forecasting

KW - Physics-Informed Machine Learning

KW - Quantile Regression Forest

KW - Santos Estuary

UR - http://www.scopus.com/inward/record.url?scp=85131700493&partnerID=8YFLogxK

U2 - 10.1109/OCEANSChennai45887.2022.9775449

DO - 10.1109/OCEANSChennai45887.2022.9775449

M3 - Artículo de la conferencia

AN - SCOPUS:85131700493

JO - Oceans Conference Record (IEEE)

JF - Oceans Conference Record (IEEE)

SN - 0197-7385

Y2 - 21 February 2022 through 24 February 2022

ER -

Enhancing the Forecast of Ocean Physical Variables through Physics Informed Machine Learning in the Santos Estuary, Brazil

Resumen

Acceder al documento

Otros archivos y enlaces

Huella

Citar esto