Enhanced Carbon Flux Forecasting via STL Decomposition and Hybrid ARIMA-ES-LSTM Model in Amazon Forest

This study presents a hybrid model, STL-ARIMA-ES-LSTM, developed to improve the accuracy of Gross Primary Productivity (GPP) forecasts in the Amazon region. The model integrates seasonal and trend decomposition using Loess (STL) with statistical methods (ARIMA and Exponential Smoothing-ES) and a mac...

Full description

Saved in:
Bibliographic Details
Main Authors: Jean A. C. Dias, Pedro H. Do V. Guimaraes, Williane G. S. Pereira, Leonardo De O. Tamasauskas, Marivan S. Gomes, Alan B. S. Correa, Karla Figueiredo, Gilson Costa, Gabriel Brito Costa, Fernando A. R. Costa, Marcos C. Da R. Seruffo
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10965627/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850268469739651072
author Jean A. C. Dias
Pedro H. Do V. Guimaraes
Williane G. S. Pereira
Leonardo De O. Tamasauskas
Marivan S. Gomes
Alan B. S. Correa
Karla Figueiredo
Gilson Costa
Gabriel Brito Costa
Fernando A. R. Costa
Marcos C. Da R. Seruffo
author_facet Jean A. C. Dias
Pedro H. Do V. Guimaraes
Williane G. S. Pereira
Leonardo De O. Tamasauskas
Marivan S. Gomes
Alan B. S. Correa
Karla Figueiredo
Gilson Costa
Gabriel Brito Costa
Fernando A. R. Costa
Marcos C. Da R. Seruffo
author_sort Jean A. C. Dias
collection DOAJ
description This study presents a hybrid model, STL-ARIMA-ES-LSTM, developed to improve the accuracy of Gross Primary Productivity (GPP) forecasts in the Amazon region. The model integrates seasonal and trend decomposition using Loess (STL) with statistical methods (ARIMA and Exponential Smoothing-ES) and a machine learning technique (Long Short-Term Memory - LSTM). Applied to GPP data from the PE-QFR site, the hybrid model achieved significantly better error metrics, with RMSE of 1.69 gC/m2/day, MAE of 1.35 gC/m2/day, and MAPE of 0.20%, compared to the standalone LSTM (RMSE of 2.16 gC/m2/day, MAE of 1.78 gC/m2/day, and MAPE of 0.27%). Furthermore, the hybrid model showed stronger agreement with the observed data, with correlation coefficient r =0.62 and <inline-formula> <tex-math notation="LaTeX">${R}^{2} =0.39$ </tex-math></inline-formula>, whereas the LSTM alone yielded r =0.26 and <inline-formula> <tex-math notation="LaTeX">${R}^{2} =$ </tex-math></inline-formula> &#x2013;0.002. The STL decomposition allowed effective separation of trend, seasonality, and residual components, enabling tailored modeling of each, which contributed to the improved predictive performance. These results demonstrate the advantage of hybrid approaches in capturing the nonlinear and seasonal patterns of GPP, supporting enhanced environmental monitoring and more informed climate change mitigation strategies in the Amazon.
format Article
id doaj-art-c40cdd06bf814b44804b0dfc4219fe3b
institution OA Journals
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-c40cdd06bf814b44804b0dfc4219fe3b2025-08-20T01:53:26ZengIEEEIEEE Access2169-35362025-01-0113847138472610.1109/ACCESS.2025.356116610965627Enhanced Carbon Flux Forecasting via STL Decomposition and Hybrid ARIMA-ES-LSTM Model in Amazon ForestJean A. C. Dias0https://orcid.org/0009-0005-7663-8072Pedro H. Do V. Guimaraes1https://orcid.org/0009-0006-2350-2451Williane G. S. Pereira2https://orcid.org/0009-0005-0611-2462Leonardo De O. Tamasauskas3Marivan S. Gomes4Alan B. S. Correa5https://orcid.org/0009-0002-9590-8347Karla Figueiredo6https://orcid.org/0000-0001-8420-3937Gilson Costa7https://orcid.org/0000-0001-7341-9118Gabriel Brito Costa8https://orcid.org/0000-0002-5254-489XFernando A. R. Costa9https://orcid.org/0000-0002-0226-7505Marcos C. Da R. Seruffo10https://orcid.org/0000-0002-8106-0560Operational Research Laboratory, Federal University of Par&#x00E1; (UFPA), Bel&#x00E9;m, BrazilOperational Research Laboratory, Federal University of Par&#x00E1; (UFPA), Bel&#x00E9;m, BrazilOperational Research Laboratory, Federal University of Par&#x00E1; (UFPA), Bel&#x00E9;m, BrazilOperational Research Laboratory, Federal University of Par&#x00E1; (UFPA), Bel&#x00E9;m, BrazilSchool of Technology, State University of Amazonas (UEA), Amazonas, BrazilOperational Research Laboratory, Federal University of Par&#x00E1; (UFPA), Bel&#x00E9;m, BrazilDepartment of Informatics and Computer Science, Institute of Mathematics and Statistics, Rio de Janeiro State University (UERJ), Rio de Janeiro, BrazilDepartment of Genetics, Institute of Biological and Health Sciences II, Rio de Janeiro State University (UERJ), Rio de Janeiro, BrazilPost-Graduate Program in Natural Resources of the Amazon (PPGRNA), Federal University of Western Par&#x00E1; (UFOPA), Santar&#x00E9;m, BrazilOperational Research Laboratory, Federal University of Par&#x00E1; (UFPA), Bel&#x00E9;m, BrazilOperational Research Laboratory, Federal University of Par&#x00E1; (UFPA), Bel&#x00E9;m, BrazilThis study presents a hybrid model, STL-ARIMA-ES-LSTM, developed to improve the accuracy of Gross Primary Productivity (GPP) forecasts in the Amazon region. The model integrates seasonal and trend decomposition using Loess (STL) with statistical methods (ARIMA and Exponential Smoothing-ES) and a machine learning technique (Long Short-Term Memory - LSTM). Applied to GPP data from the PE-QFR site, the hybrid model achieved significantly better error metrics, with RMSE of 1.69 gC/m2/day, MAE of 1.35 gC/m2/day, and MAPE of 0.20%, compared to the standalone LSTM (RMSE of 2.16 gC/m2/day, MAE of 1.78 gC/m2/day, and MAPE of 0.27%). Furthermore, the hybrid model showed stronger agreement with the observed data, with correlation coefficient r =0.62 and <inline-formula> <tex-math notation="LaTeX">${R}^{2} =0.39$ </tex-math></inline-formula>, whereas the LSTM alone yielded r =0.26 and <inline-formula> <tex-math notation="LaTeX">${R}^{2} =$ </tex-math></inline-formula> &#x2013;0.002. The STL decomposition allowed effective separation of trend, seasonality, and residual components, enabling tailored modeling of each, which contributed to the improved predictive performance. These results demonstrate the advantage of hybrid approaches in capturing the nonlinear and seasonal patterns of GPP, supporting enhanced environmental monitoring and more informed climate change mitigation strategies in the Amazon.https://ieeexplore.ieee.org/document/10965627/Gross primary productivitycarbon fluxhybrid modelsSTL decompositionLSTMstatistical models
spellingShingle Jean A. C. Dias
Pedro H. Do V. Guimaraes
Williane G. S. Pereira
Leonardo De O. Tamasauskas
Marivan S. Gomes
Alan B. S. Correa
Karla Figueiredo
Gilson Costa
Gabriel Brito Costa
Fernando A. R. Costa
Marcos C. Da R. Seruffo
Enhanced Carbon Flux Forecasting via STL Decomposition and Hybrid ARIMA-ES-LSTM Model in Amazon Forest
IEEE Access
Gross primary productivity
carbon flux
hybrid models
STL decomposition
LSTM
statistical models
title Enhanced Carbon Flux Forecasting via STL Decomposition and Hybrid ARIMA-ES-LSTM Model in Amazon Forest
title_full Enhanced Carbon Flux Forecasting via STL Decomposition and Hybrid ARIMA-ES-LSTM Model in Amazon Forest
title_fullStr Enhanced Carbon Flux Forecasting via STL Decomposition and Hybrid ARIMA-ES-LSTM Model in Amazon Forest
title_full_unstemmed Enhanced Carbon Flux Forecasting via STL Decomposition and Hybrid ARIMA-ES-LSTM Model in Amazon Forest
title_short Enhanced Carbon Flux Forecasting via STL Decomposition and Hybrid ARIMA-ES-LSTM Model in Amazon Forest
title_sort enhanced carbon flux forecasting via stl decomposition and hybrid arima es lstm model in amazon forest
topic Gross primary productivity
carbon flux
hybrid models
STL decomposition
LSTM
statistical models
url https://ieeexplore.ieee.org/document/10965627/
work_keys_str_mv AT jeanacdias enhancedcarbonfluxforecastingviastldecompositionandhybridarimaeslstmmodelinamazonforest
AT pedrohdovguimaraes enhancedcarbonfluxforecastingviastldecompositionandhybridarimaeslstmmodelinamazonforest
AT willianegspereira enhancedcarbonfluxforecastingviastldecompositionandhybridarimaeslstmmodelinamazonforest
AT leonardodeotamasauskas enhancedcarbonfluxforecastingviastldecompositionandhybridarimaeslstmmodelinamazonforest
AT marivansgomes enhancedcarbonfluxforecastingviastldecompositionandhybridarimaeslstmmodelinamazonforest
AT alanbscorrea enhancedcarbonfluxforecastingviastldecompositionandhybridarimaeslstmmodelinamazonforest
AT karlafigueiredo enhancedcarbonfluxforecastingviastldecompositionandhybridarimaeslstmmodelinamazonforest
AT gilsoncosta enhancedcarbonfluxforecastingviastldecompositionandhybridarimaeslstmmodelinamazonforest
AT gabrielbritocosta enhancedcarbonfluxforecastingviastldecompositionandhybridarimaeslstmmodelinamazonforest
AT fernandoarcosta enhancedcarbonfluxforecastingviastldecompositionandhybridarimaeslstmmodelinamazonforest
AT marcoscdarseruffo enhancedcarbonfluxforecastingviastldecompositionandhybridarimaeslstmmodelinamazonforest