Modeling the sea-surface pCO2 of the central Bay of Bengal region using machine learning algorithms

Joshi, A.P.; Kumar, Vishnu; Warrior, Hari

doi:10.1016/j.ocemod.2022.102094

Cited by 7 publications

(7 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The multiple linear regression equation to predict sea-surface p CO 2 is as follows: Artificial Neural Network (ANN) The Artificial Neural Network (ANN) is a part of artificial intelligence based on the biological neural system. It has become common practise to establish the p CO 2 for regional scales 30 , 36 – 39 . The ANN comprises interconnected neurons that interpret incoming data like how the human brain learns.…”

Section: Methodsmentioning

confidence: 99%

“…The gradient-boosted algorithm’s performance and computational speed were both expanded to produce the XGBoost algorithm. Since it performed well for the central BoB region 30 , the model’s great speed and accuracy motivate us to compare its performance to that of other models. Only the residuals are supplied to the following weaker learners once the trees or vulnerable learners have been added in sequential order.…”

Section: Methodsmentioning

confidence: 99%

“…Similar to the ANN, the XGBoost model also has tunable hyperparameters. Following previous literature 30 , we employ the Optuna optimization framework 44 to fine-tune the hyper-parameters. At https://xgboost.readthedocs.io/en/stable/parameter.html one may find the description for each of the XGB hyper-parameters.…”

Section: Methodsmentioning

confidence: 99%

“…The linearity assumption between the dependent and independent variables is not always true. Region-specific Machine Learning (ReML) algorithms showed promising results for the central BoB 30 . Hence, this study attempts to construct spatiotemporal sea-surface p CO 2 maps for the BoB using observations and advanced ML techniques.…”

Section: Background and Summarymentioning

confidence: 99%

See 3 more Smart Citations

Sea-surface pCO2 maps for the Bay of Bengal based on advanced machine learning algorithms

Joshi,

Ghoshal,

Chakraborty

et al. 2024

Sci Data

Self Cite

View full text Add to dashboard Cite

Lack of sufficient observations has been an impediment for understanding the spatial and temporal variability of sea-surface pCO2 for the Bay of Bengal (BoB). The limited number of observations into existing machine learning (ML) products from BoB often results in high prediction errors. This study develops climatological sea-surface pCO2 maps using a significant number of open and coastal ocean observations of pCO2 and associated variables regulating pCO2 variability in BoB. We employ four advanced ML algorithms to predict pCO2. We use the best ML model to produce a high-resolution climatological product (INCOIS-ReML). The comparison of INCOIS-ReML pCO2 with RAMA buoy-based sea-surface pCO2 observations indicates INCOIS-ReML’s satisfactory performance. Further, the comparison of INCOIS-ReML pCO2 with existing ML products establishes the superiority of INCOIS-ReML. The high-resolution INCOIS-ReML greatly captures the spatial variability of pCO2 and associated air-sea CO2 flux compared to other ML products in the coastal BoB and the northern BoB.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Background and Summarymentioning

confidence: 99%

See 2 more Smart Citations

Sea-surface pCO2 maps for the Bay of Bengal based on advanced machine learning algorithms

Joshi,

Ghoshal,

Chakraborty

et al. 2024

Sci Data

Self Cite

View full text Add to dashboard Cite

show abstract

“…Despite the fact that the BoB plays a major role in the global carbon budget, there have been very few attempts [25], [26], [27], [28] were made to investigate the pCO 2 variability over this region. In a recent study, Joshi et al [29] developed and tested three machine learning methods (which includes multiple linear regression (MLR), artificial neural network (ANN), and extreme gradient boosting (XGB)) over the central BoB region using SST, SSS and pCO 2 . While machine learning techniques offer numerous advantages, such as their capacity to learn from extensive datasets and provide accurate predictions, the lack of available large datasets for training a more precise pCO 2 model and validating its results presents a limitation in this study.…”

Section: Introductionmentioning

confidence: 99%

Advancements in Carbon Dioxide Modeling: An Algorithm Incorporating In-Situ and Satellite Data for Improved Understanding of pCO Dynamics in the Bay of Bengal

Shaik,

Krishna,

Begum

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Estimation of the partial pressure of carbon dioxide (pCO 2 ) in the Bay of Bengal (BoB) region plays a crucial role in better understanding the air-sea CO 2 fluxes. Complex physical and biogeochemical processes such as physical mixing, stratification, thermodynamic, and biological effects dominate the spatiotemporal variability of pCO 2 concentration over the BoB. This is difficult to estimate through in-situ platforms alone due to the time-consuming, cost-effective, and intricacies involved in water sample collection during rough oceanic weather conditions. Alternatively, remote sensing technology provides governing control parameters with high spatiotemporal resolution over large synoptic scales. Since the BoB region is influenced by the Indian monsoon system and other complex processes, existing regional and global pCO 2 algorithms are not adequate to estimate more accurate pCO 2 fields. Hence, there is a need to develop a regional pCO 2 algorithm over the BoB. To resolve this problem, in the present study, a Multi Parametric Regional Regression (MPRR) approach was developed over the BoB using satellite data such as sea surface temperature (SST), sea surface salinity (SSS), and chlorophyll-a (Chla) concentration. To train and validate the MPRR approach, required in-situ measurements were obtained from the open and coastal waters of the BoB. The validation results revealed that the present MPRR approach showed better performance with significant low errors (mean relative error (MRE) = 0.012, mean normalized bias (MNB) = 0.022, and root mean square error (RMSE) = 4.75 µatm) and a high correlation coefficient (R 2 = 0.92). Furthermore, the study demonstrated the spatiotemporal variability of pCO 2 and generated monthly, seasonal, and annual pCO 2 maps over the BoB.INDEX TERMS Ocean color remote sensing, multiparametric regional regression, SST, SSS, Chla and pCO 2 .

show abstract

A novel approach based on integration of convolutional neural networks and echo state network for daily electricity demand prediction

Ghimire

Nguyen‐Huy

AL‐Musaylh

et al. 2023

Energy

View full text Add to dashboard Cite

Modeling the sea-surface pCO2 of the central Bay of Bengal region using machine learning algorithms

Cited by 7 publications

References 43 publications

Sea-surface pCO2 maps for the Bay of Bengal based on advanced machine learning algorithms

Sea-surface pCO2 maps for the Bay of Bengal based on advanced machine learning algorithms

Advancements in Carbon Dioxide Modeling: An Algorithm Incorporating In-Situ and Satellite Data for Improved Understanding of pCO Dynamics in the Bay of Bengal

A novel approach based on integration of convolutional neural networks and echo state network for daily electricity demand prediction

Contact Info

Product

Resources

About