Three Environments, One Problem: Forecasting Water Temperature in Central Europe in Response to Climate Change DOI Creative Commons
Mariusz Ptak, Mariusz Sojka, Katarzyna Szyga–Pluta

и другие.

Forecasting, Год журнала: 2025, Номер 7(2), С. 24 - 24

Опубликована: Май 29, 2025

Water temperature is a fundamental parameter influencing range of biotic and abiotic processes occurring within various components the hydrosphere. This study presents multi-step, data-driven predictive modeling framework to estimate water temperatures for period 2021–2100 in three aquatic environments Central Europe: Odra River, Szczecin Lagoon, Baltic Sea. The integrates Bayesian Model Averaging (BMA), Random Sample Consensus (RANSAC) regression, Gradient Boosting Regressor (GBR), Forest (RF) machine learning models. To assess performance models, coefficient determination (R2), mean absolute error (MAE), root square (RMSE) were used. results showed that application statistical downscaling methods improved prediction air with respect BMA. Moreover, RF method was used predict temperature. best model obtained Sea lowest River. Under SSP2-4.5 SSP5-8.5 scenario-based simulations, projected increases could from 1.5 °C 1.7 4.7 5.1 °C. In contrast, increase by 2100 will be between 1.2 1.6 (SSP2-4.5 scenario) 3.5 4.9 (SSP5-8.5).

Язык: Английский

The effects of climate change on the thermal stratification of the Gulf of Oman DOI

Shirin Farkhani,

Nasser Hadjizadeh Zaker

Dynamics of Atmospheres and Oceans, Год журнала: 2025, Номер 110, С. 101560 - 101560

Опубликована: Май 14, 2025

Язык: Английский

Процитировано

0
Three Environments, One Problem: Forecasting Water Temperature in Central Europe in Response to Climate Change DOI Creative Commons
Mariusz Ptak, Mariusz Sojka, Katarzyna Szyga–Pluta

и другие.

Forecasting, Год журнала: 2025, Номер 7(2), С. 24 - 24

Опубликована: Май 29, 2025

Water temperature is a fundamental parameter influencing range of biotic and abiotic processes occurring within various components the hydrosphere. This study presents multi-step, data-driven predictive modeling framework to estimate water temperatures for period 2021–2100 in three aquatic environments Central Europe: Odra River, Szczecin Lagoon, Baltic Sea. The integrates Bayesian Model Averaging (BMA), Random Sample Consensus (RANSAC) regression, Gradient Boosting Regressor (GBR), Forest (RF) machine learning models. To assess performance models, coefficient determination (R2), mean absolute error (MAE), root square (RMSE) were used. results showed that application statistical downscaling methods improved prediction air with respect BMA. Moreover, RF method was used predict temperature. best model obtained Sea lowest River. Under SSP2-4.5 SSP5-8.5 scenario-based simulations, projected increases could from 1.5 °C 1.7 4.7 5.1 °C. In contrast, increase by 2100 will be between 1.2 1.6 (SSP2-4.5 scenario) 3.5 4.9 (SSP5-8.5).

Язык: Английский

Процитировано

0