Data Reconstruction for Groundwater Wells Proximal to Lakes: A Quantitative Assessment for Hydrological Data Imputation DOI Open Access
Murat Can, Babak Vaheddoost, Mir Jafar Sadegh Safari

et al.

Water, Journal Year: 2025, Volume and Issue: 17(5), P. 718 - 718

Published: March 1, 2025

The reconstruction of missing groundwater level data is great importance in hydrogeological and environmental studies. This study provides a comprehensive sequential approach for the near Lake Uluabat Bursa, Turkey. addresses both past future events using Gradient Boosting Regression (GBR) model. process evaluated through model calibration metrics changes statistical properties observed reconstructed time series. To achieve this goal, series from two observational wells lake water levels during January 2004 to September 2019 period are used. level, definition four seasons via application three dummy variables, used as inputs prediction observation wells. optimal GBR achieved by training dataset selected based on gaps series, while test-past test-future datasets validation. Afterward, models reconstructing pre- post-training sets, performance Nash–Sutcliffe efficiency (NSE), Root Mean Square Percentage Error (RMSPE) Performance Index (PI). including probability distribution, maxima, minima, quartiles (Q1–Q3), standard error (SE), coefficient variation (CV), entropy (H), propagation also measured. It was concluded that good base (the best high NSE: 0.99, RMSPE: 0.36, PI: 1.002). In particular, system one case, respectively, experienced 53% 35% rise, which found be tolerable negligible.

Language: Английский

Data Reconstruction for Groundwater Wells Proximal to Lakes: A Quantitative Assessment for Hydrological Data Imputation DOI Open Access
Murat Can, Babak Vaheddoost, Mir Jafar Sadegh Safari

et al.

Water, Journal Year: 2025, Volume and Issue: 17(5), P. 718 - 718

Published: March 1, 2025

The reconstruction of missing groundwater level data is great importance in hydrogeological and environmental studies. This study provides a comprehensive sequential approach for the near Lake Uluabat Bursa, Turkey. addresses both past future events using Gradient Boosting Regression (GBR) model. process evaluated through model calibration metrics changes statistical properties observed reconstructed time series. To achieve this goal, series from two observational wells lake water levels during January 2004 to September 2019 period are used. level, definition four seasons via application three dummy variables, used as inputs prediction observation wells. optimal GBR achieved by training dataset selected based on gaps series, while test-past test-future datasets validation. Afterward, models reconstructing pre- post-training sets, performance Nash–Sutcliffe efficiency (NSE), Root Mean Square Percentage Error (RMSPE) Performance Index (PI). including probability distribution, maxima, minima, quartiles (Q1–Q3), standard error (SE), coefficient variation (CV), entropy (H), propagation also measured. It was concluded that good base (the best high NSE: 0.99, RMSPE: 0.36, PI: 1.002). In particular, system one case, respectively, experienced 53% 35% rise, which found be tolerable negligible.

Language: Английский

Citations

0