Scholarlink

<p>Lake evaporation is a critical component of the hydrological cycle and surface energy balance, yet its response to climate change remains one of the most uncertain elements of catchment-scale water budgets. Understanding how evaporation will change under future climate conditions is essential for water resource management, particularly in regions where lakes serve as key freshwater reservoirs. Despite this importance, projected changes in lake evaporation and their underlying climatic drivers remain poorly understood, largely due to the scarcity of direct long-term measurements and the limited application of process-based models to future climate scenarios. This study addresses this gap by investigating climate change impacts on evaporation across four temperate lakes in Poland using a unique dataset of daily measurements collected from floating evaporation rafts (GGI-3000 evaporimeters). An integrated modeling framework was applied, combining CMIP6 climate projections under three emission scenarios (SSP1-2.6, SSP3-7.0, and SSP5-8.5) with machine-learning-based adaptive bias correction and process-based lake simulations. A two-stage hybrid bias correction framework combining ERA5 reanalysis with local observations was applied to ensemble projections generated using Bayesian Model Averaging. Model validation against in-situ observations yielded temperature correlations of 0.91–0.96 and evaporation correlations of 0.57–0.68, indicating satisfactory predictive skill; inter-model uncertainty was quantified using BMA-weighted spread across the five-member ensemble. Climate projections indicated substantial warming and declining humidity for most lakes, with the magnitude of change increasing from SSP1-2.6 to SSP5-8.5. Evaporation responses were lake-dependent: three lakes exhibited increasing evaporation under warming conditions, whereas Lake Sławianowskie showed a decreasing trend driven by its distinct humidity regime. The analysis of relative importance of climatic drivers, revealing two distinct lake categories: wind-dominated systems (Sławskie and Raduńskie Górne) and humidity-dominated systems (Rajgrodzkie and Sławianowskie). These findings suggest that lake evaporation responses to climate change are highly lake-specific, governed by the interaction between local climate conditions and dominant physical drivers. This lake-type classification provides a valuable framework for water resource managers to anticipate differential hydrological impacts and develop targeted adaptation strategies for lake systems under future climate change.</p> Graphical Abstract <p></p> <p>The graphical abstract provides an overview of the study on climate-driven evaporation shifts in temperate lakes, integrating CMIP6 projections with machine learning bias correction. The conceptual framework illustrates how meteorological variables, specifically air temperature, wind speed, and relative humidity, influence evaporation dynamics under future Shared Socioeconomic Pathways (SSP1-2.6, SSP3-7.0, and SSP5-8.5). Data from in situ floating evaporation rafts (GGI-3000 evaporimeters) across four Polish lakes (Sławskie, Raduńskie Górne, Rajgrodzkie, and Sławianowskie), along with ERA5 reanalysis and CMIP6 outputs, were used to derive key climate forcing variables. To explore future lake dynamics, the study employed an integrated modeling framework featuring a two-stage hybrid bias correction—combining Bayesian Model Averaging and adaptive machine learning—coupled with the process-based FLake model. Results extending to the year 2100 reveal divergent evaporation patterns; graphical representations illustrate three lakes experiencing increasing evaporation trends (+ 0.18 to + 0.50&#xa0;mm/d), while Lake Sławianowskie exhibits a uniquely decreasing trend (− 0.17 to − 0.41&#xa0;mm/d). Furthermore, statistical evaluations classify the lakes into distinct wind-dominated (55.9–72.8% importance) and humidity-dominated (77.9–79.5% importance) systems. In conclusion, these divergent responses underscore the critical role of lake-specific climate sensitivities, highlighting the value of localized modeling for effective water resource management and adaptation planning in the 21st century.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Climate-driven Evaporation Shifts in Temperate Lakes: Integrating CMIP6 Projections with Machine Learning Bias Correction

  • 期刊论文

  • 发表日期：

    2026年6月20日

查看全文
  • Mariusz Ptak,
  • Bogumił Nowak,
  • Teerachai Amnuaylojaroen

摘要

Lake evaporation is a critical component of the hydrological cycle and surface energy balance, yet its response to climate change remains one of the most uncertain elements of catchment-scale water budgets. Understanding how evaporation will change under future climate conditions is essential for water resource management, particularly in regions where lakes serve as key freshwater reservoirs. Despite this importance, projected changes in lake evaporation and their underlying climatic drivers remain poorly understood, largely due to the scarcity of direct long-term measurements and the limited application of process-based models to future climate scenarios. This study addresses this gap by investigating climate change impacts on evaporation across four temperate lakes in Poland using a unique dataset of daily measurements collected from floating evaporation rafts (GGI-3000 evaporimeters). An integrated modeling framework was applied, combining CMIP6 climate projections under three emission scenarios (SSP1-2.6, SSP3-7.0, and SSP5-8.5) with machine-learning-based adaptive bias correction and process-based lake simulations. A two-stage hybrid bias correction framework combining ERA5 reanalysis with local observations was applied to ensemble projections generated using Bayesian Model Averaging. Model validation against in-situ observations yielded temperature correlations of 0.91–0.96 and evaporation correlations of 0.57–0.68, indicating satisfactory predictive skill; inter-model uncertainty was quantified using BMA-weighted spread across the five-member ensemble. Climate projections indicated substantial warming and declining humidity for most lakes, with the magnitude of change increasing from SSP1-2.6 to SSP5-8.5. Evaporation responses were lake-dependent: three lakes exhibited increasing evaporation under warming conditions, whereas Lake Sławianowskie showed a decreasing trend driven by its distinct humidity regime. The analysis of relative importance of climatic drivers, revealing two distinct lake categories: wind-dominated systems (Sławskie and Raduńskie Górne) and humidity-dominated systems (Rajgrodzkie and Sławianowskie). These findings suggest that lake evaporation responses to climate change are highly lake-specific, governed by the interaction between local climate conditions and dominant physical drivers. This lake-type classification provides a valuable framework for water resource managers to anticipate differential hydrological impacts and develop targeted adaptation strategies for lake systems under future climate change.

Graphical Abstract

The graphical abstract provides an overview of the study on climate-driven evaporation shifts in temperate lakes, integrating CMIP6 projections with machine learning bias correction. The conceptual framework illustrates how meteorological variables, specifically air temperature, wind speed, and relative humidity, influence evaporation dynamics under future Shared Socioeconomic Pathways (SSP1-2.6, SSP3-7.0, and SSP5-8.5). Data from in situ floating evaporation rafts (GGI-3000 evaporimeters) across four Polish lakes (Sławskie, Raduńskie Górne, Rajgrodzkie, and Sławianowskie), along with ERA5 reanalysis and CMIP6 outputs, were used to derive key climate forcing variables. To explore future lake dynamics, the study employed an integrated modeling framework featuring a two-stage hybrid bias correction—combining Bayesian Model Averaging and adaptive machine learning—coupled with the process-based FLake model. Results extending to the year 2100 reveal divergent evaporation patterns; graphical representations illustrate three lakes experiencing increasing evaporation trends (+ 0.18 to + 0.50 mm/d), while Lake Sławianowskie exhibits a uniquely decreasing trend (− 0.17 to − 0.41 mm/d). Furthermore, statistical evaluations classify the lakes into distinct wind-dominated (55.9–72.8% importance) and humidity-dominated (77.9–79.5% importance) systems. In conclusion, these divergent responses underscore the critical role of lake-specific climate sensitivities, highlighting the value of localized modeling for effective water resource management and adaptation planning in the 21st century.