Hydrological Response to Climate Change: McGAN for Multi-Site Scenario Weather Series Generation and LSTM for Streamflow Modeling DOI Creative Commons
Jian Sha,

Yaxin Chang,

Yaxiu Liu

и другие.

Atmosphere, Год журнала: 2024, Номер 15(11), С. 1348 - 1348

Опубликована: Ноя. 9, 2024

This study focuses on the impacts of climate change hydrological processes in watersheds and proposes an integrated approach combining a weather generator with multi-site conditional generative adversarial network (McGAN) model. The incorporates ensemble GCM predictions to generate regional average synthetic series, while McGAN transforms these averages into spatially consistent data. By addressing spatial consistency problem generating this tackles key challenge site-scale impact assessment. Applied Jinghe River Basin west-central China, generated daily temperature precipitation data for four stations under different shared socioeconomic pathways (SSP1-26, SSP2-45, SSP3-70, SSP5-85) up 2100. These were then used long short-term memory (LSTM) network, trained historical data, simulate river flow from 2021 results show that (1) effectively addresses correlation generation; (2) future is likely increase flow, particularly high-emission scenarios; (3) frequency extreme events may increase, proactive policies can mitigate flood drought risks. offers new tool hydrologic–climatic assessment studies.

Язык: Английский

Hydrological Response to Climate Change: McGAN for Multi-Site Scenario Weather Series Generation and LSTM for Streamflow Modeling DOI Creative Commons
Jian Sha,

Yaxin Chang,

Yaxiu Liu

и другие.

Atmosphere, Год журнала: 2024, Номер 15(11), С. 1348 - 1348

Опубликована: Ноя. 9, 2024

This study focuses on the impacts of climate change hydrological processes in watersheds and proposes an integrated approach combining a weather generator with multi-site conditional generative adversarial network (McGAN) model. The incorporates ensemble GCM predictions to generate regional average synthetic series, while McGAN transforms these averages into spatially consistent data. By addressing spatial consistency problem generating this tackles key challenge site-scale impact assessment. Applied Jinghe River Basin west-central China, generated daily temperature precipitation data for four stations under different shared socioeconomic pathways (SSP1-26, SSP2-45, SSP3-70, SSP5-85) up 2100. These were then used long short-term memory (LSTM) network, trained historical data, simulate river flow from 2021 results show that (1) effectively addresses correlation generation; (2) future is likely increase flow, particularly high-emission scenarios; (3) frequency extreme events may increase, proactive policies can mitigate flood drought risks. offers new tool hydrologic–climatic assessment studies.

Язык: Английский

Процитировано

0