We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Improving Boreal Summer Precipitation Predictions From the Global NMME Through Res34‐Unet.
- Authors
Tong, Xuan; Zhou, Wen; Xia, Jiangjiang
- Abstract
Global warming and climate change have increased the frequency and intensity of floods and droughts, limiting economic development and threatening human survival. Therefore, accurate global forecasts well in advance of precipitation are essential to facilitate timely adaptation. Current seasonal forecasts are based mainly on numerical models, but raw forecasts suffer from systematic bias and under/overdispersion problems and cannot be directly used in applications. In addition, bias correction methods for global forecasts need to be further developed. Based on a fusion of ResNet34 and Unet, called Res34‐Unet, deep learning post‐processing is proposed to correct global precipitation forecasts of the North American Multi‐Model Ensemble (NMME). Compared with raw global NMME predictions, post‐processed precipitation predictions can be improved by up to 45%, which is significant at different latitudes. Feature importance analysis shows that precipitation itself, meridional wind, and sea surface temperature are key factors. Plain Language Summary: Under the influence of global warming, extreme events, especially floods and droughts, are becoming more frequent, severely limiting social and economic development and greatly affecting people's lives. Therefore, accurate global forecasts are crucial to help people adaptively respond as soon as possible. Currently, seasonal forecasts are based mainly on numerical models. The North American Multi‐Model Ensemble (NMME) is one of the most advanced multi‐model ensemble systems for intra‐seasonal to interannual global prediction. However, due to deficiencies in numerical weather prediction models and inaccurate initial conditions, raw model predictions still have a large bias compared to observations. Therefore, accurate and efficient post‐processing algorithms are essential. Traditional post‐processing algorithms have some disadvantages such as the assumption of linearity, and spatial and temporal inconsistency. In addition, most post‐processing work has a regional focus, and more research focusing on global post‐processing prediction is needed to reduce global losses from extreme events. With the rapid development of machine learning, artificial intelligence is an effective tool to improve post‐processing. The fusion of ResNet34 and Unet, called Res34‐Unet, is proposed to post‐process global NMME rainfall predictions. Compared with raw NMME predictions, post‐processed predictions can be improved by up to 45%, which is significant at different latitudes. Key Points: Accurate global seasonal precipitation prediction is essential, but the bias of raw model predictions is largeRes34‐Unet post‐processing provides accurate seasonal precipitation prediction and outperforms Unet post‐processingRes34‐Unet post‐processing can improve predictions by up to 45%, which is significant at different latitudes
- Subjects
NUMERICAL weather forecasting; OCEAN temperature; MERIDIONAL winds; CLIMATE change; PRECIPITATION forecasting; DROUGHTS
- Publication
Geophysical Research Letters, 2024, Vol 51, Issue 2, p1
- ISSN
0094-8276
- Publication type
Article
- DOI
10.1029/2023GL106391