دورية أكاديمية
Deep learning rainfall–runoff predictions of extreme events
العنوان: | Deep learning rainfall–runoff predictions of extreme events |
---|---|
المؤلفون: | J. M. Frame, F. Kratzert, D. Klotz, M. Gauch, G. Shelev, O. Gilon, L. M. Qualls, H. V. Gupta, G. S. Nearing |
المصدر: | Hydrology and Earth System Sciences, Vol 26, Pp 3377-3392 (2022) |
بيانات النشر: | Copernicus Publications, 2022. |
سنة النشر: | 2022 |
المجموعة: | LCC:Technology LCC:Environmental technology. Sanitary engineering LCC:Geography. Anthropology. Recreation LCC:Environmental sciences |
مصطلحات موضوعية: | Technology, Environmental technology. Sanitary engineering, TD1-1066, Geography. Anthropology. Recreation, Environmental sciences, GE1-350 |
الوصف: | The most accurate rainfall–runoff predictions are currently based on deep learning. There is a concern among hydrologists that the predictive accuracy of data-driven models based on deep learning may not be reliable in extrapolation or for predicting extreme events. This study tests that hypothesis using long short-term memory (LSTM) networks and an LSTM variant that is architecturally constrained to conserve mass. The LSTM network (and the mass-conserving LSTM variant) remained relatively accurate in predicting extreme (high-return-period) events compared with both a conceptual model (the Sacramento Model) and a process-based model (the US National Water Model), even when extreme events were not included in the training period. Adding mass balance constraints to the data-driven model (LSTM) reduced model skill during extreme events. |
نوع الوثيقة: | article |
وصف الملف: | electronic resource |
اللغة: | English |
تدمد: | 1027-5606 1607-7938 |
Relation: | https://hess.copernicus.org/articles/26/3377/2022/hess-26-3377-2022.pdf; https://doaj.org/toc/1027-5606; https://doaj.org/toc/1607-7938 |
DOI: | 10.5194/hess-26-3377-2022 |
URL الوصول: | https://doaj.org/article/1ef8b1f9cdd0413c9f0aea6a8ad269f1 |
رقم الأكسشن: | edsdoj.1ef8b1f9cdd0413c9f0aea6a8ad269f1 |
قاعدة البيانات: | Directory of Open Access Journals |
تدمد: | 10275606 16077938 |
---|---|
DOI: | 10.5194/hess-26-3377-2022 |