Version 1
: Received: 20 September 2024 / Approved: 22 September 2024 / Online: 23 September 2024 (11:59:18 CEST)
How to cite:
Mendoza Paz, S.; Willems, P.; Villazón Gómez, M. F. Adapting to Climate Change with Machine Learning: The Robustness of Downscaled Precipitation in Local Impact Analysis. Preprints2024, 2024091706. https://doi.org/10.20944/preprints202409.1706.v1
Mendoza Paz, S.; Willems, P.; Villazón Gómez, M. F. Adapting to Climate Change with Machine Learning: The Robustness of Downscaled Precipitation in Local Impact Analysis. Preprints 2024, 2024091706. https://doi.org/10.20944/preprints202409.1706.v1
Mendoza Paz, S.; Willems, P.; Villazón Gómez, M. F. Adapting to Climate Change with Machine Learning: The Robustness of Downscaled Precipitation in Local Impact Analysis. Preprints2024, 2024091706. https://doi.org/10.20944/preprints202409.1706.v1
APA Style
Mendoza Paz, S., Willems, P., & Villazón Gómez, M. F. (2024). Adapting to Climate Change with Machine Learning: The Robustness of Downscaled Precipitation in Local Impact Analysis. Preprints. https://doi.org/10.20944/preprints202409.1706.v1
Chicago/Turabian Style
Mendoza Paz, S., Patrick Willems and Mauricio F. Villazón Gómez. 2024 "Adapting to Climate Change with Machine Learning: The Robustness of Downscaled Precipitation in Local Impact Analysis" Preprints. https://doi.org/10.20944/preprints202409.1706.v1
Abstract
We assessed machine learning techniques (MLTs) for downscaling global climate model’s precipitation to local level in Bolivia. The skill, assumptions and uncertainty of the process are analyzed. Additionally, future projections are delivered. For that, an ensemble of 20 global climate models (GCMs) from CMIP6, with random forest (RF) and support vector machines (SVM), is used on four zones (highlands, Andean slopes, Amazon lowlands and Chaco lowlands). The downscaled series’ skill is evaluated in terms of relative errors. The uncertainty is analyzed through variance decomposition. MLTs’ skill is adequate in most cases, especially for the highlands and the Andean slopes. Moreover, RF tends to outperform SVM. Stationary assumptions are robust in the highlands and Andean slopes. Perfect prognosis assumption is poor in the highlands and Andean slopes; topographical complexity appears to be the reason. Regarding uncertainties, MLTs are the dominant source. In future projections, the highlands show shorter dry spell lengths with more frequent but less intense high rainfall events and higher annual rainfall. The Andean slopes exhibit an increased annual rainfall with a reduction in the high precipitation intensities but an increased frequency. The Amazon lowlands present a decrease in annual rainfall, and the Chaco lowlands an increase.
Keywords
Climate Change; Precipitation; Statistical Downscaling; Machine Learning; Uncertainty; The Andes
Subject
Environmental and Earth Sciences, Water Science and Technology
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.