Assessing Predictability of Marine Heatwaves With Random Forests. Issue 23 (9th December 2022)

• Record Type: Journal Article
• Authors: Giamalaki, K.; Beaulieu, C.; Prochaska, J. X.
• Is Part Of: Geophysical research letters. Volume 49:Issue 23(2022)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-12-09
• Subjects: Geophysics -- Periodicals; Planets -- Periodicals; Lunar geology -- Periodicals; 550
• DOI: 10.1029/2022GL099069
• Languages: English
• ISSNs: 0094-8276

Abstract

Abstract: Marine heatwaves (MHWs) have increased in frequency and duration over the last century and are expected to intensify in the future. Such events have become an increasing threat for marine ecosystems and subsequently the economies and populations that rely on them. Here we apply random forests to assess skill in forecasting MHWs onset and severity at multiple prediction lead times. Random forests models are trained on a range of atmospheric and oceanic conditions to identify precursors of MHWs. The best performing random forest model accurately captures (76%) MHW presence/absence in the northeast Pacific and is capable of forecasting realistic extreme sea surface temperature patterns at weekly lead times. However, the total accuracy drops to 38% when forecasting MHW severity. Machine learning algorithms affirm further exploration as forecasting tools and have the potential to accelerate our predictive ability and preparedness against upcoming extreme climate changes. Plain Language Summary: Marine heatwaves (MHWs) are defined as extremely high ocean temperatures that last for at least five consecutive days, threatening marine life and environment and the subsequent economies and populations that rely on them. MHWs are expected to become more intense and frequent in the future and therefore their predictability is crucial for the effective management of the marine environment and preparedness of coastal communities. Machine learning models based on algorithms are able to analyze and draw inference from patterns in data without following explicit instructions. These tools have shown promise in predicting a range of climate extreme events. Here we assess predictability of MHWs in the northeast Pacific using a Random Forest machine learning model. We train the model to recognize air‐sea patterns that may act as precursors to MHWs at different time lags. Our model predicts the presence or absence of MHWs with 76% accuracy on weekly lead times. Machine learning models show promise to advance our forecasting ability of upcoming MHWs. Key Points: Marine heatwaves in the northeast Pacific are predictable on weekly lead times using atmospheric and oceanic conditions Average accuracy at predicting marine heatwave presence/absence is 76%, but lowers to 38% when forecasting their severity Machine learning approaches should be considered to augment our ability to forecast marine heatwaves