Extremes · AI and machine learning
New weather and climate research on extremes with AI and machine learning as the main approach, one page per paper.
All summaries
-
ExtremesReview: deep learning helps spot potential landslides but needs fused data and physics
Deep learning is widely used to find slopes at risk of failing, but data scarcity, weak interpretability and complex landslide physics limit it, so the authors urge fusing many data sources and combining models with physical knowledge.
-
ExtremesAdding a deep-learning extreme weather factor improves WOFOST wheat simulations
Feeding extreme weather indices through an LSTM into the wheat development rate of WOFOST cut errors in heading, maturity and yield on the North China Plain, including in extreme years.
-
ExtremesA deep learning model outlines nighttime urban floods better than six other models tested
The NWseg model, built for low-light images, outlined flooded areas in nighttime street camera images more accurately than the six other segmentation models the authors compared it with.
-
ExtremesRainfall-based drought index is the top predictor of hydrological drought in the Huaihe Basin
An XGBoost model predicted next-month hydrological drought category correctly 79.9 % of the time in the Huaihe River Basin, and the precipitation drought index (SPI) mattered most, though it was weak on severe drought.