ENSO and greenhouse warming

The El Niño/Southern Oscillation (ENSO) is the dominant climate phenomenon affecting extreme weather conditions worldwide. Its response to greenhouse warming has challenged scientists for decades, despite model agreement on projected changes in mean state. Recent studies have provided new insights into the elusive links between changes in ENSO and in the mean state of the Pacific climate. The projected slow-down in Walker circulation is expected to weaken equatorial Pacific Ocean currents, boosting the occurrences of eastward-propagating warm surface anomalies that characterize observed extreme El Niño events. Accelerated equatorial Pacific warming, particularly in the east, is expected to induce extreme rainfall in the eastern equatorial Pacific and extreme equatorward swings of the Pacific convergence zones, both of which are features of extreme El Niño. The frequency of extreme La Niña is also expected to increase in response to more extreme El Niños, an accelerated maritime continent warming and surface-intensified ocean warming. ENSO-related catastrophic weather events are thus likely to occur more frequently with unabated greenhouse-gas emissions. But model biases and recent observed strengthening of the Walker circulation highlight the need for further testing as new models, observations and insights become available.

Increased frequency of extreme La Niña events under greenhouse warming

The El Niño/Southern Oscillation is Earth’s most prominent source of interannual climate variability, alternating irregularly between El Niño and La Niña, and resulting in global disruption of weather patterns, ecosystems, fisheries and agriculture1, 2, 3, 4, 5. The 1998–1999 extreme La Niña event that followed the 1997–1998 extreme El Niño event6 switched extreme El Niño-induced severe droughts to devastating floods in western Pacific countries, and vice versa in the southwestern United States4, 7. During extreme La Niña events, cold sea surface conditions develop in the central Pacific8, 9, creating an enhanced temperature gradient from the Maritime continent to the central Pacific. Recent studies have revealed robust changes in El Niño characteristics in response to simulated future greenhouse warming10, 11, 12, but how La Niña will change remains unclear. Here we present climate modelling evidence, from simulations conducted for the Coupled Model Intercomparison Project phase 5 (ref. 13), for a near doubling in the frequency of future extreme La Niña events, from one in every 23 years to one in every 13 years. This occurs because projected faster mean warming of the Maritime continent than the central Pacific, enhanced upper ocean vertical temperature gradients, and increased frequency of extreme El Niño events are conducive to development of the extreme La Niña events. Approximately 75% of the increase occurs in years following extreme El Niño events, thus projecting more frequent swings between opposite extremes from one year to the next.

Multiprobabilistic prediction in early medical diagnoses

This paper describes the methodology of providing multiprobability predictions for proteomic mass spectrometry data. The methodology is based on a newly developed machine learning framework called Venn machines. Is allows to output a valid probability interval. The methodology is designed for mass spectrometry data. For demonstrative purposes, we applied this methodology to MALDI-TOF data sets in order to predict the diagnosis of heart disease and early diagnoses of ovarian cancer and breast cancer. The experiments showed that probability intervals are narrow, that is, the output of the multiprobability predictor is similar to a single probability distribution. In addition, probability intervals produced for heart disease and ovarian cancer data were more accurate than the output of corresponding probability predictor. When Venn machines were forced to make point predictions, the accuracy of such predictions is for the most data better than the accuracy of the underlying algorithm that outputs single probability distribution of a label. Application of this methodology to MALDI-TOF data sets empirically demonstrates the validity. The accuracy of the proposed method on ovarian cancer data rises from 66.7 % 11 months in advance of the moment of diagnosis to up to 90.2 % at the moment of diagnosis. The same approach has been applied to heart disease data without time dependency, although the achieved accuracy was not as high (up to 69.9 %). The methodology allowed us to confirm mass spectrometry peaks previously identified as carrying statistically significant information for discrimination between controls and cases.