Climate-invariant machine learning.

Projecting climate change is a generalization problem: We extrapolate the recent past using physical models across past, present, and future climates. Current climate models require representations of processes that occur at scales smaller than model grid size, which have been the main source of model projection uncertainty. Recent machine learning (ML) algorithms hold promise to improve such process representations but tend to extrapolate poorly to climate regimes that they were not trained on. To get the best of the physical and statistical worlds, we propose a framework, termed "climate-invariant" ML, incorporating knowledge of climate processes into ML algorithms, and show that it can maintain high offline accuracy across a wide range of climate conditions and configurations in three distinct atmospheric models. Our results suggest that explicitly incorporating physical knowledge into data-driven models of Earth system processes can improve their consistency, data efficiency, and generalizability across climate regimes.

Towards advancing scientific knowledge of climate change impacts on short-duration rainfall extremes.

A large number of recent studies have aimed at understanding short-duration rainfall extremes, due to their impacts on flash floods, landslides and debris flows and potential for these to worsen with global warming. This has been led in a concerted international effort by the INTENSE Crosscutting Project of the GEWEX (Global Energy and Water Exchanges) Hydroclimatology Panel. Here, we summarize the main findings so far and suggest future directions for research, including: the benefits of convection-permitting climate modelling; towards understanding mechanisms of change; the usefulness of temperature-scaling relations; towards detecting and attributing extreme rainfall change; and the need for international coordination and collaboration. Evidence suggests that the intensity of long-duration (1 day+) heavy precipitation increases with climate warming close to the Clausius-Clapeyron (CC) rate (6-7% K-1), although large-scale circulation changes affect this response regionally. However, rare events can scale at higher rates, and localized heavy short-duration (hourly and sub-hourly) intensities can respond more strongly (e.g. 2 × CC instead of CC). Day-to-day scaling of short-duration intensities supports a higher scaling, with mechanisms proposed for this related to local-scale dynamics of convective storms, but its relevance to climate change is not clear. Uncertainty in changes to precipitation extremes remains and is influenced by many factors, including large-scale circulation, convective storm dynamics andstratification. Despite this, recent research has increased confidence in both the detectability and understanding of changes in various aspects of intense short-duration rainfall. To make further progress, the international coordination of datasets, model experiments and evaluations will be required, with consistent and standardized comparison methods and metrics, and recommendations are made for these frameworks. This article is part of a discussion meeting issue 'Intensification of short-duration rainfall extremes and implications for flash flood risks'.

Stable machine-learning parameterization of subgrid processes for climate modeling at a range of resolutions.

Global climate models represent small-scale processes such as convection using subgrid models known as parameterizations, and these parameterizations contribute substantially to uncertainty in climate projections. Machine learning of new parameterizations from high-resolution model output is a promising approach, but such parameterizations have been prone to issues of instability and climate drift, and their performance for different grid spacings has not yet been investigated. Here we use a random forest to learn a parameterization from coarse-grained output of a three-dimensional high-resolution idealized atmospheric model. The parameterization leads to stable simulations at coarse resolution that replicate the climate of the high-resolution simulation. Retraining for different coarse-graining factors shows the parameterization performs best at smaller horizontal grid spacings. Our results yield insights into parameterization performance across length scales, and they also demonstrate the potential for learning parameterizations from global high-resolution simulations that are now emerging.

Changing available energy for extratropical cyclones and associated convection in Northern Hemisphere summer.

The circulation of the Northern Hemisphere extratropical troposphere has changed over recent decades, with marked decreases in extratropical cyclone activity and eddy kinetic energy (EKE) in summer and increases in the fraction of precipitation that is convective in all seasons. Decreasing EKE in summer is partly explained by a weakening meridional temperature gradient, but changes in vertical temperature gradients and increasing moisture also affect the mean available potential energy (MAPE), which is the energetic reservoir from which extratropical cyclones draw. Furthermore, the relation of changes in mean thermal structure and moisture to changes in convection associated with extratropical cyclones is poorly understood. Here we calculate trends in MAPE for the Northern extratropics in summer over the years 1979-2017, and we decompose MAPE into both convective and nonconvective components. Nonconvective MAPE decreased over this period, consistent with decreases in EKE and extratropical cyclone activity, but convective MAPE increased, implying an increase in the energy available to convection. Calculations with idealized atmospheres indicate that nonconvective and convective MAPE both increase with increasing mean surface temperature and decrease with decreasing meridional surface temperature gradient, but convective MAPE is relatively more sensitive to the increase in mean surface temperature. These results connect changes in the atmospheric mean state with changes in both large-scale and convective circulations, and they suggest that extratropical cyclones can weaken even as their associated convection becomes more energetic.

Trends in continental temperature and humidity directly linked to ocean warming.

In recent decades, the land surface has warmed substantially more than the ocean surface, and relative humidity has fallen over land. Amplified warming and declining relative humidity over land are also dominant features of future climate projections, with implications for climate-change impacts. An emerging body of research has shown how constraints from atmospheric dynamics and moisture budgets are important for projected future land-ocean contrasts, but these ideas have not been used to investigate temperature and humidity records over recent decades. Here we show how both the temperature and humidity changes observed over land between 1979 and 2016 are linked to warming over neighboring oceans. A simple analytical theory, based on atmospheric dynamics and moisture transport, predicts equal changes in moist static energy over land and ocean and equal fractional changes in specific humidity over land and ocean. The theory is shown to be consistent with the observed trends in land temperature and humidity given the warming over ocean. Amplified land warming is needed for the increase in moist static energy over drier land to match that over ocean, and land relative humidity decreases because land specific humidity is linked via moisture transport to the weaker warming over ocean. However, there is considerable variability about the best-fit trend in land relative humidity that requires further investigation and which may be related to factors such as changes in atmospheric circulations and land-surface properties.

Climate research must sharpen its view.

Human activity is changing Earth's climate. Now that this has been acknowledged and accepted in international negotiations, climate research needs to define its next frontiers.

Precipitation Extremes Under Climate Change.

The response of precipitation extremes to climate change is considered using results from theory, modeling, and observations, with a focus on the physical factors that control the response. Observations and simulations with climate models show that precipitation extremes intensify in response to a warming climate. However, the sensitivity of precipitation extremes to warming remains uncertain when convection is important, and it may be higher in the tropics than the extratropics. Several physical contributions govern the response of precipitation extremes. The thermodynamic contribution is robust and well understood, but theoretical understanding of the microphysical and dynamical contributions is still being developed. Orographic precipitation extremes and snowfall extremes respond differently from other precipitation extremes and require particular attention. Outstanding research challenges include the influence of mesoscale convective organization, the dependence on the duration considered, and the need to better constrain the sensitivity of tropical precipitation extremes to warming.

Intramolecular azide to alkene cycloadditions for the construction of pyrrolobenzodiazepines and azetidino-benzodiazepines.

The coupling of proline- and azetidinone-substituted alkenes to 2-azidobenzoic and 2-azidobenzenesulfonic acid gives precursors that undergo intramolecular azide to alkene 1,3-dipolar cycloadditions to give imine-, triazoline- or aziridine-containing pyrrolo[1,4]benzodiazepines (PBDs), pyrrolo[1,2,5]benzothiadiazepines (PBTDs), and azetidino[1,4]benzodiazepines. The imines and aziridines are formed after loss of nitrogen from a triazoline cycloadduct. The PBDs are a potent class of antitumour antibiotics.

Contrasting responses of mean and extreme snowfall to climate change.

Snowfall is an important element of the climate system, and one that is expected to change in a warming climate. Both mean snowfall and the intensity distribution of snowfall are important, with heavy snowfall events having particularly large economic and human impacts. Simulations with climate models indicate that annual mean snowfall declines with warming in most regions but increases in regions with very low surface temperatures. The response of heavy snowfall events to a changing climate, however, is unclear. Here I show that in simulations with climate models under a scenario of high emissions of greenhouse gases, by the late twenty-first century there are smaller fractional changes in the intensities of daily snowfall extremes than in mean snowfall over many Northern Hemisphere land regions. For example, for monthly climatological temperatures just below freezing and surface elevations below 1,000 metres, the 99.99th percentile of daily snowfall decreases by 8% in the multimodel median, compared to a 65% reduction in mean snowfall. Both mean and extreme snowfall must decrease for a sufficiently large warming, but the climatological temperature above which snowfall extremes decrease with warming in the simulations is as high as -9 °C, compared to -14 °C for mean snowfall. These results are supported by a physically based theory that is consistent with the observed rain-snow transition. According to the theory, snowfall extremes occur near an optimal temperature that is insensitive to climate warming, and this results in smaller fractional changes for higher percentiles of daily snowfall. The simulated changes in snowfall that I find would influence surface snow and its hazards; these changes also suggest that it may be difficult to detect a regional climate-change signal in snowfall extremes.

Understanding the varied response of the extratropical storm tracks to climate change.

Transient eddies in the extratropical storm tracks are a primary mechanism for the transport of momentum, energy, and water in the atmosphere, and as such are a major component of the climate system. Changes in the extratropical storm tracks under global warming would impact these transports, the ocean circulation and carbon cycle, and society through changing weather patterns. I show that the southern storm track intensifies in the multimodel mean of simulations of 21st century climate change, and that the seasonal cycle of storm-track intensity increases in amplitude in both hemispheres. I use observations of the present-day seasonal cycle to confirm the relationship between storm-track intensity and the mean available potential energy of the atmosphere, and show how this quantitative relationship can be used to account for much of the varied response in storm-track intensity to global warming, including substantially different responses in simulations with different climate models. The results suggest that storm-track intensity is not related in a simple way to global-mean surface temperature, so that, for example, a stronger southern storm track in response to present-day global warming does not imply it was also stronger in hothouse climates of the past.

The physical basis for increases in precipitation extremes in simulations of 21st-century climate change.

Global warming is expected to lead to a large increase in atmospheric water vapor content and to changes in the hydrological cycle, which include an intensification of precipitation extremes. The intensity of precipitation extremes is widely held to increase proportionately to the increase in atmospheric water vapor content. Here, we show that this is not the case in 21st-century climate change scenarios simulated with climate models. In the tropics, precipitation extremes are not simulated reliably and do not change consistently among climate models; in the extratropics, they consistently increase more slowly than atmospheric water vapor content. We give a physical basis for how precipitation extremes change with climate and show that their changes depend on changes in the moist-adiabatic temperature lapse rate, in the upward velocity, and in the temperature when precipitation extremes occur. For the tropics, the theory suggests that improving the simulation of upward velocities in climate models is essential for improving predictions of precipitation extremes; for the extratropics, agreement with theory and the consistency among climate models increase confidence in the robustness of predictions of precipitation extremes under climate change.