A review of the experimental protocol and motivation for DYAMOND, the first intercomparison project of global storm-resolving models, is presented. Nine models submitted simulation output for a 40-day (1 August–10 September 2016) intercomparison period. Eight of these employed a tiling of the sphere that was uniformly less than 5 km. By resolving the transient dynamics of convective storms in the tropics, global storm-resolving models remove the need to parameterize tropical deep convection, providing a fundamentally more sound representation of the climate system and a more natural link to commensurately high-resolution data from satellite-borne sensors. The models and some basic characteristics of their output are described in more detail, as is the availability and planned use of this output for future scientific study. Tropically and zonally averaged energy budgets, precipitable water distributions, and precipitation from the model ensemble are evaluated, as is their representation of tropical cyclones and the predictability of column water vapor, the latter being important for tropical weather.

More than one hundred days were simulated over very large domains with fine (0.156 km to 2.5 km) grid spacing for realistic conditions to test the hypothesis that storm (kilometer) and large-eddy (hectometer) resolving simulations would provide an improved representation of clouds and precipitation in atmospheric simulations. At scales that resolve convective storms (storm-resolving for short) scales, the vertical velocity variance becomes resolved and a better physical basis is achieved for representing clouds and precipitation. Similar to past studies we find an improved representation of precipitation at kilometer scales, as compared to models with parameterised convection. The main precipitation features (location, diurnal cycle and spatial propagation) are well captured already at kilometer scales, and refining resolution to hectometer scales does not substantially change the simulations in these respects. It does, however, lead to a reduction in the precipitation on the timescales considered-most notably over the Tropical ocean. Changes in the distribution of precipitation, with less frequent extremes are also found in simulations incorporating hecto-meter scales. Hectometer scales appear more important for the representation of clouds, and make it possible to capture many important aspects of the cloud field, from the vertical distribution of cloud cover, to the distribution of cloud sizes, to the diel (daily) cycle. Qualitative improvements, particularly in the ability to differentiate cumulus from stratiform clouds, are seen when reducing the grid spacing from kilometer to hectometer scales. At the hectometer scale new challenges arise, but the similarity of observed and simulated scales, and the more direct 1 connection between the circulation and the unconstrained degrees of freedom make these challenges less daunting. This quality, combined with an already improved simulation as compared to more parameterised models, underpins our conviction that the use and further development of storm-resolving models offers exciting opportunities for advancing understanding of climate and climate change.

We discuss scientific features and computational performance of kilometre-scale global weather and climate simulations, considering the Icosahedral Non-hydrostatic (ICON) model and the Integrated Forecast System (IFS). Scalability measurements and a performance modelling approach are used to derive performance estimates for these models on upcoming exascale supercomputers. This is complemented by preliminary analyses of the model data that illustrate the importance of high-resolution models to gain improvements in the accuracy of convective processes, a better understanding of physics dynamics interactions and poorly resolved or parametrized processes, such as gravity waves, convection and boundary layer. This article is part of the theme issue ‘Multiscale modelling, simulation and computing: from the desktop to the exascale’.

Abstract. A climate model represents a multitude of processes on a variety of timescales and space scales: a canonical example of multi-physics multi-scale modeling. The underlying climate system is physically characterized by sensitive dependence on initial conditions, and natural stochastic variability, so very long integrations are needed to extract signals of climate change. Algorithms generally possess weak scaling and can be I/O and/or memory-bound. Such weakscaling, I/O, and memory-bound multi-physics codes present particular challenges to computational performance.Traditional metrics of computational efficiency such as performance counters and scaling curves do not tell us enough about real sustained performance from climate models on different machines. They also do not provide a satisfactory basis for comparative information across models.We introduce a set of metrics that can be used for the study of computational performance of climate (and Earth system) models. These measures do not require specialized software or specific hardware counters, and should be accessible to anyone. They are independent of platform and underlying parallel programming models. We show how these metrics can be used to measure actually attained performance of Earth system models on different machines, and identify the most fruitful areas of research and development for performance engineering.We present results for these measures for a diverse suite of models from several modeling centers, and propose to use these measures as a basis for a CPMIP, a computational performance model intercomparison project (MIP).

Abstract. State-of-the-art Earth system models typically employ grid spacings of O(100 km), which is too coarse to explicitly resolve main drivers of the flow of energy and matter across the Earth system. In this paper, we present the new ICON-Sapphire model configuration, which targets a representation of the components of the Earth system and their interactions with a grid spacing of 10 km and finer. Through the use of selected simulation examples, we demonstrate that ICON-Sapphire can (i) be run coupled globally on seasonal timescales with a grid spacing of 5 km, on monthly timescales with a grid spacing of 2.5 km, and on daily timescales with a grid spacing of 1.25 km; (ii) resolve large eddies in the atmosphere using hectometer grid spacings on limited-area domains in atmosphere-only simulations; (iii) resolve submesoscale ocean eddies by using a global uniform grid of 1.25 km or a telescoping grid with the finest grid spacing at 530 m, the latter coupled to a uniform atmosphere; and (iv) simulate biogeochemistry in an ocean-only simulation integrated for 4 years at 10 km. Comparison of basic features of the climate system to observations reveals no obvious pitfalls, even though some observed aspects remain difficult to capture. The throughput of the coupled 5 km global simulation is 126 simulated days per day employing 21 % of the latest machine of the German Climate Computing Center. Extrapolating from these results, multi-decadal global simulations including interactive carbon are now possible, and short global simulations resolving large eddies in the atmosphere and submesoscale eddies in the ocean are within reach.