Global climate simulations at 2.8 km on GPU with the ICON model

2020 ◽  
Author(s):  
Xavier Lapillonne ◽  
William Sawyer ◽  
Philippe Marti ◽  
Valentin Clement ◽  
Remo Dietlicher ◽  
...  
Keyword(s):  
High Resolution ◽  
Large Scale ◽  
Climate Model ◽  
Global Climate ◽  
Weather Prediction ◽  
Time Integration ◽  
Horizontal Resolution ◽  
Performance Comparison ◽  
Quasi Biennial Oscillation ◽  
Numerical Weather

<p>The ICON modelling framework is a unified numerical weather and climate model used for applications ranging from operational numerical weather prediction to low and high resolution climate projection. In view of further pushing the frontier of possible applications and to make use of the latest evolution in hardware technologies, parts of the model were recently adapted to run on heterogeneous GPU system. This initial GPU port focus on components required for high-resolution climate application, and allow considering multi-years simulations at 2.8 km on the Piz Daint heterogeneous supercomputer. These simulations are planned as part of the QUIBICC project “The Quasi-Biennial Oscillation (QBO) in a changing climate”, which propose to investigate effects of climate change on the dynamics of the QBO.</p><p>Because of the low compute intensity of atmospheric model the cost of data transfer between CPU and GPU at every step of the time integration would be prohibitive if only some components would be ported to the accelerator. We therefore present a full port strategy where all components required for the simulations are running on the GPU. For the dynamics, most of the physical parameterizations and infrastructure code the OpenACC compiler directives are used. For the soil parameterization, a Fortran based domain specific language (DSL) the CLAW-DSL has been considered. We discuss the challenges associated to port a large community code, about 1 million lines of code, as well as to run simulations on large-scale system at 2.8 km horizontal resolution in terms of run time and I/O constraints. We show performance comparison of the full model on CPU and GPU, achieving a speed up factor of approximately 5x, as well as scaling results on up to 2000 GPU nodes. Finally we discuss challenges and planned development regarding performance portability and high level DSL which will be used with the ICON model in the near future.</p>

scholarly journals Climate SPHINX: evaluating the impact of resolution and stochastic physics parameterisations in the EC-Earth global climate model

2017 ◽  
Vol 10 (3) ◽  
pp. 1383-1402 ◽  
Author(s):  
Paolo Davini ◽  
Jost von Hardenberg ◽  
Susanna Corti ◽  
Hannah M. Christensen ◽  
Stephan Juricke ◽  
...  
Keyword(s):  
High Resolution ◽  
Large Scale ◽  
Climate Model ◽  
Global Climate ◽  
Rainfall Variability ◽  
Global Climate Model ◽  
Small Scale ◽  
Low Resolution ◽  
The Impact ◽  
Stochastic Physics

Abstract. The Climate SPHINX (Stochastic Physics HIgh resolutioN eXperiments) project is a comprehensive set of ensemble simulations aimed at evaluating the sensitivity of present and future climate to model resolution and stochastic parameterisation. The EC-Earth Earth system model is used to explore the impact of stochastic physics in a large ensemble of 30-year climate integrations at five different atmospheric horizontal resolutions (from 125 up to 16 km). The project includes more than 120 simulations in both a historical scenario (1979–2008) and a climate change projection (2039–2068), together with coupled transient runs (1850–2100). A total of 20.4 million core hours have been used, made available from a single year grant from PRACE (the Partnership for Advanced Computing in Europe), and close to 1.5 PB of output data have been produced on SuperMUC IBM Petascale System at the Leibniz Supercomputing Centre (LRZ) in Garching, Germany. About 140 TB of post-processed data are stored on the CINECA supercomputing centre archives and are freely accessible to the community thanks to an EUDAT data pilot project. This paper presents the technical and scientific set-up of the experiments, including the details on the forcing used for the simulations performed, defining the SPHINX v1.0 protocol. In addition, an overview of preliminary results is given. An improvement in the simulation of Euro-Atlantic atmospheric blocking following resolution increase is observed. It is also shown that including stochastic parameterisation in the low-resolution runs helps to improve some aspects of the tropical climate – specifically the Madden–Julian Oscillation and the tropical rainfall variability. These findings show the importance of representing the impact of small-scale processes on the large-scale climate variability either explicitly (with high-resolution simulations) or stochastically (in low-resolution simulations).

Physics, Resolution and Data Assimilation: Making sense of Greenland climate and ice sheet Surface Mass Balance with HARMONIE Climate

2021 ◽  
Author(s):  
Ruth Mottram ◽  
Oskar Landgren ◽  
Rasmus Anker Pedersen ◽  
Kristian Pagh Nielsen ◽  
Ole Bøssing Christensen ◽  
...  
Keyword(s):  
High Resolution ◽  
Data Assimilation ◽  
Numerical Weather Prediction ◽  
Climate Model ◽  
Weather Prediction ◽  
Ice Sheet ◽  
Greenland Ice Sheet ◽  
Surface Mass ◽  
Mass Budget ◽  
Numerical Weather

<p>The development of the HARMONIE model system has led to huge advances in numerical weather prediction, including over Greenland where a numerical weather prediction (NWP) model is used to forecast daily surface mass budget over the Greenland ice sheet as presented on polarportal.dk. The new high resolution Copernicus Arctic Reanalysis further developed the possibilities in HARMONIE with full 3DVar data assimilation and extended use of quality-controlled local observations. Here, we discuss the development and current status of the climate version of the HARMONIE Climate model (HCLIM). The HCLIM system has opened up the possibility for flexible use of the model at a range of spatial scales using different physical schemes including HARMONIE-AROME, ALADIN and ALARO for different spatial and temporal resolutions and assimilating observations, including satellite data on sea ice concentration from ESA CCI+, to improve hindcasts. However, the range of possibilities means that documenting the effects of different physics and parameterisation schemes is important before widespread application. </p><p>Here, we focus on HCLIM performance over the Greenland ice sheet, using observations to verify the different plausible set-ups and investigate biases in climate model outputs that affect the surface mass budget (SMB) of the Greenland ice sheet. </p><p>The recently funded Horizon 2020 project PolarRES will use the HCLIM model for very high resolution regional downscaling, together with other regional climate models in both Arctic and Antarctic regions, and our analysis thus helps to optimise the use of HCLIM in the polar regions for different modelling purposes.</p>

scholarly journals The role of serial European windstorm clustering for extreme seasonal losses as determined from multi-centennial simulations of high-resolution global climate model data

2018 ◽  
Vol 18 (11) ◽  
pp. 2991-3006 ◽  
Author(s):  
Matthew D. K. Priestley ◽  
Helen F. Dacre ◽  
Len C. Shaffrey ◽  
Kevin I. Hodges ◽  
Joaquim G. Pinto
Keyword(s):  
High Resolution ◽  
Return Period ◽  
Large Scale ◽  
Climate Model ◽  
Global Climate ◽  
Global Climate Model ◽  
Geographic Region ◽  
Near Surface ◽  
Short Period

Abstract. Extratropical cyclones are the most damaging natural hazard to affect western Europe. Serial clustering occurs when many intense cyclones affect one specific geographic region in a short period of time which can potentially lead to very large seasonal losses. Previous studies have shown that intense cyclones may be more likely to cluster than less intense cyclones. We revisit this topic using a high-resolution climate model with the aim to determine how important clustering is for windstorm-related losses. The role of windstorm clustering is investigated using a quantifiable metric (storm severity index, SSI) that is based on near-surface meteorological variables (10 m wind speed) and is a good proxy for losses. The SSI is used to convert a wind footprint into losses for individual windstorms or seasons. 918 years of a present-day ensemble of coupled climate model simulations from the High-Resolution Global Environment Model (HiGEM) are compared to ERA-Interim reanalysis. HiGEM is able to successfully reproduce the wintertime North Atlantic/European circulation, and represent the large-scale circulation associated with the serial clustering of European windstorms. We use two measures to identify any changes in the contribution of clustering to the seasonal windstorm loss as a function of return period. Above a return period of 3 years, the accumulated seasonal loss from HiGEM is up to 20 % larger than the accumulated seasonal loss from a set of random resamples of the HiGEM data. Seasonal losses are increased by 10 %–20 % relative to randomized seasonal losses at a return period of 200 years. The contribution of the single largest event in a season to the accumulated seasonal loss does not change with return period, generally ranging between 25 % and 50 %. Given the realistic dynamical representation of cyclone clustering in HiGEM, and comparable statistics to ERA-Interim, we conclude that our estimation of clustering and its dependence on the return period will be useful for informing the development of risk models for European windstorms, particularly for longer return periods.

scholarly journals Increasing horizontal resolution in numerical weather prediction and climate simulations: illusion or panacea?

2014 ◽  
Vol 372 (2018) ◽  
pp. 20130289 ◽  
Author(s):  
Nils P. Wedi
Keyword(s):  
High Resolution ◽  
Numerical Weather Prediction ◽  
Weather Prediction ◽  
Space Versus ◽  
Physical Space ◽  
Horizontal Resolution ◽  
Spectral Space ◽  
Numerical Weather ◽  
Climate Simulations ◽  
Forecast Efficiency

The steady path of doubling the global horizontal resolution approximately every 8 years in numerical weather prediction (NWP) at the European Centre for Medium Range Weather Forecasts may be substan- tially altered with emerging novel computing architectures. It coincides with the need to appropriately address and determine forecast uncertainty with increasing resolution, in particular, when convective-scale motions start to be resolved. Blunt increases in the model resolution will quickly become unaffordable and may not lead to improved NWP forecasts. Consequently, there is a need to accordingly adjust proven numerical techniques. An informed decision on the modelling strategy for harnessing exascale, massively parallel computing power thus also requires a deeper understanding of the sensitivity to uncertainty—for each part of the model—and ultimately a deeper understanding of multi-scale interactions in the atmosphere and their numerical realization in ultra-high-resolution NWP and climate simulations. This paper explores opportunities for substantial increases in the forecast efficiency by judicious adjustment of the formal accuracy or relative resolution in the spectral and physical space. One path is to reduce the formal accuracy by which the spectral transforms are computed. The other pathway explores the importance of the ratio used for the horizontal resolution in gridpoint space versus wavenumbers in spectral space. This is relevant for both high-resolution simulations as well as ensemble-based uncertainty estimation.

scholarly journals Combined use of volume radar observations and high-resolution numerical weather predictions to estimate precipitation at the ground: methodology and proof of concept

2019 ◽  
Vol 12 (10) ◽  
pp. 5669-5684 ◽  
Author(s):  
Tony Le Bastard ◽  
Olivier Caumont ◽  
Nicolas Gaussiat ◽  
Fatima Karbou
Keyword(s):  
High Resolution ◽  
Weather Prediction ◽  
Ground Level ◽  
Radar Data ◽  
Horizontal Resolution ◽  
Computationally Efficient ◽  
New Approach ◽  
Numerical Weather ◽  
Simulated Precipitation ◽  
Nwp Model

Abstract. The extrapolation of the precipitation to the ground from radar reflectivities measured at the beam altitude is one of the most delicate phases of radar data processing for producing quantitative precipitation estimations (QPEs) and remains a major scientific issue. In many operational meteorological services such as Météo-France, a vertical profile of reflectivity (VPR) correction is uniformly applied over a large part or the entire radar domain. This method is computationally efficient, and the overall bias induced by the bright band is most of the time well corrected. However, this way of proceeding is questionable in situations with high spatial and vertical variability of precipitation (during the passage of a cold front or in a complex terrain, for example). This study initiates from two statements: first, radars provide information on precipitation with a high spatio-temporal resolution but still require VPR corrections to extrapolate rain rates at the ground level. Second, the horizontal resolution of some numerical weather prediction (NWP) models is now comparable with the radar one, and their dynamical core and microphysics schemes allow the production of realistic simulations of VPRs. The present paper proposes a new approach to assess surface rainfall from radar reflectivity aloft by exploiting simulated VPRs and rainfall forecasts from the high-resolution NWP model AROME-NWC. To our knowledge, this is the first time that simulated precipitation profiles from an NWP model are used to derive radar QPEs. The implementation of the new method on two stratiform situations provided significant improvements on the hourly and 6 h accumulations compared to the operational QPEs, showing the relevance of this new approach.

scholarly journals Climatology of Tibetan Plateau Vortices in Reanalysis Data and a High-Resolution Global Climate Model

2019 ◽  
Vol 32 (6) ◽  
pp. 1933-1950 ◽  
Author(s):  
Julia Curio ◽  
Reinhard Schiemann ◽  
Kevin I. Hodges ◽  
Andrew G. Turner
Keyword(s):  
High Resolution ◽  
Tibetan Plateau ◽  
Climate Model ◽  
Global Climate ◽  
Global Climate Model ◽  
Horizontal Resolution ◽  
Northwestern Part ◽  
The Tibetan Plateau ◽  
Wide Range ◽  
Tibetan Plateau Vortices

The Tibetan Plateau (TP) and surrounding high mountains constitute an important forcing of the atmospheric circulation due to their height and extent, and thereby impact weather and climate in downstream regions of East Asia. Mesoscale Tibetan Plateau vortices (TPVs) are one of the major precipitation-producing systems on the TP. A fraction of TPVs move off the TP to the east and can trigger extreme precipitation in parts of China, such as the Sichuan province and the Yangtze River valley, which can result in severe flooding. In this study, the climatology of TPV occurrence is examined in two reanalyses and, for the first time, in a high-resolution global climate model using an objective feature tracking algorithm. Most TPVs are generated in the northwestern part of the TP; the center of this main genesis region is small and stable throughout the year. The strength and position of the subtropical westerly jet is correlated to the distance TPVs can travel eastward and therefore could have an effect on whether or not a TPV is moving off the TP. TPV-associated precipitation can account for up to 40% of the total precipitation in parts of China in selected months, often due to individual TPVs. The results show that the global climate model is able to simulate TPVs at N512 (~25 km) horizontal resolution and in general agrees with the reanalyses. The fact that the global climate model can represent the TPV climatology opens a wide range of options for future model-based research on TPVs.

scholarly journals Evaluation of extreme sub-daily precipitation in high-resolution global climate model simulations

2021 ◽  
Vol 379 (2195) ◽  
pp. 20190545 ◽  
Author(s):  
Michael Wehner ◽  
Jiwoo Lee ◽  
Mark Risser ◽  
Paul Ullrich ◽  
Peter Gleckler ◽  
...  
Keyword(s):  
High Resolution ◽  
Extreme Precipitation ◽  
Climate Models ◽  
Climate Model ◽  
Daily Precipitation ◽  
Flash Flood ◽  
Global Climate ◽  
Global Climate Model ◽  
Model Performance ◽  
Horizontal Resolution

We examine the resolution dependence of errors in extreme sub-daily precipitation in available high-resolution climate models. We find that simulated extreme precipitation increases as horizontal resolution increases but that appropriately constructed model skill metrics do not significantly change. We find little evidence that simulated extreme winter or summer storm processes significantly improve with the resolution because the model performance changes identified are consistent with expectations from scale dependence arguments alone. We also discuss the implications of these scale-dependent limitations on the interpretation of simulated extreme precipitation. This article is part of a discussion meeting issue ‘Intensification of short-duration rainfall extremes and implications for flash flood risks’.

scholarly journals Tropical Cyclone Intensity Errors Associated with Lack of Two-Way Ocean Coupling in High-Resolution Global Simulations

2016 ◽  
Vol 29 (23) ◽  
pp. 8589-8610 ◽  
Author(s):  
Colin M. Zarzycki
Keyword(s):  
High Resolution ◽  
Large Scale ◽  
Climate Models ◽  
Global Climate ◽  
Computational Cost ◽  
Horizontal Resolution ◽  
Global Climate Models ◽  
Surface Boundary ◽  
Community Atmosphere Model ◽  
Slab Ocean

Abstract Tropical cyclones (TCs), particularly those that are intense and/or slow moving, induce sea surface temperature (SST) reductions along their tracks (commonly referred to as cold wakes) that provide a negative feedback on storm energetics by weakening surface enthalpy fluxes. While computing gains have allowed for simulated TC intensity to increase in global climate models as a result of increased horizontal resolution, many configurations utilize prescribed, noninteractive SSTs as a surface boundary condition to minimize computational cost and produce more accurate TC climatologies. Here, an idealized slab ocean is coupled to a 0.25° variable-resolution version of the Community Atmosphere Model (CAM) to improve closure of the surface energy balance and reproduce observed Northern Hemisphere cold wakes. This technique produces cold wakes that are realistic in structure and evolution and with magnitudes similar to published observations, without impacting large-scale SST climatology. Multimember ensembles show that the overall number of TCs generated by the model is reduced by 5%–9% when allowing for two-way air–sea interactions. TC intensity is greatly impacted; the strongest 1% of all TCs are 20–30 hPa (4–8 m s−1) weaker, and the number of simulated Saffir–Simpson category 4 and 5 TCs is reduced by 65% in slab ocean configurations. Reductions in intensity are in line with published thermodynamic theory. Additional offline experiments and sensitivity simulations demonstrate this response is both significant and robust. These results imply caution should be exercised when assessing high-resolution prescribed SST climate simulations capable of resolving intense TCs, particularly if discrete analysis of extreme events is desired.

scholarly journals Real World and Tropical Cyclone World. Part I: High-Resolution Climate Model Verification

2020 ◽  
Vol 33 (4) ◽  
pp. 1455-1472 ◽  
Author(s):  
S. Sharmila ◽  
K. J. E. Walsh ◽  
M. Thatcher ◽  
S. Wales ◽  
S. Utembe
Keyword(s):  
High Resolution ◽  
Tropical Cyclone ◽  
Geographical Distribution ◽  
Real World ◽  
Large Scale ◽  
Climate Model ◽  
Global Climate ◽  
Global Climate Model ◽  
Model Verification ◽  
State Dependency

AbstractRecent global climate models with sufficient resolution and physics offer a promising approach for simulating real-world tropical cyclone (TC) statistics and their changing relationship with climate. In the first part of this study, we examine the performance of a high-resolution (~40-km horizontal grid) global climate model, the atmospheric component of the Australian Community Climate and Earth System Simulator (ACCESS) based on the Met Office Unified Model (UM8.5) Global Atmosphere (GA6.0). The atmospheric model is forced with observed sea surface temperature, and 20 years of integrations (1990–2009) are analyzed for evaluating the simulated TC statistics compared with observations. The model reproduces the observed climatology, geographical distribution, and interhemispheric asymmetry of global TC formation rates reasonably well. The annual cycle of regional TC formation rates over most basins is also well captured. However, there are some regional biases in the geographical distribution of TC formation rates. To identify the sources of these biases, a suite of model-simulated large-scale climate conditions that critically modulate TC formation rates are further evaluated, including the assessment of a multivariate genesis potential index. Results indicate that the model TC genesis biases correspond well to the inherent biases in the simulated large-scale climatic states, although the relative effects on TC genesis of some variables differs between basins. This highlights the model’s mean-state dependency in simulating accurate TC formation rates.

scholarly journals Combined use of volume radar observations and high-resolution numerical weather predictions to estimate precipitation at the ground: methodology and proof of concept

2019 ◽  
Author(s):  
Tony Le Bastard ◽  
Olivier Caumont ◽  
Nicolas Gaussiat ◽  
Fatima Karbou
Keyword(s):  
High Resolution ◽  
Weather Prediction ◽  
Ground Level ◽  
Radar Data ◽  
Horizontal Resolution ◽  
Computationally Efficient ◽  
New Approach ◽  
Numerical Weather ◽  
Simulated Precipitation ◽  
Nwp Model

Abstract. The extrapolation of the precipitation to the ground from radar reflectivities measured at the beam altitude is one of the most delicate phases of radar data processing for producing Quantitative Precipitation Estimations (QPEs) and remains a major scientific issue. In many operational meteorological services such as Météo-France, a Vertical Profile of Reflectivity (VPR) correction is uniformly applied over a large part or the entire radar domain. This method is computationally efficient and the overall bias induced by the bright band is most of the time well corrected. However, this way of proceeding is questionable in situations with high spatial and vertical variability of precipitation (during the passage of a cold front or in a complex terrain, for example). This study initiates from two statements: first, radars provide information on precipitation with a high spatio-temporal resolution but still require VPR corrections to extrapolate rain rates at the ground level. Second, the horizontal resolution of some Numerical Weather Prediction (NWP) models is now comparable with the radar one and their dynamical core and microphysics schemes allow to produce realistic simulations of VPRs. The present paper proposes a new approach to assess surface rainfall from radar reflectivity aloft by exploiting simulated VPRs and rainfall forecasts from the high resolution NWP model AROME-NWC. To our knowledge, this is the first time that simulated precipitation profiles from a NWP model are used to derive radar QPEs. The implementation of the new method on two stratiform situations provided significant improvements on the hourly and 6-h accumulations compared to the operational QPEs, showing the relevance of this new approach.

Sign in / Sign up

Export Citation Format

Share Document