scholarly journals Automated model optimisation using the Cylc workflow engine (Cyclops v1.0)

2017 ◽  
Author(s):  
Richard M. Gorman ◽  
Hilary J. Oliver
Keyword(s):  
Cost Function ◽  
Programming Languages ◽  
Model Simulation ◽  
Wave Model ◽  
Workflow Engine ◽  
Test Case ◽  
Spatial Average ◽  
Wave Hindcast ◽  
Geophysical Model ◽  
Error Metric

Abstract. Most geophysical models include a number of parameters that are not fully determined by theory, and can be ‘tuned’ to improve the model's agreement with available data. We might attempt to automate this tuning process in an objective way by employing an optimisation algorithm to find the set of parameters that minimises a cost function derived from comparing model outputs with measurements. A number of algorithms are available for solving optimisation problems, in various software programming languages, but interfacing such software to a complex geophysical model simulation, presents certain challenges. To tackle this problem, we have developed an optimisation suite ("Cyclops") based on the Cylc workflow engine (http://cylc.github.io/cylc/ and https://zenodo.org/badge/latestdoi/1836229) that implements a wide selection of optimisation algorithms from the NLopt python toolbox (Johnson, 2014). The Cyclops optimisation suite can be used to calibrate any modelling system that has itself been implemented as a (separate) Cylc model suite, provided it includes computation and output of the desired scalar cost function. A growing number of institutions are using Cylc to orchestrate complex distributed suites of interdependent cycling tasks within their operational forecast systems, and in such cases application of the optimisation suite is particularly straightforward. As a test case, we applied the Cyclops to calibrate a global implementation of the Wavewatch III™ (v4.18) third generation spectral wave model, forced by ERA-Interim input fields. This was calibrated over a one-year period (1997), before applying the calibrated model to a full (1979–2016) wave hindcast. The chosen error metric was the spatial average of the root-mean-square error of hindcast significant wave height compared with collocated altimeter records. We describe the results of a calibration in which up to 19 parameters were optimised.

scholarly journals Automated model optimisation using the Cylc workflow engine (Cyclops v1.0)

2018 ◽  
Vol 11 (6) ◽  
pp. 2153-2173 ◽  
Author(s):  
Richard M. Gorman ◽  
Hilary J. Oliver
Keyword(s):  
Cost Function ◽  
Programming Languages ◽  
Model Simulation ◽  
Wave Model ◽  
Workflow Engine ◽  
Test Case ◽  
Spatial Average ◽  
Wave Hindcast ◽  
Geophysical Model ◽  
Error Metric

Abstract. Most geophysical models include many parameters that are not fully determined by theory, and can be “tuned” to improve the model's agreement with available data. We might attempt to automate this tuning process in an objective way by employing an optimisation algorithm to find the set of parameters that minimises a cost function derived from comparing model outputs with measurements. A number of algorithms are available for solving optimisation problems, in various programming languages, but interfacing such software to a complex geophysical model simulation presents certain challenges. To tackle this problem, we have developed an optimisation suite (“Cyclops”) based on the Cylc workflow engine that implements a wide selection of optimisation algorithms from the NLopt Python toolbox (Johnson, 2014). The Cyclops optimisation suite can be used to calibrate any modelling system that has itself been implemented as a (separate) Cylc model suite, provided it includes computation and output of the desired scalar cost function. A growing number of institutions are using Cylc to orchestrate complex distributed suites of interdependent cycling tasks within their operational forecast systems, and in such cases application of the optimisation suite is particularly straightforward. As a test case, we applied the Cyclops to calibrate a global implementation of the WAVEWATCH III (v4.18) third-generation spectral wave model, forced by ERA-Interim input fields. This was calibrated over a 1-year period (1997), before applying the calibrated model to a full (1979–2016) wave hindcast. The chosen error metric was the spatial average of the root mean square error of hindcast significant wave height compared with collocated altimeter records. We describe the results of a calibration in which up to 19 parameters were optimised.

Model-Based Development

2011 ◽  
pp. 289-312 ◽  
Author(s):  
Juan D. Lara ◽  
Esther Guerra ◽  
Hans Vangheluwe
Keyword(s):  
Programming Languages ◽  
Graphical Models ◽  
Code Generation ◽  
Simulation Analysis ◽  
Model Simulation ◽  
General Purpose ◽  
Test Case ◽  
Model Driven ◽  
Model Based ◽  
Design Activities

Since the beginning of computer science more than 50 years ago, software engineers have sought techniques resulting in higher levels of quality and productivity. Some of these efforts have concentrated in increasing the level of abstraction in programming languages (from assembler to structured languages to object-oriented languages). In the last few years, we have witnessed an increasing focus on development based on high-level, graphical models. They are used not only as a means to documentthe analysis and design activities, but also as the actual “implementation” of the application, as well as for automatic analysis, code, and test case generation. The notations used to describe the models can be standard and general purpose (for example, UML) or tightly customized for the application domain. Code generation for the full application is only accomplished for specific, well-understood application domains. A key initiative in this direction is OMG’s Model-Driven Architecture (MDA), where models are progressively transformed until executable code is obtained. In this chapter, we give an overview of these technologies and propose ideas following this line (concerning metamodeling and the use of visual languages for the specification of model transformation, model simulation, analysis and code generation), and examine the impact of model-based techniques in the development process.

scholarly journals CHIMERE-2017: from urban to hemispheric chemistry-transport modeling

2017 ◽  
Vol 10 (6) ◽  
pp. 2397-2423 ◽  
Author(s):  
Sylvain Mailler ◽  
Laurent Menut ◽  
Dmitry Khvorostyanov ◽  
Myrto Valari ◽  
Florian Couvidat ◽  
...  
Keyword(s):  
Transport Model ◽  
Model Simulation ◽  
Point Of View ◽  
Atmospheric Composition ◽  
Test Case ◽  
Polar Regions ◽  
Chemistry Transport Model ◽  
Computational Point ◽  
Improve Model ◽  
Code Complexity

Abstract. CHIMERE is a chemistry-transport model designed for regional atmospheric composition. It can be used at a variety of scales from local to continental domains. However, due to the model design and its historical use as a regional model, major limitations had remained, hampering its use at hemispheric scale, due to the coordinate system used for transport as well as to missing processes that are important in regions outside Europe. Most of these limitations have been removed in the CHIMERE-2017 version, allowing its use in any region of the world and at any scale, from the scale of a single urban area up to hemispheric scale, with or without polar regions included. Other important improvements have been made in the treatment of the physical processes affecting aerosols and the emissions of mineral dust. From a computational point of view, the parallelization strategy of the model has also been updated in order to improve model numerical performance and reduce the code complexity. The present article describes all these changes. Statistical scores for a model simulation over continental Europe are presented, and a simulation of the circumpolar transport of volcanic ash plume from the Puyehue volcanic eruption in June 2011 in Chile provides a test case for the new model version at hemispheric scale.

A new wave hindcast for the North and Baltic Sea region using COSMO-REA6

2020 ◽  
Author(s):  
Nikolaus Groll
Keyword(s):  
Baltic Sea ◽  
Extreme Events ◽  
Wave Model ◽  
Wave Climate ◽  
Good Representation ◽  
Wind Fields ◽  
Wave Hindcast ◽  
The North ◽  
Atmospheric Observations ◽  
The Baltic

<p>Wave hindcasts are still required as improved knowledge of climate variables representing the present marine climate are needed. Most long regional wave hindcasts are driven by numerically downscaled wind fields from global reanalysis. Whereas this approach gives a good representation of the regional wave climate in general, there are some deficits in the characteristics of extreme events. Using regional atmospheric reanalysis, which assimilates atmospheric observations into the numerical model, a better description of extreme events is expected. The regional atmospheric reanalysis COSMO-REA6 from the German Weather Service (DWD) showed that it is capable of a better representation of atmospheric extreme events. For the new regional wave hindcast, covering the North Sea and the Baltic Sea, we use the COSMO-REA6 to force the wave model WAM. It is shown, that his new wind hindcast leads to an improved representation of extreme wave events compared to other regional wave hindcasts and thus supports an important contribution to the understanding of the wave climate of extremes and for the design phase of offshore activities.</p>

scholarly journals CHIMERE-2016: From urban to hemispheric chemistry-transport modeling

2016 ◽  
Author(s):  
Sylvain Mailler ◽  
Laurent Menut ◽  
Dmitry Khvorostyanov ◽  
Myrto Valari ◽  
Florian Couvidat ◽  
...  
Keyword(s):  
Transport Model ◽  
Model Simulation ◽  
Point Of View ◽  
Atmospheric Composition ◽  
Test Case ◽  
Polar Regions ◽  
Eulerian Model ◽  
Chemistry Transport Model ◽  
Computational Point ◽  
Improve Model

Abstract. CHIMERE is a chemistry-transport model initially designed for box-modelling of regional atmospheric composition. In the past decade, it has been converted into a 3D eulerian model that could be used at a variety of scales from local to continental domains. However, due to the model design and its historic use as a regional model, major limitations had remained, prohibiting its use at hemispheric scale, due to the coordinate system used for transport as well as to missing processes that are important in regions outside Europe. Most of these limitations have been lifted in the CHIMERE-2016 version, allowing its use in any region of the world and at any scale, from the scale of a single urban area up to hemispheric scale, including or not polar regions. Other important improvements have been brought in the treatment of the physical processes affecting aerosols and the emissions of mineral dust. From a computational point of view, the parallelization strategy of the model has also been improved in order to improve model numerical performance. The present article describes all these changes. Scores for a model simulation over continental Europe are presented, and a simulation of the circumpolar transport of volcanic ash plume from the Puyehue volcanic eruption in June 2011 in Chile provides a test case for the new model version at hemispheric scale.

Link Cost Function and Link Capacity for Mixed Traffic Networks

2020 ◽  
Vol 2674 (9) ◽  
pp. 38-50
Author(s):  
Aathira K. Das ◽  
Bhargava Rama Chilukuri
Keyword(s):  
Cost Function ◽  
Empirical Data ◽  
Wave Model ◽  
Mixed Traffic ◽  
Vehicle Class ◽  
Link Capacity ◽  
The Road ◽  
Traffic Conditions ◽  
Link Cost ◽  
Link Travel Time

Link cost function and link capacity are critical factors in traffic assignment modeling. Popular link cost functions like the Bureau of Public Roads (BPR) function have well-known drawbacks and are not suitable for mixed traffic conditions where a variety of vehicle classes use the road in a non-lane-based movement. Similarly, capacity is generally considered as a constant value. However, in mixed traffic conditions, capacity is not constant, but a function of vehicle class composition. Toward addressing these issues, this paper proposes a link cost function in relation to link travel time and link capacity in relation to vehicular traffic flow for mixed traffic conditions. The functions are developed based on the kinematic wave model, which is popularly used for estimating traffic dynamics on the roads. The developed link cost function and link capacity use field measurable parameters that incorporate mixed traffic features. The functions are validated against empirical data obtained from 12 signal cycles from two different signalized intersections in Chennai, India, representing different scenarios of mixed traffic, and it was found that the results match well with the empirical data.

scholarly journals Wave climate in the Arctic 1992–2014: seasonality and trends

2016 ◽  
Author(s):  
Justin E. Stopa ◽  
Fabrice Ardhuin ◽  
Fanny Girard-Ardhuin
Keyword(s):  
Sea Ice ◽  
Wave Model ◽  
Wave Climate ◽  
Arctic Sea Ice ◽  
The Arctic ◽  
Wave Hindcast ◽  
The North ◽  
Marginal Ice Zone ◽  
The North Atlantic Oscillation ◽  
The Impact

Abstract. Over the past decade, the diminishing Arctic sea ice has impacted the wave field which is principally dependent on the ice-free area and wind. This study characterizes the wave climate in the Arctic using detailed sea state information from a wave hindcast and merged altimeter dataset spanning 1992–2014. The wave model uses winds from the Climate Forecast System Reanalysis and ice concentrations derived from satellites as input. The ice concentrations have a grid spacing of 12.5 km, which is sufficiently able to resolve important features in the marginal ice zone. The model performs well, verified by the altimeters and is relatively consistent for climate studies. The wave seasonality and extremes are linked to the ice coverage, wind strength, and wind direction. This creates distinct features in the wind-seas and swells. The increase in wave heights is caused by the loss of sea ice and not the wind verified by the altimeters and model. However, trends are convoluted by inter-annual climate oscillations like the North Atlantic Oscillation (NAO) and Pacific Decadal Oscillation. The Nordic-Greenland Sea is the only region with negative trends in wind speed and wave height and is related to the NAO. Swells are becoming more prevalent and wind-sea steepness is declining which make the impact on sea ice uncertain. It is inconclusive how important wave-ice processes are within the climate system, but selected events suggest the importance of waves within the marginal ice zone.

Methods of Comparing the Wave Model Simulation Data with the KA-BAND Radar Data

2021 ◽  
Author(s):  
M.A. Panfilova ◽  
A. M. Kuznetsova ◽  
Yu. A. Titchenko ◽  
D. A. Sergeev ◽  
Yu. I. Troitskaya ◽  
...  
Keyword(s):  
Model Simulation ◽  
Wave Model ◽  
Radar Data ◽  
Simulation Data ◽  
Ka Band

scholarly journals Response of OH airglow emissions to mesospheric gravity waves and comparisons with full-wave model simulation at a low-latitude Indian station

2016 ◽  
Vol 16 (9) ◽  
pp. 5611-5621 ◽  
Author(s):  
Rupesh N. Ghodpage ◽  
Michael P. Hickey ◽  
Alok K. Taori ◽  
Devendraa Siingh ◽  
Parashram T. Patil
Keyword(s):  
Model Simulation ◽  
Wave Model ◽  
Short Wave ◽  
Low Latitude ◽  
Full Wave ◽  
Short Period ◽  
Propagating Waves ◽  
The Mean ◽  
Wave Induced ◽  
Indian Station

Abstract. Quasi-monochromatic gravity-wave-induced oscillations, monitored using the mesospheric OH airglow emission over Kolhapur (16.8° N, 74.2° E), India, during January to April 2010 and January to December 2011, have been characterized using the Krassovsky method. The nocturnal variability reveals prominent wave signatures with periods ranging from 5.2 to 10.8 h as the dominant nocturnal wave with embedded short-period waves having wave periods of 1.5–4.4 h. The results show that the magnitude of the Krassovsky parameter, viz. |η|, ranged from 2.1 to 10.2 h for principal or long nocturnal waves (5.2–10.8 h observed periods), and from 1.5 to 5.4 h for the short waves (1.5–4.4 h observed periods) during the years of 2010 and 2011, respectively. The phase (i.e., Φ) values of the Krassovsky parameters exhibited larger variability and varied from −8.1 to −167°. The deduced mean vertical wavelengths are found to be approximately −60.2 ± 20 and −42.8 ± 35 km for long- and short-period waves for the year 2010. Similarly, for 2011 the mean vertical wavelengths are found to be approximately −77.6 ± 30 and −59.2 ± 30 km for long and short wave periods, respectively, indicating that the observations over Kolhapur were dominated by upward-propagating waves. We use a full-wave model to simulate the response of OH emission to the wave motion and compare the results with observed values.

scholarly journals SPECTRAL PARTITIONING AND SWELLS IN THE BLACK SEA

2017 ◽  
pp. 21
Author(s):  
Gerbrant van Vledder ◽  
Adem Akpinar
Keyword(s):  
Black Sea ◽  
Wave Model ◽  
The Black Sea ◽  
Wave Hindcast ◽  
Spectral Wave Model ◽  
Seasonal Characteristics ◽  
Individual Wave ◽  
Wind Sea ◽  
Spectral Partitioning

The swell climate of the Black Sea has been determined using a long-term 31-year wave hindcast with the third-generation spectral wave model SWAN in combination with spectral partitioning. This technique enables decomposing wave spectra into individual wave systems representing wind seas or swells and computing integral wave parameters of each partition. Results are presented of the partition technique and of spatial and seasonal characteristics of wind sea and swell systems. In addition, the average amount of swell energy and the occurrence probability of dangerous crossing sea states are determined.

Sign in / Sign up

Export Citation Format

Share Document