Professor Stephen Roberts

PhD UC Berkeley
ANU College of Science

Areas of expertise

  • Simulation And Modelling 080110
  • Numerical Solution Of Differential And Integral Equations 010302
  • Water Resources Engineering 090509
  • Numerical Computation 080205
  • Concurrent Programming 080304
  • Numerical Analysis 010301
  • Analysis Of Algorithms And Complexity 080201
  • Numerical And Computational Mathematics 0103

Research interests

Developer of efficient and robust numerical methods for the modelling of Tsunami and Flood events.

Lead developer at the ANU of the ANUGA hydrodynamic modelling software. ANUGA is a computational tool that models the impact of dam breaks, floods and tsunamis on communities. It is used extensively by councils, governments and consultant engineers. The software is developed in Python and C.

Developer of sparse grid methods for
Multi-dimensional function fitting and approximation of high dimensional probablitiy density functions.

Developing finite element approximation methods for thin plate spline functional smoothing which can scale to millions of data points.

Developer of Computational Science education program at ANU using Python.

Specialties:

Computational Mathematics, Tsunami modelling, Flood modelling, Multi-dimensional approximation, Python programming for Scientific computing.

Biography

Stephen Roberts was Head of the Department of Mathematics at the Australian National University from 2006-2012.

Stephen is heavily involved in the computational science community in Australia. From 2003-2006 he was the national coordinator of the Australian Partnership for Advanced Computing (APAC) Education, Outreach and Training program. He is currently the treasurer of the committee of the Computational Mathematics Group, a special interest group of ANZIAM.

His research area is the application of efficient and robust numerical methods for the solution of partial differential equations.

He is the lead developer at the ANU of the ANUGA hydrodynamic modeling software. ANUGA is an open source computational tool that models the impact of dam breaks, floods and tsunamis on communities. It is used extensively by councils, governments and consultant engineers.

In the area of data fitting he has developed sparse grid methods for Multidimensional function fitting and approximation of high dimensional probability density functions. He has also developed finite element approximation methods for thin plate spline functional smoothing which can scale to millions of data points.

In 1980, Stephen completed his MSc Science at Flinders Univerisy, supervised by Prof Garth Gaudry, working on A_p spaces and asymmetry of L_p operator norms for convolution operators. In 1985, Stephen completed his PhD at the University of California, Berkeley, supervised by Prof Alexandre Chorin, working on the Convergence of a random walk method for the Burgers equation.

 

 

Researcher's projects

High Dimensional Approximation and Uncertainty Quantification: 

High Dimensional Approximation and Uncertainty Quantification Uncertainty quantification using sparse grid methods. Multi-dimensional function fitting and approximation of high dimensional probability density functions.

Multi-Fidelity Solution of Tsunami and Flood Models:

We are investigating the use of multi-fidelity solutions of tsunami and flood models to improve the efficiency of uncertainty quantification studies. The multi-fidelity solutions are combined using sparse grid techniques. 

Solution of the Serre Equations:

Investigate the use of techniques from the numerical solution of conservation laws (approximate Riemann solvers) to accurately solve the Serre Equation, especially in the presence of large initial gradients in the initial conditions (e.g. in the case of dam breaks).

ANUGA Tsunami and Flood Modelling Software:

The ANUGA package is developed under a collaboration of the  Australian National University (ANU) and Geoscience Australia (GA) - hence the name. ANUGA is a Free & Open Source Software (FOSS) package capable of modelling the impact of hydrological disasters such as dam breaks, riverine flooding, storm-surge or tsunamis.

ANUGA is based on a discretization of the Shallow Water Wave Equations onto an unstructured triangular mesh using a finite-volume numerical scheme. A major capability of ANUGA is that it can model the process of wetting and drying as water enters and leaves an area. This means that it is suitable for simulating water flow onto a beach or dry land and around structures such as buildings. ANUGA is also capable of modelling difficult flows involving hydraulic jumps and rapidly changing flow speed regimes (transitions from subcritical to supercritical flows).

The package is available from GitHub github.com//GeoscienceAustralia/anuga_corewhere we average over 200 unique visitors per month, and 30 downloads/month.

Geosciences Australia has used ANUGA extensively to quantify risk from tsunamis around the coast of Australia, It was used to develop ``Last mile evacuation'' plans for Padang, in West Sumatra. It is being used in AusAID (now part of the DFAT) programs to assess risk of flooding in Manila and risk from tsunamis in Papua New Guinea.

In 2016 we developed code to integrate BoM radar rainfall data with ANUGA, and tested the method with a simulation of the full ACT catchment. Also a simple erosion module which models the erosion of sand dunes during a tsunami has been added to ANUGA.

Also we are working with Hydrata.com to incorporate ANUGA as the modelling engine for an on-line flood modelling portal as part open collaboration platform http://chennaifloodmanagement.org/en/, a community project to help mitigate the effect of flooding in Chennai, India.

 

Publications

Projects and Grants

Grants information is drawn from ARIES. To add or update Projects or Grants information please contact your College Research Office.

Return to top

Updated:  21 March 2019 / Responsible Officer:  Director (Research Services Division) / Page Contact:  Researchers