PEST - the world’s most advanced software for model calibration, parameter estimation and predictive uncertainty analysis.

These pages are jointly managed by Watermark Numerical Computing (developers of PEST) and S.S. Papadopulos and Associates, Inc., with help from friends and associates including:

• Dept of Engineering, University of Queensland

• University of Idaho

• EMSI (Environmental Modeling Systems, Inc.)

• ESI (Environmental Simulations International)

These pages are the place for modelers to:

• Find out more about what PEST does

• Download the latest version of PEST and all of its utility software

• View details on coming PEST courses by checking the Training Page

• Receive the latest information on PEST as it becomes available through the PEST mailing list

• Find out about consulting services for high-end calibration, data-processing and uncertainty analysis using PEST

• Coming in 2005 - join in a discussion group on model calibration and uncertainty analysis using PEST

What is PEST?

PEST is the most robust and comprehensive software available for model-independent parameter estimation and uncertainty analysis. And, since February 2002, PEST has been FREEWARE. Despite the fact that PEST is freeware, it continues to be improved. Active research into model parameterization and predictive uncertainty analysis in the groundwater and surface water modeling contexts is ongoing - and many of the outcomes of this research will find their way into PEST and its supporting utility software. Yet PEST is more than just the most advanced package available for model calibration - PEST enables the quantification of model predictive error in a variety of settings. PEST is a move away from the paradigm of trial and error calibration and uncertainty analysis. No other package matches PEST’s functionality, compatibility, performance, and extensive array of utility software.

PEST is a unique package that can be used with any pre-existing model for data interpretation or model calibration.  The days of manual model calibration and data interpretation are over. Now you can use all the data all the time no matter what model you are using.

Matching model-generated data to field data becomes easy with PEST.

Use of nonlinear parameter estimation techniques allows you to exercise greater control over model calibration and/or data interpretation. More complex parameterizations are possible than would even be considered under a manual calibration regime. Yet PEST can clearly indicate where further complexity is non-sustainable, given the current dataset. Contrast this with the manual calibration process where the modeler simply "gives up" when he/she no longer has the strength or the time to carry out yet another model run.

PEST is powerful - it has successfully calibrated models with hundreds of parameters on the basis of thousands of observations. Yet it is easy to use. No programming is required to interface an existing model with PEST because PEST communicates with a model through the model's own input and output files. The flexibility engendered through this approach allows ingenious calibration methodologies to be developed, for the "model" can actually be a batch file running many programs in succession; PEST can communicate with some or all of these individual programs.

Complexity is no hindrance to PEST.

PEST is continually being upgraded; it now comes with a powerful predictive analyzer and advanced regularization functionality. Parallel PEST allows calibration of large and complex models for which application of nonlinear parameter estimation methodology would have hitherto been considered impossible. Parallel PEST achieves enormous optimization-time savings by distributing model runs across a computer network.

You can download a whole suite of FREE utility software to enhance PEST's usage in groundwater and surface water model calibration. Using this software, setup time for complex calibration problems can be reduced to minutes.

For more information about PEST and ongoing software development, please contact us.