Download Instructions

Download pest10.zip which includes executables, the PEST Manual and the PEST Manual Addendum in PDF format.

Unzip the contents to a directory cited in the PATH environment variable. (This is very important as, when undertaking SVD-assisted parameter estimation, PEST runs the parcalc.exe and parcalc.exe utility programs. It is essential that the operating system knows where to find these).

Make sure that the directory holding older versions of PEST (including version 9) are no longer in the PATH string, as old versions of parcalc.exe and picalc.exe are not compatible with the new PEST.

What’s new in Version 10 of PEST

Version 10 of PEST contains the most advanced model calibration and predictive uncertainty analysis functionality by far of any available software.

Enhancements to version 10 of PEST include the following:

SVD-Assist 

PEST’s unique and powerful SVD-assist technique - that allows regularized inversion to be used with highly-parameterized models but with model run efficiencies commensurate with traditional parameter estimation - has been enhanced. Now all models can be calibrated using this methodology.

Model Predictive Uncertainty Analysis

PEST’s predictive analyzer is even more powerful. Now you can explore the full predictive range of an environmental model while maintaining that model in a calibrated state using an improved line-search algorithm, and added functionality to encompass the effects of predictive noise on predictive variability. 

Predictive Uncertainty Analysis and Regularization

In the May 2005 issue of Water Resources Research, Moore and Doherty demonstrate that traditional methods of predictive uncertainty analysis can grossly underestimate true variability through neglecting the most important contributor to predictive error variance - system complexity that is not captured in the model calibration process. Theory is presented that allows model predictive error variance to be properly quantified as an adjunct to the use of regularized inversion in the model calibration process. Version 10 of PEST provides access to these methodologies for all modelers.

State-of-the-Art Model Calibration Post-Processing

A vastly expanded range of utility software allows comprehensive post-processing of PEST results to be undertaken, including analysis and plotting of the resolution matrix and other by-products of the regularized inversion process.

Optimization of Data Acquisition

New utility software allows a modeler to quantify the contribution to pre- and/or post-calibration error variance of important model predictions by different parameter types, or even individual parameters. Other new utility programs allow the benefits of making extra measurements of system state (e.g., a water level), or of system properties (e.g., a hydraulic conductivity), in terms of predictive error variance reduction to be deduced, providing a firm basis for cost optimization of environmental data acquisition.