Next-Generation Sensitivity and Uncertainty Analysis
Variogram Analysis of Response Surfaces (VARS)
The Watershed Systems Analysis and Modelling Lab is home to the VARS sensitivity and uncertainty analysis framework. VARS is applicable to the full range of computer simulation models, including Earth and Environmental Systems Models (EESMs). VARS is under continuous development and new capabilities and features are forthcoming.
Wondering what Sensitivity Analysis and VARS are? Watch these videos:
Sensitivity Analysis Highlights
- Senstivity analysis (SA), is the study of how the ‘outputs’ of a ‘system’ are related to, and are influenced by, its ‘inputs’.
- The ‘system’ usually is a mathematical model that simulates the functioning of a real world system.
- ‘Inputs’ can be, model parameters, forcing variables, boundary and initial conditions, etc.
- ‘Ouputs’ can be, flux or state variables, production functions, cost functions, and error functions.
- SA is useful for, scientific discovery, dimensionality reduction, data worth assesement, and decion support.
- The approaches to SA are, local, derivative-based, variogram-based, and distribution-based.
VARS Highlights
- Introduces a next-generation toolbox for sensitivity and uncertainty analysis.
- Provides a multi-method approach that unifies different theories and strategies.
- Accounts for dynamical properties of Earth and environmental systems models.
- Provides various sampling strategies including progressive Latin hypercube sampling.
- Facilitates handling of high-dimensional models with hundreds of uncertain factors.