Lecture 18 The Modeling Environment.ppt

Lecture 18: The Modeling Environment CE 498/698 and ERS 485 Principles of Water Quality Modeling The modeling environment Models are an idealized formulation that represents the response of a physical system to external stimuli (p. 10) Models are tools that are part of an overall management process Rules of modeling RULE 1: We cannot model reality We have to make assumptions DOCUMENT!!!! RULE 2: Real world has less precision than modeling Precision vs. accuracy Precision Number of decimal places Spread of repeated computations Accuracy Error between computed or measured value and true value The problem with precise models… we get more precision from model than is real Modeling in management process Problem specification Do we need to model? Model selection Who will the users be? What kind of data is available? General model or specific? Use existing model or develop a new one? Modeling in management process Model development Develop/modify code Input data Determine numerical approach model resolution Timestep Spatial size Modeling in management process Model development Develop/modify code Input data Determine numerical approach model resolution Timestep Spatial size Matter Modeling in management process Preliminary application and calibration Modeling in management process Preliminary application and calibration Adjust parameters Adjust input data (where appropriate) Compare model predictions with measured Calibration measures Chapra: minimize sum of squares of residuals Calibration measures Chapra: minimize sum of squares of residuals Calibration measures Average absolute error Modeling in management process Model confirmation (validation) Use an independent data set for measured values Use same parameters/coefficients, same methods of estimating data Does model still work???? Modeling in management process Management application Verification Sensitivity analysis *Lecture 18 Management (or scientific) objectives, options, constraints Model development and