A Monte Carlo Approach to estimation and prediction

Fedra, K. (1983). A Monte Carlo Approach to estimation and prediction. In: Uncertainty and Forecasting of Water Quality. pp. 259-291 Berlin/Heidelberg, Germany: Springer. ISBN 978-3-642-82054-0 10.1007/978-3-642-82054-0_12.

Full text not available from this repository.

Abstract

The model representation of complex environmental systems requires numerous simplifications; frequently, arbitrary choices of how to formally represent the relationships between causes and effects have to be made, since these relationships are neither obvious nor easy to detect. Environmental systems in toto do not easily yield to the classical scientific tool of planned experimentation. Consequently, the analyst has to utilize whatever bits of information may be available, which as a rule are very few and not strictly appropriate in terms of the problems addressed. A priori knowledge about the structure and function of any ecosystem is generally poor, and reliable quantitative information on the governing processes and their rates and interrelationships insufficient. Consequently, building and testing models and finally applying them for predictive purposes often consists of a more or less formalized trial-and-error iterative process of estimation, testing, and improvement. The following discussion proposes an approach for formalizing this process of model building, calibration, and application; it emphasizes the interdependencies of the individual steps in the process. The approach proposed is based on the recognition of uncertainty as an inevitable element in modeling, and uses straightforward Monte Carlo techniques to cope with this uncertainty.

Item Type:	Book Section
Research Programs:	Resources and Environment Area (REN)
Depositing User:	Romeo Molina
Date Deposited:	25 Feb 2016 13:14
Last Modified:	27 Aug 2021 17:40
URI:	https://pure.iiasa.ac.at/12058

Actions (login required)

View Item