System-Analytical Modelling: 2. Assessment Of Runoff Model Sensitivity To Environmental Factor Variations.
- 1 January 2020
- journal article
- research article
- Published by L. N. Gumilyov Eurasian National University in Eurasian Journal of Mathematical and Computer Applications
- Vol. 8 (3), 67-77
- https://doi.org/10.32523/2306-6172-2020-8-3-67-77
Abstract
To evaluate the performance of deterministic models in earth science, an elementary method for assessing their sensitivity to natural variations of environmental factors and the subsequent uncertainty analysis (the evaluation of residual variance components) are proposed. Using a regional model of mountain river runoff as an example, the model sensitivity and all the components of its residual variance are calculated. It is shown that the sensitivity decreases in the sequence "precipitation - air temperature - landscape structure of the river basin" making 22-8-6 explained percent of the observed runoff variance respectively. Model inaccuracy does not exceed 34% of this variance. The most probable river runoff can be forecasted for 3-4 months ahead with a twice reduced variance as compared to the similar forecast based on the observed mean runoff.Keywords
This publication has 4 references indexed in Scilit:
- Global sensitivity analysis in hydrological modeling: Review of concepts, methods, theoretical framework, and applicationsJournal of Hydrology, 2015
- Understanding predictive uncertainty in hydrologic modeling: The challenge of identifying input and structural errorsWater Resources Research, 2010
- Survey of sampling-based methods for uncertainty and sensitivity analysisReliability Engineering & System Safety, 2006
- System-analytical modelling—Part I: General principles and theoretically best accuracies of ecological models. Soil-moisture exchange in agroecosystemsEcological Modelling, 2006