Basically in water resources (and any other filed) you have 4 types of uncertainty:

U_total = U_modeling - U_experiment

U_modeling = U_numerical +U_parameters+U_model_structure

U_total = (U_numerical +U_parameters+U_model_structure) - U_experiment

0) U_total: total uncertainty

1) U_modeling: is all uncertainties due to modeling activity

2) U_experiment: is all uncertainties due to measuring (in lab or site) the phenomenon

3) U_numerical: is the numerical error of modeling a continuum system in discrete form (discretization error, numerical algebra matrix inversion, roundoff error, ...)

4) U_parameters: uncertainty in physical and numerical parameters of your solution: for example if you have a bathymetery with resolution of 3ft by 3ft, that is an important uncertainty and you cannot ignore that, but it is not due to your numerical scheme (U_numerical) or your model_structure

5) U_model_structure: is model structural uncertainty, means if your mathematical model is really representative of the physics or not. For example people use Green-Ampt model for infiltration of water in soil, is it really the true representative of the phenomenon? if not the uncertainty of that

Terms in the parenthesis is the uncertainty of modeling (Numbers 3, 4, 5)

U experiment is the job of experimental person, and their difference is your total uncertainty.

For solved numerical example and deep discussion see this Standard by ASME

Standard for verification and validation in computational fluid dynamics and heat transfer

ASME V& V20 2009 American Society of Mechanical Engineers

Literature defines uncertainty in multiple ways. In general uncertainty refers to conditions which are not exactly quantifiable and which are beyond human control. Yen and Ang (1971) categorized into (i) objective uncertainty and (ii) subjective uncertainty. Uncertainty related to random nature and statistical sampling comes under objective uncertainty while uncertainty which is not amenable to quantitative information comes under subjective uncertainty. Generally uncertainty sources are divided into two sources, epistemic (from lack of knowledge) and aleatory (from random statistical variability). Tung and Yen (2005) classified the sources of uncertainty towards hydro-systems into natural variability & knowledge deficiency. Natural variability cannot be eliminated but uncertainty from knowledge deficiency can be reduced by more acquired knowledge.

The modelling and experimental uncertainties will propagate and accumulated in your analysis. The changing future and it projections make the analysis more complex. The impact of climate change in rainfall forecasting and uncertainty will arise from choice of GEM, RCM, emission pathways, downscaling techniques and model interdependency and non-stationary bias. The urbanization/land-use changes alters the basin roughness. Further socio-economic changes in in the basin alters the demand for the water resource management practices.

You can refer to more details.

Tung, Y.K., & Yen, B.C. (2005) Hydrosystem Engineering Uncertainty Analysis. Mc Graw-Hill Series in Civil Engineering, 1st edition, New York, ASCE Press.

Thesis Decision Making on Flood Mitigation incorporating Uncertaint...

Thank you