Boundary conditions are the physical conditions at the boundaries of a system. Two of the most common boundary conditions are rainfall and stage/flow hydrographs.

Many flood runoff generation models are extremely vulnerable to uncertainty in precipitation. The impact of uncertainty in precipitation can be most clearly seen when weather forecasts (de Roo et al., 2003) are applied to drive a model. For example, wind can be responsible for errors of up to 20% in gauged rain measurements which has to be corrected for (UNESCO, 1978). Moreover, for the most physically based distributed models, a rainfall field has to be interpolated. This is done by a variety of methods (= models) of different complexity. The quality of these interpolations and corrections will most probably not be consistent with real rainfall patterns and will depend largely on the characteristics of the storm event, which is very rarely included in any interpolation routine. Measuring rainfall by radar an alternative to rain gauges. Rainfall measured by radar is effectively a continuously updated measurement (Beven, 2001, p. 256 et sqq.) but is based on spatial correction techniques (Morin et al., 2003). Furthermore, radar measurements suffer from wind attenuation and depend heavily on the type of rainfall (drop size distribution –rain near the radar sometimes effects signal from rain further away) (Morin et al., 2003). Also ice clouds (bright band) and snow fall will affect a radar measurement and further complicate the estimation of precipitation intensities at ground level by introducing an additional source of uncertainty e.g. quantification of melting and refreezing processes (Molders et al., 2003).

Another boundary condition regularly used in flood inundation modelling is upstream discharge. The flow of water in a channel is normally estimated from stage-discharge curves. The relationship can be significantly distorted by uncertainties (Schmidt, 2002). Even when it is determined with more advanced technologies e.g. by ultrasonic gauges large uncertainties remain in the estimation of discharges for flood stages (Muste et al., 2004). Lateral and tributary inflows to a stream channel, and their dependence on conditions in and below the flood plain, will be even more uncertain.

In coastal and estuarine flooding, which is often the result of combined tidal, wind, storm surge and river discharge forcing, there may also be similar uncertainties in boundary conditions at any particular application site. The local downscaling of atmospheric model predictions to give local storm surge and wave heights will result in uncertain in water levels.

References

Beven, K.J., 2001. Rainfall-Runoff Modelling: The Primer. Wiley, New York.

de Roo, A. et al., 2003. Development of a European Flood Forecasting System. International Journal of River Basin Management, 1: 49-59.

Molders, N., Haferkorn, U., Doring, J. and Kramm, G., 2003. Long-term investigations on the water budget quantities predicted by the hydro-thermodynamic soil vegetation scheme (HTSVS) - Part II: Evaluation, sensitivity, and uncertainty. Meteorology and Atmospheric Physics, 84(1-2): 137-156.

Morin, E., Krajewski, W.F., Goodrich, D.C., Gao, X.G. and Sorooshian, S., 2003. Estimating rainfall intensities from weather radar data: The scale-dependency problem. 4(5): 782-797.

Muste, M., Yu, K., Pratt, T. and Abraham, D., 2004. Practical aspects of ADCP data use for quantification of mean river flow characteristics; Part II: fixed-vessel measurements. Flow Measurement and Instrumentation, 15(1): 17-28.

Schmidt, A.R., 2002. Analysis of stage-discharge relations for open-channel flow and their associated uncertainties, Universitz of Ollinois, Urbana, 328 pp. UNESCO, 1978. World Water Balance and Water Resources of the Earth. No. 25, UNESCO, Paris.

Go to

Risk and Uncertainty (Description and Definition)