Management of uncertainty in model predictions of long-term coastal change begins by admitting uncertainty. In the case of geometric mass-balance models, the first step is to relax restrictive assumptions to allow for open sediment budgets, time-dependent morphology, effects of mixed sediment sizes, and variable resistance in substrate material. These refinements introduce new uncertainty regarding the choice of parameter values. The next step is to actively manage uncertainty using techniques readily available from information science. The final step requires a shift in coastal management culture to accept decision making based on risk-management protocols.

Stochastic simulation was applied to manage predictive uncertainty in cases involving complications resulting from open sediment budgets, rock reefs, and seawalls. In these examples, the respective effects caused between 20% and 60% difference from conventional predictions based solely on equilibrium assumptions and substrates comprised entirely of sand.

Stochastic simulation makes it possible to establish confidence limits and determine the statistical significance of differences caused by varying effects such as substrate resistance and shoreface geometry. It also enables the likelihood of critical impacts to be specified in terms of probability. Moreover, probabilistic forecasts provide a transparent basis for coastal management decisions by revealing the consequences if quantitative estimates prove to be wrong.