Multiple Realizations and Uncertainty

One Rock, or Many

A single simulation is one plausible rock, no more likely than the next, so reporting it alone hides the uncertainty. The right object is an ensemble: many realizations, each honoring the same wells and variogram, differing only where the data cannot constrain them.

Spread Is Uncertainty

Stack the realizations and two maps fall out. Their average is a smooth, P50-style best map. Their cell-by-cell spread is an uncertainty map: zero at the wells, where every realization agrees, and largest in the poorly sampled gaps. That spread is measured from the data and the variogram, not assumed.

Carrying Uncertainty Forward

The ensemble is the point. Each realization can be turned into volumes and flowed, and the range of answers across the ensemble is the honest uncertainty on reserves and production, the P10, P50, and P90 of the next chapter. A model that ships a single realization throws that uncertainty away.