Meteorologists are increasingly using model ensembles for weather forecasting. A model ensemble is essentially a group of numerical forecasts with differing model physics packages and/or initial conditions. Because our models are imperfect, we can never define the initial state of the atmosphere perfectly, and the atmosphere is so non-linear the idea is to use the ensemble to better understand the range of possible scenarios for the future.
One such ensemble is the NCEP Global Forecast Ensemble System, which is comprised of 22 forecasts, one of which is a high resolution control run (out to 7 days), the remainder of which are lower-resolution forecasts with slightly different initial conditions.
There are many ways to visualize these forecasts, but spaghetti diagrams are amongst the most popular. With a spaghetti diagram, you plot contours of selected variables from all the ensemble members. It takes a while to get use to viewing these plots, but they can be quite useful.
With regards to Greg's future, I decided to look at a spaghetti diagram of GEFS precipitable water forecasts. I've color coded the contours so that blue is 12.5 mm, light blue 25 mm, light green 37.5, orange 50, and red 62.5. Here's a forecast loop running from 0600 UTC 19 Aug – 1800 UTC 23 Aug.
Notice the gradual growth of variability amongst the GEFS members as the forecast projection increases. Locally high precipitable water associated with Greg moves westward in all the forecasts, with the maximum decreasing in magnitude in all the ensemble members.
Those with a good eye may notice that subtropical moisture surges out of the subtropics north of Greg beginning at about 0600 UTC 21 August. Some, but not all of the ensemble members, bring precipitable water values >25 mm up the California coast and Central Valley. If such a surge occurs, thunderstorms would likely develop over the Sierra Nevada.
This GEFS ensemble provides us with some situational awareness of the possibility of a surge of moisture into California. It also tells us that there are varied predictions for the intensity of this surge. This enables the meteorologist to assess confidence and likelihood and provide better information about the range of possible weather in the future.