I don't understand your first question but the "yikes" was just because of the insane model spread (0-9", and fairly evenly across the spectrum... no clustering) for an event in the next 18 hours. Tough forecast that clearly isn't too stable in its solution.
Sure.
"HRRR" is the High-Resolution Rapid Refresh model. It runs every hour and only out to 15 hours. I'm not overly well versed on the model physics, but it is very high res (1 km) and is frequently used for its skill (sometimes lol) in forecasting radar evolution. Formerly called the RUC. http://ruc.noaa.gov/hrrr/
"SREF" is the Short-Range Ensemble Forecast. It is an ensemble of forecasts (initiated off the NAM I think) where there are three different physics schemes used (grouped by color -- ARW (pink), NMM (oranges), and NMB (blue)). Then amongst those, the initial conditions are also varied to get an idea of the sensitivity of the forecast to both the physics scheme used and the initial conditions. Clustering around a specific solution would imply a higher probability of that solution given assumed errors in model input, etc. Very often you'll see grouping by physics scheme which also shows a sensitivity to that. As you can see in the above forecast, the ARW members tend to be much higher than the rest, but otherwise there's a continuous range of possibilities along the spectrum that aren't particularly well defined or easy to pick out. What is usually posted here are the "plumes" for a given location, but traditional 2D "Weather maps" are also available -- they're just interpreted a bit differently since they are not deterministic forecasts but instead measures of spread in the solutions or the mean of all possible solutions.
Plumes - http://www.spc.noaa.gov/exper/sref/f...5&mTYP=roadmap
Maps - http://www.spc.noaa.gov/exper/sref/
Bookmarks