I don’t know if they chose to omit those, directly, versus how difficult it is to capture them and render them in a realistic manner using the existing weather generation system.
I’ve talked about this at length elsewhere: individual thunderstorm cells are fairly ephemeral phenomena. They exist for a relatively short period of time, and are caused by a fairly particular set of circumstances (moist air, lift, and a steep, unstable lapse rate). They might be individual “airmass”-type cells, they might be embedded in a larger area of rain, they might be longer-lived multicells or supercells or they might be organized into a larger-scale process like a mesoscale convective system (MCS) or a squall line. The difficulty is each of these behaves differently in terms of size, shape, structure, intensity, longevity, and general behavior. You couldn’t plop down an airmass-type cell to realistically represent a supercell or squall line, for example.
We don’t know for sure, but based on things that have been said by the dev team over the years and my observations, my understanding is that the sim’s weather system uses a blend of numerical models (forecasts) and augments this with current observations like satellite and METAR (and possibly radar) to generate weather. The numerical models are good at saying “this area might have some storms at this time of the day,” but the exact timing, location, size, etc, is fairly unpredictable until those storms actually initiate. This usually happens late in the afternoon, when the low-level capping inversion is broken by surface heating. At that point, storms are basically “free” to initiate.
However, remember the forecast models that are the source of all of this can’t predict exactly where this will happen, just a general time and place. So they’re kind of waiting for the storms to kick off so they can receive input from radar, satellite, and surface observations (basically METAR), which feeds back into the model and refines it. If the storms are of the long-duration type, they become more predictable and the models will now definitely reflect this. It’s at this point where we might actually start to see them in the sim. I’ve seen strong long-duration squall lines or MCS type storms rendered in 2020 (might not see them in 2024 yet because there really haven’t been a lot of storms where I fly this time of year).
Because they don’t last long enough to get a “bead” on their nature, and their relatively small size, the exact location of shorter-lived storms really never show up in the models, which don’t have the spatial or temporal resolution to depict them. The most rapid-refresh models are hourly at best, and have a resolution that is larger than most individual storms - remember, this is being done by some of the most powerful supercomputers in the world, taking a relatively long time to do so and they can’t even get it exactly right. So instead, they are usually depicted as an area of randomly-drawn cells as a best guess.
Now, these could be thrown into the sim as such, but is problematic to the way pilots gather and use information (radar, METAR, etc) to avoid them. It’s one thing to have an isolated storm surrounded by clear air and is visible and easy to avoid - we don’t need it to be exact for that. But when they’re embedded in other clouds and otherwise not visible, we have to use on-board radar, or ATC-based radar, or network radar (with the caveat it’s heavily delayed) to see where those cells might be and where they might go. Otherwise their nature and behavior would be anybody’s guess, which is kinda-sorta but not quite how it works in real-life - flying into an area like this definitely gives light aircraft pause, but not so much other planes, who “pick around” individual cells that may or may not be visually seen.
My best guess - I think the sim is probably at the point it could generate a “random” area of storms, but my guess is there are still technical hurdles. Either way, getting it wrong could upset as many or more people than not getting right - you shut down vast swaths of airspace and airports for ghosts that we can’t otherwise observe and avoid.
This is a long way to say the weather info feeding sim really doesn’t know where and when they’ll be, what they’ll look like, or how they’ll behave until they’re observed in real life. And most cells don’t last long enough to feed back into the model used for the sim. So that’s likely why you don’t see a lot of storms in the sim. But that only covers the existence of a storm itself.
When the sim does happen to render a storm, I don’t know what’s holding them back in terms of the hazards in the storm - icing, heavy, often frozen precip, up/down drafts and extreme turbulence, lightning, etc. Maybe they could just do a generic set of parameters - but they’d have to get it right based on the storm type.
And when the sim does happen to render a storm, I don’t know why they look so anemic. The graphical rendering capability is outside my area of expertise.
