I think “plane” events (landing, fire, failure) should be implemented like this:
Incidents have a probability to happen, e.g. 1/30 or 3%
That means around 1 out of 30 flights will have an incident.
Then the incidents have a chance that there’s really something wrong (or still wrong, when landing), e.g. around 40% (of these flights with incident). (because many issues can be addressed mid-air)
If there’s really something wrong, there’s a probability that people will die (or that one passenger will die). It has a random base value or / and a base value that is affected by the type of emergency (engine fire: 20% ±15; no gear 15% ± 10; etc)
And that probability is then influenced by:
how much time the fire trucks need to the runway while the plane is landed. (the probability (base value) for people dying grows over time while no help is near)
how many trucks are there, calculated over a limited growth function. Example: There’s an engine fire with a base-value of 25%. One fire truck would reduce that by 8%, the second by 5% the third by 3% the fourth by 2% (values have to be changed).
The growth function should also scale with the number of passengers to provide a realistic experience (larger plane = more fire trucks needed)
You are able to accept contracts with the local city to provide emergency services, but they will need much longer to get to the plane as they drive there over the public road (like delivery trucks).
You could also build you own infrastructure (fire department, vehicle depot for security). They will need much less time furthermore decreasing the fatality-risk.
Maybe the contract would cost money per emergency, while the own infrastructure costs per hour. So if you have only a few incidents, the contract would be cheaper, but if you have a lot of flights = more incidents, the own infrastrucure is cheaper to operate.
It could an option that the player has to decide by hand how many vehicles should respond when having a contract.
Of course emergency services could also influence the airport ranking, e.g. 3-star and above requires airport emergency vehicles and not the contract-type.
I think the probability-scheme would provide a good balance between having emergencies = making it fun to play, but having some incidents with fatalities, while some could still be prevented by having enough vehicles.
This design / process would also work for medical and military incidents.