You can spend lotsa money on pure simulations (only hundreds?) and get some great insights on dispersal patterns and the like. It will make you feel good, but

You can spend lotsa money on pure simulations (only hundreds?) and get some great insights on dispersal patterns and the like. It will make you feel good, but unless you validate your results against field data, you won’t know if you’ve learned enough to respond effectively during an actual event. Also, you need actual observational data to meaningfully calibrate any simulation involving potential human fatalities. There are over 100 NYC restaurants serving bean burritos, so it’s not like it’s the first time such experiments have been run.