Vertical velocities are temporarily disabled, due to issues #17 and #26 which revealed that the vertical velocity handling was introducing unexpected artifacts. The mask used to propagate water and ice from time step to time step has also been replaced. Previously, the ice particle number was used, with locations retained if it exceeded 1.0e-5 times the highest value remaining in the simulation. This had two downsides. First, if the ice particle density became somewhat uniform, then increasingly large areas would be preserved simply because the peak number was falling. Second, regions which were previously dehydrated but which crystals were not currently present in would be marked as not containing contrail, meaning that the effect of earlier passage of ice would not necessarily be preserved. Both issues are resolved by the use of an inert tracer to represent contrail influence. The inert tracer is set to 1.0 within the plume at initiation of the 2D simulation, and then allowed to advect and diffuse as a gas. At each time step it is then set to its current value plus the local number concentration in particles per meter cubed, capped at a value of 1.0. Any location which has a value of at least 0.2 is then retained when transferring data to the next time step, while all other locations are overwritten with zero ice and with water from the background meteorology at that time and altitude.
This edit can result in longer run times as ice can now persist in a just-saturated layer at low altitude. Users may wish to experiment with changing the thresholds for triggering the end of the simulation if they want to reduce run time.
Vertical velocities are temporarily disabled, due to issues #17 and #26 which revealed that the vertical velocity handling was introducing unexpected artifacts. The mask used to propagate water and ice from time step to time step has also been replaced. Previously, the ice particle number was used, with locations retained if it exceeded 1.0e-5 times the highest value remaining in the simulation. This had two downsides. First, if the ice particle density became somewhat uniform, then increasingly large areas would be preserved simply because the peak number was falling. Second, regions which were previously dehydrated but which crystals were not currently present in would be marked as not containing contrail, meaning that the effect of earlier passage of ice would not necessarily be preserved. Both issues are resolved by the use of an inert tracer to represent contrail influence. The inert tracer is set to 1.0 within the plume at initiation of the 2D simulation, and then allowed to advect and diffuse as a gas. At each time step it is then set to its current value plus the local number concentration in particles per meter cubed, capped at a value of 1.0. Any location which has a value of at least 0.2 is then retained when transferring data to the next time step, while all other locations are overwritten with zero ice and with water from the background meteorology at that time and altitude.
This edit can result in longer run times as ice can now persist in a just-saturated layer at low altitude. Users may wish to experiment with changing the thresholds for triggering the end of the simulation if they want to reduce run time.