Behind the friendly WINDOWS user interface there are a lot of nasty numerical routines doing hard work to calculate the climate in your model area. Just a few keyword, on what is used in ENVI-met:The model equations are solved in three dimensions using the ADI (alternating direction implicit) method. Using a fully implicit scheme allows ENVI-met to use time steps up to 10 sec without getting numerically unstable. Of course, different modules have to be solved one after the other in order to manage the data flow. For example, first all the surface temperatures are calculated from the energy budget and then the 3D temperature in the atmosphere is calculated using these surface temperatures.

Some modules require smaller time steps such as the turbulence, the pollutant dispersion model, especially if sedimentation processes are involved. To calculate the wind flow, the pressure is removed from the Navier-Stokes equations and an auxiliary flow field is computed first (splitting method). After that, the Poisson equation is solved to calculate the corresponding pressure perturbation field. Here, the SOR algorithm is used. It is a little bit slower than a direct method, but in general more “friendly”. Although it is possible to calculate the wind field continuously (each time step), recent computers are still to slow to do that because the calculation of the wind field requires very small time steps due to steep gradients e.g. at building walls. Therefore the usual way is to “update” the wind flow after a given time interval in order to match it with the stratification of the atmosphere. As a drawback of this method, slow thermal flows cannot be calculated with this approach.