Home

• Should match latitude and longitude between different datasets using a tolerance margin, rather than use rounded lat lon.

• Trajectory integration currently uses both start and end of the time interval of interpolation, but does not do time interpolation (on purpose, except as a first guess when the stat time is in between time steps). It iterates to find a consistent the start and end point of velocity field.

• Selected vertical part of column used for moisture weighted velocity (LowerTroposphereHumidityWeighted). Might want to detect the inversion. Roel uses velocity at a specific pressure level. Steef has implemented velocity at height.

• Mask out part of the domain which is say over the land when calculating gradients (so we can get them all the way down to the surface), currently this box is based on defining the number of degrees of the box.

• Calculate absolute height and pressure at run-time during integration of trajectory so that we have a pressure which is consistent with the thermodynamic fields, and don't have to calculate for the entire domain.

• There may be some steps at the end of creating the forcings that could be worth having as a separate, fourth, step (basically adding any parameters that do not need interaction with the domain/trajectory data)

• Convergence check for velocity integration in trajectory integration?

• Averaging area needs to be larger that LES size? SB: Physically this makes more sense, although we can allow for deriving the forcings from a smaller number of columns in sensitivity/limit studies.

• It should probably be an option to use a different area for mean profiles and gradients.