Summary of Changes to NSSL Scheme in V3.6

- Cloud nucleation: The default two-moment cloud droplet nucleation has been changed to use the Twomey 1959 estimate of maximum supersaturation (irenuc=2). This parameterization was found to have better consistency at very small time steps (dt < 0.5). The old nucleation is can still be used by setting irenuc=1 and recompiling.

- The default rain size distribution has been changed from a Gamma function of volume (imurain=3) to a gamma function of diameter (imurain=1), which is more common in bulk parameterizations. The value of imurain can be changed in the code if anybody desires to, in which case a bug fix has been made in the ventillation coefficient.

- Mass-weighted drop diameter is now used for breakup instead of mean volume diameter.

- Ice nucleation options (icenucopt). Ferrier-Meyers is still the default. Alternate options are icenucopt=2 for Cooper (as in Thompson et al. 2004) and icenucopt=3 for Phillips et al. 2007. (Must recompile if icenucopt is changed.)

- Added subroutine "calcnfromq" that runs on first time step to convert single moment initialized hydrometeors to two-moment (useful for cloud analysis initial conditions).

- Minor change in snow aggregation efficiency.

- More stringent sedimentation for large time steps to prevent problems with thin model layers