atmqty Manual: AtmQty Algorithms Used In Quantity Calculation Methods

The pydoc documentation for the class provides some details regarding the algorithms used in the AtmQty methods used to calculate derivative atmospheric quantities. Details can be found in the docstring and source code for the functions the class methods call. The module listing provides links to the pydoc documentation and source code for package modules and the procedural functions defined in those modules (and accessed by the AtmQty class methods).

Most of the algorithms used are quite standard. For instance, calculation of potential temperature uses its common definition. Here we provide details of the algorithms used for some of the more complex AtmQty methods:

ipv: Isentropic potential vorticity (IPV) is calculated using a method similar to that used to calculate IPV in the NCEP/NCAR reanalysis. The major difference is that the reanalysis IPV is not multiplied by potential temperature, whereas in the AtmQty class IPV is. Thus, IPV in the AtmQty class is given by:

IPV = (potential temperature) * (absolute vorticity) * (static stability) / (g * air density)

where the absolute vorticity is the Coriolis parameter plus the relative vorticity. Air density is calculated in the AtmQty class using virtual temperature.

rho: Air density is calculated from the ideal gas law, using virtual temperature and the gas constant for dry air. Thus, moisture effects on density are included in the calculation.

vort_rel: Relative vorticity is calculated as the vertical component of the curl of "horizontal" velocity at a given level. The vertical component is assumed relative to the lev_type levels in the AtmQty object. If SPHEREPACK is installed and accessible via the sphere package distributed with CDAT, and the domain supports the SPHEREPACK spectral algorithm, that algorithm is used to calculate curl. In all other cases (e.g. the presence of missing values in u or v, a non-global domain, etc.) an order-1 finite-difference algorithm (implemented in spherical coordinates) is used. Because of pole effects, for the spherical finite-difference algorithm, any values of relative vorticity north of 88°N or south of 88°S are set to missing.