ISOBARIC TO ISENTROPIC INTERPOLATION ERRORS AND IMPLICATION TO POTENTIAL VORTICITY ANALYSIS

Different common interpolation methods have been used for transformati on of atmospheric variables from pressure levels into isentropic level s. The accuracy of the derived variables depends both on the specific interpolation method chosen and on the vertical resolution of the pres sure levels. A most sensitive field is partial-derivative theta/partia l-derivative p, the stability factor, in which error largely determine s also that of the isentropic potential vorticity. Five interpolation methods are compared using radiosonde observations. They are examined through their capability to resolve isentropic profiles of the stabili ty factor as well as the pressure based on data from arbitrary pressur e surfaces. Comparison of the results with full resolution data reveal s, not unexpectedly, that fine vertical patterns are not resolvable. T he methods differ in their ability to resolve patterns found with 10-p ressure-level data when applied to only seven levels. However, the met hod based on the linear theta-p relation is shown to yield larger erro rs than other methods, such as that based on the linear T-ln p relatio n. In addition, the advantage of the latter method is clearly demonstr ated through isentropic and vertical potential vorticity distribution with ECMWF data.