Abstract
We have investigated the accuracies of three isentropic trajectory techniques, in particular, their dependence on the time interval between available data fields. The first trajectory technique is kinematic, using only the wind information, while the other two, the explicit method of Petersen and Uccellini and an automated version of Danielsen's implicit technique, are dynamic (use both mass and velocity fields). A primitive equation channel model in pressure coordinates was used to generate the basic flow fields. We compared the techniques for data intervals ranging from 12 min to 12 h. The overall level of agreement, even for a 12-h data interval, is very good, particularly for the dynamic techniques. This result implies that, given a spatially and temporally coherent data set, accurate trajectory calculations in the free atmosphere can be performed using these techniques even when data are available only twice daily, as is the case for routine aerological observations. Whereas more instances of important discrepancies were found for the kinematic technique with a 12-h data interval, such cases amounted to only 15 percent of the total. For relatively slowly growing systems, such as those studied here, the velocity deformation field determines the level of agreement, as the significant discrepancies occur in regions of high deformation. The implicit technique can give erroneous results near jet streaks. We investigated the accuracy for more rapidly evolving flow fields and found more serious degradation. Also, by simulating the effect of the diurnal temperature variation on the geopotential field we found a less serious but noticeable degradation of both dynamic techniques.
