-10- 
the remaining 6 probes above the CTD trace is below 10 m on average. According to 
Sippican, the depth accuracy is given as + 2 % of depth, which is + 8 m at 400 m 
depth. 
It has to be emphasized that the numbers presented are more or less raw estimates. 
Nevertheless, they show a tendency towards negative values (i.e. underestimation of 
depth fall rate), no significant depth fall rate error resulted from this test. That is in 
contrast to results from other groups, who found significant deviations towards faster 
depth fall rates than those calculated by the standard coefficients of the drop equation 
(Heinmiller et al., 1983; Szabados and Wright, 1989; Bailey et al., 1989). This could be 
due to the weak temperature gradient of less than 0.01 °C/m or due to some internal 
variability which, however, was found to be small. It may even be that the depth fall rate 
error varies with the area, i.e. with density or with the actual depth (elapsed time) where 
the comparison took place. In addition, it is very likely that different results may be 
caused by some certain differences of the probes (batches) themselves, the time of the 
probe’s storage, and the XBT units used. Finally, it cannot be excluded that the cali 
bration procedure or other factors might affect the dynamics of the probes tested 
(Green, 1984). All these points have to be proved and discussed in context with all the 
results available from all groups. 
Besides the direct effects on data quality and reliability of results, additionally an 
inaccurate depth fall rate equation has community-wide implications if corrected indi 
vidually. Thus, in October 1989, at the IGOSS Ship-of-Opportunity Meeting in Ham 
burg, F.R.Germany, a decision was made to conduct an internationally co-ordinated 
experiment to evaluate the XBT depth equation (IOC, 1989). The NOAA National 
Ocean Service and the IGOSS Task Team on Quality Control were asked to take care 
of that task. Szabados (pers. comm.) drafted corresponding procedures which should 
ensure a controlled test in as many regions as possible. Results of the collaborative 
effort are planned to be discussed during the TOGA International Science Meeting in 
July 1990 and the next IGOSS Ship-of-Opportunity Meeting in 1991. It is recommended 
that no changes should be made to the fall rate equation prior to internationally adopted 
solutions.