We agree. Henger et al. provide a more accurate location for the epicenter of the 13 January event in Novaya Zemlya - one of the examples presented in our article. We based our location on the data from the ARCESS (ARA0) and Spitsbergen (SPA0) arrays, and were unaware that signals from Norilsk (NRI) indicated a contemporaneous earthquake on the mid-ocean ridge west of Spitsbergen.
The conclusions of our article, however, are in no way altered by the refinement of the epicenter location. As we have repeatedly seen, earthquakes that occur in sensitive areas during politically sensitive times are vulnerable to misinterpretation. Independent of which location is used (all of the epicenter determinations for the 13 January 1996 event are more than 200 kilometers from the Russian test site), the earthquake raised concern over Russian compliance with the ban on nuclear testing. Given the lowering of detection thresholds and the continuation of subcritical experiments, it is not surprising that such coincidences occur. And in fact, since the writing of our article, such a coincidence has occurred again. (see news articles inset and the bannergram of this Newsletter).
To help prevent such false alarms in the future, data from all available stations needs to be examined. Henger et al.´s extraction of critical travel-time information for a magnitude 2.4 event from a single 3-component station at a distance of 1200 km is an enlightening example for those who may still debate the importance of auxiliary stations for the monitoring of the CTBT. For the more recent 16 August 1997 false alarm, critical data came from the IRIS GSN station KEV, which is not even part of the official CTBT monitoring system.
We disagree with Henger at al in their assertion that a magnitude 2.4 event, which is equivalent to a yield not exceeding 100 tons of TNT, "could hardly be termed a nuclear device in a CTBT context". From the United States´ perspective, it most certainly would. On August 11, 1996 President Clinton rejected proposals for continued sub-kiloton tests, and announced his decision to pursue a "true zero-yield comprehensive test ban." He based such a decision in recognition that the act of nuclear testing, not the threshold of such tests, is objectionable within the non-proliferation regime.
Both the United States and Russia are conducting hydrodynamic experiments at their test sites. Whether these experiments must remain sub-critical, or whether they can produce modest yields is unclear from the negotiations. Because nuclear materials such as plutonium are used in the experiments, they are conducted in a manner similar to nuclear tests. Although there is currently no precise agreed upon technical definition of the maximum explosive energy from hydrodynamic experiments, a seismic signal corresponding to 100 tons of tamped explosive would be assumed as a violation of the ban on nuclear testing by the United States.
Areas such as Novaya Zemlya, that are generally considered aseismic at teleseismic magnitude levels, will appear seismic at the lower magnitude levels detectable from regional coverage. Regional networks now provide a detection capability for the Novaya Zemlya area that is near mb 2.5 (NORSAR, 1996). Accordingly, we can expect to find events of concern every few years, such as 16 August 1997, 13 January 1996, 13 January 1995, and 31 December 1992. Open networks in the western United States, where the seismicity is greater and the coverage more extensive, provide a detection threshold below magnitude 2.2 for the US nuclear test site (Hennet et al. 1996). In fact, six days before the US´ second 1997 subcritical experiment, a seismic event was recorded on the Nevada Test Site. 
NORSAR semi-annual technical report, November 1996.