Detention at Sea: Is a 90-Day Limit Required by Law, or Is It Just Good Policy?

Julian Barnes and Evan Perez have an interesting piece today in the Wall Street Journal, suggesting that at least some military officers are increasingly concerned about the lack of a clear option for detaining terrorism suspects (other than prosecution).  This is familiar ground that I don't want to rehash right now.  Instead, I want to draw attention to the following passage in the report:

Senior defense officials said they believe they can hold a detainee up to 90 days in international waters. But after that, they fear they could be violating the Conventions. And with no long-term detention facility at the ready, they worry about missing chances to gather intelligence from suspects.

I'm not quite sure where the 90-day figure comes from.  Certainly not the Geneva Conventions, as the article suggests.  Or at least not directly from it. When the full conventions are applicable, sea-based detention is indeed forbidden.  And while some flexibility for merely transitory uses of such detention is inevitable (above all for captures on the water, obviously), there is nothing that picks out 90 days as either too much or too little as the measure of "temporariness." That said, having a 90-day limit does at least preclude the argument that detention at sea is indefinite/permanent. Of course, the full Geneva Conventions most likely would not apply for the non-Iraq, non-Afghanistan captures at issue in this article; rather, Common Article 3 would govern, assuming that there is a relevant armed conflict to which the capture is connected.  Common Article 3 says nothing about sea-based detention, as it happens, and it is unclear whether a prohibition on sea-based detention might at least be part of the customary law of armed conflict applicable in non-internationalized settings; certainly there is nothing that makes 90 days acceptable, or not, as the outer boundary imposed by international law, though again such a limit does at least foreclose arguments about permanent/indefinite reliance on vessels. In light of all this, I think the best way to understand the quote above is that the administration (or DoD at least) has determined that 90 days is the maximum amount of time for which sea-based detention should be used in a climate in which the relevant legal rule is indeterminate, but not that this is clearly legally required.  Put another way, adopting this limit may not be strictly required, but does at least support the characterization that such detention is indeed temporary in the end, and hence less in tension with the arguable prohibition on at least full-time/sustained detention at sea.  On the other hand, if you believe there is indeed a prohibition on sea-based detention in the NIAC setting, and further that the only exception is for situations of transit, you probably will not be mollified by the cap being set at 90 days.