[ESIP-all] Looking for a boat

[James Frew]

In Jan/Feb 2007 I was one of a large group of volunteers involved in the 
(unsuccessful) search for Jim Gray ( 
http://www.wired.com/techbiz/people/magazine/15-08/ff_jimgray ).

The implementation of the search is described in a CACM article ( 
DOI:10.1145/1965724.1965744 , email me if you can't access this and 
would like a copy. )

Lessons learned, short version:

* People who are lost at sea are either found by professionals, or they 
stay lost. Finding them otherwise is a bet against *very* long odds.

* A small (i.e., almost any non-commercial) boat is *very* hard to 
detect with current (unclassified) remote sensing technology.

* That said, high-resolution optical beats radar: most non-commercial 
vessels are too small, and constructed of the wrong materials, to yield 
a good radar return. (Many of these vessels have radar reflectors, but 
they're aimed sideways (to show up on commercial vessels' radars), not up.)

* Drift modeling is essential in order to know where to look now for 
something you saw "then".

* Drift modeling is hard---the CG uses hi-res wind and current fields, 
plus a library of the hydrodynamic characteristics of *anything* that 
could float and support >=1 person. Even then, the models diverge wildly 
in a matter of single-digit days.

Based on this, I opine that the "tomnod" process is your best bet for 
generating a list of candidate sightings. However, these won't be much 
use without:

* Screening for spurious hits

* Comparison with known vessel locations

* Drift modeling

Also, anytime you have a gaggle of amateurs (like me---the pros in this 
case are national search-and-rescue orgs like the CG) working at 
something like this, you need a reliable communication mechanism. For 
the Jim Gray search, this was a blog ( http://openphi.net/tenacious/ ), 
which both centralized the collection of information and limited the 
spread of misinformation.

Good luck all,
/Frew

P.S.: Apologies in advance to anyone invested in this, but I checked out 
http://evxx.com/ , and there's a lot of negativity there about various 
agencies whose assistance will be essential in any resolution to this 
crisis. Like it or not, only national militaries and SAR orgs have 
anything approaching the capacity to deploy SAR over a chunk of ocean 
the size of half a continent.

P.P.S.: SAR == search and rescue (even the terminology is a problem...)