[Taxacom] Geodetic datums do matter

[Doug Yanega]

Dave Watts wrote:

>It does matter - a geodetic datum refers to a model of world that 
>attempts to fit the shape of the earth and therefore
>is an approximation to it. Each datum uses various parameters to 
>define its shape and hence each coordinate of the same place will be 
>different for each datum.
>
>Comparing WGS84  to the Australian standard AGD96  will make a point 
>appear to move 200 metres. Most users of GPS units are unaware of 
>this and complain that the GPS unit / map is in error when they 
>compare the positions  with different datums.
>
>In reference to an earlier post, GPS horizontal accuracies for hand 
>helds is likely to be only about 20 metres. If you wish accurate 
>stuff, you need a base station and do differential measurements. 
>Heights from hand-helds are also very poor.
>
>The bottom line is KNOW YOUR DATUM when recording stuff. It will 
>save you problems in the future as you can transform from any one 
>datum to another with certainty.
>
>Cheers Dave

It matters, but how MUCH it matters (the actual practical 
implications) will DEPEND ON CONTEXT. I've gotten into many arguments 
about this point with a very good friend who is an expert on 
georeferencing; he is fond of claiming that if you were dying buried 
underneath an avalanche, a *properly* georeferenced point is one that 
could direct a team of rescuers to your position and dig you out 
before you freeze to death. In such a context, yes, the datum is 
crucial. If you're in the military and targeting "smart bombs" to 
blow up certain buildings remotely, then yes, the datum is crucial. 
However, I am conversely fond of telling him that when a team of 
entomologists goes into the field to collect specimens, we take a GPS 
reading where we park the car, and then each person is free to wander 
- maybe a kilometer or more distant, collecting dozens of specimens 
at random points anywhere within that radius. There is SO MUCH ERROR 
associated with the specimens bearing that GPS reading on their 
labels that it is no longer relevant what datum it was based on; even 
the largest possible discrepancy from one datum to another is utterly 
dwarfed by the uncertainty, not only in the location of the actual 
collecting event but especially with regards to the overall spatial 
distribution of the type of organism that was collected. An error of 
200m may seem massive if you're trying to find a single, immobile 
target, but a lot of what biologists deal with are not single 
immobile targets! If I have data from a weather station, and it's 
200m from where I think it is, it isn't going to influence the value 
of the weather data. If I collect a rare bird specimen at 33 26 13 N, 
117 19 25 W, and someone goes back 50 years later and is mistakenly 
200m from the actual spot I collected that bird, then this error is 
NOT going to negatively impact their ability to find another member 
of that same species (i.e., if they don't find one, then the *reason* 
for their failure is NOT going to be that they were 200m from where 
they needed to be). Locating a spot where you can encounter members 
of species X is not the same as locating a friend's house using a GPS 
reading alone; most species have a spatial distribution greater than 
a 200m radius, and most houses are considerably smaller. Accordingly, 
it is genuinely pointless in such a context to *worry* about the 
datum used.

Even in those very special circumstances where a habitat is along a 
boundary feature such as a shoreline or cliff, where a 200m error 
might put one across that boundary, or a cave where one might miss 
the entrance, no sensible biologist is going to take a GPS reading so 
literally as to do something like climb in a boat and start searching 
200m offshore to find a known tidal pool species, for example; those 
few incredibly rare cases where this small a discrepancy may actually 
be problematic are a negligible fraction out of all possible specimen 
labels in existence. In that case, maybe people working on shoreline 
habitats or cave-dwelling species or such might be more concerned 
than others, and place greater emphasis on knowing their datum - 
again, it's a matter of context, and for many (if not most) of us, it 
simply won't matter.

Sincerely,
-- 

Doug Yanega        Dept. of Entomology         Entomology Research Museum
Univ. of California, Riverside, CA 92521-0314        skype: dyanega
phone: (951) 827-4315 (standard disclaimer: opinions are mine, not UCR's)
              http://cache.ucr.edu/~heraty/yanega.html
   "There are some enterprises in which a careful disorderliness
         is the true method" - Herman Melville, Moby Dick, Chap. 82