[OSRS-PROJ] Goode projection discrepancy

[Morten Hulden]

Can anyone explain why PROJ in the Goode projection uses a different
latitude breakpoint for switching from Sinusoidal to Mollweide projection.
Other sources (e.g.  http://edcdaac.usgs.gov/1KM/goodesarticle.html ,
http://edcftp.cr.usgs.gov/pub/software/misc/gihll2ls.c ) have the
breakpoint at 40 44' 11.8 degrees (0.71098798993 rad).

The PROJ function uses 0.71093078197902358062 (exactly 40 44'
degrees), but then adds/substracts a correction term (Y_COR) to the
output.

Is the result the same? My limited math mind suggests it's not, though the
difference may not be significant on global maps.

I noted this when I was looking at implementing interrupted goode
homolosine in PROJ. Interrupted or not, the breakpoint should be at 40
44' 11.8, no?

Morten Hulden


Ref. from goode.c:

#define Y_COR		0.05280
#define PHI_LIM	.71093078197902358062
FORWARD(s_forward); /* spheroid */
	if (fabs(lp.phi) <= PHI_LIM)
		xy = P->sinu->fwd(lp, P->sinu);
	else {
		xy = P->moll->fwd(lp, P->moll);
		xy.y -= lp.phi >= 0.0 ? Y_COR : -Y_COR;
	}
	return (xy);
}



----------------------------------------
PROJ.4 Discussion List
See http://www.remotesensing.org/proj for subscription, unsubscription
and other information.