ProjTests

Map projection tests of PROJ.org library projections, using PostGIS.

Install

On Linux with PostgreSQL v9+.

cd ProjTests/src
cat step0*.sql | psql postgres://postgres@localhost/test

Testing DGGS

Some DGGS projections with "good fit to WGS84" claims (ref1, ref2), that are not seems realistic for Brazilian territory.

At this PostGIS SQL scripts, using ST_Area(geom,true) as reference of "WGS84 area measure", and ST_Area(geom) as projected sample, we can reproduce the projtest.vw06_proj_cmp_summary report:

projection label and definition	proj_diff_avg
BR:IBGE   classical Albers equal-area, +proj=aea +lat_0=-12 +lon_0=-54 +lat_1=-2 +lat_2=-22 +ellps=WGS84	0.0%
DGGS:001:rHEALPix   +proj=rhealpix (default +ellps=GRS80)	17.8%
DGGS:002:rHEALPix   +proj=rhealpix +ellps=WGS84	17.8%
DGGS:030:ISEA   +proj=isea	0.5%
DGGS:032:ISEA   +proj=isea +ellps=WGS84	0.5%

The column proj_diff_avg is the average of area-difference of some sample zones (Geohash rectangles with different areas) at Brazil. The VIEW projtest.vw02_proj_cmp_raw shows raw data. To see Geohash cells ("ghs" column), use viewer1 or viewer2.

The projection BR:IBGE is the official, used by Census, demonstrating that PROJ Lib can fit to WGS84.

Testing area distortion by elevation

Using projtest.vw04_area_factors_summary we can show that the "WGS84 area measure" distortion (column "km2_change") grows with elevation, and that from ~1 km elevation upwards, the distortion of ~0.01% (or more) cannot be ignored.

city_name	elevation	km2_change
Fortaleza	20	0.0003%
Uruguaiana	70	0.0011%
Manaus	90	0.0014%
Santa Maria	110	0.0017%
Trindade (ilha)	170	0.0027%
Sao Paulo	760	0.0120%
Mané Garrincha (estadio)	1200	0.0188%
La Paz	3650	0.0573%

License

CC0.