Trying to find the SRID of each official-updated ("SIRGAS 2000" or WGS84) UTM quadrant of Brazil.

I am supposing that Latitude bands can use precise local projections, so on different SRID for each quadrant. But...

PostGIS and reference websites (epsg.io or spatialreference.org) supply only SRID of the quadrant N (e.g. the SRID of quandrant "20N" is 31974 and "19N" is 31973), because its central meridians are at Equator line... And about other quadrants?.

The EPSG standard offer other (not N) SRIDs quadrants. For exemple quadrant SIRGAS 22S has SRID 31982, WGS 84 22S has SRID 32722, 25S has 32725... But not all.


At Web and PostGIS's spatial_ref_sys table I see only 3 SRIDs, that are official and updated:

SELECT auth_name,srtext FROM spatial_ref_sys WHERE srid IN (31972,31973,31974);

All srtext say something as "PROJCS... SIRGAS 2000 / UTM zone ...". So, it shows a clue to find others, and the query returns 24 rows, but no one is for Brazil (!).

SELECT srid, auth_name, substr(srtext,1,35)||' ...' srtext_clues
FROM spatial_ref_sys
WHERE srtext ilike '%SIRGAS%2000%UTM%zone%';

When looking for a specific quadrant, for example 18H (or other any 18 that is not 18N), no return, query below is empty.
PS: and other quadrants no one is SIRGAS (neither WGS84), so no one is the "official updated projection".

SELECT srid, auth_name, substr(srtext,1,45)||' ...' srtext_clues
FROM spatial_ref_sys
WHERE srtext ilike '%UTM%zone%' AND srtext~'18[HJKLM]';
  • The official source for EPSG definitions is at epsg.org. I do not find there anything about for example SIRGAS and 20M. What is your reference for those quadrants? You say they are official. EPSG database has codes for Northern and Southern zones, though.
    – user30184
    Aug 28 at 15:29
  • What happened to the nearly identical Question on this topic posted earlier today? Deleting and reposting is something that the SE automated anti-spam defenses look for, and can result in your login being locked out. Note that the Question itself has fundamental issues, since UTM runs from South Pole to North Pole, with cylindrical tangency at the Equator -- You shouldn't move the central meridian, just change the false Y origin.
    – Vince
    Aug 28 at 16:27
  • 1
    Could it be that the map does not present UTM zones at all but the Grid zone designation en.wikipedia.org/wiki/… of the Military Grid Reference System (MGRS)? If that happens be the case then your plan to insert new rows into spatial_ref_sys is wrong and you do not need to do that. Grid zone 22M is using UTM-22S and 20M is using UTM-20S.
    – user30184
    Aug 28 at 16:51
  • 2
    What you call "UTM Grid" is actually the Military Grid Referencing System (MGRS). Lots of folks conflate UTM to MGRS, but they are different (MGRS being built on top of UTM). See gis.stackexchange.com/questions/261127/mgrs-grid-in-qgis
    – Vince
    Aug 28 at 17:06
  • 1
    Then you need to change the question to not put the resolution at the top (breaking the Q&A model). Right now the Question is a bit of a mess, and difficult to understand (the extensive use of bolding is worthy of a downvote, just on the pain induced by reading it).
    – Vince
    Aug 28 at 19:09

1 Answer 1


Oops, seems that my question is "a XY Problem"... More precisely two confusions:

1- Conceptual confusion:

2- Confusion about SRID descriptions:

  • Y2 Problem: "UTM zone 18N" is the quadrant 18N of the set {..., 18L, 18M, 18N, ...}, that is, the Zone 18 Latitude bands. A wrong hypothesis.

  • X2 Problem: "UTM zone 18N" is the "zone 18 at Northern hemisphere", and "UTM zone 18S" the Southern.

Note. As explained in the question-body, a possible source of confusion is the EPSG standard, with many SRID variations, some non-complete sets of UTM variations.

The "official SRID" for UTM Grid and UTM zones

Official, and easy rule-of-the-thumb. Prefer the SRID with the title "WGS 84 / UTM zone {x}{h}",
where h=Hemisphere, Southern or Northern:

  • when h=S use SRID number326{x}
  • when h=N use SRID number327{x}

For example for Zone 22S use x=22 and h=S, so SRID=32622.

To automatize by latitude check, use h=S when ST_Y(geom)>0, see a utmzone() function here.

Grid geocode-prefixes are all based on UTM zone number, and suffixe-letters are the Latitude bands: the letter "N" is the first letter in "northern hemisphere" — so letters "A" to "M" are southern hemisphere, "N" to "X" northern.

Official SIRGAS 2000 for Brazil

SELECT srid,  substr(srtext,9,26) as title, left(proj4text,26)||'...' as proj4text
FROM  spatial_ref_sys
WHERE srtext ilike '%SIRGAS%2000%UTM%zone%' AND srtext~'(18|19|20|21|22|23|24|25)S'
srid title proj4text
31978 SIRGAS 2000 / UTM zone 18S +proj=utm +zone=18 +south ...
31979 SIRGAS 2000 / UTM zone 19S +proj=utm +zone=19 +south ...
31980 SIRGAS 2000 / UTM zone 20S +proj=utm +zone=20 +south ...
31981 SIRGAS 2000 / UTM zone 21S +proj=utm +zone=21 +south ...
31982 SIRGAS 2000 / UTM zone 22S +proj=utm +zone=22 +south ...
31983 SIRGAS 2000 / UTM zone 23S +proj=utm +zone=23 +south ...
31984 SIRGAS 2000 / UTM zone 24S +proj=utm +zone=24 +south ...
31985 SIRGAS 2000 / UTM zone 25S +proj=utm +zone=25 +south ...

WGS84_UTMzone as alternative SRID for SIRGAS2000

SIRGAS is not world-wide, is only a South American Datum, characterized by its network of ~200 stations

Therefore it has non-complete UTM Zones set at ESPG standard... But is possible to confirm that SIRGAS 2000 is de jure and de facto, since ~2015, WGS 84.

It is easy to check by implementation, using the SQL query select srid, substr(srtext,9,26)||' ...' srtext_fragment, proj4text from spatial_ref_sys where srid in (6210,32623,4326):

SRID srtext_fragment proj4text
4326 WGS 84 +proj=longlat +datum=WGS84
6210 SIRGAS 2000/ UTMzone 23N +proj=utm +zone=23 +ellps=GRS80 +towgs84=0,0,0,0,0,0,0 +units=m
32623 WGS 84/ UTMzone 23N +proj=utm +zone=23 +datum=WGS84 +units=m

The old ellps=GRS80 has minimal difference (millimettric according to Wikipedia) and only impact on high precision altitude measures. On surface measurements the WGS 84 is a synonymous, as showed by towgs84=0,0,0,0,0,0,0 parameter.

SRID equivalence:

31978 18S 32618
31979 19S 32619
31980 20S 32620
31981 21S 32621
31982 22S 32622
31983 23S 32623
31984 24S 32624
31985 25S 32625
  • You've got the XY Problem wrong. X is what you want (the root issue), Y is what you ask instead (the non-solution).
    – Vince
    Aug 28 at 19:48
  • Thanks again @Vince, correct XY order and add the "official SRID" considerations Aug 28 at 20:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.