gmt basemap polar stereographic projection, 90 degree latitude tick is funny

[mkononets]

I am making a map using a polar stereograpic projection, and wanted to get a 90 S tick and annotation on one axis. Could not get this working, and discovered that the 90 degree latitude tick is actually present on the inside.

I would like to be able to add 90 S annotation (that I did externally) and a tick, but could not get the tick. Some code:

gmt basemap -R-90/90/-90/-60 -JS0/-90/10c -BWsEN -Bxaf -Byaf --MAP_FRAME_TYPE=plain -png test

the plot, see what the tick at the pole is doing -BWSEN produces the same result:

Both ticks are actually present, W and E, gmt basemap -R-90/45/-90/-60 -JS0/-90/10c -BWSEN -Bxaf -Byaf --MAP_FRAME_TYPE=plain -png test:

I wanted the 90 S tick present on the E axis looking outside (not inside), and annotated (at least the tick present so I could manually create annotation). Bottom axis (apparently W in this specific arrangement of plot axes) tics were not needed as I used bottom axis to plot annotations in projected coordinates. S axis is degraded to the pole position, practically does not exist as I could understand.

See the 90 S tick on the plot below. The 90 S annotation was added externally, and the tick also added externally using the frame in projected coordinates with a tick step exceeding the map size so only one tick is plotted. This trick however does not work if the right axis is not vertical e.g. in case region is -R-90/15/-90/-60 -Js0/-90/-65/$scale

[joa-quim]

It looks that when it reaches the 90S the tick immediately jumps to the perpendicular to the other side. Better explained by figs.

gmt basemap -R-90/45/-90/-60 -JS0/-90/10c -BEN -Bxaf -Byaf --MAP_FRAME_TYPE=plain -png test_EN

gmt basemap -R-90/45/-90/-60 -JS0/-90/10c -BWN -Bxaf -Byaf --MAP_FRAME_TYPE=plain -png test_WN

[mkononets]

So the only thing left currently is something like drawing that tick manually using the MAP_TICK_LENGTH_PRIMARY and MAP_TICK_PEN_PRIMARY? I wonder how approximately?

I can calculate tick direction angle cause I set the stereographic projection's E and W boundaries. Is there something like a vector with no arrow that can be plotted from the point (0,-90) at the given angle having length of MAP_TICK_LENGTH_PRIMARY using pen set by MAP_TICK_PEN_PRIMARY?

[joa-quim]

Maybe a small custom axis that covers only that bit?

[mkononets]

I don't understand how to make a small custom axis exactly positioned where I need it and at the needed angle. E.g. projection is -R-90/10/-90/-60 -Js0/-90/-65/$scale so the right axis is a e.g. a 10 degree slope. I could not do this using projected coords as those are rectangular. And with -Js 90 degree the tick is turned inside and the label is not plotted.

I mean getting a 90 S tick there (plot updated, tick added using workaround shown below):

[joa-quim]

I thought more on a small Cartesian custom axis to avoid the 90 non-ploting issue. But positioning and orientation may be a problem though.
Sorry, have to leave now.

[mkononets]

Sure, thanks for answering!

[mkononets]

workaround (the plot above was updated just in case)

SCALE=1:25000000
FONT=10p,Helvetica
THETA=10 # 10 degrees
#Y 90 S tick, missing on the basemap frame/axes ticks
echo 0 -90 -$THETA 4p | gmt plot -R-90/${THETA}/-90/-60 -Js0/-90/-65/${SCALE} -Sv10.0+p -Wthinner,black -N

# vertical Y axis latitude annotations
Y_OFF=$(gmt math -Q $THETA -1 MUL SIND 20 MUL =)
X_OFF=$(gmt math -Q $THETA COSD 20 MUL =)
 gmt text -F+jMC+f${FONT} -D${X_OFF}p/${Y_OFF}p -N <<-EOF
${THETA} -60 60@.S
${THETA} -75 75@.S
${THETA} -90 90@.S
EOF

BTW read the docs but could never understand what that 1.0 in -Sv1.0+p is doing. Tried to set it to a different value like 10.0, made no visible difference on the plot. If I skipped it like -Sv+p, no vector was plotted and error message shown:

plot [WARNING]: Mismatch between actual (4) and expected (5) fields near line 1 in file <Standard Input>

UPD 1.0 in -Sv1.0+p is the size of the vector arrow head. Makes no difference once one does not need the head.

[joa-quim]

could never understand what that 1.0 in -Sv1.0+p is doing.

Did not check the docs but I think it's controling the shape of the vector head. 1.0 is maximum. Try 0.5 to see if it makes a difference.

[mkononets]

You are absolutely right, the docs clearly state that "size" is the length of the vector head. I just somehow missed it while reading the docs. Anyway I did not need vector head, just a straight line to draw my axis tick.

The docs however do not mention that 1.0 is the maximum size. Below is what's in the web docs for -Sv

-Svsize[+vecmodifiers]

    vector. The direction (in degrees counter-clockwise from horizontal) 
and length must be found after the location [and value] columns, and 
size, if not specified on the command-line, should be present as well, 
pushing the other items to later columns. The size is the length of the 
vector head. Vector stem width is set by [-W](https://docs.generic-mapping-tools.org/6.6/plot.html#w), 
with head outline pen width defaulting to half of stem pen width. 
See [Vector Attributes](https://docs.generic-mapping-tools.org/6.6/plot.html#vector-attributes) 
for specifying this and other attributes.

[mkononets]

There's an excellent section in the Technical reference:
The Drawing of Vectors

Would be excellent if it is cited in gmt plot documentation, section about drawing vectors

[joa-quim]

PR #8831 created.