Sizing a Study Area Figure for a Journal
A figure that looks right on screen at 1800 pixels wide can be unreadable in a 90 millimetre journal column. The problem is almost never resolution. It is that the figure was designed at one size and printed at another, so everything that was comfortable became small.

The three widths that matter
Most journals publish figure widths in millimetres, and most of them use some version of three sizes. Single column is usually 85 to 90 millimetres. One and a half column is around 140. Double column, full page width, is 170 to 190 depending on the publisher.
Check your target journal rather than assuming, because the difference between 170 and 190 is enough to change whether your labels fit. The instructions to authors always state it.
The arithmetic
Pixels equal width in millimetres divided by 25.4, times the dots per inch. At 300 dpi:
90 millimetres gives 1063 pixels. 140 millimetres gives 1654. 190 millimetres gives 2244.
Set the canvas to that width and export at one times. This is the part people get backwards: they export at a comfortable screen size and then scale the image up in a word processor, which adds pixels without adding detail and leaves the labels soft. Scaling down is harmless, scaling up never is.
Which format to send
SVG when the journal accepts vector art, which most now do. Every coastline stays a path, every label stays text, and a copy editor can restyle the fonts without touching the geometry. It also means there is no resolution to get wrong. A detailed panel typically runs to about a megabyte, which is fine.
PNG when they ask for raster at a stated resolution. Set the canvas as above and export at one times, or set a smaller canvas and export at two or four times, which comes to the same pixel count. Avoid JPEG for maps: the compression artefacts cluster exactly on the sharp boundaries and thin lines that a map is made of.
PDF for a supervisor, a report, or a slide deck. A single page sized to the figure is easy to hand around and prints predictably.
Font size, and the mistake that follows it
Journals usually ask for text no smaller than about 7 point in the final figure. Work out what that means in your canvas: at 2244 pixels for a 190 millimetre column, one point is about 12 pixels, so 7 point is roughly 84 pixels of cap height. That sounds enormous until you remember the canvas is 2244 wide.
When labels come out too small, raise the font size in the tool. Do not enlarge the whole figure to make the text legible, because then the figure no longer fits the column and the page layout shrinks it back down, returning you to where you started.
Two checks before you submit
Print it at final width. On paper, at 90 millimetres or whatever the journal specifies. Reading it on screen at full size tells you nothing. If a site label or a scale bar number is unreadable on the print, fix it now.
Print it in black and white. Many journals still print greyscale, and every paper gets photocopied. A teal study area and a light grey neighbour can land on nearly the same grey, at which point your highlight stops highlighting and a reviewer asks what the shading means. Change the fill rather than the outline, because a darker fill or a hatch survives conversion and a thicker border usually does not.
Multi panel figures and the extra width
A two panel figure at single column width gives each panel about 45 millimetres, which is too narrow for anything with labels. If you need a locator, ask for one and a half or double column, or stack the panels vertically instead of side by side.
A cascade of three or a grid of four should be double column, always. Squeezing four panels into 90 millimetres produces four illegible maps rather than one good one, and it is better to drop to two panels than to publish four unreadable ones.
File size and the submission system
Most submission systems cap individual files somewhere between 10 and 50 megabytes. A 2244 pixel PNG of a study area map is usually well under 5, and an SVG under 2. If you are somehow over, the cause is nearly always a raster basemap rather than the vector layers, which is one more reason study area figures tend not to carry satellite imagery.
The study area map maker takes a canvas width in pixels directly, so you can set 2244 and export at one times, and it reports the exported page size when you take the PDF.
Related Calculators and Reference
- Study Area Map Maker, where the canvas width is a field you type the journal number into.
- Study area map rules for a thesis for the print size and greyscale checks in context.
- Study area map examples to see which layouts need double column width and which fit in one.
- How to make a study area map online if you have not built the figure yet.
- Nature figure guidelines as a worked example of what a publisher actually asks for.
