How-To: Hurricane Map

Here is a run-down with all the carto nitty-gritty on how to make the following map:

It shows the locations of each recorded hurricane and tropical storm since the mid 1800s. When viewed in aggregate, the hurricanes appear to coalesce into a single great hurricane eye. It’s a basic map, with some cartographic punch. Super simple to make yourself. I’ll give links to all the data and icons, and provide a step-by-step on how to smash them together.

Get ready to live.

Data

Here are some resources that you can download, if you’d like to follow along. I hope you do that.

• Sassy Basemap Image (zip of modified satellite image and projection files)
  Starting with a NASA Visible Earth (I love these guys) cloud-free composite, I de-saturated the image and added a vignette overlay to help throw focus to equatorial regions. This derived image has been georectified (breathing geo-life into an otherwise spatially-oblivious image).
• Hurricane Point Shapefiles (zip of six shapefiles)
  The source of this storm data is curated and made freely available by NOAA, and goes back to the mid 1800s. I removed all storm points with null or 0 wind speed, then grouped them into individual shapefiles according to Saffir-Simpson hurricane strength.
• Firefly Icon Images (zip of six PNG images)
  Don’t feel stuck with the default pack of symbol markers in your GIS. You can use your own images -which is a lot of fun. This set of images will be used to represent hurricane strength.
• Coastal and Graticule Shapefiles (link to Natural Earth’s data-download page)
  These came straight from Natural Earth, a great resource for basic, commonly-needed shapefiles for map making. Here are the direct download links to the Coast shapefile and the 20-degree graticule shapefile.

That’s it for downloads, you have all the parts you need (assuming a tenacious curiosity and 10-15 minutes).

Image Enhancement and Projection

Or de-hancement, I suppose. When you download the sweet sweet satellite image from the NASA Visible Earth site, it will be just that -an image. A full-color, cloud-free unprojected portrait of Earth.

Let’s start by sucking out some of that color. I’m filled with wild hunches about the emotive and cognitive impact of color-saturation on our sense of scale and interpretative abilities. As such, I like to really de-saturate satellite imagery if I’m using it as a thematic map’s backdrop. You can do this in just about any image editing tool. I like to push it down to about a 20% color saturation.

Also, I have a vignette effect at the edges of my finished hurricane map, to draw focus to the equatorial areas and to burn in the edges for aesthetic reasons. With an understanding of what my map projection will ultimately be, I can bake it right into the image. I just overlay a rectangle, and fill it with a gradient from fully opaque black at the top, to transparent (and then a little more opacity to tame Antarctica a bit).

So far, we are working with a flat image. But we need it to bend to our cartographic will. So let’s georectify the situation. You can use pretty much any GIS tool to do this, but I’ll be using ArcGIS. Using the “Georeference” tool, you can pin control points to any pixels in your image and associate those to locations in your living breathing GIS. Now you can warp it to any crazy map projection you want!

I could puts around with projections all day. Each one provides some unique look into the nature of the data. But which is best?

Our phenomenon is a circular one, so I thought it would be effective to use a polar projection, looking up at the Earth while the hundreds of storms whoosh around it as one huge hurricane. So here it is in a South Pole Stereographic projection, with a rotated central meridian to best fit a horizontal layout:

Icons

You can poke around in the point symbols that come packaged with a mapping tool, but at the end of the day the best way to have control over what you really want is to make your own. Anyone who’s been in or near a hurricane can tell you that a rigidly precise flat symbol like a point or a line isn’t necessarily true to the broad, fuzzy-edged nature of a moving storm. Using a symbol with a blurry edge helps illustrate Tobler’s First Law, plus I think it looks really cool. I call it firefly cartography. Glowing bright symbols over a dark basemap. You can call it whatever you want.

Notice how they have a white-hot center, and only the glow effect has the color encoded? Because lightsabers.

In Arc, you can reach out to an image and use that as your point icon. So for each hurricane category, I tie it to one of my glowing PNG images. You can tweak the size of the icon in Arc, where it’s easy to make quick changes. Just try not to scale it up higher than the image’s native resolution or else you could get some pixelation.

Each storm category uses a smaller and slightly more transparent icon. Be sure to set the rendering order from biggest (draws on top) to smallest.

Once all the storms paint in, I get that awwwww yeahhhh map-making vibe.

It can be tempting to call it quits at this point. But we can tighten it up by adding helpful context layers.

Context Layers

Adding in a coastal reference goes a long way towards giving map readers some bearing when looking at an unusual, bottoms-up, perspective of the world. I gave mine a very thin width, an earthy color of 255, 211, 127 (rgb), then set the layer to 80% transparent. It’s easy to over do it with second-tier layer styling. I find that diminishing the opacity a lot helps push these things back, visually, to where they can provide context but not detract from the focus of the map.

For some reason, faint graticule lines provide a sense of order and cinematic coolness. These lines are a cool slate color of 0,132,168 (rgb), and set to 70% transparent.

Because the concept of the equator is so important to the nature of hurricanes, and such an obvious visible factor in this map, I’ve added it to the layout to lend some helpful context. I used the same earthy color of 255, 211, 127 (rgb) to help contrast it from the hurricane points, and pushed it back to 70% transparent. It also has a dashed effect to help illustrate that it is a reference, and not an actual physical feature. I then added a second version of this line, but with a much thicker stroke and a transparency of 80%. This provides a really subtle, hacked, glow effect.

I didn’t like the way that the graticule lines converge at the center over an already bright Antarctica. It was a distracting visual artifact, and didn’t really support the idea of hurricanes whipping around the belt of the earth. Darkening this area of my enhanced base map image wouldn’t help, since the lines sit on top of the image. So I made a second, purely cosmetic, georectified vignette image overlay and placed it on top of the base imagery and all of the context layers.

Now, when I turn the hurricane points on again, I have a generally complete map! All in Arc, with a couple supporting images (the icons and the enhanced basemap image) drawn in a graphic design tool.

And that’s it. In the original version of this map from a few years ago, I included a couple supplemental charts to visualize trends over time. They were made in Excel. I copied the chart graphics in Excel then pasted them into a graphic design tool.

Thanks for following along!