Tag Archives: QGIS

Mapping the Eclipse for a Citizen Science Project

Map of the continental United States showing the amateur radio grids and path of the eclipse.  Image credit: Bill Mitchell (CC-BY).
Map of the continental United States showing the amateur radio grids and path of the eclipse. Image credit: Bill Mitchell (CC-BY).

During the solar eclipse next week, I will be at the Science Museum of Minnesota with a citizen science project studying the effects of the eclipse on radio propagation. While there are many radio-related projects going on—the most accessible being a study of AM radio reception—I will be using amateur radio to make contacts and provide reception reports during the eclipse. One of the important pieces of information that will be exchanged with other amateur stations is a “grid”, which is a shorthand for rough latitude and longitude.

Amateur radio grids are 2° longitude by 1° latitude, and represented with pairs of letters and numbers. For instance, the Science Museum of Minnesota is located in EN34. Fields (20°x10°) are designated with letters, and increase from -180 longitude and -90 latitude (AA) to 160 longitude and 80 latitude (RR). Fields are further subdivided into grids using numbers, which increase from 00 at the southwest corner to 99 at the northeast. Looking again at our example, the first character, E, indicates a location between 100° and 80° W longitude, and N indicates a location between 40° and 50° N latitude. The numbers provide further refinement on that range. The 3 means the longitude is between 6° and 8° east of the west edge of the field (i.e. 94°–92° W), and the 4 after it means the latitude is 4°–5° north of the south edge of the field (i.e. 44°–45° N). Further letters (A-X) and numbers can be used to specify locations more precisely in a similar fashion. Longitude is always indicated first, and increases west-to-east; latitude is indicated second, increasing south-to-north.

For the event, I want to have a map of the continental US and southern Canada with the grids outlined on it. During the event as we hear which grid other stations are in, we can mark their location on the map. Unfortunately, I was not able to find a map that I wanted to use for this purpose, so I decided to make my own with QGIS.

For my eclipse map, I needed to gather a few datasets. First and foremost, I needed a US state map. Canadian provinces were also a high priority. Once I had those, I was still missing Mexico and other North American areas, so I found a world map as well. That covered the basics, but as long as I’m making a special map for the eclipse, I wanted to have the path of totality, which I found from NASA. I unzipped each of those files into a folder for my eclipse grid map project.

In QGIS, I loaded all the datasets (vectors). The Canadian provinces were in a different projection, so I saved (converted) it to the projection I wanted (EPSG:4269), which is a simple latitude-longitude projection. I found that the Canadian provinces included detailed coastlines and islands, so I simplified it (Vector | Geometry Tools | Simplify Geometry) using a tolerance of 0.01 or something like that. The islands cleaned up a little, but the overall shapes didn’t change much.

With the datasets loaded, I needed to make my field and grid boundaries. Using the grid tool (Vector | Research Tools | Vector Grid) I created the field grid (xmin=-180, xmax=180, ymin=-90, ymax=90, parameter x=20, parameter y=10) and the fine grid (same except parameter x=2, parameter y=1).

I looked up the coordinates for the Science Museum of Minnesota, and put them into a CSV text file. By loading in that CSV file, I put a star on the map where I will be located.

From that point, it was just a matter of adjusting colors and display properties. I gave reasonable, light colors to the US and Canada, and thickened the borders for the US states. I used a dashed line for the field lines, and a lighter grey dotted line for the smaller grids. The eclipse path I made a partially-transparent grey.

That’s about all there was to it! In the print composer I added in some of the labels for a few grids to help demonstrate the letter/number scheme.

Results (PDF): 8.5″x11″, 11″x17″.

Advertisements

Big Ben Eruption, 2017-02-04

Lava and debris flows radiate away from Mawson Peak on Heard Island.   February 4, 2017.  Image credit: Bill Mitchell (CC-BY) using data from Landsat 8 OLI (NASA/USGS; public domain).
Lava and debris flows radiate away from Mawson Peak on Heard Island. February 4, 2017. Image credit: Bill Mitchell (CC-BY) using data from Landsat 8 OLI (NASA/USGS; public domain).

On February 4th, Landsat 8 captured a clear view of the summit of Big Ben volcano, at Heard Island. Heard Island is a very cloudy location, so clear views are uncommon (I don’t have numbers, but would estimate <20%). However, the February 4th images are even more spectacular: they capture an ongoing volcanic eruption.

Observations
In the sharpened true-color image (above), four or five different lava/rock/debris flows are visible emanating from the summit. Using a false-color infrared image (below), two hot regions are apparent (red/orange/yellow), and are separated by about 250 meters. The longest of the flows stretches nearly 2 km, and drops from an elevation of roughly 2740 m to 1480 m (using 2002 Radarsat elevation data with 20 m contours). All three of the large flows to the west or southwest of the summit drop below 2000 m elevation at the toe.

False-color infrared imagery of Mawson Peak, Heard Island.  Two vents are visible in red/orange/yellow, separated by 250 meters. Data source: Landsat 8 OLI/TIRS bands 7-6-5.  Image credit: Bill Mitchell (CC-BY), data from NASA/USGS (public domain).
False-color infrared imagery of Mawson Peak, Heard Island. Two vents are visible in red/orange/yellow, separated by 250 meters. Data source: Landsat 8 OLI/TIRS bands 7-6-5. Image credit: Bill Mitchell (CC-BY), data from NASA/USGS (public domain).

Interpretation
In the sharpened true-color imagery, I have identified five rock and debris flows originating at the summit, as well as one potential avalanche. Annotation of these observations is found on the pictures below.

Annotation of lava/rock/debris flows from Mawson Peak, Heard Island, February 4, 2017.  Image credit: Bill Mitchell (CC-BY).
Annotation of lava/rock/debris flows from Mawson Peak, Heard Island, February 4, 2017. Image credit: Bill Mitchell (CC-BY).

The streaky, varying lightness of the flow areas, presence of snow and ice, and steep terrain lead me to believe that what is showing up here are mixed snow/rock/lava debris flows, rather than pure lava flows. A mix of rocky debris and snow would not be out of line for a supraglacial eruption on a steep mountain. The longest flow drops nearly 1300 m along its 2000 m horizontal path according to the 2002 Radarsat elevations. I’ll be the first to admit that I am distrustful of the specifics of the Radarsat contours due to the rapidly changing landscape and an intervening 15 years, but I think that it manges to get the general picture right.

Southwestern Heard Island is a high-precipitation area, so rocks exposed on the surface of the glaciers are likely quite fresh. It probably won’t be long before most of the deposits are covered in snow again.

Speaking of snow, it looks as though there is a faint outline of an avalanche scarp/deposit on the northeast side of the summit, which I annotated below in green.

Annotation of avalanche scarp and deposit, Mawson Peak, Heard Island, February 4, 2017.  Image credit: Bill Mitchell (CC-BY)
Annotation of avalanche scarp and deposit, Mawson Peak, Heard Island, February 4, 2017. Image credit: Bill Mitchell (CC-BY)

The two hot spots provide an interesting challenge for interpretation. Two scenarios come to mind quickly: there are two vents from which lava is issuing, or there is a lava tunnel from a summit crater down to a flow front or breakout. Analyzing the Landsat 8 OLI/TIRS infrared imagery from January 26th (most recent previous high-resolution image), only one hot spot is present—in the same place as the eastern hot spot in the February 4th infrared image. For spatial correlation without doing the whole image processing and GIS thing, use the forked flow to the south-southeast of the hotspot as a reference.

False-color infrared image of Mawson Peak, January 26, 2017.  Landsat 8 OLI/TIRS bands 7-6-5.  Image credit: Bill Mitchell (CC-BY) using NASA/USGS data (public domain).
False-color infrared image of Mawson Peak, January 26, 2017. Landsat 8 OLI/TIRS bands 7-6-5. Image credit: Bill Mitchell (CC-BY) using NASA/USGS data (public domain).

Due to a different time of day for imaging, there are significant shadows in the January image on the southwest side of ridges. It’s tricky to figure out what is going on for the flows (even in visible imagery), but the hot spot from January 26th is right on top of the eastern hot spot from February 4th.

Another thing which becomes apparent in the January image is the topography at the summit. The clouds form a blanket at an atmospheric boundary (and roughly-constant elevation), which is conveniently just below the elevation of the summit. A roughly circular hole in the clouds is present, and a conical mountain summit pokes through with the hot spot right in the center. That suggests that the second hot spot seen in the February 4th image is at a lower elevation—a possible flow front or breakout.

Excitement in the Mundane
Finding this eruption was a bit of a surprise to me: the low-resolution preview image for the Landsat data on EarthExplorer was so coarse that there wasn’t anything striking or out of the ordinary visible at the summit. Clouds covered most of the rest of the island. However, when I opened up the full-resolution color images (30 m/pixel), it was immediately evident that this was a special day. Sharpening the true-color bands with the high-resolution panchromatic band using QGIS made it pop all the more!

Upon seeing both the lava/debris flows and the thermal anomaly, I checked the MODIS volcanism (MODVOLC) site to see if the Terra and Aqua MODIS instruments had picked up thermal anomalies as well over the preceding week. They had, as shown below. Both satellites had recorded thermal anomalies at Heard on February 2nd and 3rd.

MODIS thermal events at Heard Island, in the week preceding February 6, 2017.  Image credit: MODVOLC.
MODIS thermal events at Heard Island, in the week preceding February 6, 2017. Image credit: MODVOLC.

Update:: Follow-up from February 27, 2017.

Cloud-Free Heard Island

Composite, cloud-free satellite imagery of Heard Island, being produced in QGIS.  Image credit: Bill Mitchell (CC-BY), using USGS (Landsat 8, EO-1) data (public domain).
Composite, cloud-free satellite imagery of Heard Island, being produced in QGIS. Image credit: Bill Mitchell (CC-BY), using USGS (Landsat 8, EO-1) data (public domain).

Heard Island is a pretty cloudy place most of the time. However, there are occasional times when the weather clears, particularly on the southeastern (leeward) side of the island. On rare occasions, the northwest and southwest sides of the island come out from the clouds as a satellite passes over.

For the past two years, I have been watching Heard Island using true-color imagery from four satellites: Terra, Aqua, Landsat 8, and EO-1. I have posted previously about satellite imagery from these instruments. Although every image of the Island I have seen has clouds in it covering a portion of the island, I was curious whether or not I had accumulated clear imagery of the entirety of Heard Island.

In part, this question was spurred by a follower on Twitter asking about eruptive activity at Heard. I had to admit I didn’t really know whether the activity was low-level and continuous (like Kilauea) or more intermittent. Given that our knowledge of its eruptive activity is primarily from satellite observations, do the satellite “thermal anomalies” correspond to short eruptive events, or simply a cloud-free view of the volcano?

For high-resolution imagery of Heard Island, the datasets of interest are from EO-1 ALI, and Landsat 8 OLI. The two MODIS instruments (one on Aqua, one on Terra) are moderate-resolution, and while 250-m resolution is sufficient for some purposes, this one needs more. Looking through the archives, I was able to find EO-1 ALI data primarily for Mawson Peak and points southeast, and Landsat 8 OLI covered much of the island, particularly the northwest.

Not only is having cloud-free, high-resolution data important for me, but I want the data to be recent. There has been a retreat of up to 5.5 km for some of the glaciers since 1947, and the Google Maps imagery of that area (Stephenson Lagoon) is horribly outdated. Fortunately, I found most of the island covered in large swaths with images from 2014 onward, and mostly 2016. There was even good imagery from when I was on Heard Island! Our ship, the Braveheart, is visible as a few white pixels in Atlas Roads (just north of Atlas Cove), slightly closer to the Azorella Peninsula than to the Laurens Peninsula. The tents and campsite are too small and darkly colored to be visible on this image.

Braveheart in Atlas Roads and the campsite (non-contrasting) at Atlas Cove, Heard Island.  Satellite image pixels are 15 m across, and the Azorella Peninsula isthmus (along Campsite label) is 1 km wide.  Image credit: Bill Mitchell (CC-BY), using USGS (Landsat 8 OLI) data (public domain).
Braveheart in Atlas Roads and the campsite (non-contrasting) at Atlas Cove, Heard Island. Satellite image pixels are 15 m across, and the Azorella Peninsula isthmus (along Campsite label) is 1 km wide. Image credit: Bill Mitchell (CC-BY), using USGS (Landsat 8 OLI) data (public domain).

A small portion of the island between Atlas Cove and Mawson Peak was the most difficult to find. With the topography of the island, the steady stream of wind, and the humid air, the 2.5 km by 2.5 km region was cloudy pretty much all the time. Eventually, using the EO-1 ALI instrument and going back to early 2010, I found a reasonably clear image of it.

Once I had the images (after combining true-color and panchromatic brightness data in QGIS), I needed to stitch them together. Thanks to the wonderful QGIS training manual, I was able to create vector (polygon) layers which corresponded to the clear region of each image (plus some surrounding ocean). At this point the troublesome mostly-cloudy spot became evident, and the search was on for imagery to fill the void.

Creating polygons for clipping the satellite imagery using QGIS.  Four polygons are shown here, including the small polygon of much cloudiness.  A fifth dataset was subsequently incorporated.
Creating polygons for clipping the satellite imagery using QGIS. Four polygons are shown here, including the small polygon of much cloudiness. A fifth dataset was subsequently incorporated.

Finally, I tried to put them together. This turned out to be more trouble than it was worth for my purposes, having only five images. Several of the images had differing resolutions (10 m/pixel for EO-1 ALI, 15 m/pixel for Landsat 8 OLI). Additionally, since I was handling these in their raw format, color balances/exposures were not consistent across images. I decided it best, then, to leave them separate, and sent them around to the Heard Island Expedition team.

Soon I had an email from the expedition leader: he was very interested in the imagery, but it wasn’t opening in Google Earth. Some searching later, I found that Google Earth works best with a certain map projection (EPSG:4326), and when exporting the GeoTIFF, I needed to select “rendered image” rather than “raw data”. I re-exported the images, zipped them up, and tested it out on another computer: success! This Google Earth friendly imagery is now available here (17 MB zip).

One continuation of this project would be to keep looking through the documentation on GeoTiffs to find out how to make the rendered images use a transparent, not white, border where there is no data. That would likely let me create a virtual raster catalog to load all of them in one go, rather than having to load them separately.

Argo Buoy Update

Argo buoy tracks from buoys deployed on the Heard Island Expedition.  Red dots indicate most recent position.  The first set of buoys were deployed between Cape Town, South Africa, and Heard Island (center).  The color scheme has been reused for the second set of buoys, deployed between Heard Island and Fremantle, Western Australia.  Image credit: Bill Mitchell (CC-BY).
Argo buoy tracks from buoys deployed on the Heard Island Expedition. Red dots indicate most recent position. The first set of buoys were deployed between Cape Town, South Africa, and Heard Island (center). The color scheme has been reused for the second set of buoys, deployed between Heard Island and Fremantle, Western Australia. Image credit: Bill Mitchell, using OpenStreetMap continents (CC-BY).

About ten months ago, the Heard Island Expedition team launched the first of our eleven Argo buoys into the Indian Ocean. The buoys are equipped with conductivity-temperature-depth (CTD) instruments, and spend most of their time drifting about 1 km beneath the ocean surface. Every ten days, they dive to 2 km, then record CTD data as they ascend to the surface. At the surface, they relay the data via satellite over the course of a day before returning to 1 km depth. With a large network of these buoys, scientists can gather important data on currents under the ocean, as well as changes in temperature and salinity profiles.

Over time, ocean currents move the buoys. None of our eleven buoys are where they started, and some have moved far away from where they entered the ocean. We deployed two batches of buoys: six before reaching Heard Island from Cape Town, and five more on our voyage on to Fremantle/Perth.

I have obtained the latest position data (as of Jan 14, 2017) for the eleven buoys. Their tracks are shown in the figure at the top of this post. Tracks are colored by buoy, reusing the colors for the first and second batch. Some of the buoys have moved more than 1500 km as the albatross flies, with path lengths approaching 3000 km!

The CTD data are also interesting. For instance, here are the temperature/depth and salinity/depth profiles measured by buoy 5902454 (dark blue path on second leg of map above).

Temperature/depth profiles over time for buoy 5902454.  Image credit: Coriolis Data Centre.
Temperature/depth profiles over time for buoy 5902454. Image credit: Coriolis Data Centre.
Salinity/depth profiles over time for buoy 5902454.  Image credit: Coriolis Data Centre.
Salinity/depth profiles over time for buoy 5902454. Image credit: Coriolis Data Centre.

Around December 1, buoy 5902454 encountered a different water mass with colder, saltier water throughout much of the 2 km water column.

Generally for these buoys, the surface water temperatures reflect the seasonal variations (warmer in Austral summer, colder in winter), while the deep water shows less variation—but sometimes there are shifts between different water masses.

Interested in keeping up with the latest from these buoys, or checking out their CTD data? Use these links (buoy number, in order by our deployment date):
1901910
1901812
1901811
1901813
1901809
1901838

5902451
5902452
5902453
5902454
5902455

Correction An earlier version of this article referred in several places to deploying ten buoys. The correct number is eleven.

Show Me the Data: Satellite Observations

Heard Island, March 27, 2013.  Elephant Spit imagery (at right) is from March 3, 2013.  Image has been adjusted to increase bring out detail in exposed land.  Image credit: processed by Bill Mitchell (CC-BY) using data from NASA/EO-1 (public domain).
Heard Island, March 27, 2013. Elephant Spit imagery (at right) is from March 3, 2013. Image has been adjusted to increase bring out detail in exposed land. Image credit: processed by Bill Mitchell (CC-BY) using data from NASA/USGS/EO-1 ALI (public domain).

With Heard Island being remote and uninhabited, studying it can be a bit difficult. However, as readers of this blog (and my Twitter followers) are aware, one of the ways I have been preparing for the expedition is by keeping an eye on it using various satellites and their remote sensing capabilities. Sure, there is often cloudcover at Heard, but some days it’s clearer and on a few of those days, the satellites pass over.

Most of my information comes from NASA’s MODIS instruments, aboard the Terra and Aqua satellites. These have at least every-other-day coverage of everywhere on Earth, although with a moderate resolution of 250 m/pixel. In the morning when I’m catching up on email and comics, I’ll check the near-real-time MODIS image page to see whether there are clear images of Heard Island from either instrument. Finding Heard Island can be difficult: I still usually find the Kerguelen Islands first, then look to the south-southeast. Many times there are indications such as vortices or gravity waves (not gravitational waves, those are different).

A related page is MODVOLC, which uses MODIS for volcano monitoring. In addition to visible light, MODIS can detect several wavelengths of infrared, and the signature from those wavelengths can be used to determine whether there is a likely volcanic eruption occurring at a given place.

MODIS is a great instrument in that it has daily or every-other-day coverage. However, the 250 m/pixel resolution can be quite limiting. For higher-resolution imagery, I look to the ALI instrument on NASA’s EO-1 satellite. These images are available (free registration required) from EarthExplorer, a data search portal from the USGS. ALI has a 30 m/pixel resolution on its color imagery, and 10 m/pixel resolution on the panchromatic image (total light intensity). These can be combined using QGIS into 10 m/pixel color images. By exploring the EO-1 page I found that members of the public can make requests for image targets! Imaging requests are subject to a bunch of conditions (availability of satellite, >30-day lead time, recommended >3 month window for imaging), but the request and any data generated from fulfillment of the request are free.

How did I come to know about these great resources? It takes time, searching, and some attention to detail. MODIS I learned about as a graduate student, from friends who used data products (not the true-color imagery) in their doctoral research on atmospheric chemistry. I came across EO-1 ALI from searching for images of Heard Island: I found some higher-than-MODIS resolution images from NASA which were good about indicating the source satellite/instrument. Citing image sources is incredibly useful, and I’m always disappointed when images (at least, non-screenshot images) are given without any sort of source information.

MODVOLC I learned about from the Smithsonian’s Global Volcanism Program, which cites the sources of their eruption reports. Information about the source plus a little searching yielded an interesting and useful data source.

Satellite Image Processing Revisited

Heard Island on Nov. 20, 2015, with image processing underway in QGIS.  Image credit: Bill Mitchell (CC-BY) with satellite imagery from USGS (EO-1 satellite, ALI instrument).
Heard Island on Nov. 20, 2015, with image processing underway in QGIS. Image credit: Bill Mitchell (CC-BY) with satellite imagery from USGS (EO-1 satellite, ALI instrument).

Following up on my earlier post about satellite image processing, I am happy to report that I have made progress in being able to process images myself! Through a fortunate combination of search terms, timing, and luck, I managed to come across two key pieces of information that I needed.

First, I found out how to make RGB images from raster data layers, such as different spectral bands on a satellite, fairly easily with QGIS. That was a big step forward from how I had been doing it previously, which was inelegant, inefficient, and only mostly worked. Stacking three layers (one each for red, green, and blue) into a virtual raster catalog was just a few clicks away (Raster | Miscellaneous | Build Virtual Raster (Catalog)).

Encouraged by the success with that project, I continued clicking around and stumbled across some mention of pan-sharpening (also pan sharpening), where a panchromatic (all-color) detector at high resolution is used to enhance the resolution of a colored image (sharpen). Alternately, you can think of it in the complementary way, where lower-resolution color data is added to a high-resolution greyscale image. So thanks to this blog post, I was able to find out what I needed to do to make that happen in QGIS (and Orfeo Toolbox).

Of course, it would be too easy for that to work. I didn’t have the Orfeo Toolbox installed which that needed, and ended up having to compile that from source code.* When the compiler finished and the program was installed, I went to tell QGIS where it was—but a bug in QGIS prevented me from entering the folder location. First, having just installed and compiled stuff, I attempted to get the latest version of QGIS and many of the tools on which QGIS relies. Being unsuccessful in making all of those and some of the compiler configuration software play nicely with each other, I eventually remembered I could get updated packages through apt-get, which gets pre-compiled binary files put out by the maintainers of Debian Linux. That solution worked, I added the folder location, and now I can have my pan-sharpened images.

Here for your viewing pleasure is my first properly pan-sharpened image: Heard Island on Nov. 20, 2015, seen in “true color” by the Advanced Land Imager (ALI) on the EO-1 satellite.** I’m not convinced it’s right, and I think the contrast needs to be brought down a bit, but I think it’s close.

Heard Island in true color on Nov. 20, 2015.  Image processing: Bill Mitchell (CC-BY) using data from USGS/EO-1.
Heard Island in true color on Nov. 20, 2015. Image processing: Bill Mitchell (CC-BY) using data from USGS (EO-1/ALI).

* Knowing how to compile software from source code is a rather handy skill.
** Emily Lakdawalla has written a great explanation of what “true color” means.

Geoscientist’s Toolkit: QGIS

QGIS screenshot, showing Heard Island.  Brown is land/rock, blue are lagoons, and the dotted white is glacier.
QGIS screenshot, showing Heard Island. Brown is land/rock, blue are lagoons, and the dotted white is glacier.

One of a geoscientist’s most useful tools is a geographic information system, or GIS. This is a computer program which allows the creation and analysis of maps and spatial data. Perhaps the most widely used in academia is ArcGIS, from ESRI. However, as a student and hobbyist who likes to support the open-source software ecosystem, I use the free/open-source QGIS.

QGIS can be used to make geologic maps of an area, chart streams, and note where certain geologic features (e.g. volcanic cones) are present. For instance, at the top of this post is a map of Heard Island that I’ve been playing with, from the Australian Antarctic Division. It is composed of three different layers, each published in 2009: an island layer (base, brown), a lagoon layer (middle, blue), and a glacier layer (top, dotted bluish-white).

I believe I have mentioned here previously that one interesting thing about working with Heard Island is that with major surface changes underway (glacial retreat, erosion, minor volcanic activity), the maps become obsolete fairly quickly. This week I have been learning about creating polygons in a layer, so that I can recreate a geologic map from Barling et al. 1994.[1] One issue I’ve come up against, though, is that the 1994 paper has some areas covered in glacier (from 1986/7 field work), whereas my 2009 glacier extent map shows them to be presently uncovered. In fact, even the 2009 map shows a tongue of glacier protruding into Stephenson Lagoon (in the southeast corner), while recent satellite imagery shows no such tongue.

During the Heard Island Expedition in March and April, 2016, I hope that we will have time to go do a little geologic mapping. Creating some datasets showing the extent of glaciation (particularly along the eastern half of the island) and vegetation, as well as updating the geologic map to include portions which were glaciated in 1986/7, would be a worthwhile and seemingly straightforward project.

QGIS itself is much more than a mapping tool (not that I know how to use it), and can analyze numeric data which is spatially distributed, like the concentration of chromium in soil or water samples from different places on a study site. QGIS provides a free way to get your hands dirty with spatial data and mapping, and is powerful enough to use professionally. Users around the globe share information on how to use it, and contribute to its development.

For those looking to go into geoscience as a career, I would strongly recommend learning how to use it. I didn’t learn GIS in college (chemists don’t use it much), and somehow avoided it in grad school. But I regret not having put time in to learn it sooner. There’s all kinds of interesting spatial data, and a good job market for people with a GIS skillset (or so I hear). I have only scratched the surface of QGIS’s capabilities with my use of it, but I definitely intend to keep learning. You can probably follow the day-to-day frustrations and victories on my Twitter account (@i_rockhopper).

***

[1] Barling, J.; Goldstein, S. L.; Nicholls, I. A. 1994 “Geochemistry of Heard Island (Southern Indian Ocean): Characterization of an Enriched Mantle Component and Implications for Enrichment of the Sub-Indian Ocean Mantle” Journal of Petrology 35, p. 1017–1053. doi: 10.1093/petrology/35.4.1017