This week is the American Geophysical Union’s fall meeting, which is taking place in New Orleans, Louisiana. AGU is a big conference, which recently has been drawing 20K-25K attendees each year. With the conference scientific program beginning tomorrow, many people are arriving today.
Getting 20,000 scientists together in one city when they mostly arrive on one day takes up quite a bit of the inbound capacity on transportation. My flight from Minneapolis today, for instance, was probably more than 50% meeting attendees—many of whom are are recognizable by their poster tube decked out in NASA, USGS, and other Earth science related stickers.
The community is also weirdly small. On the plane was a visiting professor from my undergraduate institution, and when I got in line for the downtown hotel shuttle, two alumni (each one year apart from me) were on either side of me.
Upon reaching the hotel, I settled in and started the process of figuring out what talks, posters, and events I should get to this week. Thursday morning will be my poster, which I’m very excited for. There are some sessions on Wednesday related to my doctoral work (and co-chaired by my doctoral adviser) which look very interesting. Several networking events will be happening as well, with varying degrees of field specificity.
Eventually I’d had enough of scheduling and decided it was time to eat food. I walked over to the French Quarter and, having again crossed paths with the prof I’d met in the Minneapolis airport, found one that had good jambalaya. A long walk and some further scheduling later, and it’s now time for bed. Tomorrow is a big day, with lots of networking to do and cool science to go learn about!
In three weeks I will be attending the American Geophysical Union (AGU) fall meeting, and on Thursday morning I will be presenting a poster about the Retreat of Stephenson Glacier, Heard Island, from Remote Sensing and Field Observations.1,2 I am very much looking forward to it, and if you will be at the meeting I hope you will stop by. There is likely to be a journal article forthcoming on this work, and the conference will be a great opportunity to discuss my project with glaciologists and get feedback on it—exactly what poster sessions at conferences are for, from what I understand.
Although my analysis is pretty much done, there is still quite a bit of work to go. Most importantly, the poster needs to be created. For that, I’ll start with a list of graphics and figures that will be needed for the poster:
Map of the world, showing the location of Heard Island
Map of Heard Island, showing the location of Stephenson Glacier
Satellite image(s) of Stephenson Glacier, showing the retreat
From that graphical outline will follow a minimal amount of text to guide a reader through the project with introduction, methods, results, and discussion sections. Once all that gets put together, it gets reviewed, sent to my co-author for further review, then changes are made until we’re satisfied and it’s sent off to the printer.
Following the conference, I hope to get a more detailed manuscript written. When it is ready for submission, I expect it will go to EarthArxiv, a new Earth science pre-print server, as well as an appropriate journal with open-access options.
Publication of that article would be the final step for this project, but there are quite a few new project ideas which have sprung up while I’ve been preparing this poster and article. One of the great things about using openly available data is that there are so many projects which could be done relatively simply and at little cost. Of course, a few other ideas have come to mind—and are perhaps more interesting—that would need further field studies.
Unless the affiliation is “Unaffiliated” for the lead author, it is incorrect. I have tried to get it corrected, but apparently the system can’t handle that.
During the Heard Island Expedition, although I was close to Stephenson Glacier I was unable to travel to that part of the island. Fortunately my co-author and several other expedition members did get there and took lots of photographs among other sampling and documentation efforts.
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.
On July 21, 2017, the Landsat 8 satellite imaged a fresh landslide on Heard Island, seen in the picture above. The slide occurred in the northeast portion of the island, on top of Compton Glacier, and I have annotated it for clarity in the image below.
This landslide is quite easy to spot because of the relatively clear conditions over Heard Island and the very high contrast between the dark, presumably-basaltic rocks and the white snow of the glaciers. Given that it is presently austral winter and Heard Island is located south of the Antarctic Convergence, the rate of snow accumulation should be quite high. It will be interesting to see how long it takes to be covered by snow.
I am fairly convinced that this is a rock- or landslide rather than an eruption. The head of the flow is along the top of a steep ridge, and the infrared imagery shows no thermal anomaly in this part of the island.
What’s interesting to me is that this slide appears to have eroded some snow on top of the glacier which then caused a secondary avalanche from a north-facing slope. I’ve annotated this in the image below.
This landslide has a run-out of about 2.5 km, an elevation drop of ~750 m, and a total affected area of ~0.8 km2. Several flow tongues are evident in the close-up image, even though the satellite imagery resolution is a modest 15 m/pixel.
From this image, it looks like the rockfall mostly happened in the portion running west-to-east, then as it turned the corner to head northeast, transitioned to a surface flow. In many ways, this reminds me of the Mt. Dixon (New Zealand) rock avalanche in 2013 (coverage by Dave Petley here and here, among others). The video below is from the Mt. Dixon (NZ) rock avalanche, but is likely similar to what occurred on Heard Island.
A fly-over after the Mt. Dixon (NZ) rock avalanche provided more video of the rock avalanche scar.
I look forward to seeing more images of this slide as they come in. Heard Island is imaged roughly every 8 days by Landsat 8, which as far as I can tell is the only publicly available high-resolution imagery for the island now that EO-1 has been decommissioned.
Recently, I was out in the area again, and this time made sure to have time to take some pictures and see some of what was to be seen. Let’s start with the quarter-mile view, which is roughly equivalent to what I saw last year.
Capitol Rock has three major parts to it: an easily eroded base, a laminated sandstone middle, and a massive sandstone top. A handy turn-out from the forest service road leads right to the base of the outcrop.
The easily eroded base is made of fine, chalky, white sediment sediment, and it remains in horizontal orientation. In several places, this unit is at least superficially porous. Surprisingly, there are occasional chert clasts in the otherwise fine sediments—I’m not quite sure how those would have been deposited or formed here.
Above the basal unit is a somewhat more resistant, coarser-grained set of beds. These strata are finely bedded, and have a tendency toward spheroidal weathering. Occasionally interbedded with the spheroidal beds are 1–3 cm thick, well-cemented strata of a white or pink color [discoloration?].
The spheroidally-weathered unit also seems to have one or more channels within it.
The upper unit at Capitol Rock has more massive sandstone (see wide view above). I didn’t notice many channels in this unit, although I didn’t get very close. A butte just to the north of Capitol Rock provided a good photograph (below).
Although I have those observations, I don’t have much for interpretation of them. The depositional environment seems to be relatively low-energy (give or take the chert clasts), evidenced by the flat strata, fine grain sizes, and relatively few cross-beds. Changes in the rock types would suggest changes in the sediment sources or the depositional environment (or both). There may be post-deposition alteration effects as well, such as cementation of the spheroidally-weathering strata.
Capitol Rock is an interesting outcrop, and if you’re in the area, I’d recommend a stop. The rocks are interesting, there are US Forest Service campgrounds nearby, and the view is quite nice. These units can probably be correlated to the Slim Buttes in South Dakota (~45 miles east).
This spring I had the privilege of being a judge at my local science fair. As a high school student, I had participated in the science fair and it was a huge part of my science learning experience. Now that I am qualified to be a judge, it is time for me to give back while avoiding the trap of turning into the dreaded Reviewer #2.*
I scored quite a few different projects, primarily in Earth & Environmental Science. I was pleased to see the large number of students involved in the discipline, and the interest they showed in environmental monitoring and sustainability. However, I was surprised to see the number of projects which focused on pH, but without understanding of pH of rainwater or the influence of carbonates.
Limited or non-existent access to instrumentation was clearly a limiting factor in many of the projects. That observation leads to a question: what can be done to address the disparity in instrument access and to improve the quality of data being used in science fair projects? I believe the long-term answer to that question is to fund our schools and support the teachers and staff who work in them.
Another solution would be to have students use and analyze publicly available data. In many cases, this cut out some of the hands-on portion of making measurements, which detracts from the overall learning goals. Using publicly-available data also means that teachers would need to be more aware of good data resources and ideas for how to go about analyzing that data—each significantly increasing the work load and responsibilities of the teachers. For research projects, it is important to have a low student:teacher ratio, so that the students can have the support they need to succeed in their project. However, publicly available data allow students to do cutting-edge research with the same tools and data used by professional scientists.
Here are a few examples of low-budget, high-quality data projects that could be interesting:
Weather forecast accuracy. Make a daily record of the National Weather Service forecast (for each day forecast) for your area, as well as the almanac data from the closest instrumented NWS station (often an airport). How does forecast accuracy change over time? How accurate is a forecast 72 hours out?
Earth-Observing Satellite data. With a constellation of Earth-observing satellites including Aqua, Terra, Landsat (7 and 8), and formerly EO-1, there are mountains of data waiting to be analyzed. Students can look at crop health locally, at glacial changes, deforestation, volcanic activity, wildfires, and a host of other things. Data are freely available, GIS software is freely available, and the data analysis skills are quite relevant in today’s job market.
Buoy data. As I’ve mentioned here before, there are several fleets of marine buoys which take various oceanographic measurements, such as conductivity-temperature-depth profiles and current measurements. Oceanography isn’t my thing, but I’m sure there are enough papers that use these data that some project ideas could be found. These projects are likely to use GIS.
* Reviewer #2 is known for being overly critical, wanting a paper that isn’t particularly close to the paper that was submitted, having unreasonable or unattainable expectations, and generally being a jerk.
Heard Island is a place of fantastic beauty. Yes, the weather can be bleak and windy, but that too gives Heard Island some of its unique character. One morning I stepped out of the tent around sunrise, and was greeted by the scene above. I stood in awe for a moment at the landscape, which looked as though it were straight out of a painting. Things change quickly on Heard Island, so after only the moment’s pause, I ran back into the tent for my camera, and managed to take a few pictures that came out looking exactly like what I saw.
Scenery on Heard Island was usually dramatic, with gloomy greys punctuated with bursts of color—often the orange patches of king penguins. The wide, featureless nullarbor added to the fantastic feeling of the island by removing any sense of scale one might have.
For more pictures from Heard Island and the voyage on the Braveheart, please see the Flickr album.
One year ago, I was on Heard Island. Over the course of the expedition I took more than 6000 photos. Although I took three images with the Gigapan (Big Ben, the Azorella Peninsula, and—my favorite—Windy City), I also took photos for stitching together manually using my own camera.
I have been slow in stitching these pictures together, but with the one-year anniversary of their production coming around, I decided it was time to finish one or two of them. This is the first, and I hope I’ll find time to finish more. Putting it together, I was amazed that this is still a relatively wide-angle compared to what I had available: 70 mm on a 70–200 mm lens. The detail came out well, as is evident at full-size. The glaciers, moraines, and hills are all more than a kilometer distant over the “nullarbor”, a broad, flat, volcanic-sand plain.
Toward the left half of the image are some penguins for scale. They look like king penguins, putting their height around 1 meter. I count at least 31 penguins in the entire image.
I think I spot some of the relatively rare basement limestones cropping out at the very left edge of the image, and their appearance is consistent with a dip of 25–35° to the south. A closer view (200 mm focal length) shows them more clearly.