(Note: This was originally posted for 12/01/2020)
Hello Planetary Pals,
It has been about a month and a half since my last research update and a lot has happened since then.
The last research update occurred just as I was starting to write the first draft of my thesis proposal. After writing the thesis proposal I have gotten a clearer idea of what is needed for my research. Despite this, there were some things that needed fixing in my thesis proposal draft. First and foremost, it became obvious to me once I opened the reviews that I had accidentally sent in the wrong version of my draft. I had sent in one of the unfinished copies of the draft, not the final pdf that I intended to submit. The five reviewers of my paper were grading and judging me on the wrong thing, and it was certainly not fun being shamed five times over for this mistake that I made. Most of the commenters still gave useful feedback on the finished portions so that I could still take those edits into account in my final draft. Some of the reviewers, however, did not give any helpful feedback whatsoever. Moving past this, there are some aspects that I improved on for my final draft. Most importantly, I changed the style of my writing. My first draft took the form of a thesis, with extensive details, background, and methodology. However, a thesis proposal differs greatly from a thesis. A thesis proposal involves much more focus on persuasion and purpose, rather than the introduction that I was making it out to be. I also made sure to take brevity into account for the proposal, as a proposal is intentionally much shorter than a thesis. With the help of the relevant editor comments, I believe I have submitted a version that more accurately portrays my writing and scientific ability.
Since the last update, I have finished the initial processing of my Sentinel-1 images. This editing took place in the SeNtinel Applications Platform. With the help of Jahnavi and Gavin, and through my own exploring, the images could be processed correctly. The first, incorrect attempts that I tried to do were to 1) apply a radiometric calibration to correct errors in the data product and to reduce noise 2) apply a multi-looking correction so that pixel size and resolution are lowered to reduce noise; 3) apply a terrain correction, using ArcticDEM as a base. This resulted in images that looked like the following:
These images are obviously very wrong. They are skewed and so not show the lava field properly. After these failures, I re-convened with Jahnavi and Gavin. It was determined that my data products (Ground-Range Detected) were already multi-looked and projected. The next, and correct way of processing images now that we figured this out was to 1) apply a radiometric calibration and 2) perform a raster flip to reorient the image to the correct direction. I also decided that it would be easiest to crop the images down at this point as well so that I would not need to import extremely large tif files into ArcGIS. Using the graph-builder function in SNAP allowed me to automate the process (though I could not figure out how batch processing worked). The images that I produced looked much better and were more accurate than what the previous attempts yielded.
With these all the images processed, I could finally import them into Arc. When I brought these tifs into Arc, they did not seem to be properly projected, so I applied a map projection to all the images. In Arc, I used the Raster Calculator to take ratios of the VH and VV radar images, making VH/VV images such as this:
and these GIFs showing changes over time of the VH images and VH/VV of the year 2019, respectively:
There are many problems that I am facing in my work at the moment, some of which are in my control, but others of which are out of my hands. First, I was instructed to obtain a facies map of Holuhraun that was made by another scientist studying the site. I emailed this individual to ask their permission to send the map for me to use. At the time of writing this blog update, I am still waiting on a response. As well, a growing side-endeavor of this project is to understand how weather (and in particular precipitation) affects the radar images. To do this, I need to obtain historical data for the weather on the days that the radar images were taken. This, however, is proving to be a very difficult task. Holuhraun is remote and does not have weather stations nearby. As well, the historical records for weather are rather spotty, and I have yet to find daily information. I have emailed the Icelandic Meteorological Office to inquire about historical records but have not yet heard back from them. Lastly, I have been having problems georeferencing. Arc has not given me the option to georeferenced the images. I have used Earth Explorer to find a Landsat8 image to georeferenced to, but I am still waiting for the website to send me the image. After I get the Landsat, I can georeferenced all of the VH and the VV images, and then ratio them. This will be better than trying to georeferenced the VH/VV images that I have already made.