The Geography of Mars

Getting Your APXS Clusters into Google Earth

• To have you create a CSV file from your APXS K-means clustering spreadsheet that can be imported into Google Earth Pro
• To show you how to import CSV files sorted by cluster and format placemarks during the process of mass database import (colors and icons) or modify folders all at once after the fact
• More generally, show you how to move fairly large amounts of data into Google Earth Pro at once for mapping, rather than having to go placemark by placemark. To do this requires Google Earth Pro (which is freely available at https://www.google.com/earth/desktop/), but, once a CSV has been converted into KML or KMZ form, it is usable by plain old Google Earth.
• To have you sort out the geochemical "personalities" of the four rovers' working areas

Deliverables:

In the last lab, we did the K-means clustering in PAST, brought the clusters back into the original Calc spreadsheet, counted the targets that fell in each cluster, and then did charts showing the relative abundance of each oxide and element in each of the four clusters. I had you hang onto all that for mapping work in this lab. So, the deliverables are going to be a compendium of work from both labs.

• KMZ file that you create during this process to display your APXS target clusters (uploaded to Dropbox for Lab 9)
• the CSV spreadsheet you used to import your data into Google Earth Pro (Dropbox for Lab 8)
• your original Calc spreadsheet, including a small table of target counts by cluster and your four graphs. The table and graphs may be turned in as hard copy (always appreciated).
• a stacked bar chart of relative prevalence of each cluster at the four rovers' work sites
• a brief lab report on the character of the four clusters and how they may or may not reflect the metaclusters described at the top of last week's lab, as well as comparing and contrsting the geochemical "personalities" of the four regions the rovers work or worked in.

Preparing the CSV

Fire up OpenOffice and open the spreadsheet in which you did your K-means clustering of the merged APXS data. Your four clusters should be listed down column AP if you followed directions from Lab 8 exactly and didn't improv. Now, we need to get those four clusters into a form that Google Earth Pro won't mangle, which it will if we leave them as numbers. So, we'll trick it.

You'll notice an empty column I didn't have you use in Lab 8: Column A, labeled Folder. We'll use this column to convert the cluster numbers into cluster letters to browbeat Google Earth into doing what we want, which is mapping each cluster with a different symbol that we can edit en masse without having to hand-edit each of the 901 APXS targets individually.

In cell A2, type in the following if-then-else code:

if(ap2=1;"A";if(ap2=2;"B";if(ap2=3;"C";"D")))

This will classify all your cluster numbers into letters, assuming you have every little bit of punctuation in exactly the right order. Drag that successful formula all the way down the column.

Save that file!

Now, create a new file. Save it as a CSV (comma-separated values file). To do this, Save As and then use the Save as type drop-down menu to choose the Text CSV option. Please name it LastnameLab9.csv (so I know which masterpiece belongs to which of you. We need to customize it so Google Earth will only display relevant information about each APXS target.

The new file should contain only the following columns:

• Folder
• Mission
• Name
• Latitude
• Longitude
• Sol
• Description

Highlight and copy only those, in that order, into your new file. Save it.

Now, close the file and we're ready to move into Google Earth Pro.

Before we actually go there, Download (Save Link as ... whereever you save stuff) the following file:

• Fernando Nogal's traverse map for Curiosity (Google Earth Mars' is all messed-up).

Importing your CSV into Google Earth Pro

Fire up Google Earth Pro from the Desktop. After it loads and you shut down its little helpful hints, click on the Tools tab on the top toolbar and then Options. A dialogue box comes up. On the 3D View one, click on Deimal Degrees under Show Lat/Long. Under Units of Measurement, pick Meters, Kilometers. Now, under the Navigation tab, pick Do not automatically tilt while zooming, which tends to mess up the map function we're interested in here. All that done, hit OK.

Click on the Planets icon up top and select Mars (sort of obvious!). Now, hit the File tab up on the left. Our first chore, optional but extremely useful, given Google Earth screwing up Curiosity's traverse, is to Open and navigate to wherever you stowed Fernando Nogal's Martian Way Curiosity traverse. All you-know-what will break out, so you may be shutting down the Mars clock and then doing a search on someplace else, such as Spirit, to get back into normal nadir-looking perspective!

Now to go get YOUR data: From the File tab, this time pick Import (NOT Open) and navigate to wherever you put your LasstnameKmeans4.csv file. A dialogue box comes up so you can specify that it is a Delimited file (sometimes it comes up delimited by commas, sometimes delimited by tabs: If the preview window looks like your file, just hit Next (if it doesn't, try the other delimiter).

In this new dialogue, make sure that there is no check by "Does not include latitude and longitude" (because it does -- this option is for use of addresses on Earth for mapping instead of the geographic grid). Verify that the latitude and longitude field names are okay. Hit Next.

Now you can Specify field types. Folder, Mission, Name, and Description should all register as string variables. Latitude, longitude, and Sol should be floating point. Your original K4 numbers will probably be identified as integers. Change any that aren't guessed correctly (unlikely). Then, hit Finish.

Another small dialogue box comes up: Apply style template? Choose Yes. another dialogue box comes up with four tabs.

• Under Name, set the Name field as, well, the Name field
• Under Color, select Set color from field and choose the Folder field. Very important: Click the create sub-folders for each bucket option!!! Be sure to copy the same letters of your clusters to the Folder name column that appears (next to the numbers of targets that showed up in each of your clusters).

  Now, start fiddling around with the Palette start color and end color options. Use the color picker to pick two colors you like that are very contrasting, one for the Start and then one for the End. You can, if you're familiar with RGB, go nuts and design end-member colors. Probably not wise your first time through all this!

• Under Icon, you again specify Set icon from field and choose Folder. If you picked a color palette under the Color tab, you don't need to create sub-folders here. If you didn't use the Color tab, you will have to create sub-folders here. You can choose the blank icons (circles, squares, stars, triangles, diamonds) toward the middle of the drop down menus that show up under each folder's icon button and have them filled with the color palette you picked under the Color tab. Alternatively, you can just pick one of Google's pushpin icons (so many goofy choices!).

• Under the Height icon, it's probably best to pick Clamp features to ground. Unless you really like that circus-balloon effect of icons on strings floating above your APXS target! No accounting for taste!

On trying to Finish up, the program will ask you if you'd like to save your artistic endeavors as a template (KST file) you could recycle later for some future quick and dirty map job. Probably a good idea, just in case. We can't have enough fashion accessories for lazy future mapping.

On hitting Finish, Google Earth will ask you if you want to stay on Mars: Yes, you do.

Admire your handiwork!

Go to the bottom of the Places navigation bar on the left side of Google Earth, down to Temporary Places. Be sure to click on your imported CSV there to light the map up with your icons.

Also, be sure to expand that file by clicking the triangle next to the checkbox. This lets you see the KMZ file it created from your CSV. It should be expanded now, too. You will see your four lettered folders. You can expand those to see all the individual targets they each contain (or not).

At this point, surf around to see your rover traverses decorated with icons marking each spot one of the rovers nuked a target. To do this, use the Search button on the upper left and type in "Spirit," "Opportunity," "Curiosity," or "Pathfinder." Scroll around and change your scale to peer at the landscapes the rovers crossed and the specific spots where someone thought it would be an interesting idea to take an APXS spectrum.

If you click on one of the icons, a "thought balloon" opens up and lists all of that site's attributes from the CSV (the seven columns are now seven rows).

If, for some reason, you don't like the way your placemarks look for a particular folder, you can touch the folder name, right-click, and choose Properties. You can do all sorts of stuff in there using the Style, Color tab. To edit the folder there, you have to click Share Style and then touch the box in the upper right (with a radio button in it). Touch that box to re-affirm your icon choice -- or pick another. Now, you can change its scale and color. Hit okay, and you can then change the size of the labels

Once you're happy with your artistry, Save the KMZ. Click on it to highlight it, then click File -- Save Place As and put it on your flash drive as LastnameLab9.kmz and then upload that (named that way) to BeachBoard's Dropbox.

Getting at the "personalities" of the four regions

In the spreadsheet you created your CSV from (4rovers441541.ods, NOT the csv), let's set up a table and graph to convey the regional variations in the dominance of our four clusters, using a nifty spreadsheet function.

Somewhere in there, write your four cluster "names" (1, 2, 3, 4 or A, B, C, D) in four cells. For convenience, I'll put them in row 1, starting in column K. So, K1:N1 will contain your cluster names.

In cell I2, put MPF; in I3, put MER-A; in I4, put MER-B; and in I5, put MSL, the technical names of the rovers. For pretty graphing, put their popular (Google Earth search) names in column J (Pathfinder, Spirit, Opportunity, Curiosity).

Now, we can use a cross-tab counting function in Calc to count the number of occurrences of each cluster in each of the four rover work areas. In cell K2, write in =countifs($b$2:$b$902;$i2;$a$2:$a$902;k$1). CountifS will classify all occurrences of each Mission by the mission specified in I2:i5 by the number of cases of each cluster in the Folder column. If you're careful about all the dollar signs, colons, and semi-colons, you'll get a cross-tabbed table suitable for analysis. You can also sum the four columns and the four rows of the table to make sure you have 901 records, 11 for MPF, 220 for MER-A, 370 for MER-B, and 300 for MSL and that the column counts match the numbers of targets that you got in your K-means cluster.

Assuming everything's okay, you can now create a graph to visualize all this. Highlight cells J1:N5 and ask for a graph. Calc defaults to a bar chart. Pick the third type of bar chart, where each column is stacked and the stacks are the same height, so that you can visualie relative abundance. Hit Next, Next, Next, and then fill out your titles and axis labels. You should leave the legend in to interpret the colors by cluster. Hit Finish and then Format Chart Area to put a neat line around the graph so you can see it when you're moving it around. h

Analyzing Your Clusters

So, now you have your clusters mapped and graphed. You figured out what the four clusters reflect in the last lab by making four graphs of their average relative abundances. We talked about those graphs:

• the commonest cluster doesn't diverge from the Mars average relative abundances in anything, so it's no doubt basalt or ST1.
• another cluster shows enrichment in felsic oxides (such as those of sodium, aluminum, silicon, phosphorous, potassium, and titanium) and is depleted of the mafic oxides (those of iron and magnesium), while signs of water alteration are depleted or Mars-typical). This is probably evolved igneous material (leaning in the andesitic direction or ST2).
• another cluster shows depletion in those felsic oxides and Mars-typical levels of mafic oxides and halogens, but is highly enriched in sulfur trioxide, probably a signal of acidic water altering rocks and soils, which was most likely during Hesperian times after massive volcanism and all the associated sulfur species cranked out by volcanic eruptions.
• a very rare cluster was the crazy-looking one due, no doubt, to small sample effects with a rare process. This suggests signs of water alteration but in the form of neutral-ish pH waters. Such waters readily liberate chlorine and bromium, as well as some susceptible oxides (e.g., those of magnesium, potassium, and manganese) and zinc. These get concentrated during evaporation, creating salt flats, rinds, and joints in rock, as well as clays.

Granted, Pathfinder and its little rover, Sojourner, only got eleven APXS measurements out there on the southeastern side of Chryse Planitia. Which two target types were found there?

Opportunity is a stand-out in heightened abundance of which of the four clusters? What does that tell you about that region? Which other region is most similar to Opportunity's Meridiani Planum in that cluster? Which of the four clusters found at all four sites is rarest around Opportunity's work site? What does that tell you about the kind of volcanism in its area?

Which of the four sites has the richest and, within limits, the best-balanced of the four clusters? Which type is uniquely visible there? What does that tell you about its hydrological history?

Considering only Spirit, Opportunity, and Curiosity, which of the other two does Gale Crater (Curiosity's work space) most resemble?

If you're morbidly curious about the 15 cluster analysis I did en route to developing this lab, feel free to visit: https://home.csulb.edu/~rodrigue/mars/apxs/GE/.