The Geography of Mars

Constructing Longitudinal Profiles along Channels
Using IDL Virtual Machine, Gridview, and Calc
and MOLA Data to Evaluate Proposed Drainage Channels

This lab has the following objectives:

• to familiarize you with MOLA data
• to introduce you to IDL Virtual Machine software and the Gridview application for viewing the MOLA derived digital elevation model for Mars
• to let you continue developing proficiency in spreadsheet analysis and data visualization in OpenOffice Calc
• to give you practice in the fine art of transforming data formats for use in different programs (converting .txt files into .csv files in Notepad and then importing them into OpenOffice Calc for conversion into .ods format)
• to have you do all this in order to evaluate the hypothesis that many features on Mars that look like fluvial drainages in fact once carried water or other fluids
• you'll do this by constructing longitudinal profiles of proposed drainages to see if they could have drained fluids from higher elevations to lower elevations (each student will do two of these)

Background

Mars has several features at a variety of scales, which look like fluvial drainages of various types.

• Some of them sport dendritic networks of the sort we see on Earth (first order, second order, third order, etc.), which develop from precipitation, overland flow, underground flow and springs, and eventually channelized flow that takes water from highlands to oceans, lakes, and playas.
• Other systems feature long main trunks with few and rather short tributaries culminating upland in alcoves. On Earth, these "theater-headed valleys" develop in arid lands from groundwater seepage undermining strong caprock layers.
• Still other systems are almost incomprehensibly huge and seem to have carried massive amounts of liquid great distances, possibly in single catastrophic outflows. On Earth, that has happened when large ice dams gave way suddenly at the end of the Pleistocene. A terrestrial example is the Channeled Scablands of Idaho, Washington, and Oregon, which were scoured out when Lake Missoula in Montana disgorged after a branch of the Cordilleran Ice Sheet gave way. These apocalyptic flood events are called "jökulhlaups," and they can be triggered by climate change undermining an ice dam or by volcanic eruptions or magma movements under an ice cap or perhaps even under permafrost.

A recurrent controversy that bears on this issue is the one surrounding a possible ocean on Mars in the Northern Lowlands, which was proposed by CSULB Geological Sciences alumnus Tim Parker. One of his arguments was that several of these drainage features seem to debouch at the same general elevation on the shores of one of his proposed shorelines.

For more information:

• if you would like a copy of Gridview to explore its many features at home, it is freely available (with IDL Virtual Machine) at http://denali.gsfc.nasa.gov/mola_pub/gridview/gridview.html.
• an image of a dendritic network on Mars (Viking Orbiter image 606A56, in the Thaumasia Highlands, -42.5° and 92.6° W, and you can now use Gridview to find it!): https://home.csulb.edu/~rodrigue/mars/MarsDendriticChannels.gif
• about gigantic Pleistocene floods in North America: http://hugefloods.com/
• an overview of jökulhlaups more generally: http://www.antarcticglaciers.org/glacial-geology/glacial-landforms/jokulhlaups/
• a video of the jökulhlaup triggered in Iceland in April 2010 when the Eyjafjallajökull volcano erupted under the glaciers covering it since its last eruption in the 1820s: http://www.youtube.com/watch?v=fJII-u-41Lg&feature=player_detailpage
• about alcove-headed tributary systems on Mars: http://www.niu.edu/landform/papers/luo_groundwater_erosion_2007JE002981.pdf

Your data

Your data consist of elevation readings collected by the Mars Orbiter Laser Altimeter or MOLA. MOLA was one of the instruments carried on the Mars Global Surveyor orbiter, which operated from 1997 to 2006. It worked by sending laser beams to the surface of Mars and recording the time of return of the reflected beam (kind of like a satellite-borne total station). These laser returns eventually generated hundreds of millions of discrete elevation estimates, with a elevation uncertainty about + 3 m, and these are the basis of the Mars digital elevation model (for a cool, but huge representation of the MOLA DEM as a contour map, hypsometrically tinted differently than the MOLA maps you've already seen, do check out http://geopubs.wr.usgs.gov/open-file/of02-283/. The metadata explains uncertainty levels).

The MOLA data are saved as a "grid" in a .sav file for use in IDL and the Gridview application. To get at them, you will need to be able to fire up the IDL Virtual Machine and Gridview and then use Gridview to open this grid. You will eventually produce tables of MOLA readings for particular points, consisting of latitude, longitude (westing and a means of calculating easting), elevation in meters, and distance in kilometers from an arbitrary starting point.

IDL VM and Gridview

To get to your data, open the Statistics folder on the desktop and double-click on the IDL VM Gridview 6.3 icon. A globe grid will come up.

Click on File and then Load (.sav) grid. Navigate to C:\RSI\IDL 63\ Gridview\Grids, selecting the mola_04.sav grid. This will populate your gridded globe with MOLA elevation data, hypsometrically tinted in a blue and red scheme.

To move around Mars, you need to type in various latitude and longitude coördinates and then hit Reset. This is the only way to move around in Gridview: There is no panning function. The geographic grid used here represents latitude by plusses and minuses, not N and S. North is positive latitude, while south is negative latitude. Longitude here is the cartographically optimal westing, or planetographic, system (based on center-of-figure, rather than center-of-mass of the planet). Longitude goes from 0°: full circle to 360°. On Earth, our custom is to have longitude values increase in both directions from Greenwich, differentiating the hemispheres with W and E Here, all longitudes are westings, the values increasing going west. So, 30° E would be shown as 330°. This may take a little discomfort getting used to.

So, you need to figure out the latitude and longitude (westing) of the starting point of your assigned feature. Find it, perhaps on Google Mars or in Google Earth Mars in our lab. Estimate lat./lon. Then, in Gridview, type in your estimate and hit Reset and see if Mars is recentered roughly where you estimated your feature was. You can fiddle around with this until you have your feature's region in the middle of the globe.

You will need to zoom in now. Use the cross-hair cursor to define a rectangle encompassing your feature and hopefully not too much else. If you zoom in too far, the image pixelates pretty badly. Once the box appears on the globe, go to Tools and select Zoom in. Now that you can see what you're doing, you get to collect topographic profiles. You may need to string several of them together because the "stream" features meander around, and the profile feature only does straight line segments.

In Tools, select Profile. Then, go back to the globe and carefully pick your starting point to represent the "headwaters" of your drainage feature and click on it. Now, pick a point "downstream," trying to stay more or less on the "thalweg" of your "stream," or in the middle of the deepest part of the channel there.

A window will pop up, showing a longitudinal profile along the path you selected. You can move the cursor, a vertical and horizontal crosshair line, around and it will give you elevation readings for various latitude and longitude points along the profile. You can also select Save as and create a .txt file of the points defining the profile. You can ask it to save the text file with the .csv extension, instead of .txt, which will save you a little work later on. You should probably create a folder for these text files with a name that you will be able to link with the feature you're doing. You can also save a .bmp graphic of your profile, if you like.

Once that text file is safely stowed, go back and use Tools -- Profile to do another one. You will see the previous one as a black line on the globe. Try to get the next segment started exactly where the previous one left off (you will almost certainly not get it exactly but close enough for the purpose at hand). Then, pick another end point somewhere farther along the "thalweg" of your feature, and repeat the whole process of Save as.

You will be going back and forth from Profile-Start-End and Save as .txt (.csv) several times, depending on how straight your feature or subsections of it are and how closely you try to move "downstream" along the meanders of the valley's course. This may get a little tedious. About six to twelve segments should do most of these features justice.

Preparing the Data for a Spreadsheet

If you were to try opening these .txt files at this point, you would probably activate a word processor in the lab or at home. That would be useless for further analytic processing. You need to get the data reformatted to work in a spreadsheet. If you remembered to save the text file with the .csv extension, your spreadsheet should recognize the format and open it. If you forgot and saved it as .txt, don't worry. Here's a workaround.

To get a .txt file ready for opening by a spreadsheet, go to Start -- Programs -- Accessories -- Notepad. Use it to navigate to the folder you used to save your profiles. Open the first profile .txt file. Now, ask Notepad to Save as.

Here's a tricky part. In the dialogue box, you have to go into the Save as type box and select All Files, NOT Text Documents (.txt). Now, erase the .txt extension and rename the file so it has a .csv extension (and make sure it doesn't save as .txt or .csv.txt). Very important.

Once you see the file is somethingorother.csv, save it and close it.

Now, however you created your .csv file, fire up OpenOffice and ask it to open the .csv file you just created. You will get a dialogue box asking you about the characteristics of this .csv file, part of which you can see in the box. Under Other options, click on Detect special numbers. Under Separator options, click on Fixed width. A ruler bar will show up above the preview of your file. Touch it and you will see a black line dropping through it into the file below. Move it so that it JUST clears the rightmost number in the first column and click. You will see a column has been created for that column of numbers. Now, do the same to the immediate right of the next three columns. Now, hit Okay.

OpenOffice Calc will now import the file as five columns of data and a messed up header. Immediately, save the file as an ODF spreadsheet (.ods extension) from the Save as type bar. Voilà! You have just successfully reformatted your file for use in a spreadsheet!

Now, in OpenOffice Calc, do a File -- Open and open the next .txt (.csv) file the same way. This time, however, highlight all the data record rows, hit Control-Copy, and then put it in the first .ods spreadsheet and Paste it there just after the first pile of data. While it's still highlighted in the .ods spreadsheet, you might want to change the color of this second pot of data, just so you can keep track of which segment this is (very important later). Go back and do the same thing, over and over, until you have appended (and saved) all the separate .csv files into a common .ods file. And save! Once you've saved the master .ods file, go on and close out all those .csv files.

Further Pre-Processing the data in Calc

You now have a spreadsheet with five columns. The variable names are sitting, for some reason, only in cell A1. Please use the following variable names in row 1:

• Column A is "Latitude"
• Column B is "W Lon"
• Colmn C is "E Lon" (actually, it isn't, exactly. To use eastings, you would need to create a new column in which you would have =360+C# to conver the longitude into eastings. Let's not go there).
• Column D is "Elev (m)"
• Column E is "Distance " -- or distance in kilometers from the beginning point in that particular profile.

Now, while you're at it, pretty things up. Click on the grey box in the upper left corner of your spreadsheet to highlight the whole spreadsheet. Click the right-justify button up top (so the variable names will line up with the numbers below them). Also, click Format, then Cells, then Numbers, and, under Decimal places, pick a common decimal place format, perhaps 2 or 3 or 4.

Unfortunately, to create a consolidated longitudinal profile, we need to do a scatterplot of Elev (m) against Dist (km), but the Distance variable is messed up. We need to get the distances appended to one another, not starting anew from 0 for every single (differently colored?) profile you have.

To do that, we need to create a new column and cook this mathematically. Insert a new column between E lon and Elev. Call it "Dist (km)" (in the new cell D1). In cell D2, type =f2. Copy this cell down to the end of the data from the first .csv file (in other words, until cell f# is 0. Now, we have to deal with the new zero problem.

Let's say, your first batch of data ended in row 156. In this example (your numbers may be different), row 157 is the first record from the second batch of data. So, in cell D 157, type =D$156+(D$156-d$155)/2+f157 -- and pay attention to every dollar sign (and lack thereof). What this does is keep the distance increasing monotonically when you copy this formula all the way down the second profile (using that color-coding to know where to stop). But there's a plot complication.

Uhhh, there may be another plot complication, which came up in the S/16 class: What if you keep getting error messages every time you try typing in this formula, even though you checked every letter and number? It may be that you imported the numbers as text. This would happen in certain versions of OpenOffice, if you did not click on the Detect special numbers box when you imported the .csv file into the spreadsheet. Not to worry: Here's another workaround. You can copy the six variable names into cells G1 through L1. Yes, you're going to create six duplicate variables. Now, in cell G2, type =value(a2) and hit Enter. This will convert the "text" value for Latitude into the correct number value. Now, put your cursor back in cell G2, in the lower right corner, right-click, and drag it to L2. Now, go to the dot at the lower right corner of the highlighted cells and right-click and drag from row 2 all the way to the last row of your spreadsheet. For the rest of the lab, you'll have to convert directions for, say, D2 to J2 and E2 to K2.

Now, back to our regularly scheduled plot complication. I can guarantee that the first point of your second profile is not exactly on the last point of your first profile. It is a virtual impossibility to get them lined up perfectly. So, there is probably some distance between these two points. If you didn't factor that in, you would be creating two different elevations for the same point in space.

So, what I'm having you do is insert a distance placeholder, using the distance between the last two points of your first profile as a guesstimate of the distance between the two profiles themselves. That's what that formula at the beginning of each profile is all about. It's a little bogus, but the small error possibility is too small at this scale to make any significant difference.

Now, the unhappy part is you have to do the same sort of little correction between your second and third profiles, and your third and fourth profiles, and so on down the line. This is why I asked you to color code the rows as you imported them, so you'd know where one profile ended and the next started.

Now, the fun part: Chart your data

Once you have Dist (km) set up on the same continuum from the first point on the first profile, you can construct a common longitudinal profile for your drainage feature. Highlight columns D and E (or J and K, if you had to convert text values to numerical values) and click on the garish little bar chart button at the top of Calc. Select XY (Scatter). Now, select Lines Only (the third button from the left). Click on Smooth lines below and then hit Next. And Next, Next, and Next. Here, come up with a title for your longitudinal profile, Distance in km for the X axis title, and Elevation in m for the Y axis title. Uncheck Display legend. Hit Finish.

Not quite ready for your next art show, is it? To pretty it up, first, stretch that graph out in the X direction, so it looks a little more like a landscape. Click on the X axis (up there in the middle of your graph somewhere) and right-click and select Positioning -- Axis line -- Cross other axis at Start (under Value). Also, under the Numbers tab, unclick Source format and then change Decimal places to 0. Click on the Y axis and also change that to 0 decimal places.

You can fiddle with the thickness of the line (thinner would look better), colors of the line, chart area, chart wall, whatever. Pretty soon, it will start to look like one of your signature graphs -- and will convey the longitudinal profile quite effectively to help you analyze the potential movement of liquid across this landscape.

Analysis

You might want to print a working draft of this graph, so you can scribble on it for analytical purposes. Find the highest point on the profile and then follow it to the lowest point. Probably a pretty bumpy ride.

On Earth, streams drive toward a dynamic equilibrium state, where steeper slopes concentrate the degradation or erosive work of the stream and gentler slopes concentrate the aggradation or deposition work until, eventually, a graded longitudinal profile is approached. If you plotted stream elevation against distance from the headwaters, a graded stream would produce a nearly perfectly smooth concave-upwards curve. It would resemble half of a parabolic curve, with the steeper parts of the curve at the headwaters and the nearly flat parts where the stream's floodplain approaches its base level (usually the ocean here).

Perfect grade is an ideal condition, an equilibrium state towards which the stream adjusts its erosion and deposition work. In the real world, there will be imperfections.

• Perhaps the stream is flowing over a cliff of resistant material, forming a waterfall.
• Perhaps the stream's course is interrupted by lakes supported by one or another irregularity in its course, resistant bedrock, a landslide dam, an ice dam. The stream will concentrate erosion against the lake outlet and deposition within the still waters of the lake. The lakes will disappear as the stream approaches grade.
• Perhaps base level has risen (for example, as the Pleistocene glaciers melted and raised sea levels; sea levels are rising now in the wake of global warming, slowly raising base level). When the base level rises, the balance of a stream's work shifts towards aggradation.
• Perhaps the base level has fallen, as happened repeatedly during the Pleistocene, as ice ages withdrew ocean water for storage in glacial ice sheets. When the base level drops, the balance of a stream's work shifts towards degradation, incising into its landscape.
• Perhaps this is a newly raised or lowered landscape due to tectonic uplift or subsidence.

On Earth, with our abundant water, streams take a long while to attain a reasonably graded profile, thousands to millions of years, depending on the magnitude of the disturbance and the amount of water available in a system.

On Mars, streams clearly formed very early in its history, back in Noachian times (before ~3.7 Ga). Whether many of them lasted long enough to attain a reasonably graded condition is a question of interest. If we find landscapes with graded profiles, that would be a pretty clear evidence for fluvial processes.

Look at your landscape profile. Does it show that kind of concave upward profile? Remember that the profile you constructed has extreme vertical exaggeration, so even a perfectly graded profile might look bumpier than it really is. And Noachian Mars was hammered by impacts, leaving behind craters of varying sizes, which could become lakes or seas in a drainage course. Impacts could even fall in an active or long since dried up drainage network and disrupt the longitudinal profile, even if that had been a graded profile once upon a time. Trying to factor in the vertical exaggeration and the presence of craters, does the profile tend toward the upward concave profile?

Something you might try if you have a seemingly hopelessly ragged profile (perhaps due to craters), is use a pencil and fill in imaginary waters in each depression up to the top rim of each crater on the downhill pointing side. Look at the resulting string of "paternoster" "lakes" and see if the outlet of each is at or below its inlet upstream and if the outlet of one crater "lake" is above the inlet of the next one downstream. Does the whole system connect like a stairway of lakes leading liquid from the highlands towards the base level? Or, alternatively, is it possible that one or another of the craters served as the base level for a much smaller drainage basin, an interior drainage like the playas of the arid Southwest? It is possible that early martian fluvial landscapes had pretty deranged drainage patterns, with lots of interior drainage systems that may have functioned at very different times and no functional connection into longer systems debouching into the Northern Lowlands.

Lab report

Write a brief lab report interpreting the longitudinal profile of your two landscapes, giving due consideration to relevant factors mentioned in the Analysis section. Please include the two longitudinal profiles you constructed in OpenOffice (it's okay to turn in the original spreadsheet with the graphs in it or copy the graphs into your word processed report).

You can deposit the report in the Dropbox for the course in BeachBoard (Lab: Longitudinal profiles in Gridview).