Stratigraphic Photogrammetry: Part 2 – A Quick Proof of Concept


I’ve been moving along, making progress on my methodology for stratigraphic photogrammetry. I don’t remember just off of the top of my head, but I think that the actual title of the paper is something like A Method for the Use of a Non-Metric Digital Camera in Stratigraphic Photogrammetry. I present the paper (which I will post here as soon as I am able, although I think that I want to have it edited by someone else before I post it here) on Wednesday morning. As part of finishing the paper, to show that my method can overcome the basic obsticles in using a non-metric camera to make metric photographs, I set up a flight-line on one of the public trails in the area and took a series of stereo photographs (digital images, actually) of a stratigraphic profile, as a proof-of-concept.

I set up on the John Wayne Trail in Eastern Ellensburg, Washington. The John Wayne Trail is a public jogging trail created along an old railroad prism and cuts. After weeks of plotting and figuring and calculation, I really didn’t know what to expect when I went out to take my images. craig2s.jpgI was pretty confident in my calculations for scale and coverage, but I really didn’t know. It turns out that my biggest problem wasn’t the photogrammetric method at all. The biggest problem was that most of the tasks of setting up a flight-line require two or three people to do them. I should have known that it would take more people. All of the things I needed to do, like pulling level lines and measuring grids, etc. are things that I have done before on monitoring projects or at the mammoth dig this summer. I really should have guessed that I wouldn’t have much luck trying to dig my toes into a slick clay slope, pull a metric tape and a level line with one hand, place a grid-corner nail with the other before hurrying to grab my hammer and try to pound the nail in before it fell. I eventually gave up trying to build a reference grid onto the wall, which would have been a necessary step if I was planning on orthorectifying my images to overcome radial distortion caused by the short focal length and cheap lense of my camera. craig3s.jpgLuckily, I was trying to create stereo images, which no longer show stereo if they are orthorectified, and so had no real need for the reference grid. Overall the images turned out even better than I expected. I couldn’t get right at the wall, so I had to pull back away from it to 10m. This put the visible stratigraphy at the very top of the image: much smaller that I had hoped for, but more than enough under the circumstances. The images do indeed show up in 3D under a stereoscope.

If you click on the thumbnails you can see bigger versions of the images, but even those are small, poor quality versions. I reduced the resolution to save bandwidth. If you have any interest in the full-resolution pictures, just drop me a note, either by e-mail, or in the comments to this post, and I will send them to you.

Cultural Palimpsest

I have been working on a Land Use Land Cover (LULC) mapping project of the Yakima River delta area for the last few weeks. The task assumed that the only information available to us on the area was the set of four aerial photographs we were given, and included making a detailed map (accurate to within 51m) showing the patterns of land use and land cover in the 8mi2area. A part of the task was a 3-page essay giving an explanation of the patterns of development shown by the new map. Writing the essay for the project, I had an idea of development in the area as making something of a cultural palimpsest.


The word palimpsestcomes to us from more ancient times, when paper and skins for writing were very dear, and very expensive. Everything was reused. To reuse a skin, the writing on it had to be scraped off with a knife. The process of scraping off writing removed most of the ink, but little ghosted stains remained, very much like the ghosted writing on a dry-erase marker board. These ghosted writing stains are called palimpsests.

Although the term originally refers to a remnant of writing, palimpsest has been used in many different fields to mean an original form that has been covered up, but some of it still remains. That is the meaning of the word in geomorphology. Cultural palimpsest, though is generally used in the field of literature, and refers to something like pieces of one culture shining through into the writings of a different, later culture. When I talk of a cultural palimpsest, I am not talking about literature, I am talking about land use and land development.

The land development that got me thinking was a number of clean, geometrically-shaped housing and commercially developments located around the delta of the Yakima River, where the Yakima meets up with the Columbia River. Considering my essay on the patterns of development in the area, it seemed to me that there seemed to be a correlation between the density of a development and the disregard of that development’s creators for the physical context of the development. In an area dominated by two rivers, many of the housing and commercial areas, formed at highway intersections and at nodes of transportation corridors, might as well have been built in Nebraska, for all the regard they paid to the rivers. Assuming a negative correlation between development density and regard for physical context in the area, I hypothesized (in my essay) about a succession of development, in layers, from a river-context-sensitive bottom layer of farm land, to early urban development, to layers of increasingly dense urban and suburban development, and that small patches of farm land that have survived to the present day actually represent a cultural palimpsest, preserved in gaps between the increasingly more geometrically-shaped developments.

I hypothesized a cultural palimpsest, but I had no real way of knowing, based only on the air photos that I had available to me. However, our newest assignment in the airphoto lab is an analysis of historical change in the area using the LULC maps (which were made from 1996 or 1997 air photos) and an air photo from 1940. I took one look at the 1940 air photo and came really very close to a penalty for excessive celebration. The original settlement of the area was in large sheets (an entire layer) of river-context-sensitive farm land. As development occurred over the years, it increased in density and disregard for the cultural and physical context of the area, which meant that small slivers of farm land were left in what I have been calling a cultural palimpsest. If you have a better name for this phenomenon, please let me know

Stratigraphic Photogrammetry: Part 1 – Calculations for Photometrics on the Olympus C-750

For the last few months, I have been working on a methodology for using terrestrial photogrammetry techniques in capturing stratigraphic profiles. The original idea for this came on the Wenas Mammoth dig (which I wrote about on 10 August 2005), when we were not sure, at one point, if our trench, cut into a wall of loess, would survive over the weekend. At that point, we were only about halfway through our stratigraphic analysis, and a colapse of the trench over the weekend would have set us back very far. I took systematic close-up photographs of the trench wall, hoping to capture some of the features of the stratigraphy in case we lost the wall.

The trench didn’t fall over the weekend, which was a very good thing: in my ignorance, I made many mistakes taking my photographs, and they were completely unusable. Since then, I have been working on the techniques to make measureable photographs of a wall. Because I am using a digital camera to make digital images, the first step in the methodology is to calculate image size, scale, coverage, resolution and spacing for the camera. Because the focal length and sensor size, as well as the pixel resolution, of each digital camera is different, the calculations have to be made for each model of camera (although cameras with the exact same focal length, pixel resolution, and sensor size would have the same calculations). I have an Olympus C-750, which has a focal length of 7.8mm, and a sensor size of 5.67mm x 4.39mm. I am including here a table of all the different calculations for my camera, assuming a screen and print resolution of 200dpi, which would give a photo size of 8″ x 10″.

Update (11/12/05): Going out in the field today to actually try making metric photographs with my camera, using the calculations from Table 1.1 below, I realised that the calculations I made were actually for the Olympus C-140 and C-160 3.2 MP UltraZoom cameras, and that the model that I actually own is the Olympus C-150 4MP. The origianl table of calculations is of no use to me because I don’t own those cameras, but I will leave it here on the chance that someone might be able to make some use of it. I have added the calculations of the Olympus C-150, the camera that I actually own, as Table 1.2, below.


Update (11/14/05): I have posted Part 2 of my discussion of stratigraphic photogrammetry using a consumer-level digital camera.


Table 1.1 Coverage, Resolution, and Scale Calculations for the Olympus C-740/760 3.2MP Digital Cameras
Distance from Wall DV Coverage (m) DH Coverage (m) Photoscale (200dpi) Photo Resolution (mm/px) Stereo-pair H-spacing Stereo-pair V-spacing
0.50 0.28 0.37
1 /
1.4 0.18 0.15 0.11
1.00 0.56 0.74
1 /
2.8 0.36 0.30 0.22
1.50 0.84 1.11
1 /
4.3 0.54 0.45 0.33
2.00 1.13 1.48
1 /
5.7 0.72 0.60 0.44
2.50 1.41 1.85
1 /
7.1 0.90 0.75 0.55
3.00 1.69 2.22
1 /
8.5 1.08 0.90 0.66
3.50 1.97 2.58
1 /
9.9 1.26 1.05 0.77
4.00 2.25 2.95
1 /
11.4 1.44 1.20 0.88
4.50 2.53 3.32
1 /
12.8 1.62 1.35 0.99
5.00 2.81 3.69
1 /
14.2 1.80 1.50 1.10
5.50 3.10 4.06
1 /
15.6 1.98 1.65 1.21
6.00 3.38 4.43
1 /
17.0 2.16 1.80 1.32
6.50 3.66 4.80
1 /
18.5 2.34 1.95 1.43
7.00 3.94 5.17
1 /
19.9 2.52 2.10 1.54
7.50 4.22 5.54
1 /
21.3 2.70 2.25 1.65
8.00 4.50 5.91
1 /
22.7 2.88 2.40 1.76
8.50 4.78 6.28
1 /
24.1 3.06 2.55 1.87
9.00 5.07 6.65
1 /
25.6 3.25 2.70 1.98
9.50 5.35 7.02
1 /
27.0 3.43 2.85 2.09
10.00 5.63 7.38
1 /
28.4 3.61 3.00 2.20
Table 1.2 Coverage, Resolution, and Scale Calculations for the Olympus C-750 4MP Digital Camera
Distance from Wall DV Coverage (m) DH Coverage (m) Photoscale (200dpi) Photo Resolution (mm/px) Stereo-pair H-spacing Stereo-pair V-spacing
0.50 0.35 0.46
1 /
1.6 0.22 0.15 0.11
1.00 0.70 0.91
1 /
3.1 0.45 0.30 0.22
1.50 1.05 1.37
1 /
4.7 0.67 0.45 0.33
2.00 1.39 1.83
1 /
6.3 0.89 0.60 0.44
2.50 1.74 2.29
1 /
7.9 1.12 0.75 0.55
3.00 2.09 2.74
1 /
9.4 1.34 0.90 0.66
3.50 2.44 3.20
1 /
11.0 1.56 1.05 0.77
4.00 2.79 3.66
1 /
12.6 1.79 1.20 0.88
4.50 3.14 4.11
1 /
14.2 2.01 1.35 0.99
5.00 3.48 4.57
1 /
15.7 2.23 1.50 1.10
5.50 3.83 5.03
1 /
17.3 2.46 1.65 1.21
6.00 4.18 5.49
1 /
18.9 2.68 1.80 1.32
6.50 4.53 5.94
1 /
20.5 2.90 1.95 1.43
7.00 4.88 6.40
1 /
22.0 3.13 2.10 1.54
7.50 5.23 6.86
1 /
23.6 3.35 2.25 1.65
8.00 5.57 7.31
1 /
25.2 3.57 2.40 1.76
8.50 5.92 7.77
1 /
26.7 3.79 2.55 1.87
9.00 6.27 8.23
1 /
28.3 4.02 2.70 1.98
9.50 6.62 8.69
1 /
29.9 4.24 2.85 2.09
10.00 6.97 9.14
1 /
31.5 4.46 3.00 2.20

Hardness of Glacial Ice

The following table is adapted (a little) from a handout given in my Geomorphology class. No source was given on the handout. The table shows (suggests, really) that the hardness of glacial ice increases as its temperature decreases. I think the point that is trying tomake, though, is that even very, very cold ice (-78.5°C, which seems very unlikely for a glacier to get so cold, especially considering the heat generated, and the thin layer of luricating water resulting from the basal sliding that would be required for abrasion) isn’t hard enough to scratch rocks as hard as quartz,making the case for most of theabrasive work of glaciers being done by debris that has been picked up movement of the glacier.


Table 1 – Hardness of Glacial Ice in Comparison to Representative Minerals and Common Items
Glacial Ice at 0°C
Copper Penny, Glacial Ice at -40°C
Knife Blade, Glass Plate, File, Glacial Ice at -78.5°C

Definitions of Landslide Features

I have spent a lot of time lately learning about mass wasting. No, that doesn’t have anything to do with drugs. Mass wasting is the downslope movement of slope material under the influence of gravity. Mass wasting shows up as rock falls, landslides, and various types of flow (mud flow, earthflow, debris flows, etc.). I am reproducing here a table from a handout that I got in a recent lecture on landslides. The handout has on it a hand-written reference to Turner & Schuster (eds) 1996, Landslides: Investigation & Mitigation.  On closer examination, an almost identical table can be found at the wikipedia entry for Landslide Clasification.  The wikipedia version, attributed to Varnes (1978), Cruden and Varnes (1996), Hutchinson (1988), and Hungr et al. (2001), is given below.

Type of movement Type of material
Bedrock Engineering soils
Predominantly fine Predominantly coarse
Falls Rockfall Earth fall Debris fall
Topples Rock topple Earth topple Debris topple
Slides Rotational Rock slump Earth slump Debris slump
Translational Few units Rock block slide Earth block slide Debris block slide
Many units Rock slide Earth slide Debris slide
Lateral spreads Rock spread Earth spread Debris spread
Flows Rock flow Earth flow Debris flow
Rock avalanche Debris avalanche
(Deep creep) (Soil creep)
Complex and compound Combination in time and/or space of two or more principal types of movement

Desirable Watershed Conditions

The capture, storage, and beneficial release of water.
The Watershed Manager’s Mantra

To achieve the watershed manager’s goal of capturing, storing, and releasing water in a safe and beneficial way, he must:

  1. Maintain vegetation on the site sufficient for absorbing the energy of precipitation (so raindrops hit leaves and dissipate their energy – raindrops splatter soil, but they also plug pore spaces, which lowers infiltration, and increases runoff, which means an increase in erosive potential.).  *Enhances infiltration
    • (organic soil has higher infiltration) *Vegetation delays movement of water toward and into drainage pathways.  1in rainfal on 1ac exerts 900ft-tons of energy.
    • Increasing velocity increases the capacity of water to do erosive work.
    • A slower raindrop has much more time to infiltrate.
    • The more water that plants absorb, the less water to run over the surface.  Less surface water equals less erosion.
  2. Maintain minimal drainage density.  The higher the drainage density, the better the drainage (more risk of flash flooding, water leaves the system quicker).  Achieve highest possible sinuosity.
  3. Optimize temporary water storage.

In Search of a Better Scraper

One of my main pedological field tools (pedology is the science of dirt), other than my trusty shovel, is the “scraper.”  The search for the perfect scraper is the focus of all my gear-hunting efforts.  The scraper is used when looking at a soil profile to get a clean, flat, clear soil face to look at.  I have tried, or seen tried as scrapers: old army shovels; the head of a Pulaski firefighting tool on a shorter handle, like a hammer handle; garden hoes; garden trowels; masonry trowels; barbeque spatula; and an ice scraper.  I even used a pie knife for a while.  The very best scraper tool I have used, though, and perhaps the most widely used and deeply loved, is a pointed Marshalltown Trowel.

Wenas Mammoth Dig: 2005 Field Season

There have been no updates on the site for quite a while because I have spent the last six weeks (June 28-August 05, 2005) working at the Wenas, WA mammoth excavation. This was the first season of excavation at the site. The dig was set up as a field school through Central Washington University’s Office of Continuing Education, with participation from the University’s Geography and Anthropology Departments. My participation in the dig was as a member of the field school.

Mammoth left humerus next to the same bone from an adult cow

In February of this year, a construction crew was building a private road to a house on the hill on the south side of Wenas Valley, when the backhoe they were using to create the road-cut struck bone. The bone that they found was determined, after examination by a member of the faculty at CWU, to be the left humerus of an elephant-sized mammal. The excavation field school began in the end of June, with nine students (only two of us from the Geography Department: the rest were archaeology students from the Anthropology Department), three CWU professors, and Bax Barton, a paleoecologist with the University of Washington’s Quaternary Research Center. At the time, the primary objective of the dig was to establish a geologic context for the left humerus that had been found, with finding more bone, and identifying the species of the animal as secondary objectives. One of my tasks, part of establishing the geologic context of the humerus, was to work on the stratigraphy and sedimentology of the site.

Because of the multi-disciplinary nature of the field school, much of the first week was spent in the classroom, where principles and techniques from the various disciplines involved (geomorphology, paleontology, archaeology, stratigraphy, geology, ecology, biology, etc.) were presented to the members of the field school. The time not spent in the classroom during the first week was spent preparing the excavation site for a ground penetrating radar (GPR) survey. The GPR machine only has a clearance of a few three or four inches, so all the brush and grass had to be removed from the site in patches large enough to create two survey grids, something like 20m x 15m, each.

The GPR machine is walked back and forth across the survey site at 50cm intervals, both laterally and horizontally, measuring the rate of reflection of radio waves at depths, in slices across the grid. When the lattice of slice, lateral and horizontal are put together in advanced software, and interpolated, the result is a three-dimensional image, and map slices at various depths, showing spots where reflected returns are higher, or lower than average, suggesting buried materials that reflect radio waves better or worse than the surrounding soil matrix.

Based on vague returns from the GPR survey, an excavation grid was laid out, and two backhoe trenches were dug, at an average depth of 2m, and with a total length of something like 35m, in an L-shape around the East and South of the excavation grid. I spent the better part of five weeks in those two trenches, with Dr. Karl Lillquist, a geomorphologist, and chair of my department at CWU, working out the stratigraphic story of the site, in a way that is both understandable, and defensible.

Mosaic of the North trench stratigraphic drawings.  Click to view a PDF of the Wenas site stratigraphic drawings.  Adobe Acrobat or Acrobat Viewer is required.


The field school was originally set to be four weeks long. However, Jake Shapley, the field assistant, and the greatest champion of the Wenas dig, pushed for a six-week field school, and it was approved, extending to six weeks before it was advertised. The thinking for the four-week school was that, with only one bone, and no real reason to believe that there was any more at the site, it might be a stretch to fill up even a four-week course. The first three weeks of the excavation seemed proof that there just wasn’t enough to fill up the time. While Dr. Lillquist and I systematically described the site stratigraphy in the trenches, using orange-flagged nails on a 50cm grid to mark boundaries between layers, and while Ryan Murphy, the other member of the field school from the Geography Department, used a total-station (an advanced piece of surveying equipment that uses lasers and prisms to plot three-dimensional coordinates of surveyed points) to map the topography and excavation geography of the site, the archaeology students (with participation from the rest of us: everyone doing their part, especially when large volumes of dirt had to be moved) proceeded to open 2m x 2m excavation units, and dig them down, in a controlled, scientific manner, using archaeology techniques, 10cm at a time, screening all of the dirt from their units, looking for bones or artifacts. The excavation units really turned up nothing but shattered fragments of bone, mostly spongy material, for the first three weeks.

View Southat the Wenas site, up the North trench.

During week four, everything changed. All of a sudden, starting on Tuesday of that week, all of the excavation units started to have large bone. From the fourth week on, a right humerus (possibly a mate to the original left humerus), some possible rib bone, and what might be large pieces of cranial bone, were all uncovered at the site. The finds started to be of a great enough volume that the dig, which was only originally scheduled for four weeks, has been approved for a second field season, and there is talk of third and fourth field seasons.

The last few days of the dig were spent marking the stratigraphies of the excavation units, and marrying them to the stratigraphy of the trenches.

Speaking of the significance of the dig, Bax Barton mentioned that this might be the largest scientific excavation of Pleistocene mega-fauna that has been done in the Northwest. While that may seem to be a rather focused and qualified achievement, it is some kind of contribution.

Photos, diagrams, maps, and more explanation of the results of the dig will be posted here, in a revision to this article, as they are available.