Tuesday, July 30, 2013

Surveying the Spruce Mill Site

Over the past weeks, teams of students have been learning archaeological survey skills at the site of the World War I Spruce Mill, a significant home-front site associated with the processing of old growth Sitka Spruce cut from the forests of the Coast Range of Oregon and Washington and the Olympic Peninsula. 
Besides learning basic pedestrian survey techniques, the crews have also conducted subsurface surveys searching for remnants of the old cut-up mill building, the shops, and tent city for the 3,000 troops that manned the mill. An excellent history of the site by Ward Tonsfeldt helped to guide probing along with remote sensing work by Kendal McDonald reported in an earlier post.
Students excavated probes at locations defined by the magnetometer anomalies recorded earlier in the summer and at locations discovered during monitoring of the removal of some hangars a few years ago. The results confirm that there is abundant evidence of the WWI facility including concrete foundations, concentrations of wire nails and railroad spikes, and even an axe head. Notably, we have been able to see some of the vegetation anomalies associated with the Spruce Mill appearing more distinctive as the ground dries up.

Dry grass in more closely cropped areas suggest where Spruce Mill foundations sit. The ones inside the fort denote where the loading docks to the East and South of the main mill building were located.

The foundations of the sawdust burner are clearly visible as linear dry grass areas adjacent to the reconstructed bake house. 


House Floors, Imaging, and Modelling with Tablet Computers

Over the past two weeks, we have exposed the house floors in both the Little Proulx House and the 1 X 5 m test trench in the WilliamKaulehelehe House.  In both cases, the floors contained evidence of burning consistent with the destruction of both houses by the US Army ca. 1860.  The floors are rich with Hudson’s Bay Company-era artifacts, including beads, buttons, ceramics, vessel glass, window glass, square nails and many other items. There is a very distinctive burnt surface in the eastern half of the Kaulehelehe House site with some evidence of north-south running wooden planks which may indicate a wooden floor that burnt in place or a fallen wall.  In the midst of this rich deposit of artifacts and charred wood and charcoal, a door pintle was found.  This object is of particular importance as it is reported that the U.S. Army removed the windows and door from the Kaulehelehe House prior to torching it on March 20, 1860.  This door pintle and the associated burn layer adjacent to it may be a direct link to the destruction of the house site and the eviction of the Native Hawaiian preacher.
The William Kaulehelehe House site showing the hearth and some preserved planks and charcoal staining.  A door pintle is located in the northern portion of the image.  A water screen sample was previously removed from the southwest corner of the unit and the floor around the hearth has already been excavated.
We had two groups who came to tour the site from the Ke Kukui Foundation.  It was amazing to share these finds with people who have such an interest in Hawaiian heritage and the story of the Hawaiian diaspora.

Dr. Bob Cromwell and I interpret to the Ke Kukui Foundation tour at the Little Prouxl House Site. 

Following the E'se'get Archaeology Project, I decided to capture the hearth using the Autodesk 123D Capture program with my iphone.  As we do not have a 4G connection or wireless access in the field for the iPads, I did not use the tablets for this experiment.  I was quite surprised at the resolution and ease with which 3-D models could be generated in the field.  I took 36 photos of the feature from different angles and submitted them to the Autodesk server that crunched the data in about 15 minutes.  The results were later converted into a video animation on my lap top that I have posted on my University YouTube channel:

YouTube animation of the William Kaulehelehe House Site Hearth (Feature 406)
 As we use the tablet computers, we have begun to discuss improvements to field recording that are facilitated by the concentration of many tools associated with one device.  An obvious improvement would be to take a photograph of the floor of each level as a background for drawing things like rocks, sediment variation, feature boundaries, etc.  Without tablet computers, this is difficult as there are generally only a few cameras on each project.  As each tablet contains a high-resolution camera, it is much easier to collect photographic data on the floor plans (and profiles).  Both the iDraw and pdf Expert apps can import images, although the iDraw app is more sophisticated.  

One issue has been with correcting the distortion caused by cameras that were not placed directly above the floor.  The use of photo-processing apps that remove the distortion (orthorectify the photo) and create a planimetric view may allow a resolution to this problem.  There are a variety of apps available that straighten and flatten images, including programs designed to capture the text and images from whiteboards.  We will be experimenting with some of these apps to improve the capturing of these data.  Simplifying the drawing process may help to streamline the archaeological recovery of data and allow for more sophisticated data to be collected and processed directly in the field. 

Once the distortion is removed, then the picture can be cropped to the size of the unit floor (usually 1 x 1 m) and then dropped into and registered to the image space on the recording form.  Annotations can be placed on top of the image.  For those annotations that were placed prior to the completion of the plan, a translucent image can be generated that will show the earlier details underneath the image.  Theoretically this will free up time drawing things like rocks that are obvious in images, while leaving the ability to annotate those aspects of the floor plan that are not as easy to discriminate with photography.

Tuesday, July 16, 2013

More on Battery Life

After a week of use, a few statistics can be developed regarding the average use of the iPad and its battery capacity during archaeological excavation.  Table 1 below identifies my tracking of the battery life used each day for the ten iPads. iPads 5, 9 and 10 were only used for the cemetery survey work which only occurred on Thursday and Friday last week.  In addition, Friday afternoon (7/12) was truncated by a lecture and the water screening demonstration so only represent about a 5-hour day.  Due to a glitch, iPads 1 and 2 were not charged the night of 7/9/2013 and the students started 50x50-cm shovel tests (with traditional paper forms) the morning of 7/10/2013 while the units were charging.  iPad 4 was at 8% at 1 pm on 7/10/2013 when it was recharged, therefore only representing about a 4.5 hour work day.

TABLE 1. iPad BATTERY LIFE USED OVER A FIVE-DAY FIELD TEST IN JULY

  7/9/2013 7/10/2013
iPad# Start End Battery Life Used Start End Battery Life Used
1 78 42 36 100 45 55
2 68 3 65 100 84 16
3 100 86 14 86 75 11
4 100 61 39 61 8 53
5 100 100 0 100 100 0
6 81 56 25 100 55 45
7 76 55 21 100 100 0
8 100 59 41 100 67 33
9 100 100 0 100 100 0
10 100 100 0 100 100 0
7/11/2013 7/12/2013
iPad# Start End Battery Life Used Start End Battery Life Used
1 100 73 27 100 97 3
2 100 50 50 100 79 21
3 100 93 7 100 100 0
4 100 42 58 100 82 18
5 99 24 75 100 37 63
6 100 60 40 100 88 12
7 100 37 63 100 50 50
8 100 68 32 100 48 52
9 100 14 86 100 41 59
10 99 22 77 100 38 62
7/13/2013
iPad# Start End Battery Life Used
1 100 60 40
2 100 79 21
3 100 99 1
4 100 98 2
5 100 100 0
6 100 40 60
7 100 2 98
8 100 19 81
9 100 100 0
10 100 100 0      

As shown in the Table, the cemetery monument recording work, which generates many more forms and digital images, uses up a lot more battery life.  For the two days in which they recorded cemetery headstones, the student's iPads used between 59% and 86% of battery life.  Because Friday was only a five-hour day, and Thursday was a training day when the new crews were getting oriented to the recording strategy, the average battery use might be a bit more for a fully trained crew over an 8 hour workday.  The average for Thursday of 79.3% is probably conservative.  

Ignoring the partial days and charging glitches, the other seven iPads measurements over the five days is 35 observations of battery use.  The average use was 34.0% of the battery with a standard deviation of 24.26.  There was a significant range between no use (the iPad was not used) and 98% of battery use. As iPads are tied to particular excavation units, no use indicates that those units were not excavated that day.  Excavation tasks included filling out the four-page level form, taking images, and some intensive use of iDraw to map the World War I railroad grade. I think a fully trained professional field crew (excavating at a rate of 2 10-cm levels per person per day) will likely generate more forms and perhaps use more battery life depending on the complexity of the site.   Even so, there appears to be plenty of capacity for excavation forms and digital images. This is consistent with the results from other projects, such as the Pyla-Koutsopetria Archaeological Project on Cyprus and the E'se'get Archaeology Project in Nova Scotia.

Monday, July 15, 2013

House Floors, the Spruce Mill, and Other Excavation Notes

Concentration of beads being discovered at the Little Prouxl House Site.

Hearth exposed below house floor
(about 70 cm top to bottom -- west is up)
Over the past two weeks we have been exploring the house floor of the Little Proulx House Site and clearing more units to expand on these excavations.  There have been a number of notable finds, including a cluster of about 70 white beads (what appear to be small, hot-tumbled tube beads) in the southwest corner of one of the excavation units.  The location of this cluster is mapped using the annotation tool on the PDF Expert program. Just to the west of this, a concentration of baked clay (or "bisque" as it is used in the Pacific Northwest, probably from the house fire when it burned), a charred beam, and concentrations of  bisque and charcoal likely associated with the destruction of the house.  In one area nearby, a nicely-defined hearth with a charcoal-black rim and reddish-brown interior was discovered below the clay floor.  This may be a hearth that preceded the construction of the house or an earlier iteration of the structure before the clay floor was
laid down.



We are also shovel probing the Spruce Mill Area, exploring some of the magnetic anomalies identified by Kendal McDonald a few weeks ago.  The probes are about 40 cm in diameter and are dug to at least 50 cm in depth.  The first week of probing we found many interesting strata tied to fill associated with the World War I mill site and later uses of the property for aviation and other U.S. Army and City of Vancouver uses. We found surprisingly few metal objects or other sources that could be tied to the anomalies.  This past week we used a metal detector to really home in on the magnetic sources and were rewarded with many metal objects, including cast spikes, a likely aviation fuel line, and other objects.
Excavation of a shovel probe at the Spruce Mill cut-up mill site.
Pearson Air Museum, including the repurposed World War I hanger  in the background.
Back at the dig site, we used iDraw to map the top of one of the World War I Spruce Mill spur lines that rest above the Little Prouxl House Site.  Besides being an excellent training excercise, we are troubleshooting the use of iDraw for illustration of plan and profile maps.

Monday, July 8, 2013

Battery Life in the Field


One of the concerns raised by some with the use of tablets in the field was battery life and if the tablets would last through an entire day of field recording. I am happy to report that battery usage during excavation is about 20-40% during the day depending on the number of forms worked on. While the school is digging more slowly than a professional crew in the field, it appears that there is plenty of juice to record excavation data without having a supplemental source or having to swap out units.

In contrast, the cemetery recording project uses up an entire battery each day. One tablet ran out of juice before the final photos of monuments were recorded. This is undoubtedly due to the much larger number of forms created each day during cemetery recording.

By recording battery life each day, we should be able to get some metrics on the number of forms generated contrasted with battery expenditure. This week I will collect some data.

Wednesday, July 3, 2013

Survey Crew Starts Work at Spruce Mill

As part of the Field school, the students learn skills in field surveying to discover and record archaeological sites, including basic orienteering and gps, pedestrian and subsurface shovel probing, and mapping and site recording.  This week, a quarter of the field school started this portion of the course under the direction of Tony Hofkamp.  The site of this course is the field east of the reconstructed fur-trade fort and west of Pearson Air Museum.  This was the location of the World War I Spruce cut-up mill.

Last week Kendal McDonald (Z-Too Archaeogeophysical Prospection and Applied Archaeological Research) collected remote sensing data from the Spruce Mill area using her gradiometer (see Remote Sensing post from last week).  She was able to get us some preliminary results from the 160 m long north-south by 20 m wide transect we set up across the Cut-up mill. There are some dramatic patterns in magnetic anomalies that the survey crew will explore over the next four weeks.

While we excavated over a dozen probes this week, we have not had good fortune hitting the source of some of the anomalies. Next week I plan to use a metal detector to refine the locations for probes

Kendal McDonald shows Field School the results of the
Gradiometer Survey in the Field Laboratory
Magnetometer Results laid out on the Laboratory Floor.  Some of the strong positive-negative
linear anomalies may be buried cables for aircraft ties from the later airport.

Bag Catalogs and the Field Laboratory

Yesterday we started our field laboratory portion of the course.  Students will cycle through the field lab to gain basic skills in organizing, cleaning, and error-checking the collections and archival materials. When we were designing the digital forms we made the conscious decision to only maintain a bag catalog tied to the level and feature forms and to not have a separate bag catalog for artifact bags collected from each unit.  While this removes an aspect of redundancy which is designed to avoid errors, it was thought that the digital forms would facilitate the creation of a bag catalog in the field laboratory to assist in error checking, and building a site-wide artifact catalog.  On Monday, while the Field School rested, I explored how to go from the adobe forms to excel to create a usable bag catalog. The pdf forms can be extracted to an excel spreadsheet with ease and it is simple to extract the data from each level form as a row (single line on a relational database) tied to individual units.  To then use these data to create an inventory of bags collected, I wrote a short macro using Visual Basic which primarily uses a For Next routine to collect the six variables collected for each bag, including its bag number, type of artifact (e.g., beads, transferprint ceramic, etc.), the field count of artifacts in the bag, and if it was point-provenienced, the northing, easting, and elevation of that object.  The routine simply copies these data for each level form listed on the excel worksheet as a line to a new bag catalog worksheet.  At this point there is still a bit of cleanup necessary to sort the bags to remove blank spaces between levels and I had to add a "level" field that we are adding by hand in the lab during error checking.  When I get a free moment I will work on the macro to attempt to further streamline these processes.  We are starting to use these "bag catalogs" in the field lab to create lot and specimen numbers to track the collections through the cleaning, analysis, and cataloging process.  Stay tuned for how it will work out!