Environmental Diagnostics at Fort Union National Monument, Part 3: Pilot Monitoring Methodology

Since my last visit to Fort Union National Monument, I was tasked with devising a long term monitoring methodology to record critical data on, around, and within the adobe ruins. No small task.

Select walls had been pre-selected in previous conservation plans, particular from the early ’90s, as well as from our last visit and rapid assessment survey, as ones being of higher priority. These walls typically had visible features such as leaning, slumping, warping, or red-flag superficial characteristics such as failure of the shelter coats, basal erosion etc.
In the weeks preceding this recent visit, which would only leave five days/four nights of boots on the ground, I compiled a proposal for various variables to be monitored, means for monitoring them, as well as the requisite installation protocols.

A photo of the Onset weather station in the distance through the early morning New Mexico fog.
A photo of the weather station in the distance through the early morning New Mexico fog.

Of paramount importance was the installation of a high resolution weather station on site. Although the park prided itself on collected weather data since the mid nineteenth century, the hand-written format in which this was done rendered it valuable but ultimately useless when it came to what was needed moving forward. Readings taken once a day at noon, including sky cover, temperature, relative humidity, and others, have been invaluable in looking at large scale trends over the last hundred years, however the proposed monitoring proposal requires a temporal resolution, at a minimum, of one reading/15minutes–an industry standard.

Therefore it did not makes sense to discuss a monitoring methodology for the walls that would gather high resolution microclimate data without having the corresponding macroclimate data at the same resolution. The weather data from the Las Vegas Municipal Airport, over forty miles away, simply does not provide the level of accuracy required. With the temperature, relative humidity, wind and gust speed, wind direction, and rain depth readings for the station, it is possible to correlate these factors with readings taken from the wall. An oversight, the lack of a solar radiation sensor (which will be installed during the next site visit in March), was temporarily resolved by also logging the weather station system voltage, which is fed from a 3.5W solar panel. By knowing the voltage levels it is possible, in the time being, to deduce cloud cover.

Having secured the funding to acquire and install the weather station, the primary variable to be monitored, however, is moisture content of the walls, which in adobe is not easily done. And as it turns out, very little has been published on long-term, embedded, moisture content monitoring in solid historic adobe masonry walls. Two methods were selected for a pilot test: capacitive moisture readings and ambient RH/temp. These methods are commonly used to measure moisture content of soils, and to varying degrees can be adaptive to adobe walls. Capacitive sensors work essentially by recording changes in resistant between two poles of a sensor ( a + and a -); the presence of moisture increases conductivity, resulting in a lower resistance which can be measured. Furthermore, in soils this can be calibrated to a specific type of soil, and volumetric moisture content can be calculated. The sensors on the market presently that provide this capability and compatibility with the Onset loggers that we have at the ACL don’t however also record temperature. One will note that conductivity will change with respect to temperature, and thus a second sensor would need to be installed to measure simultaneously record temperature at that location. Not optimal. Furthermore, these sensors–for obvious reasons–do not function below freezing temperature, a condition that occurs during the winters at Fort Union.

The test wall segment and snow depth gauge in the background with the timelapse camera in the foreground.
The test wall segment and snow depth gauge in the background with the timelapse camera in the foreground.
Quick lab sketch of sensor installation and positioning schematic.
Quick lab sketch of sensor installation and positioning schematic.
The other, more promising option involves recording the ambient RH and temperature within a core made inside the wall. In this case an ambient RH/temp sensor made by Onset was used and installed within a 3/4″ PCV housing 8″ long. At the end of the PVC a moisture permeable mesh mebrane was installed to prevent the ingress of liquid water and debris but allow the free flow of moisture. The hole was carefully drilled at a depth of 8″ and at a diameter of 1″. The PVC tube was installed to a depth of 6″ leaving 1″ air gap between the end of the hole and the PVC housing, and 2″ of pipe sticking out from the wall. Two sensors of this type were installed an 8-bit and 12-bit variation. Three capacitive sensors were installed as well. Two were installed at grade, adjacent to the foundation stone (one on each side, east and west), and one was installed at the same height as the other embedded sensors, and dry-packed in with loose, wet adobe mud (to allow for proper contact between the core walls and the sensor). THe three capacitive sensors and one eight-bit temperature/RH sensor are then connected to a 4-channel Onset microstation datalogger. The remaining temp/RH sensor is a standalone installation, and is connected to a solo Onset pendant logger. This, along with all excess cable slack are housed within a custom PVC housing (fabricated in the isles of the Hacienda in nearby Las Vegas–the equivalent of Home Depot) to prevent rodent interference. As rodents, particularly pack mice, are notorious for chewing threw data cables, the cables themselves were encased in light gauge plastic sprinkler line. Split down the middle and wrapped around all exposed cables, this as well as the PVC housing–we hope–will be enough to deter an unexpected rodent malfeasance.

The advantage of the core, ambient method is that with proper lab testing, it is possible to establish the sorption and desorption curves for that particular species of adobe, which correlate equilibrium moisture content of the material with the relative humidity of its environment. In this way, the moisture content can be calculated and approximated using a regression of these curves just by knowing, to some accurate measure, the RH and temperature. This is the strategy that will be proposed moving forward, and samples are being sent back from the site to execute the appropriate lab tests.

An example of a EMC curve for generic materials. Such charts are empirically formulated, and as such are few and far between for adobe masonry. Source: https://permies.com/t/43637/Breathable-Walls
An example of a EMC curve for generic materials. Such charts are empirically formulated, and as such are few and far between for adobe masonry. Source: https://permies.com/t/43637/Breathable-Walls

Also installed as part of the pilot study was a means to record snow drift/bank patterns at the test wall site (and a bit beyond) using time lapse photography and an embedded scale stick. In this case a Brinno time lapse, weatherproof camera was installed on a segment of embedded 3/4″ galvanized conduit approximately 42″ above grade. The camera is focused on the wall and the immediate foreground, with some resolution capable of picking detail into the background, where a corresponding segment of 3/4″ galvanized conduit was wrapped with DOT reflective marker tape in 6″ increments. As compared to an ultrasonic pinger (typically used on mountainsides and in remote locations for snow depth measurement) or a capacitive depth recorder, the time-lapse method allows for the gauging of not only snow depth, at a reasonable level of accuracy, as well as snow deposition pattern at the test wall and its immediate surrounding context.

The next post will discuss alternatives for datalogging and open source methods for fabricating my own sensors using an arduino-based platform–and at a fraction of the cost. Lab testing is underway at the moment at the architectural conservation lab to put my system through its paces and prepare it for installation during the next site visit in March. Also, I plan on devoting a separate post to ongoing development of machine-learning algorithms to correlate the weather and wall data into a predictive protocol of risk assessment. I also hope to cover ongoing research into developing open source image processing methods to detect minute changes in time lapse photos for wall movement detection, crack formation detection, moisture detection, as well as snow depth and pattern detection. in the Stay tuned!

Teaser of some lab testing I am doing with an arduino-based datalogging platform.
Teaser of some lab testing I am doing with an arduino-based datalogging platform.

 


Leave a Comment