As you may have seen in our three part “Running your Picarro GasScouter and eosAC Off-Grid” series, we have been busy integrating our eosAC autochambers and eosMX-P Portable Multiplexer with the new Picarro GasScouter. With the deployment at BEEC complete and development work wrapping up, Chief Scientist Nick Nickerson and I decided to take the equipment out for a weekend drive to gather some survey-style flux measurements.
Polly’s Cove is situated just East of its more famous relative, Peggy’s Cove, as shown in this very Nova Scotian walk-through. These coastal barrens are quite striking; featuring a mix of stunted evergreens, low and dense vegetation, small bogs and exposed granite bedrock.
At the end of a hot and dry summer, the drought stress in the local plant life is obvious. Local naturalist groups have been observing dieback and early fall colors in vegetation across the province. Fens, bogs and marshes appear to have dried up in some locations, with the margins of many lakeshores or bogs, which are not usually traversable, now accessible. Instead of typical brilliant red and purple colors common of the barrens later in the fall; leaves have changed to to the crumpled browns, yellows and grey which indicate plant stress.
Biologist Caitlin Porter at Saint Mary’s University‘s E.P.I.C lab is studying the effects of the drought on barrens plant communities, and first noticed impacts of the extreme conditions while visiting a rock barren site at Tupper Lake. “At Tupper Lake, it was apparent that entire plant communities were experiencing stress related to the drought. We observed large areas where the dominant species were changing color two months early or in some cases entire plants have been killed.” Her experience at Tupper Lake sparked a study on the impacts of drought across Nova Scotian barrens which included our Polly’s Cove field site. “Polly’s Cove is part of the iconic Peggy’s Cove Barrens complex, a field site frequented by our lab for research on plant ecology for more than a decade. None of us have seen drought conditions this severe there in the past. Species that are specialists at tolerating drought are exhibiting stress responses; something we haven’t seen before. In the large bog area that [Eosense] surveyed, we were shocked when we recently visited and observed open water pools drying up and submergent aquatic vegetation fully exposed. Those areas have been covered by water for as long as we have worked at this location.”
Given the diversity of vegetation available in a relatively small area, it was a perfect place to gather some interesting chamber measurements on a sunny Saturday. Along with the Picarro GasScouter and Eosense eosAC autochamber, we took several PVC collars to ensure a good seal on the uneven terrain. We deployed the collars at five locations, as shown on the map below, between 10:00-10:45 AM.
Collar depths varied between 4-6 cm, depending on soil conditions. Site One was situated approximately 5 m from the bank of a nearby pond, in an area of coastal dwarf heathland dominated by Black Crowberry, Cinnamon Fern and Cloudberry. Numerous other species were present in abundance, as dwarf heaths are known for their high species richness. Site Two was located 1.5 m inside an adjacent tree island of Black Spruce trees, with a sparse undergrowth of tall heathland plant species. This community exhibited significant near-surface root systems. Site Three was at slightly higher (perhaps 3 m) elevation than the other locations, near an apparent border between several different dominant vegetation types and near the edge of a sizable lithomorphic community (Common Juniper, Peat Mosses and Hoary Rock Moss).
Sites Four and Five were located next to the small peat bog that was showing signs of severe drought. Site Four was on the upland edge, about 2.5 m from the bog, while Site Five was located on a patch of dried peat. These sites were characterized by an abundance of Peat Mosses and Deer Grass, among other wetland species. All locations were situated in direct sunlight, with the exception of Site Two (Spruce Trees), which was in partial shade.
Once all five collars were installed, we took a brief hiatus before deploying the eosAC to give the subsurface some time to recover from the disturbance. The effective time between collar and chamber deployment at each site was approximately two hours, or enough time to take in some coastal scenery.
In addition to the Picarro G4301 GasScouter, we used one of our eosAC autochambers to measure gas fluxes. With a bit of hardware and software tweaking, we were able to power the chamber from the GasScouter‘s internal battery, and modify our existing eosAC software to collect on-demand chamber measurements†. With a combined power draw of less than 40 W, this system could operate in the field for up to six hours on a single charge.
Starting at 11:52 AM, we began collecting CO2 and CH4 flux measurements at each site, with two 7 minute chamber closures (three at Site Three) separated by 3 minutes open. Given the relatively short overall measurement window (~ two hours), sequential measurements were taken at each site in order, rather than alternating or randomizing locations. Based on the shape of the chamber accumulation curves, an exponential fit was used for both gas species to determine the rate of flux from the surface.
CO2 average fluxes varied between 1.21 and 6.45 μmol m-2 s-1, while CH4 average fluxes were between 0.43 and 5.06 nmol m-2 s-1, with the highest producing sites being the spruce tree and bog locations, respectively. The repeat measurements at each site showed very low percent variation in estimated flux rate (<6%) in general, though there were a few exceptions (Pond: 14.2% for CO2, Spruce Trees: 15.2% for CO2 and 21.4% for CH4, Bog Bank: 18.7% for CH4). In each instance of large variation, the measurement(s) after the first showed a decrease in emission rate, indicating that the initial higher value could be due to soil disturbance from placing the chamber.
CO2 emissions were fairly consistent across sites, with the exception of the high production among the spruce trees at Site Two; possibly due to the extensive near-surface root structure. CH4 fluxes were significantly higher near the water bodies, particularly the bog (Sites Four and Five) where there was a substantial amount of peat and decomposing organic matter. Interestingly, methane emissions were almost identical between the bank and directly on top of the dried peat, though CO2 fluxes at the latter were less than half, possibly due to the near-saturation water levels.
Given the large variance in subsurface composition and plant population at each location, comparing the relative fluxes between sites during a normal (not drought-stressed) period could shed further light on these data. “Part of understanding drought impacts and ecosystem stress is more than just how plants and communities respond to them but how whole ecosystems respond.” Porter says. “I think the tools and techniques your (Eosense) researchers are implementing could be useful in quantifying impacts that we can’t see or maybe as an early warning detection for ecosystem stress. It’s been established that peatlands serve as a carbon sink and I think quantifying these processes and response to stressful weather such as drought should be important in that context as well.”
All in all, we collected 11 flux measurements over the span of about two hours, once the collars and our lunch had settled. Even with the GasScouter supplying power for the eosAC chamber, we headed back to the office with over half of the battery capacity remaining. For similar day trips, we recommend installing the collars well ahead of time (> 48 hours) and planning your measurements around expected soil temperature effects.
Special thanks to Caitlin Porter for recommending this field site and for her assistance in identifying and interpreting plant species and possible moisture stress effects. Caitlin is a co-coordinator of the Acadian Heathlands Ecosystem classification project, which aims to better characterize biodiversity of heathland ecosystems across Nova Scotia.
Chance is Eosense’s VP Research and Development. In addition to product design and refinement, he also creates Eosense’s interface software and numeric models.
† Eosense does not currently offer a commercial version of the eosAC that operates without the use of an eosMX multiplexer. If you are interested running a similar single-chamber system, please contact us for more information.