Progress 02/15/24 to 02/14/25
Outputs
Target Audience:Target audiences reached include K-12 teachers and their students engaged in education and outreach, the scientific community, and the broader public. Changes/Problems:Co-PI Dr. John Campbell passed away in December of 2024. The PhD student working on the development UAV-GPR methods for detecting soil frost across the landscape decided to change his degree focus and has therefore left the project. What opportunities for training and professional development has the project provided?Opportunities for training and professional development include training seven undergraduate students in field measurements of snow depth and soil frost (for the winters of 2022-23 and 2023-24) and an additional three students in soil sampling and processing (to examine soil physicochemical drivers of frost depth and duration) during the current reporting period. Two undergraduate students assisted with installing our soil frost manipulation experiment and collecting samples for our planned lab incubations in the fall of 2023. We have also contributed to training two PhD students throughout the previous reporting period who have been developing our FroSen soil frost sensor prototype.. In September of 2024, we welcomed a postdoctoral researcher onto the project team who has been leading the efforts to address Objective 3 and understand the ecological consequences of changing soil frost regimes. Beyond the university environment, the project has provided training and professional development of five K-12 teachers and their students who have been collaborating in making manual and sensor-based measurements of snow depth, soil frost, and near-surface microclimate in outdoor classrooms in New Hampshire and Maine. How have the results been disseminated to communities of interest?We disseminated our research findings to scientific audiences in multiple presentations at the American Geophysical Annual Meeting. Beyond educational outreach described above, we also communicated our results to the K-12 community of teachers and students through the GLOBE training we conducted in collaboration with NH Sea Grant in fall of 2024 and direct engagement with Moharimet Elementary School students School . Members of the project have also shared project findings on popular media. Contosta and Burakowski were featured in a NOVA PBS video about trends in loss of snow cover and potential ecosystem impacts: https://www.youtube.com/watch?v=GAMFj0IkbrQ as well as a ScienceNewsExplores article about ecological impacts of a disappearing subnivium: https://www.snexplores.org/article/subnivium-life-under-snow-climate-risk What do you plan to do during the next reporting period to accomplish the goals?During the next reporting period, we will continue to address our research objectives. For Objective 1, the next reporting period will focus on optimizing sensor calibration, advancing data analysis, and disseminating our findings. To enhance freeze-thaw sensitivity, we will refine sensor material selection, adjust firmware settings, and integrate additional environmental sensors such as soil temperature and moisture sensors to improve data interpretation. A machine learning-based prediction model will be developed to analyze real-time sensor data alongside meteorological variables, including air temperature, wind speed, and precipitation, to improve frost depth forecasting. Additionally, correlation analysis will be conducted to assess how environmental conditions influence soil freezing and thawing dynamics. To ensure wide dissemination of our work, we will prepare and submit a manuscript based on 2024-25 sensor deployment data to peer-reviewed journals such as IEEE Transactions on Geoscience and Remote Sensing, ACM Transactions on Sensor Networks, or Springer's Environmental Monitoring and Assessment. We will also present our findings at conferences such as IEEE International Geoscience and Remote Sensing Symposium (IGARSS), or ACM International Conference on Embedded Networked Sensor Systems (SenSys) and the American Geophysical Union (AGU) Annual Meeting. We will also engage with stakeholders in agriculture, transportation, and environmental monitoring to explore real-world applications of the FroSen sensor system. These efforts aim to improve sensor accuracy, develop predictive models, and establish the practical relevance of our research. For the GPR work, we are in the process of analyzing radiograms for snow depth estimation and frost depth estimates, with the intention of sharing our findings in a peer-reviewed publication. For Objective 2, we will finalize the manuscript examining drivers of spatiotemporal variation in soil frost and will submit it for publication in a peer-reviewed journal. Regarding Objective 3, we plan to continue exploring the ecological consequences of soil frost through the field experiment we established in the winter of 2023-24 and are continuing in the winter of 2023-24. We will present our findings at the American Geophysical Union Annual Meeting in 2025 and will publish the results of our analysis in a peer-reviewed journal. We will also continue to pursue our collaboration with PNNL to examine how freeze-thaw cycles impact soil structure, and what the implications are for soil carbon losses. To do this, we will combine X-Ray Computed Tomography (pore-scale measurements) with core-scale CO2 efflux measurements for soils incubated at two contrasting water contents (low vs. high moisture) that are subjected to three freeze-thaw cycles. After each freeze-thaw cycle, the soil cores will be analyzed for (a) CO2 efflux and (b) pore structure using XCT (representing physical stability vs. disruption of the soil). As with other research activities, we anticipate that this work will yield both presentations to the scientific community and peer-reviewed publications. Finally, we will continue to address Objective 4, training undergraduate students in field measurements of snow depth and soil frost, lab-based soil frost sensor development, and field and lab-based soil sampling and analysis. Our project will also include ongoing training of one PhD student developing new automated soil frost sensor prototypes and one postdoctoral researcher examining the ecological consequences of changing soil frost regimes. In addition, we will continue our engagement with K-12 schools, ranging from elementary through high school, in student-led research in outdoor classrooms centered around measurements and monitoring of changing winters.
Impacts
What was accomplished under these goals?
Objective 1, focused on developing and refining wireless, in situ sensors. Between late January and early April, we deployed three frost sensors at our local field site, adjacent to manual soil frost measurement locations (Figure 1). The sensor functions by detecting a change in air pressure as water inside a sealed tube undergoes a phase change from liquid to ice. Data was collected at five-minute intervals. Analysis revealed a strong correlation between air pressure inside the tube and manual frost depth measurements, with Pearson and Spearman correlation coefficients of 0.80 and 0.78, respectively. This high accuracy, combined with the sensor's frequent data collection intervals, makes it well-suited for detecting rapid freeze-thaw transitions. Findings from this study were presented at the IEEE International Conference on Computing, Networking and Communication (ICNC) in February 2024 and at the American Geophysical Union (AGU) Annual Meeting in December 2024. Additionally, we have submitted a manuscript detailing our FroSen sensor to the journal IEEE Sensors, where it is currently under review. . Figure 1: Deployment pictures from the 2024-25 season. Figure 2: UAV GPR system (left); Soil frost tube and buried aluminum rod for comparison to GPR (right). The 2023-24 field season also featured continued measurements of soil frost using GPR at the University of New Hampshire's Kingman Farm. We developed a UAV platform for flying a GPR system. We made a UAV attachment for our GSSI 900 MHz antenna, SIR4000, and the Juniper Geode. We installed nine frost tubes to manually measure frost depth and also buried an aluminum rod to assess the ability of the system to detect impenetrable layers present in frozen ground (Figure 2). We conducted seven field campaigns in the winter 23/24 and have completed one field campaign in the winter 24/25. Winter 23/24 campaigns consist of a spiral pattern that covers the extent of a satellite SAR pixel while minimizing flight battery consumption. In addition to the spiral, we flew three small flight lines that were over the buried aluminum rod. We also made lidar flights following the GPR flight to obtain snow depth estimates over the area (Figure 3). We are in the process of analyzing radiograms for snow depth estimation and frost depth estimates. Challenges to deployment include equipment malfunctions during cold conditions and disturbance of the snow by UAV rotors. Figure 3. Flight lines for a UAV GPR survey on February 6, 2024, with locations of the buried aluminum rod and frost tubes (left); Lidar flight to monitor snow depth following the UAV GPR flight (right). For Objective 2, we have completed data analysis and are writing a manuscript describing how land cover, soil physiochemical properties, and climate conditions affect the depth and duration of soil frost across the landscape (Figure 4). Our results show that land use was the strongest predictor of soil frost spatial variation in seasonally snow-covered ecosystems in the Northeast region of the United States. Furthermore, soil temperature, air temperature, and precipitation were the main drivers of temporal variation in soil frost. Figure 4. Time series of average soil frost depth per land cover with 2021-22 represented by dashed lines and 2022-23 depicted by bold lines Project activities that address Objective 3 include conducting a second year of the manipulative field experiment initiated during the winter of 2023-24 (Figure 5). Each plot contains a manual frost tube (measured weekly in 2023-24 and daily in 2024-25), a soil respiration collar (for bi-weekly measurements of soil greenhouse gas fluxes), and an iButton for continuously measuring soil temperature. Bi-weekly sampling of carbon dioxide, methane, and nitrous oxide fluxes occurred from February through April of 2024, resumed in December of 2024, and will continue through April of 2025. Soil samples were collected from each plot in April of 2024, and will be collected again in April of 2025 to understand how differences in soil frost depth and duration impact soil microbial and biochemical properties and processes. Figure 5. Experimental design for soil frost manipulation experiment. Results from the first year of our experiment indicate that treatments successfully manipulated soil frost. The average soil frost depth was higher with snow removal compared to control and insulated treatments. Surprisingly, the number of freeze-thaw cycles was not different between treatments but was significantly higher in the pasture than in the deciduous forest. Pasture exhibited higher CO2 emissions than forest, but the soil frost manipulation treatments did not differ in greenhouse gas (GHG) emissions over the season. Data will be integrated with soil parameters (e.g., texture, density, and soil carbon content) to identify the drivers of soil frost and GHG emissions after the second year of the experiment. We presented these findings at the American Geophysical Union Annual Meeting in December of 2024. We have also had the opportunity to collaborate with a colleague at the Pacific Northwest National Lab (PNNL) at no additional cost to the project to gain additional insights into how soil freeze-thaw cycles can affect soil greenhouse gas emissions by altering soil pore structure. To do this, we collected intact soil cores in late November of 2024 from locations at our nearby field site and shipped these soils to PNNL, where our colleague and his team have conducted preliminary analyses and incubations. Activities that aligned with Objective 4, included training ten undergraduate students in field measurements of snow depth and soil frost (for the winters of 2023-24 and 2024-25). One undergraduate student has been actively involved in prototyping the wireless FroSen sensor and its field deployment. We have contributed to training two PhD students who have been developing our FroSen soil frost sensor prototype. We continue to mentor the postdoctoral researcher leading the efforts to address Objective 3. The project has continued to engage K-12 schools in Global Learning and Observations to Benefit the Environment (GLOBE, globe.gov) soil frost tube protocols. Participating classrooms have included elementary through high school grades who are participating in student-led research in outdoor classrooms, as well as middle and high school teachers in a professional development training in GLOBE protocols (Figure 6). Figure 6. Dr. Elizabeth Burakowski assists local middle and high school teachers in soil frost tube construction at a recent teacher professional development training hosted by NH Sea Grant at the University of New Hampshire. At Moharimet Elementary School in Durham, NH, first graders learned how to repair the frost tube that has been installed in their school's maple sugarbush since 2021 (Figure 7). Dr. Burakowski is also mentoring a high school student for the New Hampshire Science and Engineering Exposition (NHSEE) International Science and Engineering Fair (ISEF). As a mentor, Prof. Burakowski has taught the student how to construct frost tubes, assisted with installation at the student's local field site and has been providing guidance on data analysis and interpretation. Figure 7. Dr. Elizabeth Burakowski teaches 1st graders at Moharimet Elementary School in Durham, NH, how to repair their CRELL-Gandahl soil frost tube.
Publications
|
Progress 02/15/23 to 02/14/24
Outputs
Target Audience:Target audiences reached include K-12 teachers and their students engaged in education and outreach, the scientific community, and the broader public. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?Opportunities for training and professional development include training seven undergraduate students in field measurements of snow depth and soil frost (for the winters of 2022-23 and 2023-24) and an additional three students in soil sampling and processing (to examine soil physicochemical drivers of frost depth and duration) during the current reporting period. Two undergraduate students assisted with installing our soil frost manipulation experiment and collecting samples for our planned lab incubations in the fall of 2023. We have also contributed to training two PhD students throughout the previous reporting period, one of whom has been developing our FroSen soil frost sensor prototype and the other who has been devising methods to use ground-penetrating radar mounted on UAVs to spatially detect frost depth. In September of 2024, we welcomed a postdoctoral researcher onto the project team who has been leading the efforts to address Objective 3 and understand the ecological consequences of changing soil frost regimes. Beyond the university environment, the project has provided training and professional development of five K-12 teachers and their students who have been collaborating in making manual and sensor-based measurements of snow depth, soil frost, and near-surface microclimate in outdoor classrooms in New Hampshire and Maine. How have the results been disseminated to communities of interest?We disseminated our research findings to scientific audiences in a presentation at the American Geophysical Annual Meeting. Beyond educational outreach described above, we also communicated our results to the K-12 community of teachers and students in a recorded interview describing our research about changing winters and shifting seasonality posted on YouTube as part of the Global Learning and Observation to Benefit the Environment North America Phenology Campaign. Two members of the project team, Contosta and Campbell, also spoke about our research results in an invited public talk at the Mount Washington Observatory Science in the Mountains seminar series. Our research findings were also featured on our local ABC affiliate, WMUR, as part of their Forecasting the Future series describing how climate change affects New Hampshire. What do you plan to do during the next reporting period to accomplish the goals?During the next reporting period, we will continue to address our research objectives. For Objective 1, we will evaluate data collected from the FroSen sensors during the winter of 2003-24. We will also compile and analyze associated data, including air temperature, relative humidity and snow depth. In the next reporting period we will have results from the machine learning analyses and will use this information to prepare a manuscript for submission to a peer-reviewed journal. We will continue GPR development necessary to produce research quality data. We anticipate modest enhancements to the technology because it is performing as desired. More effort is needed to finalize the GPR radiograms processing workflow and to extract quantitative, spatially distributed soil frost depths and snow properties. There is limited recent research, especially when incorporating the UAS component. For Objective 2, we will continue to evaluate drivers of soil frost across the landscape by completing and submitting our manuscript examining the timing, duration, severity, and spatial heterogeneity of soil frost across two winters and three land cover types. We will also finish analyzing soil samples for total carbon and nitrogen and will statistically compare soil physicochemical properties such as soil texture and carbon with maximum observed frost depth. We will evaluate the overlap between measured soil variables and those estimated from the POLARIS spatial data product that we used to establish plots for monitoring soil frost across a gradient of soil conditions. We plan to summarize the results of this analysis in a publication to be submitted in a peer-reviewed journal. Regarding Objective 3, we plan to continue exploring the ecological consequences of soil frost through the field experiment we established in the winter of 2023-24 and in the lab incubation that we will carry out in the spring of 2024. We aim to present our findings at the American Geophysical Union Annual Meeting in December of 2024. Finally, we will continue to address Objective 4, training undergraduate students in field measurements of snow depth and soil frost, lab-based soil frost sensor development, and field and lab-based soil sampling and analysis. Our project will also include ongoing training of one PhD student developing new automated soil frost sensor prototypes, one PhD student developing ground-penetrating radar techniques for detecting soil frost across the landscape, and one postdoctoral researcher examining the ecological consequences of changing soil frost regimes. In addition, we will continue our engagement with K-12 schools, ranging from elementary through high school, in student-led research in outdoor classrooms centered around measurements and monitoring of changing winters.
Impacts
What was accomplished under these goals?
For Objective 1, we field deployed and tested FroSen sensors, experimenting with two different prototypes that detect changes in pressure due to the phase change of water in a sealed tube.The tubes are buried in the ground and the ice that forms in them is consistent with the frozen soil that surrounds them.The first prototype measures changes in air pressure above the frozen portion of the tube.The second has a pressure transducer at the bottom of the tube that measures changes in water pressure in the unfrozen portion.Both have been tested in a freezer prior to field testing. Data also collection includes air temperature, relative humidity, and snow depth. We are manually collecting soil frost depth data adjacent to our FroSen sensors and are using machine learning methods to develop algorithms that ingest pressure, temperature, humidity, and snow depth data to estimate frost depth. We have installed vertical soil temperature arrays to detect soil temperatures throughout the soil profile to provide data for interpolating the frost line and serve as an independent measure of frost depth. A conference paper describing development, testing, and deploying of FroSen sensors to date has been accepted for the upcoming IEEE International Conference on Computing, Networking and Communication (ICNC). Our efforts to use GPR in frost detection involved establishing two, 120-m, N-S transects with 22 frost tubes per transect. Manual measurements of soil frost occurred twice per week, and GPR measurements occurred 18 times throughout the winter along these transects. Two airborne, UAS-flown GPR flights were conducted throughout the winter with variable snow conditions. The GPR unit was able to capture variation in snow depth, suggesting its potential to detect soil frost when present. The difficulty in maintaining a maximum 1-m flight altitude (FCC regulations) and in georeferencing the location of the GPR unit inspired improvements for the 2023-24 field season: enhanced terrain following, a high-precision GPS unit, and flying the UAV in a spiral pattern that eliminates abrupt turns during which data are not collected. We have relocated GPR sampling efforts to the University of New Hampshire's Kingman Farm and have established two grids, one in a cleared forest opening and the other in an adjacent open field. Three of the four non-baseline flights to date identified that frost was present in the soil, and we are working to process data collected to date and schedule our next set of flights. For Objective 2, we continued to characterize the spatiotemporal dynamics of soil frost across the landscape by conducting a second field season measuring the depth and duration of soil frost at the Thompson Farm Ecological Observatory in Durham, NH, USA. We conducted weekly measurements of snow depth and soil frost at nine locations within each of 15 plots. We are in the process of writing a manuscript summarizing our results. To establish how soil physicochemical characteristics affect the depth and duration of soil frost, we also conducted a spatially explicit soil sampling. Soils were collected adjacent to frost tubes that were established during the fall of 2021, with cores taken to 30 cm at three depth increments: 0-5 cm, 5-15 cm, and 15-30 cm. These increments were chosen to correspond with the depths used in the POLARIS dataset that we used to inform our experimental design. Soils were weighed, sieved (2 mm), and analyzed for moisture, bulk density, texture, and pH. We are preparing the samples for total carbon and nitrogen analysis. This extensive sampling will enable us to compare hypothesized drivers of soil frost with maximum soil frost measurements and to compare ground-based measurements with spatial data products such as POLARIS that are important for scaling. Project activities that address Objective 3 include establishing a manipulative field experiment and collecting samples for a laboratory incubation, both of which will examine the ecological consequences of changing soil frost regimes. We established six new experimental plots that overlap with are where we had previously monitored soil frost and snow depth. These six plots were selected to capture the range of soil textural classes and organic matter content present at the site, and thus align with our other efforts to understand drivers of soil frost across the landscape. Each of the six plots included four treatments that modify soil frost depth and duration by manipulating ground insulation: snow removal (no insulation), control (insulation present from ambient snow cover), one layer of insulation (mimics moderate snow cover), and two layers insulation (mimics deep snow cover). Insulated plots are covered with bags filled with styrofoam beads to allow for water and gas exchange. Each plot contains a manual frost tube (measured weekly), a soil respiration collar (for bi-weekly measurements of soil greenhouse gas fluxes), and an iButton for continuously measuring soil temperature. The experiment was installed in Nov and Dec 2023, and bi-weekly sampling of carbon dioxide, methane, and nitrous oxide fluxes will begin in Feb 2024 and continue through the growing season (May-Oct of 2024). Soil samples were collected from each plot at the time of experiment establishment for an lab incubation that will modify the intensity and duration of soil frost and assess the same soil biogeochemical responses being measured in the field. Activities that aligned with Objective 4 included training seven undergraduate students in field measurements of snow depth and soil frost (for the winters of 2022-23 and 2023-24) and an additional three students in soil sampling and processing (to examine soil physicochemical drivers of frost depth and duration) during the current reporting period. Two undergraduate students assisted with installing our soil frost manipulation experiment and collecting samples for our planned lab incubations. We have also contributed to training two PhD students, one developing our FroSen soil frost sensor prototype and the other devising methods to use ground-penetrating radar mounted on UAVs to spatially detect frost depth. Most recently, a postdoctoral researcher joined the project team who has been leading the efforts to address Objective 3 and understand the ecological consequences of changing soil frost regimes. The project has continued to engage three schools in student-led research in outdoor classrooms. At Moharimet Elementary School, both first and fourth graders were involved in collecting snow depth, soil frost depth, sap flow, and vegetation green-up data for the second year in a row, with first graders measuring physical variables of snow depth and soil frost and fourth graders monitoring the biological phenomena of sap flow and vegetation phenology. Project personnel visited Medomak High School in Waldoboro, ME, where they worked with students and teachers to install an environmental monitoring station alongside frost tubes that were deployed in 2021. They also measured and mapped all the trees in the plot. Students engaged in Medomak are in an alternative education program that helps them to engage in science outside of a traditional classroom setting. Engagement at Old Town High School in Old Town, ME consisted of continuing collaboration to measure snow depth, soil frost, and vegetation phenology at several sites that differ in forest composition near the school.
Publications
|
Progress 02/15/22 to 02/14/23
Outputs
Target Audience:Target audiences reached include K-12 teachers and their students engaged in education and outreach, undergraduate students, graduate students and the broader public. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?Education and outreach activities also addressed Objective 4 of the project, to train the next generation of scientists and stakeholders in the theory and practice of understanding the local and global impacts of changing winters. During the reporting period, the project trained six undergraduate students in field measurements of snow depth and soil frost. One of the students used our field experiment at Thompson Farm as the basis for her senior capstone project. In addition, two PhD students were involved in developing new automated soil frost sensor prototypes, and one PhD student was engaged in both analyzing data collected during the 2021-22 field campaign to detect spatiotemporal variation of soil frost across the landscape and developing ground-penetrating radar techniques for detecting soil frost. Some of the unanswered questions from our 2021-22 field season inspired the PI to conduct a semester-long research project with her introductory soils course to investigate the soil physico-chemical drivers of soil frost. This course featured 58 undergraduate students who visited Thompson Farm in the fall 2022 semester to delineate soil profiles at three soil pits, one in each plant function type at the site (open pasture, deciduous forest, and coniferous forest). Students compared the presence, depth, color, and texture of the soils they observed on the ground to to NRCS soil survey and found some discrepancies between mapped soil series and their observations. Students also collected surface soils in the area surrounding the pit, brought them back to the UNH campus, and analyzed the samples for moisture, texture, pH, inorganic N, micro- and macro-nutrients, and organic matter. The soil physico-chemical measurements they made did not align with the values provided by the gridded POLARIS dataset that we used for establishing our plots along a gradient of organic matter content and drainage. However, student measurements of soil clay content and soil organic matter content were positively correlated with maximum frost depths recorded at the site during the winter of 2021-22. The data and insights that the students in the introductory soil science course generated inspired a soil sampling campaign that we will undertake during the next reporting period. In addition to undergraduate and graduate training, the project has engaged four schools, ranging from elementary through high school, in student-led research in outdoor classrooms. These include Old Town High School, Old Town, ME, Waldoboro High School, Waldoboro, ME, Bishop Guertin High School, Nashua, NH, and Moharimet Elementary School, Madbury, NH. Each of these schools collected about 30-50 measurements during the winter of 2021-22 field season. Moharimet Elementary School offered an especially unique opportunity to train students across a range of ages, as both the first and fourth graders were involved in collecting snow depth, soil frost depth, sap flow, and vegetation green-up data, with first graders measuring physical variables of snow depth and soil frost and fourth graders monitoring the biological phenomena of sap flow and vegetation phenology. Because measurements were made adjacent to the school's sugar house, students were encouraged to think about how conditions during winter and spring affect traditional food resources such as maple syrup. Through the Global Learning and Observations to Benefit the Environment (GLOBE; globe.gov) program, elementary school teachers and classrooms in Alaska and Moharimet were able to connect online to exchange data findings on soil frost, snow, and phenology. Education activities also included training a cohort of middle and high school teachers from Vermont and Maine during a week-long professional development workshop in July of 2022 to make frost tubes. Teachers constructed tubes at the workshop that they took home as science souvenirs, along with a kit for constructing tubes with their students in the ensuing school year. How have the results been disseminated to communities of interest?In addition to education and outreach activities described elsewhere in the report, our ongoing research has been communicated to the broader public through both print and radio media. Print media consisted of a story in the Valley News, March, 2022, that explored how soil frost may have contributed to the historic mud season experienced last spring (https://www.vnews.com/Mud-season-45600850). Two radio broadcasts, one on New Hampshire Public Radio, March 21, 2022 and the other on Vermont Public Radio, March 30, 2022, also featured descriptions of the project and its relevance for understanding the impacts of changing winter climate on infrastructure (https://www.vpr.org/show/vermont-edition/2022-03-30/mud-season-is-bad-this-unh-professor-says-climate-change-could-make-it-worse What do you plan to do during the next reporting period to accomplish the goals?We will continue to address Objectives 1, 2, and 4. For Objective 1, we will deploy prototypes of our wireless, in situ sensors at our local field site at Thompson Farm in the same locations as our manual soil frost measurements for direct comparison. We will also be testing the GPR unit from Geophysical Survey Systems Inc. along three transects in the open pasture at Thompson Farm that vary in drainage class. Of interest is the ability of the unit to detect frost beneath the snowpack. Regarding Objective 2, we will conduct a second field season to measure snow depth and soil frost to better understand variability of soil frost across the landscape while also supporting activities that address Objective 1 (i.e., comparing manual measurements of frost depth with estimates from prototype in situ sensors). We also aim to conduct a spatially explicit soil sampling across all of our plots at Thompson Farm. In late fall of 2022, soils will be collected adjacent to frost tubes that were established during the fall of 2021, with cores taken to 30 cm at three depth increments: 0-5 cm, 5-15 cm, and 15-30 cm. These increments were chosen to correspond with the depths at which the POLARIS dataset, which we used to inform our experimental design, reports soil physicochemical variables. Soils will be weighed, sieved (2 mm), and analyzed for moisture, bulk density, texture, pH, organic matter, and total carbon and nitrogen. This extensive sampling will enable us to compare hypothesized drivers of soil frost with maximum soil frost measurements. It will also allow us to compare our ground-based measurements with spatial data products such as POLARIS that are important for scaling. For education and outreach (Objective 4), we will continue to collaborate with partner K-12 schools, to train undergraduate and graduate students, and to communicate findings to broad audiences.
Impacts
What was accomplished under these goals?
Activities during this reporting period focused on addressing Objectives 1, 2, and 4. As part of Objective 1, the development of wireless, in situ sensors largely consisted of testing and refining the frost tube design devised during the previous reporting period. This design consists of a pressure sensor installed within a rigid clear pipe capped on both ends and housed within an outer tube made of PVC. The theory behind the design is that the pressure sensor can detect soil frost because the pressure in the tube increases as liquid water becomes ice and expands. Initial experiments, which were conducted in freezer mesocosms, indicated that pressure initially displays a linear increase with decreasing temperatures. Following the initial expansion with freezing, the water within the tube seems to contract before expanding again, with pressure in the tube mirroring this pattern. Follow up experiments have focused on quantifying the predictability of pressure fluctuations as water expands and contracts in the tube, correlating these fluctuations with manual measurements of frost depth, and training a non-linear, machine-learning model that can both simulate fluctuating pressure conditions while also ingesting other environmental variables for accurately measuring frost depth. We are tracking additional variables using our "iBUG" system, which is a microprocessing system developed in-house that tracks temperature, relative humidity, light levels and carbon dioxide concentrations just above the soil surface at hourly intervals. The iBUG performs signal processing at the hardware level, including machine-learning algorithms that can correct measurements of carbon dioxide concentrations with simultaneous temperature measurements. Our test deployment of the iBUG on the UNH campus during the winter of 2021-2022 was promising and elucidated the need to optimize power consumption by programming a sleep cycle into the system. Alongside soil frost sensor development, one of the graduate students on the project has been conducting a review of soil frost sampling strategies for submission in a peer-reviewed journal to help place sensor development activities within a broader context. Other work over the most recent reporting period that aligned with Objective 1 centered around piloting ground penetrating radar systems for detecting soil frost. This work progressed from proof-of-concept bench-top measurements of the dielectric constant using soils collected from a local field site to in situ experiments during the summer of 2022 that prepared us for the winter 2022-23 field season. Bench-top measurements were made by constructing small troughs of frozen, wet, and dry soils within a walk-in freezer to conceptualize the GPR signatures of each soil condition. Additionally, soils of all three conditions were measured for their dielectric constants to detail which sensor properties were adept for replication throughout the winter field season. In situ experiments occurred at a private property in Newmarket, NH, USA and focused on determining whether the GPR unit from Geophysical Survey Systems Inc. (GSSI) could detect differences in radiometric signatures between the soil matrix and both frozen soil and metal plates buried beneath the surface. In addition, we have devised an unpiloted aerial system (UAS) for flying our GPR unit ~1m above the soil surface. This airborne platform was tested for safety and ease of maneuverability at the Thompson Farm Ecological Observatory in Durham, NH, USA throughout August 2023. To characterize the spatiotemporal changes of soil frost across the landscape (Objective 2), we conducted a spatially-explicit study of soil frost depth and duration throughout the 2021-22 field season. Our research took place at the Thompson Farm Ecological Observatory in Durham, NH, USA (43.107°N, 70.949°W). This site has a wide range of physical and biological features that may drive variability in soil frost, making it an excellent model system for our work. It has three distinct plant functional types, open pasture, deciduous broadleaf forest, and deciduous needleleaf forest. These plant functional types have distinct microclimates, with air temperatures often showing greater diurnal variation in open pastures as compared to forests. The plant function types at Thompson Farm also influence snowpack and development and melt in different ways. For example, evergreen needleleaf trees intercept more falling snow than deciduous broadleaf trees, resulting in a shallower snowpack in evergreen-dominated forest stands. The soils present at the site represent a variety of textural classes, from loamy sands to silt loams. They also vary in organic matter content, with forested areas having higher amounts of soil organic matter due to the presence of an organic horizon. The combination of texture, organic matter, and bulk density can also influence saturated hydraulic conductivity (Ksat), which, because of its relationship with soil moisture, may be related to the thermal conductivity of soils. Soil frost and its hypothesized drivers were measured at 15 plots at Thompson Farm that capture the range of plant functional types and soil variables at the site. In the fall of 2021, five, 9 m2 plots were established within each plant functional type (open pasture, deciduous broadleaf, and coniferous needle leaf). Areas were selected within each plant functional type that represent the median, interquartile ranges, and 95% confidence intervals of Ksat and soil organic matter as determined from the probabilistic mapping of SSURGO data product, POLARIS (Chaney et al., 2019). Each plot was separated into nine, 1 m × 1 m quadrants, with a Gandahl frost tube installed to 30 cm depth in each, for a total of 135 frost tubes. Plots were sampled weekly from early December 2021 through late March 2021. Seasonal soil frost development commenced in early-January, experienced maximums in mid-February, and continued to persist until late-March after rapidly melting in mid-March. The open pasture land cover class displayed the highest maximum soil frost depth at 36.0 cm. The deciduous broadleaf and needleleaf coniferous plots displayed maximum depths of 25.9 cm and 29.4 cm, respectively. Soil frost depths rapidly increased from early-January to maximum soil frost depths in mid-February due to a combination of consistent below-freezing temperatures and limited to no snow cover to insulate soils. Following maximum soil frost depths in mid-February, soil frost depths remained nearly constant due to consistently cold temperatures sustaining frost conditions. Statistically significant differences in soil frost were observed across sampling dates within each land cover type and between the two forested land covers and the open pasture class, wherein the open pasture consistently displayed higher soil frost depths from early-January until thawing in mid-March. Soil frost measurements were not correlated with physical or biological drivers such as surface soil organic matter, soil drainage, and in-situ snow depth. The lack of correlations between frost depth and soil characteristics may have been due to the soil data themselves. The POLARIS product from which we obtained the soil organic matter and texture layers may not represent conditions on the ground given the spatial scale at which this data product was developed versus the scale at which measurements were collected.
Publications
|
Progress 02/15/21 to 02/14/22
Outputs
Target Audience:Target audiences reached include K-12 teachers and their students engaged in education and outreach. The research was also communicated to a broad public audience through an interview on New Hampshire Public Radio broadcast on April 21, 2021: https://www.nhpr.org/post/new-hampshire-warms-unh-studies-effects-more-freeze-thaw-cycles#stream/0 Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?Other activities during Year 1 of the project also aligned with Objective 4, training the next generation of scientists and stakeholders in the theory and practice of understanding the local and global impacts of changing winters. During the reporting period, the project trained three undergraduate students in field measurements of snow depth and soil frost. Two PhD students were involved in developing new automated soil frost sensor prototypes, and one PhD student, who began his graduate program at the end of the reporting period, started focusing on the development of ground-penetrating radar techniques for detecting soil frost across the landscape. In addition to undergraduate and graduate training, the project has engaged four schools, ranging from elementary through high school, in student-led research in outdoor classrooms. These include Old Town High School, Old Town, ME, Waldoboro High School, Waldoboro, ME, Bishop Guertin High School, Nashua, NH, and Moharimet Elementary School, Madbury, NH. We met with teachers from each of these schools to plan the winter 2021-2022 field season of making manual measurements of snow depth and soil frost. Other education activities included developing open-source, low-cost sensor arrays for deploying at each school to complement manual data collection. How have the results been disseminated to communities of interest?We disseminated some of our initial research findings to broad audiences through an interview on New Hampshire Public Radio broadcast on April 21, 2021: https://www.nhpr.org/post/new-hampshire-warms-unh-studies-effects-more-freeze-thaw-cycles#stream/0 What do you plan to do during the next reporting period to accomplish the goals?During the next reporting period, we will continue to address Objective 1, focusing on the development of several soil frost sensing methods that we plan to initially test in the lab and then deploy in the field for comparison with manual soil frost measurements. We will also test different strategies for measuring soil frost using GPR systems at our local field site in Durham NH. At the same time, we aim to conduct a field campaign that tracks spatiotemporal changes in soil frost across the landscape, which addresses Objective 2. Again, we will leverage our local field site at Thompson Farm, Durham NH, as it features a wide range of physical and biological features that may drive variability in soil frost, making it an excellent model system for our work. We will also continue to address Objective 4, training undergraduate students, both in field measurements of snow depth and soil frost and lab-based soil frost sensor development. Our project will include ongoing training of one PhD student developing new automated soil frost sensor prototypes, and one PhD student developing ground-penetrating radar techniques for detecting soil frost across the landscape. In addition, we will continue our engagement with four schools, ranging from elementary through high school, in student-led research in outdoor classrooms centered around measurements and monitoring of changing winters.
Impacts
What was accomplished under these goals?
Most of our activities during the first reporting period addressed Objective 1, particularly the development of in situ sensors for making spatiotemporally continuous measurements of soil frost. During the first part of Year 1, we focused on field testing prototype FroSen soil frost sensors at three locations within the study region: the UNH Thompson Farm Observatory, Durham, NH, the Hubbard Brook Experimental Forest, North Woodstock, NH, and the Sleeper's River Research Watershed, Danville, Vermont. Three prototype FroSen sensors were deployed at each site in November of 2020 in tandem with three traditional frost tubes to facilitate comparison between automated and manual measurement techniques. Each site also included a soil temperature sensor array, with sensors installed at 1, 5, 15, 30, and 50 cm depths for determining the isoline of soil frost and comparing this isoline with both automated and manual measurements of soil frost depth. The FroSen protype consisted of a dye-filled tube that was similar to the traditional Gandahl-CRRELL frost tube. There was an LED light at the bottom of the tube that was intermittently turned on and off with a microcontroller. A color sensor measured the intensity of the light, which was proposed to be commensurate with the amount of ice that formed in the tube (i.e., as the ice forms, the dye would become more concentrated). At all three field sites, the colorimetric signal declined over time, even as manual frost measurements and soil temperature measurements indicated soil freezing. To troubleshoot the decline of the colorimetric signal over time, we conducted a freezer study during the summer of 2021. We constructed new FroSen (automated) and Gandahl (manual) frost sensors and installed them in an insulated soil-filled box (0.6 cm x 0.6 m x 1 m) housed in a walk-in freezer. Thermocouples were installed at the five soil depths used in the field study (1, 5, 15, 30 and 50 cm) , and air temperature was monitored within the freezer. We conducted multiple soil freezing experiments, including tests for: 1) determining if the manual frost tubes performed as expected (i.e., froze from the top down); 2) comparing manual tubes with two different dyes (fluorescein vs. methylene blue); 3) at room temperature, testing the sensitivity of the FroSen color sensor with dyes of different concentration to mimic freezing; 4) determining if the dye was stable with exposure to LED over time by replacing the dye solution; 5) determining if the LED was stable over time by measuring light intensity without the dye solution; 6) determining if the battery was stable over time by replacing batteries; 7) comparing frost depths measured with the FroSen vs. manual tubes and thermistors. Although we were successful in developing a system for freezing soil that mimicked field conditions and will be useful for future sensor tests, the optical sensor did not perform as expected and showed the same declining signal in the laboratory as it did in the field. We were able to rule out some potential problems; however we ultimately decided to pursue more promising soil frost sensing methods. One of our new candidate sensors uses a 25.4mm clear (rigid) pipe (inner tube), capped on both ends, that sits inside a 50mm PVC pipe (outer tube). The top cap on the inner tube consists of a pressure sensor and sealed with silicone. Our theory is, when liquid water is cooled, it contracts until a temperature of approximately 4 degrees Celsius is reached. After that, it expands slightly until it reaches the freezing point, and then when it freezes it expands by approximately 9%. This results in increases in pressure inside the rigid pipe (inner tube). We conducted an experiment to capture this increased pressure and form a mathematical relationship with the depth of soil freezing. After a few calibrations, we achieved promising results during the lab tests, however, we still need to do more testing (lab and field) for a concrete conclusion.
Publications
|
|