The Todd-Brown Lab is pleased to present our first Research Symposium. This year we will be sharing ongoing research on a variety of computational biogeochemical topics including: a land carbon stock estimation for regions in northern Canada, a land carbon model with integrated fire component, a new GUI and documentation for a coastal carbon model, best practices for compiling research driven soil data products, and a summary of soil data product structures.
When: August 20, 2020, 5-7PM UTC (1-3PM US-Eastern time)
Location: Register here to get the zoom link https://forms.gle/VVC1AtPCrjRQPWJGA
Boreal forest projects
Introducing land carbon stock of proposed protected areas within the Northwest Territories of Canada. Quentin Ashley, Kathe Todd-Brown
For current and future protected lands, it is important to understand all ecological and anthropological features of the area to better inform management decisions. One of these important ecological features is land carbon stock. Quantifying land carbon stock and the effect wildfires have on the release of this carbon is important in understanding climate change. Climate change can be tied to the amount of carbon released into the atmosphere and information on land carbon stock can help slow the atmospheric accumulation of carbon. The Northwest Territories project has been focused on estimating this land carbon stock and how wildfires can affect it. In order to communicate the full scope and impact of land carbon stock in proposed protected areas within the Northwest Territories, we need to understand why land carbon stock matters and why these lands were proposed for protection.
In order to provide context about the proposed protected lands, we utilized a comprehensive review of various publicly available documents that included: Canadian legislature, Canadian constitutions, Canadian treaties, Indigenous legislation, and other working group reports written by various governing bodies. This revealed that Indigenous Governments and Communities in the Northwest Territories were instrumental in lobbying and providing motivation for protecting these lands. These motivations include historical, cultural/spiritual, educational, and ecological reasons. In an effort to address Indigenous Groups’ wishes, the Government of the Northwest Territories and the Government of Canada are working with Indigenous Governments and Communities to protect these lands at the territorial and/or federal level.
Uncertainty Analysis for Soil Carbon Stock Estimates. Ignacio Sastre, Emily Miller, Kathe Todd-Brown
Quantifying variation and uncertainty is critical to risk assessment and decision making. Here, we provide estimates of the soil organic carbon stock for several proposed protected areas in the Northwest Territory of Canada. The proposed protected areas contain significant wetland regions that have the potential to contain high amounts of soil carbon stocks. In this analysis, we visualize the sources of uncertainty in soil carbon stock estimates for the proposed protected areas and develop a set of equations as a first step to quantify the relative impact of different drivers of variation and uncertainty.
In this talk, we will link field data to area estimates. Starting with loss on ignition, organic carbon fraction, and bulk density quantifications at the layer level, we gap-fill missing values of organic carbon and calculate soil organic carbon density for each layer. Then we aggregate the layers to calculate soil organic carbon area-density for each core. Cores are then aggregated across a transect and further aggregated across land cover class. At each of these steps, there is a fundamental question of uncertainty and variation to be addressed. The relative amounts of uncertainty associated with each of the sources mentioned above are still in discussion, and it is expected to have more accurate and precise results once these sources of error are found in its mathematical form.
Vegetation carbon stock estimates and uncertainties. Emily Miller, Ignacio Sastre, Kathe Todd-Brown
Vegetation carbon stock in the boreal forest ecosystems will become increasingly important with the rise of climate change and more frequent forest fire events. The estimation of the carbon stock in specified proposed protected areas in the Northwest Territories in Canada will inform the policies of the Government of the Northwest Territories, both in terms of economic pursuits and approval of new governmentally protected areas. Previous work by collaborators measured diameter at breast height across N transects in the Northwest Territories. These measurements have been fed into allometric equations to generate aboveground biomass estimates.
In this study, we upscale these biomass estimates to proposed protected areas using land cover class maps with particular attention to variation and uncertainty. In all calculations, there is error due to variations in experimental data, missing data points, estimations, and other aspects. It is important to identify and correct the error to ensure informed land management decisions. Here, we first discuss uncertainties across different allometric equations to extrapolate diameter at breast height to aboveground biomass. Then, we discuss propagating transect aboveground vegetation area-density averages to land cover classification. Finally, we discuss extrapolating both the vegetation biomass and the soil carbon stocks for the land cover classes across the proposed protected areas.
Future works will build off of this by accounting for uncertainty in the land cover classification assignment.
Visualizing the Grow Burn Decay Model. Emily Hetherington, Raichel Gulde, Kathe Todd-Brown
Fire is a critical controller of the land carbon cycle in the Northwest Territories. To examine the effects between fire and land carbon stocks, the Todd-Brown Lab developed the Grow Burn Decay model. Unlike previous models, this proposed model examines the effect of pyrogenic carbon and decomposition in addition to the land carbon lost during a burn event. For this new model, the Growth-Decay and Burn diagrams are essential to communicating the processes represented in the codebase. On the larger scale, the diagrams outline how carbon flows through the ecosystem and illustrate the effects of burn events on the defined carbon pools. This in turn identifies specific parameters that are necessary for the model and how those parameters fit into the overall system.
In this talk, we will present a new set of model diagrams and walk through the governing processes of the Grow Burn Decay model. The relative simplicity of these diagrams not only provides clarity to those working on the project but are used to convey our research clearly to those interested, regardless of their familiarity on the topic.
Literature synthesis on carbon-related variables in the Grow Burn Decay Model for the boreal forest. Raichel Gulde, Emily Hetherington, Emily Miller, Kathe Todd-Brown
The Grow Burn Decay model represents how carbon moves through an ecosystem where fire suppresses both decomposition and reburn potential of pyrogenic carbon, while also releasing carbon dioxide with each burn event. To accurately represent these processes, we need realistic values for the net primary production inputs during fire and ‘mature’ years, fire recovery times, carbon pool turnover times, and burn loss rates. In this talk, we will present the suitable ranges for the variables derived from collaborator field surveys and various literature sources. This parameter research is a key component for developing better models for land management decisions concerning the proposed protected areas of the Northwest Territories.
Marsh carbon modeling
Increasing Functionality for the Marsh Equilibrium – Cohort Theory Model: Jessica Hicks, Farah Aryan, Kathe Todd-Brown
Coastal marshes provide numerous benefits to society, such as water quality improvement through nutrient filtration, coastal protection from storms and erosion, carbon sequestration, and habitats for fish and wildlife. Rising sea-levels threaten to inundate these important coastal ecosystems. While these marshes can sometimes keep up, it is not clear when this will be the case. The Marsh Equilibrium Model was created to try to explore this relationship between sea level rise and marsh accretion rates.
The Marsh Equilibrium Model (MEM) and the Cohort Theory Model (CTM) are computational models that quantify marsh accretion rates, or changes in surface elevation, by simulating the addition of inorganic surface sediment and below-ground organic matter to the soil column. One key feature of these models are soil cohorts, which represent parcels of annual mineral deposits. These soil cohorts are tracked as they pass through the soil column where organics are added via root turnover and removed via decay. In this project, we reviewed and refactored a recent implementation of the MEM and CTM models in R, created a graphic user interface, created conceptual diagrams, and generated equations to better explain how MEM/CTM works.
From this work, we hope to have made the model more understandable and accessible to scientists and the public as well as laid the groundwork for future model sensitivity studies.
Soil informatics
Creating common vocabularies: harmonizing soil data for global scientific insight. Kristen Frederick, Lillian Heran, Marisa Younger, Kathe Todd-Brown
Soil carbon is a large natural source of carbon dioxide. As global warming continues, fueled by the anthropogenic release of this greenhouse gas, soils are expected to release more carbon dioxide, causing a vicious cycle. A number of datasets exist that contain soil carbon data that will allow for a better understanding of how quickly this process is occurring and may change in the future. A main frustration with all these data efforts is the lack of common vocabulary leading to inordinate amounts of time and effort spent matching variable names across data providers.
For this project, we are describing the soil data structure across nine data products to create a prioritized list of candidate variables that would broadly describe soil biogeochemistry data. We matched the column names within them, reducing the number of reported variables from 2,734 unique columns to 1,016 referenced variables.
This work creates the basis for a common soil biogeochemical data model structure laying the foundation for future data harmonization. Additionally, it creates the groundwork for future ontologies that can inform machine learning, creating a common control vocabulary within the soil community.
Current and best practices for data harmonization: harmonizing soil data for global scientific insight. Lillian Heran, Kristen Frederick, Marisa Younger, Kathe Todd-Brown
Soil data harmonization projects need an organized workflow in order to merge datasets from various sources to a common data model and a format that encourages usage of the data product. A workflow is an organized series of activities that guide a project from start to finish. A soil data harmonization project, for example, will most likely include the following major steps in its workflow: collecting and identifying the project data, merging the project data points into a common data format, filtering data for quality assurance and control, and generating a data product.
We are conducting one on one interviews with the principal investigators of nine current soil harmonization projects in order to access current data practices and recommend best practices for future soil data syntheses. The initial results highlight the value of good people management, a formalized project workflow, and clear goals driving the creation of the data product. From these interviews, most groups used manual data entry into a project Excel template then used a scripted QA/QC to deliver a harmonized data product. A common frustration across projects was that manual data transcription is error prone, time consuming, and occasionally not reproducible.
A paper is being drafted on best practices to individual research and harmonization projects for the soils research community.
Community perception of soil data products and harmonization: harmonizing soil data for global scientific insight. Marisa Younger, Lillian Heran, Kristen Frederick, Kathe Todd-Brown
Meta-analysis is widely used in the scientific community. Most researchers will carry out a meta-analysis aggregating other people’s data at least once in their careers to either contextualize their study site or draw broad global conclusions. Very few people believe this process is easy or come away from the experience confident and satisfied with their work.
We are conducting a community survey of soil scientists to determine current practices and frustrations with conducting meta-analysis. The survey consists of eight questions asking the members of the community how they compile and harmonize data products, what some of their frustrations were, and what they thought could be done better in the future. While other factors did come into play, we found the lack of a control vocabulary and diversity in data models to be the most frequent issues.
This survey provides a starting assessment of the current state of meta-analysis in the soils community. Future work will draw on informatics best practices and make suggestions for how soil scientists can leverage work like ontologies to make these studies more organized and extendable in the future.