Statement_of_Work_2.pdf

Southwestern Natural Resource Management: Assessing vulnerability of vegetation and wildlife communities to post‐fire transformations

PERFORMANCE BASED STATEMENT OF WORK

I. GENERAL INFORMATION

A. Introduction:

The United States Geological Survey (USGS), Western Geographic Science Center (WGSC), has a requirement for work to be performed by a Data Scientist. The USGS Land Change Science project Remote Sensing for Resource Management ‐ Southwestern Natural Resource Management is broadly concerned with issues of wildfire, forest management and fire hazards in the transboundary Madrean Archipelago ecoregion, covering parts of Arizona and New Mexico in the US and Sonora and Chihuahua in Mexico, and the adjacent Northern Sierra Madre Occidental in Mexico and north to the Mogollon Plateau in Arizona. The current research objective is to 1) define ecosystem vulnerability to mega‐fires and post‐transformation at local and regional scales, and (2) link multi‐scale bird and vegetation responses to fire regimes across the vulnerability framework to assess factors and mechanisms that drive vulnerability and resilience. This contract builds upon research conducted in FY20 to FY22 and is in support of WGSC’s FY23‐24 research project funded by the USGS Southwest Climate Science Center.

B. Background:

Wildfires have and always will be an essential part of forested systems given most species have evolved with fire for millennia. In the southwestern U.S. and northwestern Mexico, forests are vital for economic growth, recreation, clean water, and many other ecosystem services. In the last three decades these forests have been impacted by mega‐fires that have fundamentally changed ecological functions and the structure and composition of communities. Therefore, it is vital to understand how both our actions and inactions contribute to the vulnerability and sustainability of forest ecosystems. To better manage our forests, the main goal of the project is to understand how recent fires have affected vegetation and animal communities by assessing mega‐fires in a historical and ecological context. Doing so will require models based on remote sensing and field data at different spatial scales, including information on vegetation and wildlife. By integrating these data using spatial and temporal models our project will not only develop new knowledge but also transfer this information to resource managers in both the U.S. and Mexico, including those from First Nations by collaboratively developing sets of best management practices so that information can be applied.

Mega‐disturbance events and ecological transformations are major concerns in southwestern forests and woodlands, but effective management strategies depend on understanding variability in responses of vegetation and wildlife to changing climate and shifts in fire regimes at multiple scales. Mega‐disturbances such as fires, fundamentally change forested landscapes, including changing the community type, structure and composition of vegetation. The degree and consequences of changes driven by mega‐disturbance events varies both spatially and temporally, depending on severity, patch size, spatial arrangement, and time since fire. Plant and animal species have evolved with fire, but the severity, extent, and timing of recent fires are occurring in a context of un‐naturally high fuel loads and shifting climates. It is critical to evaluate the effects of recent mega‐fires, from contemporary, historical and future perspectives. Bringing traditional and non‐traditional perspectives to the table from a bi‐ national region of high biodiversity is needed for landscape management that meets multiple goals, and identifying where it is possible to resist change, improve capacity for resilience, or facilitate post‐fire transitions. Our objectives are to:

1. Define ecosystem vulnerability to mega‐fires and post‐fire transformation at local and regional scales. The foundation of our approach is a vulnerability framework defined by fire regime characteristics across climatic environments and associated vegetation communities. Proposed work will build upon our recent studies of human‐altered fire regimes in the Madrean Sky Islands (MSI) of the U.S. and Mexico and extend our approach to a wider geographic area and range of climatic environments on the Mogollon Plateau (MP) and northern Sierra Madre Occidental (SMO).

2. Link multi‐scale bird and vegetation responses to fire regimes across the vulnerability framework to assess factors and mechanisms that drive vulnerability and resilience. We will integrate remote sensing and spatial modeling with field‐based data on vegetation and bird avian communities using data we have already gathered from a network of more than 1700 points in Mexico and the U.S., and expand sampling to cover a broader range of environments. Our framework will identify potential tipping points based on vulnerability factors and mechanisms to identify management actions.

II. WORK REQUIREMENTS

A. Technical requirements: Several technical issues must be considered to successfully complete the project goals within the larger context of the USGS Land Change Science/National Land Imaging/ Southwest Climate Science Center project.

Performance requires on‐site or cloud‐based capacity for computer‐intensive data processing; assembly of spatial datasets in georeferenced formats, and their use in model construction in conjunction with available field datasets; and provision of metadata in standard formats for all data used in the project. Analysis functions designed to complete project tasks must be tested for performance and reliability;

and delivery of computer code, georeferenced and tabular products must be delivered in accessible formats for use in commonly available GIS, spreadsheet, and statistical software.

B. Specific requirements: The project has specific technical and background knowledge requirements. Demonstrated experience working with large spatial data sets in combination with field data is necessary to completion of tasks. Experience in the use of generalized linear and flexible modeling techniques, multivariate statistics, as well as machine learning approaches for spatial modeling are prerequisite. Modeling experience will preferably be documented by published studies. The R Project free software environment will be used for statistical and spatial analysis; prior experience and proficiency in R programming is essential. Successful completion of the project will require a PhD with emphasis in ecology, fire ecology, and/or landscape ecology. Good communication and professional relationship skills will ensure successful participation on a research team.

C. Tasks and Deliverables Study design and database development: linking bird and vegetation communities with contemporary fire regimes

Task 1. Repeat characterization of fire regimes conducted previously (Villarreal et al.

2020), including additional vegetation types encountered in northern Sierra Madre Occidental and Mogollon Plateau. The analysis will use the latest available fire data for the region including Landsat‐based (Villarreal and Conrad 2020 or updated via USGS Burned Area algorithm); potential to use MODIS data will also be considered.

The fire regimes, derived from spatial fire history data, will be used in conjunction with tabular historical fire history summaries developed by team members.

A) Graphical and tabular summaries of fire characteristics for vegetation types, ownership classes and climate gradients.

B) Maps/spatial models of climate gradients, fire characteristics (times burned, severity, return interval), vegetation types and ownership classes.

Task 2. Construct a multidimensional environmental data space for selection of field sample plot locations that complement locations of existing field data. The primary axes of the data space will represent climate gradients (multivariate spatial climate and/or topographic microclimates). Identification of data gaps will consider location in the climate space, along with fire history, aiming for a broad range of climate and fire history characteristics.

Deliverables:

A) Geographical coordinates and spatial attributes in format for upload to field navigation system.

B) Archive of computer code with documentation of algorithm and selection criteria.

C) Archive of spatial data used in constructing environmental space.

Task 3. Develop candidate predictor variables to include in the analysis of bird and vegetation response to fire. The candidate variables will be identified through an analysis of modeled relationships between remote sensing indices and change in forest structure and composition through time.

Deliverables:

A) Write up of the process employed to identify candidate predictors, including data description (field and remote sensing indices), steps in the analysis and results.

B) Archive of data and code.

References

Villarreal, Miguel L and Caroline Conrad. 2020. Differenced Normalized Burn Ratio (dNBR) data of wildfires in the Sky Island Mountains of the southwestern US and northern Mexico from 2011‐2017. USGS Data Release https://doi.org/10.5066/P99S0I9W

Villarreal, Miguel L, José M Iniguez, Aaron D Flesch, Jamie S Sanderlin, Citlali Cortés Montaño, Caroline R Conrad, and Sandra L Haire. 2020. Contemporary Fire Regimes Provide a Critical Perspective on Restoration Needs in the Mexico‐United States Borderlands. Air, Soil and Water Research 13 (January): 1178622120969191. https://doi.org/10.1177/1178622120969191.