brglea [CODING CHALLENGE] Habitat Suitability

[github-actions]

Our changing climate is changing where key grassland species can live, and grassland management and restoration practices will need to take this into account.

In this project, you will create a habitat suitability model for Sorghastrum nutans (or a plant species of your choice), a grass native to North America. In the past 50 years, its range has moved northward. The model will be based on combining multiple data layers related to soil, topography, and climate. You will also demonstrate the coding skills covered in this class by creating a modular, reproducible workflow for the model.

You will create a reproducible scientific workflow

Your workflow should:

1. Define your study area: If you are using Sorghastrum nutans, you can download the USFS National Grassland Units and select at least 2 as study sites.
2. Fit a model: For each grassland:
   1. Download model variables as raster layers covering your study area envelope, including:
      • At least one soil variable from the POLARIS dataset
      • Elevation from the SRTM (available from the earthaccess API)
      • At least one climate variable from the MACAv2 THREDDS data server. Your project should compare two climate scenarios of your choice (e.g. different time periods, different emission scenarios). You can find a tutorial on how to access these climate data on earthdatascience.org
   2. Calculate at least one derived topographic variable (slope or aspect) to use in your model. You probably will wish to use the xarray-spatial library, which is available in the latest earth-analytics-python environment (but will need to be installed/updated if you are working on your own machine). Note that calculated slope may not be correct if you are using a CRS with units of degrees; you should re-project into a projected coordinate system with units of meters, such as the appropriate UTM Zone.
   3. Harmonize your data - make sure that the grids for each of your layers match up. Check out the ds.rio.reproject_match() method from rioxarray.
   4. Build your model. You can use any model you wish, so long as you explain your choice. However, if you are not sure what to do, we recommend building a fuzzy logic model (see below).
3. Present your results in at least one figure for each grassland/climate scenario combination.

If you are unsure about which model to use, we recommend using a fuzzy logic model

To train a fuzzy logic habitat suitability model:

1. Research S. nutans, and find out what optimal values are for each variable you are using (e.g. soil pH, slope, and current climatological annual precipitation).
2. For each digital number in each raster, assign a value from 0 to 1 for how close that grid square is to the optimum range (1=optimal, 0=incompatible).
3. Combine your layers by multiplying them together. This will give you a single suitability number for each square. Check out this article about raster math for more info.
4. Optionally, you may apply a threshold to make the most suitable areas pop on your map.

You will be evaluated on your code AND how you present your results

I will use the following rubric:

Description	Maximum Points
GITHUB REPOSITORY	30
Project is stored on GitHub	3
The repository has a README that introduces the project	5
The README also explains how to run the code	5
The README has a DOI badge at the top	5
The repository has a LICENSE	2
Repository is organized and there are not multiple versions of the same file in the repository	5
Repository files have machine and human-readable names	5
CODE	120
The code runs all the way through using the instructions from the README	10
The code follows the PEP-8 style standard	10
The code is well-documented with comments	10
The code uses functions and/or loops to be DRY and modular	10
Any functions have numpy-style docstrings	10
The code makes use of conditionals to cache data and/or computations, making efficient use of computing resources	10
The code contains a site map for the US National Grassland(s) used (1 ugrad, 2+ grad)	10
For each grassland (ugrad 1, grad 2+), the code downloads at least model variables as raster layers: soil, elevation, and climate (ugrad 1, grad 2 scenarios)	10
The code correctly calculates a derived topographic variable	10
The code harmonizes the raster data	10
For each climate scenario (1 ugrad, 2+ grad), the code builds a habitat suitability model	10
For each grassland/climate scenario combination, the code produces at least one (sub)figure displaying the results	10
Any unfinished components have detailed pseudocode or a flow diagram explaining how they could be finished in the future, and or a complete bug report explaining the problem	up to 90 points, in place of other categories
WRITTEN ANALYSIS	50
The notebook contains a project description	10
The notebook contains a researched site description	10
The notebook contains a data description and citation for each data source	10
The notebook contains a model description	10
The notebook contains a headline and description for each figure	10

Keep your eyes out for videos!

I won’t release a full demo of this, but you will have videos on writing pseudocode, accessing data sources, and any tricky problems that come up.