Assignment 4: Conservation priorities using multicriteria decision analysis

ESM 263 (Frew), Winter 2018

Due at 8am on Monday, 19 March 2018 (hard deadline! we need time to grade it!)

NOTE: We weren't kidding. This is a REALLY complicated assignment! Read it through, CAREFULLY, all the way to the end, before you start working on it!


  1. Implement a multicriteria analysis (MCA), and produce a model and results data.
  2. Parameterize the model
  3. Design a map to communicate your results.
  4. Submit your work to course website.


This assignment is worth 20 points—5 points each for your map and data, and 10 points for your model.


Once upon a time, a Bren School group project (Boland et al. 2005) developed a multicriteria analysis (MCA) framework to evaluate conservation potential of land in Ventura County based on viewshed, habitat, and public access criteria. In this assignment, we will use a simplified version of their approach to evaluate the conservation potential of lands in the southern region of Santa Barbara County.

Your task is to advise a land trust on which watersheds have the best overall conservation potential. The land trust wants to conduct conservation efforts in watersheds that would help preserve both riparian habitat and scenic viewsheds through conservation easement purchases on inexpensive, but developable land. You will use an MCA framework to produce your results.

Available data

The data for this assignment can be downloaded as HW4.7z.

The following data come from a variety of sources, and are available in the HW4.gdb geodatabase. Except where indicated, we've projected these data into the Teale Albers (NAD 1983) projection and clipped them to our Region of Interest (ROI), which is a collection of watersheds in the southern region of Santa Barbara County.


Normalized Difference Vegetation Index (NDVI)

In Landsat imagery, Band 3 is red, Band 4 is near-infrared (NIR), and band values are 8-bit unsigned integers (0 to 255). Once the band values have been converted to floating-point numbers (0.0 to 255.0), the NDVI formula is:

(NIR − red) ÷ (NIR + red)

which yields a floating-point number between −1.0 and 1.0.

Task 1:  Implement MCA

Use a raster-based index model, or "scoring" approach to perform an MCA that will prioritize watersheds based on your client's criteria.

All MCAs follow four basic steps:

  1. identify criteria
  2. standardize factors
  3. assign weights
  4. combine

Your task to develop an MCA model that will implement each of these steps (Figures 1 & 2).

Figure 1: A conceptual model for the complete analysis. Your task is to replace each of these "Run" tools with the geoprocessing workflow you develop for that step. To simplify your task, please use a single model for all of your workflows.

Figure 2: Example workflow for scoring the watersheds in the "Run MCA" step.

The output of your model will be a Scores table that has a score for each watershed. The last step in Figure 2 is to use the Zonal Statistics by Table tool to generate this table using CALWNUM as the unique key for the watersheds. (You can use the output table's MEAN field for the watershed's score.)


Before you start your analysis, download the HW4.7z file and setup your project with a project folder (H:\ESM263\HW4) and a map document (HW4.mxd). The HW4.7z file includes, in addition to the usual geodatabase, a Python script ( and a toolbox with a pre-loaded link to that script. Note that you must have the actual Python script and the link in your toolbox to properly use this tool. In the HW4 toolbox, create a new model in which you'll perform all of your analyses.

When working with your model in ModelBuilder, remember to mark all of your intermediate (green) outputs to be Managed (not Model Parameter).

Set your model properties as follows:

Recommended tools

When designing the geoprocessing workflow for each step of your analysis, we recommend using these ArcToolbox tools (note: steps may use tools more than once):

  1. Watersheds rasterization: Polygon to Raster
  2. NDVI calculation: Divide, Float, Minus, and Plus
  3. Riparian land constraint: Euclidean Distance, Extract by Mask, and Reclassify
  4. Priority viewshed factor: Reclassify and Viewshed
  5. Developable land factor: Divide, Extract by Mask, Polygon to Raster, Reclassify, Select, Slope, and Times
  6. Scoring the watersheds: Calculate Quantiles, Reclassify, Weighted Sum, Zonal Statistics, and Zonal Statistics as Table

See "Homework 4 notes" for more tips about specific tools.

Task 2: Create a model interface

Create a user interface for your model by designating model environment settings, input datasets, output table, reclass tables, and score weights as model parameters. Specifically, set the following as model parameters (variable type in parentheses):

* Asterisked parameters will need to be set by creating new variables (see: Setting model parameters). Non asterisked parameters can be set by simply right-clicking an object already in your model and setting to Model Parameter.

Task 3: Map

Design a map that communicates your results, and clearly identifies your prioritization of watersheds. Also, include a brief (50–100 words max) description of your analysis on your map.

Task 4: Turn in your work

You will upload three files: your map, your model, and your results data:

  1. Map: Export your map as a PNG image (use options of 300dpi and 8-bit color palette), and name it HW4Lastname_map.png
  2. Model with user interface: Organize and name your processes and data to clearly describe your model. Make sure model parameters are clearly labeled in the user interface and that your model works from that run dialogue. Also, ensure that your model properties and current/scratch workspaces are set as specified and the only model dependencies are to the default and HW4 geodatabases (otherwise we won't be able to run your model locally). Rename your toolbox HW4Lastname.tbx and make sure it contains a single, run-ready model.
  3. Data: Put your results from the Scores table into a dBASE table with only two fields:
    Name Type Description
    CALWNUM Text CalWater watershed identification code (Interagency)
    SCORE Float watershed score from your model, where 1 ≤ score ≤ 5

    To do this, use the Table to Table tool with these parameters:

  4. Upload your three files to the course website:

back to ESM 263 home