Lab Session:Conceptualization

Time Limit:45 minutes

There are four primary steps involved in model development: (1) conceptualization, (2) quantification, (3) evaluation, and (4) application. Over the next two days, your group will work through each of these steps to create a model for your restoration project. The following notes are a guide to help you move rapidly. Don’t get too wrapped up in meeting each step explicitly. Model development is a judgment-laden process, and there is no single right answer.

Each team should document their model development process. A recommended form of documentation is a powerpoint presentation with a few slides per model development step. This “documentation” can be a few key bullet points to convey assumptions, limitations, and team notes. This will, by necessity be a crude, unpolished product, but it can serve to key track of information (to avoid the “why did I do that” problem), quickly communicate status (which you will be required to do each day), and serve as a preliminary outline for more thorough documentation (e.g., a model manual).

Tasks for Your Group:

  • Develop a conceptual model of your ecosystem using the six model development steps:
  • State the model objectives.
  • Bound the system of interest.
  • Identify critical model components within the system of interest.
  • Articulate the relationships among the components of interest.
  • Represent the conceptual model.
  • Describe the expected pattern of model behavior.
  • Review and revise as needed.
  • Document your model for presentation to the larger group
  • Where did you derive this information?
  • What assumptions did you make?
  • Are there good references (literature, data, people, etc.) to support your model?
  • Iterate this process as time permits
  • A few notes and thoughts:
  • Did you start with a more complex model and then simplify it? Presenting both models can be a useful communication strategy.
  • Would it be useful to present the conceptual model in more than one format (e.g., box and arrow, pictures, etc.)?
  • Sometime it is helpful to develop a first model as individuals and then combine them
  • Who should be involved in the next round of model adaptation?
  • What model components should be added in the future?
  • How do restoration actions play out in the model? For instance, if restoration involves manipulating a hydrograph, where is the “flow” variable in the model?

Lab Session:Quantification

Time Limit:60 minutes

This task is a continuation of the model development exercise for your restoration projects. The overarching goal is to develop a quantitative model (i.e., a numerical tool) to distinguish the pros and cons of restoration alternatives at your site. This “tool” can be a few calculations on a napkin, a table of best professional judgments (e.g., each team member’s score on a scale of 0-100), or a simple Excel file. Remember, a first model may be VERY SIMPLE (e.g., a couple of variables estimated from professional judgment), so don’t get too bogged down in the details. Even a qualitative comparison of the alternatives (e.g., big, bigger, biggest) is a model. BUT, identify key areas for improvement (i.e., better data, mechanistic hydraulic model inputs, etc.) as the project progress.

Tasks for Your Group:

  • Translate the conceptual model to a quantitative model
  • If needed, simplify your previous conceptual model into a version that may be quantified. If it helps, feel free to iterate through the conceptual model development steps from the previous exercise.
  • Linking to the conceptual model: Use the conceptual model as a template. What type of model most satisfies the model objectives and represents the conceptual model?
  • Choice of model complexity: What levels of complexity are appropriate for the applications under consideration? Does the model need to simulate multiple types of processes/interactions (e.g., hydrologic, hydraulic, ecological)? What levels of complexity is needed to distinguish between alternatives?
  • Selecting the general quantitative structure of the model: time unit (length of analysis, time step), spatial scale (project, watershed, regional), tools to implement this structure
  • Does the proposed structure adequately represent the conceptual model or were additional assumptions and/or simplifications made?
  • Identifying functional forms of model equations: How does x relate to y? What is the shape of the response curve? What equations meet this conceptual response curve? For each process in the model, what analytical equations or formulae will be applied? What literature or data support these equations (or response shapes)?
  • Estimating the parameters of the model equations: Are data available? Are professional judgments required? What are the “book ends” of parameter estimates? What are reasonable minimum and maximum parameter estimates? Are there physical or ecological impossibilities in parameters (e.g., maximum brood size)?
  • Executing the baseline model: Does the model produce an output?
  • Documentation and communication (Develop presentation for larger group).
  • What did you do?
  • What assumptions were made (e.g., functional form, parameters)?
  • What limitations exist (e.g., ecosystem type, depth, etc)?
  • What are key areas for model improvement?

Lab Session:Evaluation

Time Limit:45 minutes

Tasks for Your Group:

Evaluate your model using the following techniques.

  • Model testing:
  • Do the outputs of the model make sense (e.g., Does depth go up when discharge increases?)
  • Play with your model. What are the inputs to the model? Are the outputs more or less sensitive to a particular parameter?
  • Are there any numerical errors in the model (i.e., Does 1+1=2?)
  • Document your test cases!
  • Model verification
  • Model validation
  • Model calibration
  • Sensitivity analysis
  • Uncertainty analysis
  • Model documentation:
  • How did you evaluate your model?
  • How certain are you the numerical outputs are correct?
  • What did you leave out of your conceptual model? Did the model become too simple?
  • Identify data gaps and research needs that may not have been obvious during conceptual model development

Lab Session:Application, Communication, and Iteration

Time Limit:45 minutes

Tasks for Your Group: Apply your model to assess the merits of restoration alternatives.

  • Model application
  • Identify at least 5 restoration alternatives (including “no action”). Can you apply the model to each alternative to estimate outputs? Can you present your results as an output matrix comparing alternatives (i.e., alternatives on the rows and different outputs as the columns, example below)?
  • Define scenarios of interest. Are there scenarios beyond the alternatives that you would like to consider (e.g., climate change, land use change, urban development, increased water withdrawal, etc.)?
  • Model Iteration: Refine the conceptual and/or quantitative model as needed.
  • What sort of documentation would you produce?
  • Apply model to each scenario (i.e., execute the model, conduct experiment).
  • Analyze and interpret results.
  • Scenario interpretation
  • Conduct a preliminary cost-effectiveness analysis?
  • Documentation: Continue to develop a presentation summarizing the modeling process by adding details on model application.

Alternative / Description / Cost ($K) / Ecological Output
Alt-0 / No action / 0
Alt-1 / Remove Dam-A / 500
Alt-2 / Remove Dam-B / 750
Alt-3 / Remove Dam-C / 250
Alt-4 / Remove A+B / 1,100
Alt-5 / Remove B+C / 900
Alt-6 / Remove A+C / 700
Alt-7 / Remove A+B+C / 1,250