Intro to learning with Computational Models

Sugat Dabholkar
Environmental Science, Self-directed
40-50 min
Middle School / High School
v1

Overview

This is an introductory lesson for using certain types of computational models designed using a software called NetLogo.

In this lesson, students will learn:

  • how to computationally study spread of wildfire

  • how to engage in the scientific inquiry practices of constructing knowledge in the context of an "emergent systems microworld" (ESM)

  • how to engage computational thinking practices in the context of an ESM. We will focus on four computational thinking practices, namely: data practices, modeling and simulation practices, computational problem solving practices, and systems thinking practices

Standards

Next Generation Science Standards
  • NGSS Crosscutting Concept
    • Systems
  • NGSS Practice
    • Analyzing Data
    • Constructing Explanations, Designing Solutions
    • Using Models
    • Arguing from Evidence
    • Conducting Investigations

Credits

Thanks to Kevin Hall and Connor Bain for their contributions to this Lesson.

Activities

  • 1. Using models to learn science
  • 2. A not-so-sneak peek into the code behind the model
  • 3. Systematically investigating spread of wildfire
  • 4. Constructing knowledge by engaging in scientific inquiry practices

Student Directions and Resources


Many lessons in CT-STEM curricula use computational models designed using an agent-based modeling language called NetLogo. In this lesson, we will understand what these models are and how to use them.

This lesson specifically focuses on learning science with such computational models. These computational models are of emergent natural phenomena. Emergent phenomena are the ones in which simple interactions between autonomous agents and environment result into complex patterns, for example, a flock of birds. 

                    

In case of a flock of bird, each bird follows simple rules regarding alignment, coherence and separation with neighboring birds and that results in formation of complex patterns of the flock, such as V-shape formation. 

Learning Goals:

  • In this lesson, we will use a NetLogo model about wildfire to learn about how to computationally study spread of wildfire.
  • We will learn how to engage in the scientific inquiry practices of constructing knowledge.
  • We will learn how to engage computational thinking practices. We will focus on four computational thinking practices, namely, data practices, modeling and simulation practices, computational problem solving practices, and systems thinking practices.

Let's get started!

1. Using models to learn science


Scientists use scientific modeling approaches to construct knowledge about the world. In this section, we will explore ideas about scientific models.

 


Question 1.1

It only became widely accepted knowledge that all matter in the world is made up of tiny elementary particles in the early 19th century.

Let's look at the the picture below. Can you tell what it is a model of?



Question 1.2

Some of you would say it's a model of an atom. Some would say it's a model of 'Neon atom', because it has 10 electrons. It could also be one of several ions, since we don't know the number of protons, correct?

Anyway, the point is that these representations in a model allow us to think about a model and various natural phenomena that are associated with the model in certain way. Can you explain what could this model be useful for?



Question 1.3

Let's look at a computational model of wildfire. Imagine that you are observing a forest using a drone that is hovering on top of a forest. So what you see in the model is the top view of a forest. Each patch represents a tree. A red patch represents a burning tree. 

Play with the model and make some observations.

Describe what could this model be useful for researchers or other people using it.



Question 1.4

Change density of trees in the model and observe spread of the fire.

In such models, trees are agents. Behavior of agents is programed in the model. 

An example of such rule would be - a tree would not move.

Can you guess, a rule that trees might follow in this model regarding catching fire?



Question 1.5

Based on your exploration of the model, can you guess how density of tree affects spread of the fire in the forest?



Question 1.6

This 'fire model' is an example of an emergent systems microworld. It is modeled in terms of interactions between the agents (trees) and it allows us to observe emergent patterns regarding the spread of fire in the forest. We can change parameters such as density of trees and study how that would affect the spread of fire in the forest. It also allows us to make some predictions regarding the spread of fire in a forest.

However, this model does not include all the factors that would affect spread of fire in a forest. Can you suggest some factors that can be added in this model?

 

2. A not-so-sneak peek into the code behind the model


Here's a model of wildfire that a team of researchers tried to modify. But it does not run as they expected. In fact, it broke, it does not run at all.

Can we help them fix it?


Question 2.1

Setup the model. What is the mistake (or what computer science people call 'bug') in the model?

 



Question 2.2

Did you observe that after you press 'setup', you see blue colored trees. Maybe the mistake in the code is about color of the tree. Maybe the color of the tree is set to 'blue' instead of 'green' by mistake.

Click on the blue bar that says 'NetLogo Code'. You can find it below the big square in the model.

Where does it say 'blue' in the code? And there does it say 'green' in the code?



Question 2.3

You can fix the bug! Go to a line that say 'blue' where it should have been 'green'. Change the code.

Click on "Recompile code" and run the model again. Does it work now?

Can you explain why it did not work before?



Question 2.4

These NetLogo computational models allow users to change the code and observe the effect of those changes.That is one very important feature of these Emergent Systems Microworlds, that you can play with those by changing the parameters such as 'density' or by changing the code as well.

Now try to read the code. Write a line in the code below and explain how it affects the way the model would behave. 

For example, 

set initial-trees count patches with [pcolor = green]

This line above, sets a value for a variable 'initial-tree' by counting the patches that have pcolor (patch-color). This line is written in NetLogo language that the NetLogo compiler understands.

NetLogo language is designed to be easy to understand for humans. Pick another line in the code and explain what it could mean.   

3. Systematically investigating spread of wildfire


Let's investigate how the density of the tree affect the spread of wildfire. 

We will first generate some data using the model and then represent it using another computational tool called CODAP.

Let's follow an experimental design that is described below. 

Research Question: How does density of tress in a forest affect spread of wildfire?

Hypothesis: As the density of trees in the forest increases, percentage of forest burned will increase linearly. That means, if density of trees doubles, the percentage of forest burned will also double.

Let's test our hypothesis.

Change the values of density systematically. Record the value of 'percentage forest burned' in the data table. Make sure that you press 'setup' button every time you run an experiment.

Run each experiment twice. Make sure you record values for each experimental trial. 

The software will plot average of the two values that you will record. 

  


Question 3.1

Describe your observations of the graph of 'density' vs 'percentage burned'.



Question 3.2

Do you think that evidence that we gathered with our experiment support our hypothesis?



Question 3.3

Explain your answer to the previous question.



Question 3.4

Spread of wildfire is an emergent phenomenon. Below certain density the fire does not spread much, however when the density crosses a 'tipping point' or threshold, the fire engulfs almost all the forest. 

Tipping point in this model falls within which of the following ranges?

  Between 30 and 40
  Between 40 and 50
  Between 50 and 60
  Between 60 and 70


Question 3.5

Can you give an example of another such phenomenon with a tipping point? 

4. Constructing knowledge by engaging in scientific inquiry practices



Question 4.1

Explore a more detailed version of the fire model above.

Explain what "probability-of-spread" might mean in the model and how it would affect the behavior of the model.



Question 4.2

Write a questions that is of an interest to you which can be answered using this model.

An example of such question would be, how does wind speed affect spread of fire?



Question 4.3

Based on your exploration of the model, guess an answer to your question and state it in the form of a testable statement (hypothesis) - something that you can test using the model.



Question 4.4

Design an experiment to test your hypothesis. Explain your design.



Question 4.5

Perform the experiment. Describe your observations and explain whether those support your hypothesis or not.