There are 8 activities in this module. Each activity will require 1-2 class periods (assuming 50-minute class periods) to complete.
This module introduces two types of data sources: data derived from observations, and data produced by models. Students will explore simulated marine environments using a computer-based virtual reality (VR) data visualization tool called VES-V (The Virtual Ecosystem Scenario Viewer). VES-V displays marine environments as though one was SCUBA diving in those habitats. The numbers and types of fish and other aquatic creatures displayed in different locations in the VES-V simulations are based on actual abundances of marine creatures.
Some data used in VES-V are the output of scientific computer models, while other data comes from actual observations made in each specific environment. Students will learn about differences and similarities between these two data sources, how to recognize each, and how each type of data is used by scientists to make predictions.
Students will first learn how models are used to study weather and climate, then investigate model-based datasets of marine species populations using the VES-V software. Students will create a pair of extremely simple models to estimate temperature values in places or times where/when observational data is not available. They will then use these simple temperature models to learn why and how scientists validate and improve models by comparing data from models with observational data.
Students will also explore the methods biologists, ecologists and other scientists use to estimate the sizes and compositions of populations of living creatures by observing limited samples from those populations. Students will investigate these methods in potentially familiar environments, including a strawberry patch in a garden and birds at a feeder, before applying the concepts to marine habitats.
Finally, students will apply the concepts of observations, models, model validation, and population sampling to marine environment datasets presented in VES-V. Students will gather data from VES-V about two organisms that people eat, tuna and lobsters, and estimate future population sizes for each. Students will explore the question: “Can the data from models and observations help us predict future population sizes of fish and other marine organisms?”.
Modern techniques and technologies make fishing so efficient that it is possible for humans to severely deplete populations of marine organisms. People must therefore monitor and manage populations of marine creatures to ensure that they are sustainable. This usually involves placing upper limits on the number of fish or other marine organisms allowed to be caught each year. Government agencies, such as NOAA Fisheries, strive to set catch limits that balance peoples’ desire for fish as a food source with the need to keep fish stocks above critical thresholds that ensure sustainable populations.
Fishery management experts need data about fish populations to assess the size and health of those populations, so they can set appropriately balanced catch limits. It is impossible to count all the fish in the sea, so fishery scientists use different methods to estimate populations of marine organisms. Two major sources of data about fish populations are data derived from observations, and data generated by scientific models.
Observational data is gathered by directly observing marine organisms at specific times and places. Data collection might involve towing a net behind a ship, SCUBA diving, using underwater cameras or robotic submarines. In each case, scientists note the numbers and types of organisms captured in a net or otherwise observed. The number of observations are limited; it isn’t possible to monitor every place in the ocean 24 hours a day, 365 days a year. Scientists must make assumptions based on a limited number of observations about populations in places or at times where direct observations are not available. Also, scientists must be careful to make sure their sampling techniques don’t introduce biases. If a net with a large mesh size is used to capture samples, smaller organisms might slip through it and fail to be counted. Observations made near the surface could fail to detect a large population of bottom-dwelling species. Schools of fish can be highly localized; observations near a large school might detect thousands of fish, but the same type of observations made less than a mile away might not detect any. Scientists must use clever sampling techniques and be cautious when extrapolating observations to other places and times, to minimize sampling biases and errors. Sometimes observations of the types and numbers of fish caught by fishing boats can supplement other data. However, such fishing data generally has a strong bias driven by the types of nets used and where in the water column those nets are placed.
Observational data helps scientists create computer models of expected populations. If we want to know the abundance of a specific species - like fish, or another marine species, we can look at factors such as water temperature, availability of food, and the prevalence of predators. All of these might affect a species’ population size. By comparing these factors with past observations, scientists can create models that may predict current and future populations. Data from models can fill in the gaps from observations, providing “educated guesses” about population sizes in places where, and at times when, no observations are available. As is the case with observations, the models are not perfect - at best, they provide reasonable estimates, but scientists must always use caution when interpreting data derived from models.
Observational data is the starting point for generating scientific models. Observations can also be used to test the quality of a model. Scientists validate, or test the accuracy, of a model by comparing it with observational data. For example, imagine we had observational data for tuna populations over the past thirty years. If we also had a model for tuna populations, we could use the conditions from 30 years ago as the starting input for the model, and then run the model for 30 years of simulated time. If the model’s prediction of tuna populations today was very close to the observational data, we would have high confidence in using that model to predict the future of tuna populations. If, however, the model’s prediction of current tuna populations was quite different from observations, our model would need to be refined and improved before we could trust its predictions. Scientists use this validation process to test the quality of models, and to make changes and improvements to those models.
If we want to better understand the populations of organisms, we need a somewhat accurate estimate of their sizes and what affects them. Both observations and models have limitations in their ability to produce accurate counts of fish and other marine organisms which are needed if we want to sustainably manage their populations.
Using observational data and data generated by models, scientists can get the best possible estimate of actual populations, which helps fishery managers decide how large of a catch they should allow to maintain a sustainable population.
Predictions about weather and climate also rely heavily on a combination of observations and models. This lesson includes weather-related examples, which may be more intuitively familiar to students, as a comparison to the approaches used with fisheries data.
Introduction: Introduce the Observations and Models Module with a Big Question: “How do Individuals, communities or governments ensure that there is enough seafood for people to eat in the future?”