Grand Challenge 3:

What approaches are effective for students to understand various models (numerical and analytical) that are used for prediction and research in atmospheric, oceanic and climate sciences, including model limitations?

Rationale

The study of the atmospheric, oceanic, and terrestrial systems is based on models that are used for prediction and for the conceptual understanding of these complex systems (Figure 6). Knowledge of computer programming and advanced math is needed to create, validate or understand these models, making the field less accessible to the broad student population (Ledley et al., 2011; Hamilton, 2015; Hamilton et al., 2015). One possible approach to reduce the mystery in the 'black box' approach to computer models is through the use of simple, familiar models like flow charts, graphs, and pictures, and physical models, like sand tanks for groundwater flow (Harrison and Treagust, 2000; Schwartz et al., 2009).

×

Figure 6: This image shows the concept used in climate models. Each of the thousands of 3-dimensional grid cells can be represented by mathematical equations that describe the materials in it and the way energy moves through it. The advanced equations are based on the fundamental laws of physics, fluid motion, and chemistry. To "run" a model, scientists set the initial conditions (for instance, setting variables to represent the amount of greenhouse gases in the atmosphere) and have powerful computers solve the equations in each cell. Results from each grid cell are passed to neighboring cells, and the equations are solved again. Repeating the process through many time steps represents the passage of time. The complexity inherent in these models make them conceptually challenging for undergraduate students. Image source: NOAA.

Another challenge to the use of systems models in atmospheric science is the fact that uncertainty is inherent in them, yet education research shows that novices are not comfortable with uncertainty. This requires a simplification of the models to adapt them to the student population and the implementation of targeted approaches (e.g., Gold et al., 2015).

Unanticipated changes in the forcing functions of the system resulting from unpredictability of human behavior (Konikow, 1986) that commonly involve activities such as increased water use and land conversion further demands continuous upgrade and creation of new models (Oreskes, 2003). Therefore, time-to-time update in our modeling curriculum makes it challenging for students to grasp completely new materials.

Recommended Research Strategies

1. Two working groups (Cognitive - Spatial and Temporal Reasoning, and Cognitive - Problem-Solving, Quantitative Reasoning, and Models) are focusing on the cognitive understanding of complex systems. Other DBER communities (like ecology) have conducted research in educational approaches that are effective for the understanding of models. The science education community has studied extensively how best to teach students about models (Gilbert, 2011) and we can apply what they have learned to weather and climate models. If the difficulty is related to understanding the concepts of deterministic vs. probabilistic models, perhaps research in statistics education can provide valuable information. We recommend that education researchers refer to contributions of these groups to identify research paths for the fluid Earth community.
2. An important aspect of teaching models is to be able to minimize, or even eliminate, the widespread skepticism students have about outcomes of the models. We recommend expanding research on learning impacts of various models that can be broadly divided into two groups: i) models that have their validation index reported or that can be validated with existing data, and ii) models that lack validation measures. What is the learning impact of one vs. the other group within the realm of weather and climate models?
3. Research students' attitudes towards models and modeling, and the efficacy of different approaches to stimulating students' interest to learn about models. For example, one could show and test the use of models as decision-support tools in the context of resource management.

Show References

Hide