Integrate NEON & GLEON Data into your Classroom using Macrosystems EDDIE!

1. Integrate NEON & GLEON Data into your Classroomusing Macrosystems EDDIE! Cayelan C. Carey Associate Professor Virginia Tech @CareyLab, cayelan@vt.edu

2. Agenda for today • Introductions • Overview of Macrosystems EDDIE and Project EDDIE • Pedagogical framework • Module descriptions • Demo of a shortened version of a Macrosystems EDDIE module • Cross-Scale Interactions: Explore how climate and land use change may interact to affect the water quality of your favorite GLEON or NEON lake! • Lessons learned & tricks of the trade in the classroom • Next steps for Macrosystems EDDIE • Discussion & exploration of modules on your laptops

3. Ecology is now a ‘sensored’ science How can we prepare undergraduate students to understand and use sensor/big data in ecology? Image: denniskneale.com

4. A challenge: Undergraduates need quantitative reasoning skills to analyze large datasets and tackle ecological problems, yet… Image: letsgraph.com

5. Project EDDIE: Environmental Data-Driven Inquiry & Exploration Project EDDIE EDDIE: Environmental Data 2013-2017 EDDIE: Earth & Ecosystems 2018 - ongoing EDDIE: Macrosystems 2017-ongoing http://projectEDDIE.org

6. http://ProjectEDDIE.org

8. Pre-packaged, ready to use EDDIE Module = • Instructor lesson plan & PowerPoint • Pre-class readings • In-class activities & datasets • Homework & answers Overall learning objectives: • Build data manipulation and analysis skills using real, messy ecological data • Use large datasets to improve ecological understanding • Develop graphing and statistics skills using Excel Photo: Grace Hong Photo: Cayelan Carey http://ProjectEDDIE.org

9. Macrosystems EDDIE: teaching local to continental-scale ecology Overall objectives: • Develop & test hypotheses about complex effects of global change on lakes • Run simulation ecosystem models • Learn basic computer programming in R http://MacrosystemsEDDIE.org

10. Macrosystems EDDIE modules • Teleconnections • Climate Change Effects on Lake Temperatures • Macro-Scale Feedbacks • Cross-Scale Interactions

11. Macrosystems EDDIE approach Flexible & adaptable • A-B-C structure • Plug and play Tools we use • R + RStudio • Simulation modeling • Sensor data

12. Learning outcomes in four dimensions (4DEE) Core Ecology Concepts • Ecosystem structure, productivity, and nutrient cycling Cross-Cutting Themes • Systems thinking:local to continental interactions • Complex interactions occurring across multiple spatial and temporal scales Human Dimensions • Human-accelerated climate and land use change Science Practices • Computer skills; R programming, modeling & simulation • Ecological inquiry; evaluating claims and communicating ecology

13. Climate Change Effects on Lake Temperatures • How can simulation models help us understand local vs. regional drivers of ecosystem processes? • Students design and test a climate change scenario to understand how lake thermal structure may change • Students use distributed computing tools to scale up and run hundreds of climate scenarios for their lake to detect tipping points Photo: Wikimedia Commons Carey & Gougis 2017 J Sci Education & Tech

14. Cross-Scale Interactions • How do local and regional processes interact to affect water quality in lakes? • Students set up a lake model for a GLEON or NEON lake of their choice and force it with climate, land use, and interactive scenarios. • Students test hypotheses about how local (nutrients) and regional (air temperature) drivers interact to affect phytoplankton blooms in different lakes Photo: NOAA Great Lakes ERL

15. Teleconnections • How do connections to distant systems affect local ecosystem dynamics? • Students run ecosystem models to simulate water temperatures and ice cover in multiple NEON and GLEON lakes • Students compare lake responses to an El Niño scenario and predict how lakes may respond to changes in the intensity of global meteorological phenomena Photo: Wikimedia Commons

16. Macro-Scale Feedbacks • How do local and regional processes amplify each other? • Students choose a GLEON or NEON lake, run their lake models to simulate CO2 and CH4 fluxes to the atmosphere, and calculate if the lake is a carbon sink or source • Students explore how lake greenhouse gas fluxes may both respond to and amplify changing climate Photo: Cayelan Carey

17. Continental datasets allow students to model lakes across ecoregions GLEON.org engr.wisc.edu

18. Modules build computational literacy • >500 students and ~50 instructors from ~20 universities have completed pre/post-module assessments • Modules increased self-reported proficiency and confidence using R software and ecosystem modeling Simulation modeling R software Farrell & Carey, 2018 Ecology & Evolution

19. Modules reduce performance gap • Largest post-module gains seen in students who initially reported lowest level of knowledge • Significantly greater understanding of ecological concepts • Evidence of greater “systems thinking” skills after completing one module Farrell & Carey 2018, Ecology & Evolution Carey & Gougis 2017, J Sci Education & Tech

20. Long-term goal Excitement!

21. So, how do you get ready to EDDIE? • Let’s walk through what teaching a module might look like in your classroom • We’ll go through the module as if we’re teaching it to a classroom but at an accelerated pace and show the instructor materials • Please ask if you have any questions!

22. Let's dive in to a module! http://module2.MacrosystemsEDDIE.org

23. Pre-packaged Teaching Materials http://module2.MacrosystemsEDDIE.org

24. Ready, Set, EDDIE! • Overview of pre-module technology preparation • General background about the lake model used in the modules • Developed for instructors with minimal R experience

25. Ready, Set, EDDIE! • General background about the ecosystem model used in the modules • File types included • Example model output

26. Instructor Manual • Key elements: • Pedagogical background • Learning objectives • Module overview & workflow • PowerPoint slide notes (also embedded in the slides!) • "Common stumbling blocks" for each activity (& how to work through them!) • Answer key with model output examples for each activity

27. Instructor PowerPoint

28. Getting Started "Bonus Slides" • Optional to include in PPT (or in a previous class) • Provide step-by-step guidance on downloading & unpacking module files • Orient students to RStudio • Provide detailed troubleshooting for common challenges

29. R You Ready for EDDIE? • File to walk students through: • R + RStudio installation & setup • Installing R packages needed for the module • Downloading & unzipping files to run the module

30. Student Handout • Guides students through module activities • Questions to elicit hypotheses, compare model output to predictions

31. Files for Running the Module

32. Files for Running the Module

33. Files for Running the Module If students ignore ##!! indicators that they need to change something in the script (e.g., if they're running the code quickly without reading! ), they will get an error message

34. Files for Running the Module • In Activity B and C, students have to modify the model glm2.nml file for their climate and land use scenarios It's important that they double-check that they've saved their modified glm2.nml files before they re-run the model!

35. Files for Running the Module • Students generate multiple heat map and line plots to directly compare scenarios at the end of Activity C! • (Note that the color scale can differ between model runs and lakes!) Baseline Land use + climate

36. Files for Running the Module • At the end of the module, have student teams share with their scenarios, hypotheses, and model outputs with the rest of the class Photo: K.J. Farrell • We provide some potential questions in the instructor's manual to use as a starting point for class discussion

37. Do you want to teach aMacrosystems EDDIE module?Lessons learned & tricks of the trade

39. Pitfall 1: Intimidation by computational tools • Programming has a steep learning curve Image: https://norcalbiostat.github.io/MATH130/01_intro.html

40. Pitfall 1: Intimidation by computational tools • Programming has a steep learning curve • BUT many students recognize the importance of knowing how to program– they just need help getting started Most Least Data from Farrell & Carey 2018, Ecology & Evolution

41. Solutions: • Module activities assume no prior knowledge of R or programming • Students modify and run ready-to-use scripts and models • Modules break down complex activities into short, do-able chunks of code to reinforce developing skills • Use of real-world tasks makes programming relevant • "I had very limited computer modeling experience prior to this activity. This was my first time truly modeling an ecosystem.” • Patience and understanding is really important – the risk of failure seems very high

42. Faculty tester feedback: "The module was really useful in engaging the students in an exercise using R without overwhelming them with details about how to actually write code." “I think that it is cool to introduce R to students in this way. It is much less intimidating than a completely blank slate and also gives students a taste for what is possible in R and any programming language.”

43. Pitfall 2: "Digital natives"? • Despite ready access to technology tools, students' individual computing experiences vary dramatically • "[my students] had essentially no coding (or, in some cases, computer) experience before attempting the module" Image via: https://www.comaround.com

44. Solutions: • Work with a partner to equalize experience levels • Near-peer helpers  students who finish a section early can help fellow classmates • Use worksheets and discussion questions to check-in and keep students engaged with the module materials • Walk around classroom to touch base on students’ progress http://module2.macrosystemseddie.org

45. Student & faculty feedback: "I enjoyed getting to fill in the worksheet along with using R so that way we had to understand the material, and not just plug in the code." “My students had almost no experience with R or Excel... The modules were very thoroughly prepared and with my experience as a regular R user, I was able to follow the workflow easily.”

46. Pitfall 3: Trouble with troubleshooting • Many students lack experience troubleshooting software issues • "App-ification" of day-to-day computing tools tends to hide error messages and underlying source code https://xkcd.com/1024/

47. Solution: • Detailed step-by-step troubleshooting, with screenshots from Windows and Mac operating systems, help students resolve problems on their own • Use of an on-campus computer lab may help streamline troubleshooting, as instructors only have to juggle one type of operating system and version of R/Rstudio

48. Faculty tester feedback: “Students were able to follow the instructions in the R script and complete the module. We used a computer lab with R and RStudio preloaded which helped facilitate using the module.” “One group crashed the module towards the end… but we didn't have time to solve the issue. This type of thing is bound to happen and is awesome that it only happened to one pair! In the future, I'll have clean files on a flash drive to quickly replace.”

49. Pitfall 4: Are you also intimidated? • Instructors may lack experience and comfort with advanced computational tools, like R • Hard to troubleshoot if you don't know what you're doing! Photo: C.C. Carey Photo: C.C. Carey

50. Solution: • Prepared assuming no prior faculty experience with ecosystem modeling and very limited exposure to R • Recommend faculty do a full run-through of module on their own so they know what students will encounter! • Complete modules are ready-to-use 'as is' or modified by faculty • Pre-packaged lesson plan for instructors • Pre-class readings • In-class activities & datasets • Homework & answers • Testing with faculty from range of institutions and experience levels • Catch areas that need more detailed explanations