Project Waterhole

---

Downloads

• Browse source (trunk; corresponds to revision 303)
• Revision snapshots (in ZIP archives): 303, 302, 300, 250

Abstract

Analysis and 3D Visualization of Multiple Trophic Cascade Scenarios

Brian J. Krent, Joshua M. Cook, Matthew J. Knox, Randy R. Appleton* (rappleto@nmu.edu)
Northern Michigan University, Department of Mathematics and Computer Science
Marquette, MI 49855, United States

In this work we analyze the behaviors of multiple tropic cascade scenarios in simulated ecosystems. Extending upon previous work (cf. "3D Visualization of Simulated Trophic Cascades", Christopher E. Wells, 2008), we exhaustively search numerous subsets of the problem space by adding or removing different organisms, adjusting initial quantities of interacting agents, and altering several other environmental aspects. We accomplish this by refining and parallelizing the existing codebase for a cluster computer. Further, we examine the resulting data through extensive visualization.

*Advisor for 2009 Argonne Undergraduate Symposium

Tasks

1. Review over existing codebase (and possibly make tweaks, etc.)
2. Clearly specify the problem space(s)
   • Adding elements (more complex environments)
   • Removing elements (simpler environments)
3. Refactor into a parallelizable model
4. Run experiments on the cluster ([an attempt at] exhaustive search of each problem space)
5. Gather data and analyze (statical and other mathematical analysis)

Timeline and Log

Week	Dates	Tasks and Notes
Week 1	Sept. 27 – Oct. 3	Tasks: Get comfortable with breve and the steve language. Review over existing codebase (waterhole2.1.tz and exponentialgrowth.tz) and possibly make tweaks, etc. Identify parameters and re-factor code to read in a given simulation's parameter values from file. Notes: Standard input parameters (over 16) = { wander-range = (x,y,z), pond-size = (x,y,z), initial-population-size = (phytoplankton,zooplankton,hydra), initial-nutrient-level = (patch,phytoplankton,zooplankton,hydra), initial-death-clock = (phytoplankton,zooplankton,hydra), ... }
Week 2	Oct. 4 – Oct. 10	Some readings and miscellaneous reference material to go over this week: The breve Simulation Environment Documentation (Class reference) New documentation section (#1 above seems better than the terse stuff on spiderland wiki) Tutorials and Class Materials
Week 3	Oct. 11 – Oct. 17	Tasks: Read about plankton, hydra, and trophic dynamics. (Know scientific classification in biology.) Finish parameterizing and start doing runs on the cluster. Notes: Source code available from brevecluster.nmu.edu/projects/waterhole. To get the arguments (for use of setting the initial parameters for each process run on each cluster node), use the Control class's get-argument-count and get-argument at-index n (int) methods. We get the ID of the cluster node (the hostname) via the NetworkServer class's get-url method.
Week 4	Oct. 18 – Oct. 24	Tasks: Finish doing runs/collecting all data.
Week 5	Oct. 25 – Oct. 31	Tasks: Analysis on data and visualizations and start making presentation.
Week 6	Nov. 1 – Nov 7	Tasks: Nov. 2 deadline: submit initial presentation (PPT). Finish analysis and polish presentation.
Week 7	Nov. 8 – Deadline	Tasks: Presentation practice and review from Randy.