Virtual Prairie

From BOINC Projects
Revision as of 15:15, 26 July 2026 by Al Piskun (talk | contribs) (add links)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search











Virtual Prairie
Project
StatusCompleted
CategoryEcology
ComputeCPU
RequiresNone
Development
DeveloperUniversity of Houston Department of Computer Science; University of Rennes 1 (UMR CNRS ECOBIO)
AuthorMarc Garbey, Cendrine Mony
SponsorAgence Nationale de la Recherche (ANR), CNRS, Cemagref, INRIA
Initial release2008
Completed2011
Metadata
Websitehttp://vcsc.cs.uh.edu/virtual-prairie/ (defunct)

Virtual Prairie (also referred to as ViP) was a completed BOINC volunteer computing project that studied the ecology of prairie and grassland ecosystems through large-scale individual-based simulation of clonal plant growth.[1] The project was jointly led by Professor Marc Garbey of the University of Houston Department of Computer Science and Assistant Professor Cendrine Mony of the University of Rennes 1, and combined high-performance computing with BOINC-based volunteer computing to model how agricultural and land-management practices affect plant competition and genetic structure within prairies.[2]

The project's goal was to help design new grassland systems that balance water purification, biodiversity conservation, and carbon storage.[2] At the time it was described by its organizers as the world's largest ecological simulation.[2]

Prairie grassland ecosystems, like this restored tallgrass prairie in Illinois, provide ecological services such as water purification, biodiversity support, and carbon storage that the Virtual Prairie project sought to help optimize.

Background

Environmental policy developments of the 2000s, including the European Common Agricultural Policy reforms of 2005 and France's Grenelle Environment Forum of 2007, focused new attention on the ecological services provided by grassland ecosystems, particularly the availability of unpolluted fresh water and the regulation of carbon emissions.[1] Natural prairies had historically been studied mainly for their agricultural productivity, but researchers increasingly sought to understand and reproduce their broader ecological functions, including nitrate pollution remediation, erosion prevention, and habitat stability for vulnerable species.[3]

Virtual Prairie approached this problem computationally: since a prairie is a complex system with emergent properties arising from the interaction of many individual plants, the project's researchers built an individual-based model (IBM), sometimes described as an agent-based model, to simulate the growth of clonal plants at the level of individual ramets (the repeated modular units of a clonal plant network) and to test which combinations of plant traits and management strategies produced the best ecological outcomes.[4]

Research design and computational methods

Individual-based clonal plant model

The core simulation, referred to in publications as the CLONAL model, represented a clonal plant as a network of interconnected ramets that could share and store resources.[4] The model incorporated nine growth rules governing metabolic processes, plant architecture, and resource sharing and storage, controlled by nineteen input parameters.[5]

Rather than testing only a small, expert-selected subset of parameter combinations, as was typical in earlier ecological modelling literature, the project's researchers chose two to four candidate values for each of the nineteen parameters and tested essentially all resulting combinations, an approach only made tractable through volunteer computing:[5]

N=∏i=119ki≈2×107

where ki is the number of candidate values tested for parameter i. Sources close to the project reported the resulting simulation count slightly differently depending on the run being described, with the February 2009 project press release citing more than 22 million completed simulations in the first phase of the project[2] and the project's 2011 journal paper describing roughly 20 million parameter combinations tested overall.[5]

Two-phase simulation strategy

The project ran in two broad phases. In the first phase, researchers modeled individual plant space colonization through clonal reproduction to generate initial hypotheses about competitive strategies between plant types; this phase alone produced more than 22 million simulation runs and found that prairie growth was best explained not by a single dominant strategy but by a small number of distinct viable strategies.[2] The second, more ambitious phase aimed to run thousands of combinations of plant types together in a single master simulation across an entire growing season, in order to identify optimal species combinations for long-term prairie sustainability under varying conditions.[2]

To manage the scale of the second phase, the team developed a parallel multi-objective genetic algorithm (PGA) that could run on top of BOINC to fine-tune the optimization process, described in a 2009 conference paper by PhD student Malek Smaoui and Garbey.[6]

Computing infrastructure

The University of Houston, whose Department of Computer Science and Research Computing Center hosted Virtual Prairie's local computing infrastructure.

Early, computationally intensive runs used local high-performance computing (HPC) hardware at the University of Houston. The department's third such system, provided by SiCortex, was a single-cabinet 2.1 TFLOPS cluster built around 648 64-bit processors, acquired as a joint venture between the university's Department of Computer Science and its Research Computing Center (RCC).[2] SiCortex, which marketed itself on the energy efficiency of its systems, ceased operations in May 2009, a few months after this system was highlighted in connection with the project.[7]

The Beaulieu campus of the University of Rennes 1 in France, home to the UMR CNRS ECOBIO laboratory that led the project's ecological modelling work.

For the larger, second-phase master simulations, the project moved to BOINC-based volunteer computing, distributing work to home and lab computers around the world.[6] By 2011, the university reported that more than 10,000 volunteers in 90 countries had contributed computing time to the project.[3] The project's BOINC application was hosted at http://vcsc.cs.uh.edu/virtual-prairie/, run as a Virtual Campus Supercomputing Center (VCSC) project of the University of Houston.[8]

Project team

Virtual Prairie was a joint effort between two university research groups:

  • At the University of Houston, principal investigator Marc Garbey, chair of the Department of Computer Science, worked with PhD students Malek Smaoui, who focused on volunteer computing, evolutionary algorithms, and Monte Carlo methods, and Waree Rinsurongkawong, who worked on parallel data mining methods.[1][9] Multicore and software optimization researcher Barbara Chapman also contributed to the project's computing infrastructure work.[2]
  • At the University of Rennes 1, Cendrine Mony led a team that included Professor Bernard Clement and PhD candidates Marie-Lise Benot and Anne-Kristel Bittebiere.[3] Benot later became a post-doctoral researcher at the Laboratory of Alpine Ecology in Grenoble, France.[3]

Funding

The project was funded by several French national research agencies, including the Agence Nationale de la Recherche (ANR, under grant ANR-08-SYSC-012), the Centre national de la recherche scientifique (CNRS), Cemagref (the Institute for Research in Science and Technology for the Environment, later merged into Irstea), and INRIA (the National Institute for Research in Computer Science and Control).[1][3][10]

Results and publications impact

The project's central ecological finding, drawn from its large-scale parameter study, was that high plant performance in prairie conditions was reached through several distinct combinations of traits, such as high metabolic gain paired with low resource storage, or short branch spacing, rather than through a single optimal strategy.[5] Interactive effects between more than half of the tested input parameters were also demonstrated, providing a level of insight into clonal plant growth strategies that the researchers argued would not have been feasible without volunteer computing.[5]

Current status

Virtual Prairie's BOINC servers stopped issuing new work sometime around mid-2011, shortly after the project's main results were published in Ecological Modelling; a project administrator confirmed on the BOINC message boards in June 2011 that the servers appeared to have gone silent.[10] The project's original website, vcsc.cs.uh.edu, no longer hosts Virtual Prairie content.[11] The BOINC project maintains an entry for Virtual Prairie in its historical list of retired projects.[12]

Publications

The project produced the following peer-reviewed and conference publications, drawn from the project's archived publications page:[13]

Using BOINC-computed data
Other project-related publications
  • Marc Garbey, Malek Smaoui, Waree Rinsurongkawong, Cendrine Mony. "The Virtual Prairie Project." Invited plenary lecture, 2nd International Conference on Parallel, Distributed, Grid and Cloud Computing for Engineering, Ajaccio, Corsica, April 2011, Civil-Comp Press.[16]
  • Cendrine Mony, Marc Garbey. "Updates on the Virtual Prairie Project." 5th Pan-Galactic BOINC Workshop, 22–23 October 2009, Barcelona, Spain.[17]
  • Marc Garbey. "Virtual Prairies with BOINC." 4th Pan-Galactic BOINC Workshop, 11–12 September 2008, INRIA, Grenoble, France.[18]
  • Cendrine Mony, Marc Garbey, Malek Smaoui. "Using modelling to understand clonal strategies in optimal growing conditions." 51st Annual Symposium of the International Association for Vegetation Science, 7–12 September 2008, South Africa.[19]

Archived copies of several of these papers and a related workshop video are held via the Wayback Machine on the project's former publications page.[13]

See also

External links

References

  1. ↑ 1.0 1.1 1.2 1.3 (22 April 2009).Virtual Prairie Project With Garbey as PI Taking Off. University of Houston. Retrieved 26 July 2026.
  2. ↑ 2.0 2.1 2.2 2.3 2.4 2.5 2.6 2.7 Template:Cite press release
  3. ↑ 3.0 3.1 3.2 3.3 3.4 (7 March 2011).Protecting Ecosystems, Pollution Remediation Goals of Research at UH. University of Houston. Retrieved 26 July 2026.
  4. ↑ 4.0 4.1 4.2 (2011).Large scale parameter study of an individual-based model of clonal plant with volunteer computing. Ecological Modelling. pp. 935–946. DOI: 10.1016/j.ecolmodel.2010.10.014.
  5. ↑ 5.0 5.1 5.2 5.3 5.4 Research works (profile excerpt discussing the CLONAL model parameter study). ResearchGate. Retrieved 26 July 2026.
  6. ↑ 6.0 6.1 6.2 (2010})."Parallel Genetic Algorithm Implementation for BOINC".In International Conference on Parallel Computing (ParCo), 1–4 September 2009, Lyon, France.pp. 212–219.
  7. ↑ (28 May 2009).SiCortex Meets an Untimely End. HPCwire. Retrieved 26 July 2026.
  8. ↑ VirtualCampusSupercomputerCenter. BOINC wiki (GitHub). Retrieved 26 July 2026.
  9. ↑ Improving volunteer computing scheduling for evolutionary algorithms (author biography excerpt). ScienceDirect. Retrieved 26 July 2026.
  10. ↑ 10.0 10.1 (June 2011).Virtual Prairie results published (forum thread quoting ANR-08-SYSC-012 grant acknowledgement). BOINC. Retrieved 26 July 2026.
  11. ↑ vcsc.cs.uh.edu site scan. Sur.ly. Retrieved 26 July 2026.
  12. ↑ old_projects.inc. GitHub. Retrieved 26 July 2026.
  13. ↑ 13.0 13.1 Virtual Prairie: Publications (archived). Wayback Machine. Retrieved 26 July 2026.
  14. ↑ (2008})."Virtual Prairie: Going green with volunteer computing".In APSCC, 9–12 December 2008, Yilan, Taiwan.pp. 427–434.
  15. ↑ (2011})."Fluid flow, agent based hybrid model for the simulation of virtual prairies".In Parallel Computational Fluid Dynamics Conference (ParCFD), 18–22 May 2008, Lyon, France.pp. 369–376.
  16. ↑ (April 2011})."The Virtual Prairie Project".In Ajaccio, Corsica.
  17. ↑ "Updates on the Virtual Prairie Project".In 5th Pan-Galactic BOINC Workshop, 22–23 October 2009, Barcelona, Spain.
  18. ↑ "Virtual Prairies with BOINC".In 4th Pan-Galactic BOINC Workshop, 11–12 September 2008, INRIA, Grenoble, France.
  19. ↑ "Using modelling to understand clonal strategies in optimal growing conditions".In 51st Annual Symposium of the International Association for Vegetation Science, 7–12 September 2008, South Africa.
Retrieved from "https://boincsynergy.ca/wiki/index.php?title=Virtual_Prairie&oldid=1786"