Dear Virtual Prairies Explorers,
This is our fourth news letter on the Virtual Prairie project, and as usual we would like to update you on our progress.
Currently, we have got two main activities around the ViP project: first we keep improving our model of the virtual prairie and run large parameter studies with more and more complex configurations. One of the most exciting one has been to examine the dynamic of a population with two types of clonal plant classified as Phalanx and Guerrilla according to Lovett-Doust (1981). This classification applies to clonal plants that have network architecture and is characterized by two opposite strategies of invading space. Guerillas invest resources to explore and conquest space rapidly through long and unramified stolons or rhizoms. Conversely Phalanx are compact and produce dense clouds of ramets on a highly ramified network to extract all resources on site. This work has been done with the help of S. Pennings from the University of Houston. We call this study episode VII of the clonal war…. And you may think about many analogies to these clonal warriors!
Second with the help of Huy Nguyen, we start using Genetic Algorithm (GA) to optimize the prairie for various ecological functions such as nitrate or herbicides depollution, maintaining species richness, etc… the GAs are optimization techniques that mimic life evolution with its cycle of selections and mutations to let the best fitted survive. In principle, this algorithm can be parallelized easily. However things get more complicated with volunteer computing, because clients may keep switching on and off the ViP share of their system to run other tasks. In this context we need to come up with imaginative algorithmic way of speeding up the parallel processing no matter the status of the pool of clients. This work is developed through the PhD of Malek and has been presented at the last conference ParCo 2009 conference in Lyon.
All this activity in computing, thanks to your participation, is developed hand by hand with an ambitious program of experiments. Anne Kristel has set up a large system of controlled experiments with over one hundred frames containing chosen clonal plants (see Figure Experimental design in the common garden of the University of Rennes1). This experiment, together with the complementary master work of Nolwenn, will help us compare real plant interactions with virtual ones. This interplay between experiments and simulation allows us to develop more accurate models of virtual prairies.
|
Figure: Experimental design in the common garden of the University of Rennes1
Florian Dang and Marie-Lise Benot have worked together on a third campaign of computing. This project aims at testing how clonal plants growth may be impacted by herbivores.
Furthermore, we are ready for a campaign of data acquisition in two long term ecology sites, one in French Brittany, one in Georgia, USA, where we are going to build fairly detailed digital map of prairies and marsh to get extensive data on clonal plant spatial distributions. Thanks to Remi Salmon, we will be able to produce dozens of thousands of digital images that will be processed by volunteers...
We keep going full speed on this ViP project and expand our portfolio of publications from computer sciences to ecology, through applied mathematics.
We are thankful to David Anderson, from the Berkeley Space Lab and PI of the BOINC project, for his invaluable help. Actually, an update on the ViP project was presented at the BOINC workshop in Barcelona.
Well that's all for now! We'd like to thank you all, and we will do our best to take advantage of these BOINC simulations. Next log should come in about few months...
Be the force of the Silicon with you!
Marc Garbey and Cendrine Mony