Novel contributions in Ecology (WP1) Benoit Gauzens & Cendrine Mony Review, March 20, 20141 Content Ø Working within the Diversify consortium Ø Exchanges with the ecological board Ø Ecological projects undertaken within Diversify WP1: Ecological Modelling Ø Task 1.1. Ecological board coordination -> Organization of workshops 1 & 2 Ø Task 1.2. Knowledge transfer for diversity and specialization synthesis -> Literature surveys and meetings Ø Task 1.3. Knowledge transfer for reasoning and adaptation mechanisms -> Literature surveys and meetings Working within the Diversify consortium Knowledge on ecological processes Computer scientists Ecologists Bipartite graphs Ecological processes Measuring diversity Measuring resilience … Working within the Diversify consortium Knowledge on ecological processes Ecologists Computer scientists New questions in ecology How implementing diversification processes in graphs ? What type of plastic responses to environmental variations ? (…) Exchanges with the ecological board Involvement of the ecological board in WP1 B. Kunin, Univ. Leeds C. Melian, EAWAG M. Hutchings, Univ. Sussex E. Thébault, CNRS Project meeting (July 2-6tth, Oslo) Invited stay (Oct. 17-18th, Rennes) Invited stay (Feb. 3 & 4th, Rennes) Exchanges with the ecological board Involvement of the ecological board in WP1 B. Kunin, Univ. Leeds C. Melian, EAWAG M. Hutchings, Univ. Sussex E. Thébault, CNRS Application 1 Application 2 Methodological issues on graphs Review on ecological graphs Conceptual Framework and main goals The importance of graphs in Ecology Ø A set of nodes and edges that can depict any set of components interacting together Little Rock Lake food-web (Martinez, 1991) Co-occurrence of soil taxonomic units (Barerán et al., 2011) Landscape graph of a carabid species (Vasas et al., 2009) Conceptual Framework and main goals The importance of graphs in Ecology Metacommunities Communities Individuals Landscape graphs Biogeography Species networks Co-occurrence Mating graphs Social graphs Architectural graphs Conceptual Framework and main goals Complexity vs. Stability Ø Graph structure may determine the system stability to local environmental disturbance Architectural graphs -> Plant structure vs. performance Species graphs -> Complexity Stability debate (May 1972, Allesina and Tang 2012, Thébault and Fontaine 2010) Landscape graphs -> Connectivity vs. species conservation Conceptual Framework and main goals Complexity vs. Stability Ø Positive impact of evolution - Species diversification and evolution increase stability - Adaptation of species behaviour has stabilising effects Loeuille & Loreau, 2005; Kondoh, 2003; Heckmann et al., 2012 Conceptual Framework and main goals Main questions Ø Adaptivity in food webs What functional implications? Which response to global warming? Conceptual Framework and main goals Main questions Ø Local plasticity in architectural graphs in response to heterogeneous environments What response ? In which environmental conditions ? Application 1: Adaptivity in trophic food webs Introduction Size of prey (Handling time)-1 Energy intake Modelling studies in food webs: invariant trophic relationships ≠ Optimal foraging theory Size of prey Brose et al. 2008, Petchey et al. 2008 Application 1: Adaptivity in trophic food webs Introduction Ø Foraging increases stability in food webs With foraging Persistence Without foraging Connectance Connectance Kondoh, 2003 Application 1: Adaptivity in trophic food webs Introduction Persistence Ø Foraging is dependent on web allometric structure Allometric structuration Ø Foraging increases stability in food webs Foraging effort Kondoh, 2003 Application 1: Adaptivity in trophic food webs Introduction Ø Predation impacts species size at short term (days) Rutilus rutilus Abramis brama Daphnia F0: no fishes F1…3: increasing fish biomass Borcic et al. 1998 Application 1: Adaptivity in trophic food webs Main question What are the consequences of size adaptation on food web functionality ? Range of size evolution Mechanisms Convergence toward steady state Generation of realistic food web structure underlying food webs assemblage Impact on stability Response to temperature increase Functional properties Application 1: Adaptivity in trophic food webs Methods Ø Use of an allometric model Yodzis & Innes 1992, Heckmann et al. 2012 Application 1: Adaptivity in trophic food webs Methods Ø Use of an allometric model Ø Species are defined by mean size and variability (normal law) Predator Preys 103 106 Size Yodzis & Innes 1992, Heckmann et al. 2012 Application 1: Adaptivity in trophic food webs Methods Ø Use of an allometric model Ø Species are defined by mean size and variability (normal law) Eats on (10-3 predator prey size ratio, Brose et al. 2006) Predator Preys 103 106 Size Yodzis & Innes 1992, Heckmann et al. 2012 Application 1: Diversification in trophic food webs Methods Size Greater individuals consumed => decrease of species size Smaller individuals consumed => increase of species size Application 1: Diversification in trophic food webs Preliminary results Ø Food web persists only with unrealistic parameter values without size adaptivity In agreement with results from Heckmann et al. 2012 23 Application 1: Diversification in trophic food webs Next steps Ø Implementation of the adaptivity rules Ø Comparison of the resulting topology to empirical results Ø Implementation of temperature effects 24 Application 2: Local plasticity in architectural graphs Introduction: Plasticity in clonal networks Ø Due to their sessile life, plants face high temporal variability and spatial heterogeneity in environmental conditions Ø Modular vertical and horizontal growth forms Ø High plasticity at the local scale Application 2: Local plasticity in architectural graphs Introduction: The particular case of clonal plants ramet spacer connection Elytrigia repens Application 2: Local plasticity in architectural graphs Introduction: Foraging vs. Specialization Hutchings & de Kroon, 1994; Kroon & Hutchings, 1995; Alpert & Stuefer, 1997 Application 2: Local plasticity in architectural graphs Local plastic response Foraging behavior Ramet specialization (division of labour) Hutchings & de Kroon, 1994; Kroon & Hutchings, 1995; Alpert & Stuefer, 1997 Application 2: Local plasticity in architectural graphs Local plastic response Network function Foraging behavior Modification of the connexion traits Spacer length Branching degree Branch orientation Ramet specialization (division of labour) Modification of the ramet traits Root:Shoot ratio Space exploration Network Performance Space exploitation Hutchings & de Kroon, 1994; Kroon & Hutchings, 1995; Alpert & Stuefer, 1997 Application 2: Local plasticity in architectural graphs Environmental heterogeneity nb of bad patches patch size patch contrast Application 2: Local plasticity in architectural graphs Main questions patch size Foraging Plasticity type ? patch contrast Specialization Application 2: Local plasticity in architectural graphs Hypothesis Patch size ? Foraging ? Specialization Application 2: Local plasticity in architectural graphs Hypothesis Patch contrast ? Specialization/ Foraging Application 2: Local plasticity in architectural graphs Hypothesis Difficulty to argue for good hypothesis due to: Gain vs. Cost of the two types (no experimental or empirical data) Potential trade-off between the two plastic responses Benefit of these responses may be relative: degree of the response vs. spatial scale of the heterogeneity Things may be disentangled through modelling Application 2: Local plasticity in architectural graphs Modelling CLONAL Elementary unit: the clone Basic metabolism Plant Structural Blue-­‐Print Ontogeny Plas6city Simulations 11 input parameters 2-­‐4 values tested each X Non Plas9c, Foraging, Specializa9on, F+S Environmental Heterogeneity Scenario selec6on Number and size of the patches Patch contrast Literature survey on experimental data Application 2: Local plasticity in architectural graphs Simulations CLONAL Elementary unit: the clone Basic metabolism Plant Structural Blue-­‐Print Ontogeny Plas6city 11 input parameters 2-­‐4 values tested each X Non Plas9c, Foraging, Specializa9on, F+S Environmental Heterogeneity Number and size of the patches Patch contrast Network performance Optimization Application 2: Local plasticity in architectural graphs Methods: 1. Foraging behavior – Effect of grain size Length = L0 + µ*[d0,d1] µ € [0,1] d0 -> poor d1 -> rich Total Nb of ramets Total Biomass Application 2: Local plasticity in architectural graphs Methods: 1. Foraging behavior – Effect of patch contrast dBiomass/dt(ramet) = B0(1-rs)B(t)(1-B(t)/Bmax)) Bm0 -> poor Bm1 -> rich B1=Bm0/C C: Contrast poor vs. rich Total Nb of ramets Total Biomass Application 2: Local plasticity in architectural graphs Preliminary results: Effect of grain size (Strong contrast) Fine grain Total Biomass Total Biomass Coarse grain d0 d1 d0 d1 Application 2: Local plasticity in architectural graphs Preliminary results: Effect of patch contrast Strong contrast Total Biomass Total Biomass Low contrast d0 d1 d0 d1 Application 2: Local plasticity in architectural graphs Next steps Ø Implement 2-D simulations Ø Implement Specialization process Ø Large-scale simulations with Foraging and Specialization together Graph simplification information loss and ecological meaning Introduction: Grouping species in graphs Modularity Trophic groups Groups of species with rather similar interactions Groups of species interacting more between themselves than other species Graph simplification information loss and ecological meaning Introduction: Grouping species in graphs Modularity Trophic groups Groups of species with rather similar interactions Groups of species interacting more between themselves than other species - Strong mathematical background - Few ecological significance - No clear definition - Little mathematical background - Strong ecological significance Graph simplification information loss and ecological meaning Main questions Ø Develop a method for trophic group detection Ø Analyse of information loss by aggregation Ø Comparison of relationships between different clustering methods Graph simplification information loss and ecological meaning Methods Ø Development of a model for trophic group detection in food webs based on modularity maximisation approaches Difference between proportion of common trophic links between species of a group and its random expectation E G( E ) = ∑ g =1 1 g ∑ (T (i, j) − T (i, j)) r i , j∈g i< j Gauzens et al. Sub. (arXiv) Graph simplification information loss and ecological meaning Preliminary results Creteil lake food web: colors set trophic groups, grey square modules Graph simplification information loss and ecological meaning Next steps Ø Functional implications of a trophic group structure Ø Species traits underlying group compositions Ø Link between trophic groups composition and key species detections Toward a synthesis on ecological graphs Main goals Ø Need for more general overview on graphs in ecology in order to detect common patterns to transfer to computer scientists Common and divergent characteristics Structure vs. Functionning Structure vs. Stability Response to environmental variations Toward a synthesis on ecological graphs Main goals Ø Find a new methodology for coupling different types of graphs Spatial, temporal, organisational scales Available data Toward a synthesis on ecological graphs Main goals Ø Examples that could be addressed through this coupling: Conservation issues Response to climate change Ecosystem services (…) Publications and Communications Ø Gauzens et al. 2013 Intermediate predation pressure leads to maximal complexity in food webs. Intecol, London. Ø Gauzens et al. Submitted. Trophic groups and modules: two levels of group detection in food webs. arXiv $ u o y k n Tha n o i t n e t t a r u o y for