Slide Terrestrial Trophic Cascades

Terrestrial Trophic Cascades Shurin et al (2002) Across ecosystem comparison of the strength of trophic cascades Meta-analysis of 102 studies reporting plant biomass Cascades strongest in marine benthos>lakes and streams>grasslands …but some striking examples of terrestrial trophic cascades McLaren and Peterson (1994) Study on 500 km2 Isle Royale National Park in Lake Superior Producer: Balsam fir Herbivore: Moose (59 % of winter diet of Moose is fir) Carnivore: Wolf (colonized island in 1959) Observation: Over 30 years fir trees from across the island show cyclic intervals of ring growth suppression Periods of growth suppression show no correlation with climatic data but follow periods with high moose density Suppression of fir Release of fir Wolf population fluctuation 1960 - 1990 Moose population fluctuation Fir tree mean ring width (mm) western Isle Royale Fir tree mean ring width (mm) eastern Isle Royale Actual evapotranspiration Adding behavioral ecology to trophic cascades: Wolf behavioral response to climate (Post et al 1999) How would the change in behavior of a top predator cascade through a community? Isle Royale: Fluctuations in North Atlantic Oscillation (NAO) results in changes in winter snow accumulation Annual aerial surveys of wolves show close correlation between wolf pack size and the state of the NAO (and therefore snow depth) Lower NAO index = greater snow depth correlates with larger wolf packs Snow Kill rate per pack increases with wolf pack size Wolf mortality rate also lower in snowy years Climatic effects on hunting behavior cascade, influencing moose population size and finally fir growth effects of NAO on firs: i) drop in moose density reduces browsing intensity year after heavy snow ii) additional effect in current NAO year due to protection from browsing by snow cover Climate has multiple effects - on sociality of wolves - on vulnerability of moose to predation (deep snow impedes moose locomotion and susceptibility to predation) - on accessibility of forage to moose Trophic cascade results from higher predation rate by wolves Changes in wolf behaviour have ecosystem level effects in this community because moose influence primary production, litter production and edaphic nutrient dynamics (Pastor et al 1993) Sheriff et al (2009) J Animal Ecol 78:1249-1258 Also show that predator effects mediated through physiological stress Faecal cortisol metabolite (FCM) concentrations in female hares - inversely related to litter size - offspring birth mass and foot length - highest in individuals that did not produce viable young Increased cortisol increases survival under stressful conditions? Or Increased cortisol in the mother increases stress response of offspring, increasing offspring survival Trophic cascade without prey consumption (Beckerman et al 1997) Behavioural mediation: the mere presence of predators could alter prey foraging behaviour by reducing herbivore feeding time, or by inducing a shift in herbivore diet selection Also observed for invertebrate herbivores… Grasshoppers and spiders Study in old field community in NE Connecticut Constructed trophic levels: Grasses and herbs; generalist leaf chewing grasshopper (Melanoplus femurrubrum); and nursery web hunting spider (Pisaurina mira) Problem: Need to create a predator that does not have the capacity to consume prey, but, can display hunting behaviour or signal ‘risk’ to its prey… Solution: Glue the spider’s mouth parts together! - No effect on spider hunting behaviour - Glued spiders survive up to months Experiment: Test whether indirect effect of predators on prey arises from density or behavioural responses Made mesh enclosures for trophic levels: - plants only - plants + juvenile or adult grasshoppers - plants + grasshoppers + glued or unglued spiders Significant effect on grass biomass consistent with trophic cascade - Treatments containing spiders significantly less herbivore damage than treatments with just grasshoppers - No significant difference between glued and unglued spiders on grass herbivory and biomass When can we expect trophic cascades to occur? What factors determine when and where predators or resources dominate in regulating populations? Power (1992) predicted: Cascades influenced by consumer efficiency - shape of the predator numerical response and functional response Predators may not regulate their food resources efficiently if they: -fight with each other (interference competition) -are limited by or compete for resources other than food, -cause prey to hide -induce defenses in their prey or, if significant time lags exist between prey consumption and predator population response Additional feedback pathways across trophic levels (unrelated to consumer efficiency: Non-linear herbivore effects on primary productivity - nutrient recycling from grazers can increase productivity at intermediate levels of consumption for terrestrial plant and phytoplankton communities - Highly non-linear relationships between grazing pressure and productivity could decouple top-down effects Plants provide cover for predators and prey Changes in plant biomass could alter the consumption rates of herbivores by predators Interference competition (-) (-) Predators (+/-) Induced defense, Retreat to refuges Consumers (+/-) Cover from (for) predators Stimulation of Primary productivity (+) Plants Mechanisms modulating top-down and bottom-up forces in food chains (modified from Power 1992) Borer et al (2005) Meta-analysis of trophic cascades hypotheses for why trophic cascades (sometimes) occur: 1.  High spatial heterogeneity weakens cascades because refugia reduce the search efficiency of predators (Polis 2000) Food webs that deviate from a linear food chain weaken cascades (deviations can arise from competition, intraguild predation, or from omnivory)(Agrawal 1998) High resource availability and quality promotes cascades (consumption rate of herbivores increases; larger impact on producer organisms)(Polis 1999) Cascades decrease with increasing study duration if plants recover from browsing or become less palatable (Polis 1999) High predator or herbivore efficiency will increase cascade strength (e.g strong numerical or functional response) via high consumer conversion efficiency (Polis 1999) Methods: 114 studies that include some measure of plant community biomass in presence/absence of predators Dependent variable: log ratio of plant biomass with and without predators Results Most support for hypothesis 5: Efficiency of predators/herbivores Predator taxonomy, thermal regulation, and herbivore mass specific metabolic rates account for 31% of variation in cascade strength No evidence for hypothesis 3: that high productivity results in stronger cascades Possibly because of feedbacks between productivity, antiherbivore defense and plant regeneration?? No evidence for hypothesis 2: community regulation of cascades through competition, omnivory etc Species diversity was not related to cascade strength, but may not be sufficiently variable to evaluate Hypothesis (heterogeneity) and (duration) could not be assessed in this study Methodological biases in the data (most studies short and small) What about bottom-up effects? Might they be as strong, but less investigated? Johnson (2008) Bottom-up effects of plant genotype on aphids, ants, and predators Ecology 89:145-154 Explored how plant genotype (n=28) affects aphid population growth rate, abundance of aphid-tending ants, and richness of predators Plant genotype explained 29 % of variation in aphid growth rate (ants and aphid population density only %) Several heritable plant traits (e.g., trichome density, leaf nitrogen, lwc) explain ~50% of variation in aphid growth rate and ant density Would you choose to study trophic cascades in a system where you think it would be unlikely to occur?? Would reviewers/funding agency support a proposal to study a question in a system where the effect of interest is expected to be weak or non-existent? Would you be able to publish your results of non-existent effects (*yes in PLOS ONE)? “The problem arises from the fact that ecology is still in a narrative phase of its science, collecting stories about this or that system The narrative approach produces very weak tests of theory through meta-analyses of case studies, whose criteria for selection are often not stated, or worse, are based on unconscious but biased decision rules (such as whether one likes or doesn't like a given hypothesis), and which can therefore easily descend into largely confirmatory or "rejectatory" \non-objectivity We frequently seem totally oblivious of the potentially huge selection bias that we introduce when we choose which study systems in which to ask our questions” Steve Hubbell, pers.comm Some of this issues discussed in: Travis (2006) Is it what we know or who we know? Choice of organism and robustnesss of inference in ecology and evolutionary biology American Naturalist 167:303-314 Are our/your inferences robust to our/your choice of organism? Summary Many striking examples of trophic cascades - often involving aquatic or benthic communities Biological processes important in driving strength of cascades (taxonomy and metabolism are important predictor variables) Invertebrate herbivores generate the strongest cascades Duration of study may be important - but terrestrial studies restricted to single plant generation More general question: can meta-analyses generate unbiased interpretations of determinants of cascade strength? [...]... food chains (modified from Power 1992) Borer et al (2005) Meta-analysis of trophic cascades 5 hypotheses for why trophic cascades (sometimes) occur: 1.  High spatial heterogeneity weakens cascades because refugia reduce the search efficiency of predators (Polis 2000) 2 Food webs that deviate from a linear food chain weaken cascades (deviations can arise from competition, intraguild predation, or from... our/your choice of organism? Summary Many striking examples of trophic cascades - often involving aquatic or benthic communities Biological processes important in driving strength of cascades (taxonomy and metabolism are important predictor variables) Invertebrate herbivores generate the strongest cascades Duration of study may be important - but terrestrial studies restricted to single plant generation... grasshoppers - No significant difference between glued and unglued spiders on grass herbivory and biomass When can we expect trophic cascades to occur? What factors determine when and where predators or resources dominate in regulating populations? Power (1992) predicted: Cascades influenced by consumer efficiency - shape of the predator numerical response and functional response Predators may not... exist between prey consumption and predator population response Additional feedback pathways across trophic levels (unrelated to consumer efficiency: 1 Non-linear herbivore effects on primary productivity - nutrient recycling from grazers can increase productivity at intermediate levels of consumption for terrestrial plant and phytoplankton communities - Highly non-linear relationships between grazing... whether indirect effect of predators on prey arises from density or behavioural responses Made mesh enclosures for 3 trophic levels: - plants only - plants + juvenile or adult grasshoppers - plants + grasshoppers + glued or unglued spiders Significant effect on grass biomass consistent with trophic cascade - Treatments containing spiders significantly less herbivore damage than treatments with just grasshoppers... cascades (deviations can arise from competition, intraguild predation, or from omnivory)(Agrawal 1998) 3 High resource availability and quality promotes cascades (consumption rate of herbivores increases; larger impact on producer organisms)(Polis 1999) 4 Cascades decrease with increasing study duration if plants recover from browsing or become less palatable (Polis 1999) 5 High predator or herbivore efficiency... account for 31% of variation in cascade strength No evidence for hypothesis 3: that high productivity results in stronger cascades Possibly because of feedbacks between productivity, antiherbivore defense and plant regeneration?? No evidence for hypothesis 2: community regulation of cascades through competition, omnivory etc Species diversity was not related to cascade strength, but may not be sufficiently... aphid population density only 2 %) Several heritable plant traits (e.g., trichome density, leaf nitrogen, lwc) explain ~50% of variation in aphid growth rate and ant density Would you choose to study trophic cascades in a system where you think it would be unlikely to occur?? Would reviewers/funding agency support a proposal to study a question in a system where the effect of interest is expected to be... offspring survival Trophic cascade without prey consumption (Beckerman et al 1997) Behavioural mediation: the mere presence of predators could alter prey foraging behaviour by reducing herbivore feeding time, or by inducing a shift in herbivore diet selection Also observed for invertebrate herbivores… Grasshoppers and spiders Study in old field community in NE Connecticut Constructed 3 trophic levels:... areas Since wolf reintroduction willow and cottonwood height has doubled and return of beaver colonies to Lamar Valley of Yellowstone Kauffman et al (2010) Ecology 91:2742-2755 Behaviorally mediated trophic cascade (BMTC) of wolves -> elk -> aspen never adequately tested - Detailed demographic data for aspen using tree ring data - Browsing levels +/- exclosures along gradient of predation risk Tree