Food webs

Question 1

What are food webs?

Answer 1

Ecological (antagonistic!!) networks

Describe feeding relationships (who eats who) across trophic levels

They ignore non-trophic relationships (interactions that don’t involve feeding) e.g. mutualism like pollination, interference

Recent food webs also incorporate information on the frequencies of feeding interactions among species (energy flow web)

Question 2

Why study food webs?

Answer 2

Help us to understand

• How ecological communities are structured (which species occur in a community, and why some are rare and some are common - ie. the factors that influence their abundance)
• The dynamics of ecological communities! (what happens when we interfere with them)
• How changes to the abundance of one species can propagate

Summarise complexity of community interactions
o more realistic than models of a few interacting species
- they comprise large sets of communities
- different trophic levels, linked together through complex direct and indirect interactions

Ecologists have long argued that the complexity of food webs will be crucial for determining community stability

Question 3

What are the 3 main ways food webs can be studied?

Answer 3

a. Models
look at the properties of computer-generated webs. Can use these to see how they respond to perturbations

b. Observation
what patterns can be seen in ʻrealʼ webs? Are there particular rules in the way webs are organised?

c. Experiments
test food web theory in the laboratory or in the field. Eg. add or remove species, or controlled lab experiments where we set up micro food webs

Question 4

What structures food webs?

Key factors arising from food web studies that may structure communities
i.e. determine which species occur and their relative abundances

Answer 4

a. Indirect effects eg. apparent competition
b. Keystone species
c. Anthropogenic (human) disturbance

Question 5

Darwin’s tangled bank - what is it trying to convey?

Answer 5

• Species do not exist in isolation, they depend on each other in complex ways

o Complex networks of antagonistic (predation, herbivory and parasitism)
o Or mutualistic interactions (e.g. pollination)

Question 6

Name the different types of food webs?

Answer 6

– Linkage, energy flow, functional/interaction

– Source (prey-based) and sink (predator-based) webs

Question 7

Linkage web

Answer 7

Circles (NODES) usually indicate species

Lines indicate species interacting through predation (EDGES)

Question 8

Energy flow web

Answer 8

Quantifies fluxes of energy/ frequency of interactions between nodes along links between a resource and a consumer

Weighted network that indicates how much species are interacting!

Question 9

Functional/interaction strength web

Answer 9

o Ask what the most important interactions are in terms of population dynamics - some interactions have greater bearing on community organisation than others, so have more energy flow pathways
o Frequency of interactions doesn’t necessarily mean importance in terms of dynamic effects
o Require experiments to conduct

Functional webs have compartments, which are sub-groups in the larger network where there are different densities and strengths of interaction.

Functional webs emphasise that “the importance of each population in maintaining the integrity of a community is reflected in its influence on the growth rates of other populations

Question 10

Source web

Answer 10

BASED ON A PREY SPECIES

Source web - one or more node(s), all of their predators, all the food these predators eat, and so on.

Question 11

Sink web

Answer 11

BASED ON A PREDATOR SPECIES

Sink web - one or more node(s), all of their prey, all the food that these prey eat, and so on

o Start with an apex predator eg. starfish
o What does it consume?

Question 12

Turning a binary food web into a matrix

Answer 12

o Binary matrix of interactions
o ABC etc are codes for different species
o Interaction = 1, no interaction = 0

Question 13

Arrow goes towards…

Answer 13

the thing that is doing the eating

Question 14

Quantitative food web as matrix

Answer 14

o Networks can be represented graphically in many ways
o Information on frequency of interactions/biomass
o ie. use an energy flow web and assign numbers to the ‘strength’ of the interaction

Question 15

How do you represent interactions between predators/prey or parasitoids/hosts?

Answer 15

Graphical representations of networks

Can represent in 3D

Gets v complicated v quickly

Question 16

Non-ecological examples of networks

Answer 16

• Banking networks
• Social networks (e.g. Facebook, scientific collaborations, etc.)
• All organised in fairly similar way
• In each case, we are trying to describe the way in which “nodes” (species) are connected by “edges” (links)

Question 17

Attributes of food webs

How can we describe food web organisation in a simple way?

Answer 17

You can use the metric ‘connectance’

S = The no. of species in the web
L = The no. of observed links or connections (solid lines)
S(S-1)/2 = the no. possible links (solid lines + dotted lines)
C = Connectance (the fraction of possible links in the web that actually occur) 

Connectance allows for the possibility that any species can interact with any other species in the network

Connectance = Actual Links / Possible Links
C = L/[S(S-1)/2]

Question 18

Observation - documenting food webs in the field

Early studies

Modern studies

Answer 18

Loads of effort has been put on to try and document food webs in the field

• Joel Cohen found patterns in published food webs
• Many of patterns now thought to be artefacts - can’t use old research! due to…
o Taxonomic bias
o Lumping (e.g. ʻplanktonʼ as one unit)
o Omnivory (tendency to feed on more than one trophic level) and ʻrareʼ links underestimated

• Newer analyses use ʻpurpose builtʼ webs
• Some patterns are robust and consistent
o But they are different from the ones that emerged from early studies!
o Food chains are short (3-4 levels)
o Omnivory can be common

Question 19

Modelling - how do you make computer-generated webs?

What can this tell you about web stability?

Answer 19

ʻDynamic Modelsʼ suggested that more diverse communities are less stable (completely opposite to conventional wisdom)

❖ May extended Lotka-Volterra models to sets of interacting species
❖ In May’s 1973 book he constructed randomly assembled model webs with

different no. of species (S)
connectance (C)
interaction strengths between predators and prey (β)

❖ He disturbed the webs in different ways (S, C, and no. trophic levels) then analysed their stability in response to these perturbations to see which combinations were stable
o This allows you to identify food web properties that allow sets of interacting species to persist
❖ Perturbations are likely to propagate further in food webs with more species or with a greater connectivity among species

o Rule of thumb – webs were stable (pops returned to equilibrium after a small disturbance) IF you satisfy β(SC)^1/2 <1
• So complexity (more species and/or more connections) appears to decrease stability

Question 20

Conclusions about how to get stable food webs from models

Answer 20

Conclusions from May’s models

❖ He showed that individual populations become less stable with:

• More species (S)
• More links between species (C)
• When interactions between species are stronger

Therefore..

1. Food chains should be SHORT (longer chains were not stable)
2. Complexity may reduce stability (contradicts conventional wisdom)
3. Species feeding on more than one trophic level (ʻomnivoresʼ) should be rare
   a. Omnivory seemed to be destabilising
   b. But in the few cases omnivores present and were stable, they were very stable

Question 21

Why model and do experiments on food webs?

Answer 21

Modelling and experiments provide insights about food webs that are not obvious from observation alone

• run combinations of properties to infer stability conditions

Question 22

What are the 2 main hypotheses for why food chains are short

these are not mutually exclusive, could be a combination of the two!!

Answer 22

1) Productivity (energy attenuation) hypothesis (Elton, MacArthur)
2) Trophodynamics (Instability of long chains) explanation (Pimm, Lawton)

Question 23

Does complexity reduce stability?

Conventional wisdom in the 50s and 60s and how things changed in the 70s

Answer 23

Link to diversity&stability lectures!

YES
Conventional wisdom by OBSERVATION that complexity ie diversity “begets stability” (Elton, 1950s)

o Diverse communities have more routes for energy flow, and more negative feedback loops to control outbreaking populations
o Also more species capable of taking over the ecological roles of others (species redundancy);

Idea that a species that feeds on many prey species is ʻbufferedʼ against chance fluctuations in prey abundance - if a predator loses one prey species, it can switch to another (having many links to diff prey increases stability)

NO
Mayʼs MODELS contradicted this (1970s)

ʻDynamic Modelsʼ suggested that more diverse communities are less stable (completely opposite to conventional wisdom)

❖ May extended Lotka-Volterra models to sets of interacting species
❖ In May’s 1973 book he constructed randomly assembled model webs with

different no. of species (S)
connectance (C)
interaction strengths between predators and prey (β)

❖ He disturbed the webs in different ways (S, C, and no. trophic levels) then analysed their stability in response to these perturbations to see which combinations were stable
o This allows you to identify food web properties that allow sets of interacting species to persist

❖ He showed that individual populations become less stable with:

• More species (S)
• More links between species (C)
• When interactions between species are stronger

❖ Perturbations are likely to propagate further in food webs with more species or with a greater connectivity among species

o Rule of thumb – webs were stable (pops returned to equilibrium after a small disturbance) IF you satisfy β(SC)^1/2 <1
• So complexity (more species and/or more connections) appears to decrease stability

Question 24

Omnivory

• what is it?
• rare or common?

Answer 24

Omnivory = Consumers that eat prey from more than one trophic level

Rare or common?
o No – it is overlooked
• Underestimated in early webs made from observations in nature (Cohen)

Question 25

• types of omnivory

Answer 25

Different types of omnivory
 Temporal (seasonal prey - species use different food at different time of year)
 Opportunistic switching
 ʻLife History Omnivoryʼ
E.g. many insects such as mosquitos have larvae live in aquatic habitat and feed on algae, adults feed on blood/nectar
 ʻIncidental Omnivoryʼ ? E.g parasites

Question 26

• implications of omnivory

Answer 26

Implications
• Seems to have important implications for dynamics of communities
o Increases food web complexity and connectivity
o More omnivory = more connectance (the fraction of possible links in the web that actually occur)

Question 27

Using diagrams and examples, explain the difference between a ‘source’ food web and a ‘sink’ food web

Answer 27

Source web

• based on a prey species
• one or more node(s), all of their predators, all the food these predators eat, and so on.

Sink web

• based on a predator species
• one or more node(s), all of their prey, all the food that these prey eat, and so on.

o Start with an apex predator eg. starfish
o What does it consume?

Question 28

The benefit of using source/sink webs

Answer 28

o It’s an extreme challenge to document all interactions (if even possible?)
o Source webs limit webs – eg. focus on a primary species eg. tree
o Look at predators, pathogens, parasites
o Keeps it manageable for study and valuable

Question 29

Explain the Productivity (energy attenuation) hypothesis for why food chains are short

(Elton, MacArthur)

• predictions from the hypothesis
• seen in reality?

Answer 29

• Energy is lost at each tropic level (heat, respiration, etc.) through inefficiencies
• 2nd law of thermodynamics
• Eventually you will run out - there is not enough energy for a further trophic level

Predictions from productivity hypothesis
- Communities with more efficient energy transfer and higher productivity should have longer food chains

Seen in reality? Hmm…
 Sometimes communities with more efficient energy transfer do have longer food chains, e.g. ectotherms vs . endotherms
 At very low productivities, slight increases do give longer food chains but there is a limit (few communities have > 3 or 4 trophic levels)

THEREFORE little evidence that food chain length is increased by increased resources, except at very low productivity

Question 30

Explain the Trophodynamics (Instability of long chains) hypothesis for why food chains are short

(Pimm, Lawton)

• predictions from the hypothesis
• seen in reality?

Answer 30

• Chance variations in population size are amplified up the food chain leading to unpredictable dynamics for the top predator (small change at bottom of stack of plates will propagate up chain)

Reality? hyp is correct

• Food chains are indeed shorter in unpredictable environments
• Once you get up chain, environmental stochasticity has greater effect

Question 31

What is ‘apparent competition’? Give an example

Answer 31

o Form of indirect interaction
o Looks like two species are competing, even though they are NOT directly competing for resources. Have a shared natural enemy!
o Occurs when two species (N1 and N2) share one or more natural enemy species P (e.g. two herbivores share predators or pathogens)
o Population dynamics of the two species are linked so that increases in the abundance of A reduce populations of B, via the action of the shared enemy, rather than through exploitative competition for shared resources (this is intra anyway)
o Can be symmetrical or asymmetrical
o Students should draw a food web module diagram to illustrate the concept, showing species as nodes with edges to represent the direct and indirect linkages (including signs showing positive and negative effects):
o Apparent competition is often asymmetrical

Eg. grey and red squirrels (share squirrel pox virus)
• Red squirrels now confined to Scotland, NE, Ireland
• Everyone assumed they were competing
• Seems to be more complex
• Effects mediated by a virus that affects both species of squirrel (squirrel pox virus)
• Red squirrels are much more vulnerable, hence its recent decline

Question 32

What extra information is needed to quantify a food web?

Answer 32

• frequency of interactions
• most important interactions in terms of how they affect pop dynamics (interaction strengths between predators and prey (β))

in order to build energy flow webs and functional/interaction webs

TO quantify the overall complexity of a food web you need to calculate connectance using no. species in the web and no. links/connections

Question 33

What is an ‘indirect interaction’ in the context of food webs?

Answer 33

• Most food web models consider direct
(trophic) interactions, e.g. A eats B but neglect the importance of indirect
• Indirect effects occur e.g. when B influences the population
size of C via the action of A
• An example of this is apparent competition (red and grey squirrels share susceptibility to squirrel pox virus)

Question 34

What is the importance of weak interactions?

Answer 34

• In models, weak to intermediate strength links are important in promoting community persistence and stability
• Weak interactions are like elastic bands – allow food web to come with perturbations
• Weak links dampen interactions between consumers and resources, keeping population densities further from 0 and hence decreasing the chance of population extinction
• A rocky shore field experiment provides support for the models

Question 35

What’s the standard measure of food web complexity?

Answer 35

Connectance (C)!!

Actual Links / Possible Links
C = L/[S(S-1)/2]

Question 36

Explain experiments on food chain length to test which of the productivity or trophodynamics hypotheses were correct

Answer 36

Productivity hypothesis
o Experimentally manipulate resources
o Look over time at stability of population dynamics (Colpidium??) not sure if this is PH. Or TDH
o Little evidence that food chain length is increased by increased resources, except at very low productivity

Trophodynamics hypothesis
o experimentally manipulate food chain length
o Food web architecture and population dynamics in laboratory microcosms of protists.
o Long food chains show more variable dynamics

Question 37

Explain experiments on food chain length to test which of the productivity or trophodynamics hypotheses were correct

Answer 37

Productivity hypothesis
o Experimentally manipulate resources
o Look over time at stability of population dynamics (Colpidium??) not sure if this is PH. Or TDH
o Little evidence that food chain length is increased by increased resources, except at very low productivity

Trophodynamics hypothesis
o Experimentally manipulate food chain length
o Food web architecture and population dynamics in laboratory microcosms of protists.
o Long food chains show more variable dynamics

Question 38

Explain experiments on food chain length to test which of the productivity or trophodynamics hypotheses were correct

Answer 38

Productivity hypothesis - LITTLE EVIDENCE
o Experimentally manipulate resources
o Little evidence that food chain length is increased by increased resources, except at very low productivity

Trophodynamics hypothesis - CORRECT
o Experimentally manipulate food chain length
o Look over time at stability of food web architecture and population dynamics in laboratory microcosms of protists.
o Long food chains indeed show more variable dynamics

Question 39

What are two key factors arising from food web studies that may structure communities

i.e. determine which species occur and their relative abundances

Answer 39

• Indirect Effects

* Keystone Species

Question 40

What’s a keystone species?

• how would you recognise one?
• relevance to conservation
• give an example

Answer 40

ʻA Species whose impacts on the community or ecosystem are large, and much larger than might be expected from its abundance [or biomass]ʼ

Key here is that the effects must be SURPRISING ie. Not in proportion to abundance or biomass

• Often but not always near the top of food web (predators)

How would you recognise one?

• Usually recognised when they are missing, has a marked knock on effect
• If you remove all the oaks, pretty clear will be big -ve effect

Relevance to conservation – if these species go extinct then population dynamic effects and/or functional consequences will be large and widespread

Eg. sea otters
• Pacific NW of America
• Main prey of otters are sea urchins (that graze heavily on kelp)
• Dramatic decline due to over-hunting led to increases in sea urchins, and a decrease in kelp (seaweed) and all the marine organisms dependent on kelp (trophic cascade initiated!)
• Losing single species has marked knock on effects

Eg. Wolves in yellowstone, Aldabra giant tortoise, fig trees (and fig wasps)

Question 41

Why might looking for keystone species be a little simplistic?

Answer 41

Keystone species are context dependent

as we erode biodiversity by causing extinction, ecosystem becomes less and less stable, and relies on different individuals

Question 42

Human uses of webs

Answer 42

Webs can be used to look at the consequences of human actions for the structure of communities

o Invasive (introduced) species 
o Habitat loss/fragmentation 
o Overkill
o Climate change 
(these are all things that cause species extinction - Jared Diamond!!)

• Use webs as replicated units for hypothesis testing
• Changes to web structure may inform us about stability or fragility
• Can draw habitat modification gradients - no. nodes doesn’t change much as you degrade habitats. What’s changing is the organisation of food webs

Question 43

What is an ecosystem cascade? Provide an example

Answer 43

oAn ecosystem cascade occurs when a change in the distribution of an apex predator results in dynamic change in the size or structure of populations at lower trophic levels. It is important that the cascade is flagged as being down the trophic pyramid.

Eg. Yellowstone wolves. A really good answer would distinguish between density and trait-mediated cascades. A density-mediated one occurs when a change in the distribution of an apex predator generates direct effects on the numbers of lower levels by altering predation rates. A trait-mediated cascade is via indirect effects, perhaps via the landscape of fear.