This lecture will discuss food webs. Read chapter 14 in Molles or chapter 22 in Begon et al. These lectures will contain material not included in Molles.
I will discuss the idea of bottom-up control, and look at questions
- The chief difficulty with the study of food webs is that very difficult mathematics has to be combined with a vast amount of ecological observations.
The chief question addressed in this chapter is the stability of food webs. Stability is a key concept depends upon
important techniques in applied mathematics.
- Top-down control - structure (abundance, biomass, diversity) of lower trophic levels depend upon the effect of consumers from higher trophic levels.
- The world is green because predators keep herbivores in check.
- Bottom-up control - structure depends upon factors, such as nutrient concentration and prey availability, from lower trophic levels. Example:
The world is green because "it is prickly and tastes bad".
- Keystone species - Species, usually predators, that cause local extinctions if they are removed. Example:
- Paine demonstrated that if a dominant starfish,
Pisaster, were removed from a rocky shore then species diversity was greatly
- Connectance - The fraction of all possible pairs of species that interact directly.
Empirical Properties of Food Webs
- Few food webs seldom have more than 3 or 4 levels.
- Webs are not too complex. Connectance (C) usually declines with species richness (S). This is consistent with theoretical models.
- Ominvores are relatively scarce. Typically food chains have one omnivore
per top-down predator.
- Omnivores feed on species in adjacent trophic levels.
- Insects, detritivores, and scavengers are exceptions to the above two generalizations.
- The ratio of prey to predators is relatively constant. For freshwater
invertebrate communities we find that there are 0.36 predators per prey.
- The ratio of basal, intermediate and top predators is relatively constant
Theoretical Properties of Food Webs
Models are used to interpret (read "think about") real ecosystems. Notions
such as stability cannot be even defined without rather complex mathematics.
The following is a brief discription of the mathematical ideas behind stability analysis. It is NOT required that you understand the following for the exam, but I will
discuss it in class because these are very important ideas.
- Start with a system of nonlinear differential or difference equations that describe one conception of a food chain.
- Find the equilibrium solutions.
- Linearize the system of equations around each equilibrium. These equations describe the behaviour of the model close to the equilibrium.
- The behaviour of the solutions to a linear system of equations is determined by the eigenvalues of the corresponding system of algebraic equations.
- For differential equations, an equilibrium is stable if the Real part of the largest eigenvalue is negative.
- For difference equations, an equilibrium is stable if the absolute value of the largest eigenvalue is less than 1.
What models tell us
- Complexity does not beget stability.
- Models predict that long food chains should not be stable.
Why do Food Webs only Have 3 or 4 Levels?
Most food webs comprise only 3 or 4 levels. Why?
- Traditional Explanation: Only a maximum of 10% transfer efficiency between
levels means rapid attenuation and short chains.
- But if this explanation holds, then highly productive environments or communities with cold- rather than warm- blooded species should have move levels. This is not so.
- However, in extremely low productivity systems, freshwater lakes in
Antartic, no predators exist. Thus, energy flow does set a lower limit
- Size and other design constraints. A predator is usually twice as
big as his prey, and sets a design constraint on the length of food chains.
- No reason to think that such constraints impose important limits to
- Dynamical constraints: in simulated webs with Lotka-Volterra interactions, long chains lead to severe population fluctuations which could not persist.
- This hypothesis is born out by some model results.
- But are these model predictions reliable? Long chains may be more stable because the doubling time at the top of the chain is much slower than the bottom.
Trophic cascades are examples of top-down control.
Removal of large predators can have impacts all out of proportion to their
Trophic Cascades are characterized by 3 features.
- Top-down influences control the assemblages structure.
- Indirect effects two or more links distant from the primary one are
- Alternative community states can persist.
Example: Sea Otters in the North Pacific Coastal Ecosystem
- Their prey is a broad range of invertebrates and fish, e.g. urchins and
- Exploitation for pelts led to near extinction by 1911.
- During early years population grew at 15% a year.
- Otters are a keystone species, they indirectly regulate algal biomass and
- Otters -> low urchin density -> large biomass of
kelp -> large amounts of detridal material ->
large biomass of suspension feeders.
Meta-analysis of Freshwater Trophic Cascades
- The generality of the trophic cascade has been an intensely debated topic among ecologists.
- Brett and Goldman (1996) conducted a meta-analysis of 54 separate enclosure and pond experiments that measured the
response of the zooplankton and phytoplankton to zooplanktivorous fish treatments.
- These results
provide unequivocal support for the trophic cascade hypothesis in freshwater food webs.
- Zooplanktivorous fish treatments resulted in reduced zooplankton biomass and increased
- The trophic cascade was weakly dampened at the level of the phytoplankton.
- Additional experiments and meta-analysis showed that zooplankton biomass was under strong consumer
control but was weakly stimulated by nutrient additions.
- Phytoplankton biomass was under strong
resource control with moderate control by fish.
- Citation - Brett, M.T., Goldman, C.R. 1996. A meta-analysis of the freshwater trophic cascade PROC. NATL. ACAD. SCI. USA . 93: 7723-7726. (You do not have to read this for the exam.)
>Example: Bottom Up Control in the Serengeti
The Serengeti ecosystem of east Africa is probably the best studied example of an ecosystem with its large animal component present.
- Big animals fundamentally alter their environment.
- Long distance migration (wildebeeste) and growth to very large size (elephant, hippopotamus, rhinoceros and buffalo) result in a virtual escape from predation.
- This results in a "bottom up" regulation by food limitation.
- Wildebeeste grazing alters the protein content of grass, stimulates new shoots, and increases the food supply for smaller grazers.
- Strong evidence for alternative stable states (woodland versus grassland), maintained by grazing and disturbance by elephants and by fire.
>Example: How many Green Sea Turtles Lived in the Caribbean?
A short explanation of meta-analysis
- 13,000 turtles per year were caught on the Cayman Islands and brought back to Jamaica to feed slaves between 1688 and 1730.
- Only females were caught, thus we multiply by 2 to get total population.
- Females reproduce every 2.5 years.
- Thus, the Nadult = 2 * 2.5* Nnesting females
- Assume 1% of nesting females were captured in 1688 (based upon descriptions of the islands being covered with turtles. This is probably conservative.
- Nadult = 5*13,000 / 0.01 = 6.5 million adults.
- Check: Females mature at age 50, thus harvested females could not be replaced for 50 years. A take of 13,000 over 42 years would result in a population of around 2.5 million with the above correction.
- 5 other large populations existed: Total population in Caribbean around 35 million. This is 15 to 20 times the abundance of large ungulates in the Seregenti today!
- Alternative calculations based upon carrying capacity: Shelf area 660,000 square km, 10% seagrasses.
- An adult green sea turtle requires 72 m square per year (we round up to 100 m square).
- Nadult =660,000 sq. km. * .1 * 10,000 turtles/sq. km. = 660 million turtles.
This calculation is based upon extension of the idea of bottom up control of herbivores from the
- Similar stories, but not as well documented, exist for other turtles, sharks, groopers, manatees, and the extinct Caribbean monk seal.
- "Studying grazing and predation on reefs today is like trying to understand the ecology of the Serengeti by studying termites and the locusts while ignoring the elephants and wildebeestes."
- Herbivores and predators were reduced to very small fish and sea urchins by the time scientific investigations begin.
- Citation: Jackson, J. B. C. 1997, Reefs since Columbus, Coral Reefs: 16, Suppl. s23-s32. (You do not have to read this for the exam, but you may want to read it anyway sometime.)
I will expect you to know the basic ideas of meta-analysis but you do not have to remember