Ecosystem Trophic Structure Part 2 / Overkill v Keystone Hypotheses

Ecosystem Trophic Structure: Continuation

Last week we had a look at how megafaunal carnivores can affect ecosystem trophic structures, focusing on a case study of the reintroduction of wolves into Yellowstone National Park. Also readily apparent in that study were the crucial roles that megafaunal herbivores have, both on ecosystem physical structure (to be covered in greater detail later), as evinced by the direct limiting effects elk had on cottonwood, aspen, willows and other woody browse, and in terms of suppression of smaller herbivores through competition for resources, as evinced by the rise in beavers, ravens, magpies etc, subsequent to the fall in elk numbers. 

Interestingly, this seems to paint megafauna as both ecological hero and villain, with increases in wolf populations exerting trophic pressure to maintain a healthy, diverse and robust ecosystem while increases in elk populations had the opposite effect. This may be too simplistic a viewpoint, the general consensus being that loss of megafauna leads to increased abundances of smaller animals and thus simpler ecosystems with fewer interactions between species, shorter food chains and species occupying the same functional roles, and therefore less resilience to change and undue pressure (such as those exacted by human encroachment for example) Malhi et al 2015, but perhaps this aspect warrants further exploration in a future post. The magnitude of megafaunal influence on ecosystem structure is, however, undeniable.

The Overkill Hypothesis

Another interesting prospect indirectly raised links back to a couple of posts ago when the architects of late-Pleistocene megafaunal demise were being discussed. One prevalent explanation for human-caused megafaunal extinction is the overkill hypothesis, essentially postulating that the rate of human hunting of megafauna exceeded the megafaunal birth rate, leading to megafaunal extinction. This was once well thought of, and reasonably compelling when considering the evidence of megafauna going extinct in the presence of modern humans and largely only surviving to the Anthropocene in areas less inaccessible to humans. Nowadays this hypothesis is increasingly seen as outdated and unsophisticated, assuming, contrary to more modern zoological evidence, that naïve prey don’t avoid novel human predation, among other issues. 

In the case of the wolves of YNP, humans were in fact responsible for their localised extirpation, believing them to be undesirable predators and hunting them without legal restraint well into the 20^th century. However, this is an unrepresentative and localised case, and a more credible explanation for anthropogenic-linked megafaunal loss lies in the not dissimilar concept of loss of keystone species.

Loss of keystone species, also variously described as foundation species or ecosystem engineers, can induce collapse of trophic cascades, leading to habitat change, altering species abundance and leading to further extinctions. 

Keystone Case Study: Extinction of Stellar's Sea Cow

Estes et al convey a strong argument for this hypothesis in their exploration of the extirpation of Stellar’s sea cow from the Commander Islands. Occurring not long ago, between 1741 (when Stellar’s sea cows were abundant in the Commander Islands, although largely extinct in other areas previously peopled with human hunters) and 1768 (when they were completely extinct from the islands) in a thoroughly researched and well understood coastal kelp forest environment, and under the observation of modern humans, this extinction event potentially entails a representative model for past megafaunal extinctions occurring as a result of indirect human effects on keystone species.

Previous arguments for complete overkill of Stellar’s sea cow were not based on robust evidence (presence of fur-hunters spuriously used as indication of sea cow hunting instead of solid evidence) whereas kelp forest collapse (the main source of food for Stellar’s sea cow) after sea otter loss has been well proven. 

Estes et al instead propose that extinction of the sea cows was due to a myriad of factors rather than just human intervention. Their argument is backed by; contemporary parallels of sea-otter–sea urchin–kelp dynamics mirrored elsewhere in the north pacific oceans and robust evidence of extensive otter-hunting for the Pacific maritime fur trade beginning in 1743. There is also data on the responses of dugongs (the closest living relatives of Stellar’s sea cow) to food reduction showing that extinction of the sea cows would have occurred even without any human predation (although this likely did occur to a certain extent). The process proposed is as follows:

-> Hunting to virtual extinction of sea otter (keystone species) in the Commander Islands in 1700s 

-> Reduced otter predation on sea urchins 

-> Increased sea urchin densities 

-> Consumption of kelp forests above critical threshold 

-> Kelp forest collapse 

-> Replacement of kelp forests by urchin barrens ~ loss of food base for Stellar’s sea cow 

-> Complete extinction of Stellar’s sea cow within decades

Closing Thoughts 

Such keystone species trophic cascades and habitat changes resulting from the loss of an ecological engineer are likely to be a widespread feature of extinctions and co-extinctions both past and future. This has been demonstrated by with numerous other examples recently proposed, such as the extirpation of wolves in the continental United States as responsible for increased coyote numbers, thus decreasing snowshoe hare numbers and in turn leading to localised extinctions of lynx. It is clear that such interaction dynamics between species can be crucial enough to lead to significant changes in species abundance, even extinctions, if significantly disturbed. Detailed knowledge of such interactions in late-Pleistocene ecosystems is not readily available, but such chain effects seem as/more worthwhile to consider than overkill hypotheses when faced with evidence of megafaunal extinction in temporal proximity to modern human arrival.

Ecosystem Trophic Structure, Part 1: Top Carnivores

Having established some basic concepts and had an introductory look at the prehistory of megafauna and megafaunal loss (and the culprits) in the last post, the plan moving forward is to explore the salient impacts this megafaunal perdition could entail, first via the investigation of the roles megafauna play in the environment on a topic by topic basis.

Perhaps the most conceptually accessible effect that megafauna have is on ecosystem trophic structure and this is the aspect we will be investigating over the next couple of posts.

Top Carnivores and the Collapse of Trophic Cascades 

Top carnivores (carnivores situated at the top of a food chain) as a subsection of megafauna, have a substantial impact on the trophic structure of ecosystems, regulating the abundance and behaviour of smaller herbivores and mesopredators (mid-trophic level predators that both prey and are preyed upon - mostly omnivorous so not reliant on any one food source – making them superpredators for smaller prey if left unchecked) and creating inhospitable landscapes that herbivores try to avoid (often referred to rather forbiddingly as ‘landscapes of fear’).

A well-known and well-researched example highlighting this megafaunal effect is encapsulated in Ripple & Brescha 2011, a study of the effect of 1995/96 reintroduction of grey wolves to Yellowstone National Park following their extirpation (localised extinction) more than 70 years before.

The wolves’ extirpation caused the collapse of a three level (predator = wolves, prey = elk, plants) trophic cascade, opening the door for elk (no longer under significant predation pressure) to significantly alter habitat, soils and woody plants in the park. Over the 70-year absence of the wolves, herbivory (essentially plant-eating) of the booming elk population inhibited the recruitment (addition of new organisms to a species population) of plants such as aspen, cottonwood and willow. Given the rare opportunity to study a naturally occurring experiment outlining the effects of large predators on an ecosystem, Ripple & Brescha did the following over the period following the reintroduction of wolves into the park. 

Method

• Carried out aspen/cottonwood recruitment surveys (taking into account other factors that might affect tree recruitment)
• Summarised trends in wolf, elk, bison, beaver populations
• Summarised temporal trends in willow stem cross-sectional ring area growth (higher areas indicating more willow growth and less browsing suppression)
• Synthesised data from other literature concerning vegetation changes in YNP from 1996 to 2010 (year before publication of their article)

Results

Summarised in the figure below. With the reintroduction of wolves in 1995/96, elk populations decreased; consequentially the percentage of aspen leaders browsed decreased whilst aspen height, willow ring area and cottonwood, beaver and bison populations all rose.

The overall changes since wolf reintroduction were as would be predicted by ecological theory - put simply; more wolves, fewer elk (with altered behaviour), and more plant biomass.

Examining some consequential ‘butterfly’ effects instigated by wolf reintroduction in more detail:

• Under threat of predation by wolves (back to landscape of fear idea) YNP elks exhibited altered habitat use, movements, group sizes and vigilance – all likely contributing to reestablishment of the historical wolves-elk-browsing plants cascade and a return to recoupling of behavioural and density effects of wolves on elk with fire disturbance expediating further plant recruitment. Furthermore, new and continuing recruitment of such woody browse species (species with parts to be browsed upon (eaten)) has potential rippling effects on abiotic processes (such as decomposition of nitrogen into plant accessible forms) and biotic functions. Decreases in elk herbivory, resulting in decreased coyote numbers and increased cover and forage could lead to increases in small herbivore populations, in turn significantly affecting the prey base for mid-size predators such as foxes and badgers.
• Resurgent willows providing a more structurally complex habitat for songbirds, leading to an increase in songbird population richness. The resurgence of willows, being a major component of beaver diet in YNP was also at least partially responsible for the increase in beaver populations, the implications of which are huge. Production of dams and ponds by beavers decreases streambank erosion / increases sediment retention, raising wetland water tables, modifying nutrient cycling, all of which play an important role in plant and animal diversity in riverside ecosystem. Recovering riparian (relating to wetlands adjacent to rivers / streams) vegetation can work in tandem to help stabilise eroding channels and add complexity to river/pond shapes.
• Wolf-killed carcasses could benefit scavengers like ravens, magpies, eagles.
• Less browsing on berry-bushes by elk -> higher berry production -> more food for birds and bears -> increased shrub establishment through seed dispersal after consumption, transport and defecation of berries by birds and bears (linking to another idea to be looked at in its own right later).

Conclusions

The reintroduction of wolves seems to have initiated a restructuring of YNP’s ecosystems, re-establishing a historical trophic cascade. Beyond direct implications for wolf management as a conservation tool in ecosystem restoration (a positive step in itself), this study evinces the crucial role that large predators can play in establishing and maintaining resilient, robust wildland ecosystems. As Ripple and Brescha put it ‘Predation and predation risk associated with large predators appear to represent powerful ecological forces capable of affecting the interactions of numerous animals and plants, as well as the structure and function of ecosystems’.

History of Megafaunal Loss

What are we talking about?

Before embarking on this exploratory journey in earnest, let’s first establish some loose definitions for the temporal and descriptive characteristics that will be referenced moving forward.

What: Fundamentally derived from the Greek megalo (meaning large), and the Latin fauna (meaning animal), megafauna are simply defined as large animals. Defining ‘large’ in this context is then where complexity arises. Delimitation of megafauna tends to be according to some threshold mass of the adult animal, commonly >44 kg but varying greatly between >5 kg and >1000 kg in different studies. Elsewhere, megafauna are defined and stratified according to a trophic (relating to feeding and nutrition) herbivore-carnivore cascade. Malhi et al. divide megafauna into megaherbivores (>= 1000 kg), megacarnivores (>=100 kg), large herbivores (45-999 kg), and large carnivores (21.5-999 kg) in order to incorporate intra-megafaunal trophic interactions.

When: The timeframes to be dealt with are likely to be variably and sometimes interchangeably referred to as the late Pleistocene (from ~60,000 years ago to the beginning of the Holocene (~11,700 years ago)), the late Quaternary (from ~60,000 years ago to the present), or in terms of thousands of years before present. For the uninitiated (including myself until very recently), the following table may be of some help: 

Having, in characteristically (and arguably unnecessarily) loquacious fashion, established that megafauna are large animals and that we’re dealing with a timeframe of ~60,000 years to the present, let’s get stuck in.

History of Megafaunal Loss

Aside from the Cretaceous-Tertiary mass extinction event, infamous for its termination of the dinosaurs ~65 million years ago, megafauna experienced no significant losses until ~1 million years ago in Early Pleistocene Africa, when, coincident with the evolution of Homo Erectus into the carnivore niche space (increased human consumption of meat, perhaps in relation to expanding brain size), several carnivore lineages were locally cut short (some continued to flourish elsewhere for thousands of years) and there was a substantial decrease in proboscidean (order of mammals containing Elephants, Mammoths, similar) diversity.

The speed and magnitude of these losses was to be dwarfed by the ~1 billion land megafauna individuals lost globally in the late Pleistocene. ~50,000 years ago (50 kya) the earth hosted more than 150 genera of megafauna greater than or equal to 44kg. By  ~10 kya, at least 97 were extinct. The figure below, from Barnosky et al summarises the number and suspected causes of megafaunal extinctions on each continent in the context of human arrival and climate change by genera:

In Eurasia (northern Europe, Siberia and Alaska), megafaunal extinction occurred in 2 pulses:

•          From 45-20 ky RCBP (thousands radio carbon years before present), coincident with decreasing temperatures and dispersion of Homo Sapiens
•          From 12-9 ky RCBP, coincident with increasing temperatures and increasing Homo Sapiens populations

In North America, megafaunal extinction coincided with climate change and the arrival of advanced Homo Sapiens armed with a repertoire of advanced stone hunting tools between 11.5-10 ky RCBP, with more than 15 species becoming extinct during the Younger Dryas (a shift from cold glacial to warmer interglacial state) between 11.4 and 10.8 ky RCBP.

In the southern hemisphere (where a very significant proportion of the extinctions occurred), pending further research, chronology is less certain. Current indications point to megafaunal extinctions loosely around 45 kya in Australia, coincident with believed timeframe of Homo Sapiens expansion in the region but interestingly predating regional Last Glacial Maximum climate change. In South America, the current consensus is that humans arrived between 12.5-12.9 ky BP and that megafaunal extinction occurred sometime thereafter, probably coincident with changing climate.

Outside of these major regions, with the notable exception of Africa and Central Eurasia where hominids have been present for hundreds of thousands of years, megafaunal extinctions coincided with the global expansion of Homo Sapiens; ~30 kya in Japan, ~6 kya in the Caribbean, ~1-3 kya in the pacific islands, ~2 kya in Madagascar, ~0.7 kya in New Zealand.

Throughout the world, megafauna that persisted into the mid-Holocene and beyond tended to be found in areas where human populations never grew large – although some species, such as horses and mammoths in mainland Alaska, became extinct without significant human populations present, instead being the likely result of substantial regional climate change.

Humans as omnivorous, generalist super-predators capable of effecting and maintaining predation pressure on even the very largest animals (who had been under very little predation pressure before their arrival) had the ability to wreak havoc on megafauna, particularly vulnerable to increases in predation pressure as a result of their long lives and slow reproduction rates. Additionally, domestic animals associated with humans may have introduced diseases into endemic populations as well as competing with endemic carnivores for food. Resultant loss of keystone (to be covered in detail later) megaherbivores, leading to changing vegetation and fire regimes and loss of a food-base for megacarnivores, could then have compounded losses, setting into motion subsequent cascades of extinctions. Yet in Africa, central Asia and various other regions around the world, humans and megafauna coexisted without such drastic megafaunal extinctions as elsewhere.

Climate change has been proven to affect animals principally through the elicitation of significant, abrupt vegetation change. However, while in some areas, the timing of vegetation change coincided with various stages of extinctions, in others it did not.  Furthermore, climate shifts in the late Pleistocene are believed not to have been unusual enough to effect considerable ecological change, being neither faster nor of greater magnitude than other such shifts in the past 700 kya that didn’t result in such megafaunal devastation.

Neither human nor climate effects appear to be sufficient in themselves, but the above seems compelling evidence that a combination of human effects and climatic change (likely leaning further towards humans than climate change in the majority of cases), interacting to apply theretofore unprecedented pressure, is the likely culprit of the late Pleistocene megafaunal extinction event, characterised by short term, regional pulses.

Much of the information conveyed in this post was derived from a pair of fascinating papers, Barnosky et al. 2004 and Malhi et al. 2015, both well worth a read in their original entirety.