Extinction.
Extinction is a hot topic, of central interest to conservation, environmentalism and also to academic ecology. Ecologists are increasingly realizing that extinction has shaped the communities and ecosystems we see today, and that it operates on a range of timescales and spatial scales.
Extinction doesn't just involve the complete loss of species, vanishing from the Earth. Extinction of local populations is often of key importance in shaping communities, even though other populations of the same species may exist in the same general area.
Local extinction of populations.
In any area of forest, or patch of swamp, or coral reef, populations of animals and plants will be coming and going. If you were to compile lists of (say) all the bird species nesting in an fairly small isolated area of forest in Connecticut in one year then come back 20 years later, although most of the species in the list would be the same, the chances are that some of them would be absent (but quite likely replaced by other species). The ones that have come and gone would likely be the ones that had relatively few nesting pairs there anyway.
Why would they have gone extinct? Well, most often it would simply be that the population had a run of bad luck. Maybe there were 3 pairs nesting, and just by chance there for several years in a row there were late frosts that killed all the insects so the parents had nothing to feed their young, so the nestlings all died. Small populations are at the mercy of such 'bad luck', known as stochasticity. Even an initially abundant population (of say 150 nesting pairs) might still suffer a series of bad luck events (an exceptionally cold winter knocking back the population, followed by the late spring frosts) that can take it down from abundance to nothing at all in a few generations.
Another problem in a small population might be inbreeding. In species that normally have quite large populations and lots of exchange of genes, a range of harmful recessive genes can be present in the population without often showing themselves. Only when the population shrinks due to bad luck is there a significant chance of close relatives mating (e.g. if there are only 2 nesting pairs in the forest each year, the young are bound to end up mating with close relatives). When close relatives mate, they each have a good chance of carrying some recessive lethal genes in common; the result may be high mortality among the offspring. Experiments with repeated brother-sister crossings in domesticated animals (for several generations in a row) show that within 3 or 4 generations the young almost never survive to maturity because they have so many things wrong with them. This inbreeding effect is thought likely to hammer many isolated populations as their numbers get low, adding to the bad luck, although there are no well-documented cases of it occurring in wild populations. An apparent example of inbreeding effects appears to be the cheetah (see below), which has very low genetic variability apparently due to its numbers crashing to only 4 or 5 individuals at some unknown time in the last 10,000 years, then expanding rapidly again. Cheetahs have low fertility rates (e.g. very low viable sperm counts), and their cubs have high death rates that seem to be the result of its recent history of inbreeding. In this case, the population remained viable and recovered despite the inbreeding effects, but one can imagine that the problems might equally well kill off less lucky populations/species that reach very low numbers.
Observations suggesting the importance of inbreeding in causing near-extinction:
1) Gray Wolves of Isle Royale, Canada. The population was established by 1 breeding pair of wolves around 1949. By 1980 the population had grown to approximately 50 individuals, but by 1990 it had dropped to 14. Few pups were sighted by investigators who also noted that females were not breeding. Further study indicated that the population had probably had no more than 2 to 3 breeding pairs of wolves for from 5 to 7 generations. Based on previous studies, there was an expected loss of from 39 to 65% of the populations genetic variation, which was later confirmed at 50%. All the wolves were as genetically similar as siblings and were probably descended from a single mother. The decline in the wolves was attributed to inbreeding depression, with the lack of reproducing females attributed to the unwillingness of the wolves to mate with close relatives. It is important to note that this example is considered to be controversial in that recent reports indicate that the population has rebounded, leading some to suggest that the original population crash may have been due to some cause other than inbreeding depression.
2) The Florida Panther ( Felis concolor corgi). This subspecies of mountain lion has suffered an extreme reduction in population size due to habitat fragmentation, with numerous lions killed by cars. The current population is less than 50, and the results of inbreeding are clearly illustrated in data compares the panther’s sperm motility, sperm abnormality, etc. to other non-inbred subspecies.
3) The cheetah (Acinonyx jubatus) also displays evidence of a genetic bottleneck in the past. Studies indicate that the cheetah displays very low levels of genetic variability in two isolated populations. These low levels of genetic variability are about what would be expected after 10 to 20 generations of inbreeding. Consequences of this loss of heterozygosity in the Cheetah include: a male sperm count only about 10% that of related felid species, over 70% of sperm are aberrant in some way, and an unusual vulnerability to disease in captivity (loss of genetic flexibility).
Populations or species may thus suffer an 'extinction vortex' if their population is reduced to low levels; inbreeding and stochastic population variation may easily take the species down to extinction.
A paper on inbreeding risk in small populations: Brook, B. W., D. W. Tonkyn, J. J. O'Grady, and R. Frankham. 2002. Contribution of inbreeding to extinction risk in threatened species. Conservation Ecology 6(1): 16
Local extinctions and metapopulations.
If you look on the narrowest scale, extinction is much more common that if you look on a somewhat broader scale.
Fundamental predator-prey equations (see the earliest lectures in this course) suggest that local populations of predator and prey will often crash together, but that they prey may always be one step ahead of the predator, colonizing new patches of habitat where the predator is absent (or recently extinct!). Thus if this predator-prey model is correct, if you looked at any local population on a small enough scale (which will vary from one organism to another; say for example it is a single area of woodland in Connecticut) you would see the prey population going extinct frequently. But if you stand back and look at the whole set of interconnected populations (say on the scale of the whole of Connecticut), the prey population is always present, it just shifts around a bit. It would be the same story for the predator too of course. The 'broad scale' population is known as the metapopulation (from the Greek word meta- meaning beyond).
No one really knows if this predator-prey extinction pattern really goes on in nature. It is too difficult to study really, especially with all the 'noise' of all the pests and predators that affect each prey species. But the local predator-prey 'crash' might also be another cause (a third cause) of a local extinction of a particular species (say those particular species of nesting birds in a patch of woodland).
Sometimes, the temporary existence or long-term survival of a population can only be explained with reference to other nearby populations in the metapopulation. At the edge of a species' broad scale geographical range (e.g. where the climate becomes too cold in summer for it to complete its lifecycle), local establishment but then extinction of populations may be frequent as the weather fluctuates from one year to the next; after extinction, the populations are re-established by immigration from the nearest long-term viable populations. Also, on a very local habitat scale, populations which can't sustain themselves by reproduction may be maintained by seeds or individuals blowing in from a 'strong' population where lots of reproduction occurs. This is for example the case in some landward populations of dune plants; the constant seed rain ensures some individuals are always present, but each individual (even though it grows and may look healthy) never quite replaces itself (e.g. each individual may have only 0.9 viable offspring) so the population is always declining, and depends on new input from outside to avoid extinction.
Or say that the plant colonizes gaps just after they have formed, and quickly forms local populations, but can only survive a couple of years before competition for light from other species gets too intense and it dies out. But by then, the seeds have moved on and established a population in another gap elsewhere. The whole setup is a 'metapopulation', which shifts its boundaries from time to time as the more local populations come and go; local populations may die out, but the metapopulation lives on.
Local species richness as a balance between extinction and colonization
In a patch of forest or any other isolated patch of habitat, including an island surrounded by water, the balance between local extinction of small populations and chance 'new arrivals' from other populations elsewhere, will determine the overall species richness. Thus for example, on a very small island where all populations are very small in numbers, stochastic extinction will be very frequent. Regardless of how frequently new arrivals can establish new populations, on a small island (or patch of forest) extinctions will be more common than on a large island.
This idea was established in a tentative sense by MacArthur & Wilson's theoretical work in the 1960s (see the lecture on broad scale variations in species richness). The general idea was backed up by some classic experiments on insects of mangrove islands in Florida, although what they really studied was recolonization rates not extinction rates (after all, they had made everything go extinct by fumigating the islands).
The 'local extinction' mechanism (and also the counterbalancing mechanism of how often new populations colonize or recolonize) can thus help us to understand some of the more local-scale species richness differences we see in nature, between (say) different patches of forest, or between different nearby islands.
The minimum viable population concept, and its importance in conservation.
Quite often, the local population of some desirable species of bird or mammal that is being conserved goes extinct not long after a nature reserve is set up specifically to protect it. The likelihood of this happening has been found to be related to nature reserve size. The smaller the reserve, the more likely that extinction will occur. This is related not to size of the reserve as such, but instead the size of the population of the species that can potentially exist within it, and the chance of random fluctuations in population level which become more important as the maximum population gets smaller.
Even if you have a huge nature reserve to conserve a population that is already down to very small numbers (e.g. due to severe over-hunting) there is still a major chance that the species will go extinct before it can expand.
In a rough sort of way, it may be possible to calculate the chance of a population (starting at any given level) declining to extinction just through random bad luck. You can make a mathematical model, in which you input the 'r', and the 'K' (Logistic Equation, remember), and some plausible assumptions about how much year-to-year variability there will be in offspring survival, adult offspring production and adult death rates.
The calculation is complicated by the choice of whether you fix the starting population only, and let it drift up or down after that point, or if you assign a maximum carrying capacity and hold the species down below that level. Generally minimum viable population size calculations assume a fixed carrying capacity (just as in a nature reserve of limited size).
So usually, for a particular species, you can run the model with any starting number of individuals (e.g. 50, 25, 2500) and a similarly sized carrying capacity, and see how long it is likely to be before the population just goes extinct through pure bad luck (stochasticity). Generally, when it comes to mammals and birds, one finds that populations held below about 200 individuals are quite likely to go extinct within at most a couple of hundred years through random fluctuations in death rate and recruitment. Moving up to populations of around 1000 individuals, the survival time of the population increases dramatically (to tens of thousands of years).
Calculating the long term survival time for a population is crucial in conservation biology. Isolated populations are often what conservationists are trying to preserve; they may be the last remnants of their whole species. In deciding how much land to buy for a reserve to ensure the survival of a species, conservation agencies must generally take into account that they need to buy enough land to support a population of several hundred to ensure long term survival (though as an interim measure, enough land to support only 50 or 75 individuals will probably allow the species to survive for a few decades).
The minimum viable population concept has often been criticised as simplistic (e.g. how do you realistically allow for the average probability of all the 'bad luck' events that affect birth rate and death rate variability from year to year), but it may still be a rough indicator of what to aim for in conservation. And its most important point is that a population consisting of only a handful of individuals cannot ensure survival of the species for very long.
Whole species extinctions.
It is thought that almost all the species that have ever existed are now extinct. This is because of those species which turn up in the fossil record beyond about 5 million years ago are now almost all extinct (though many have been replaced by other somewhat similar forms). The fossil record of complex multicellular life forms extends back some 700 million years, and we can assume that for any time frame in the past, only a fraction of the species existing turn up in the fossil record at all. We assume this because looking in the present-day world, we see that many species are too rare to stand a chance of being fossilized (a rare event in itself) and found (an even rarer event), or live in environments from which hardly any fossils will ever be preserved in the long term over millions of years (e.g. in tropical rainforests, where long-term preservation rarely occurs).
If you look at (say) the entire mammal fossil record of the last 65 million years since the dinosaurs went extinct, you see species coming and going at different and fairly random intervals. On average, mammal species seem to survive about 2 million years before they go extinct, although speciation is continually producing new species from existing ones which tends to counterbalance this exinction in terms of overall diversity.
What causes individual species to pop out of existence after a few million years? Most probably it is just a combination of broad-scale bad luck events. For example, the climate may change so the potential habitat range of a species is shrunk down from covering most of a continent to being confined to a small peninsula, and then a volcano happens to go off and wipe out all the remaining population on that peninsula. Or a competitor species may evolve a trait that means it can push this particular species down close to extinction; then a random run of cold years finishes it off.
If we look back at detailed climate indicators from the geological past, we see lots of signs that the environment was always changing, offering plenty of opportunities to obliterate species, even if we have no fossil record of those species. On a local scale a river shifts its course and floods the soil on which the last population of a particular species of plant was living, or a tidal wave obliterates a coastal forest containing the only population of a particular bird species.
On the broad continental scale on the timescale of tens of thousands of years, we see abundant signs that temperature and rainfall have fluctuated up and down, often involving sudden climate changes (perhaps only taking a few centuries or even decades to occur). In the last 2 million years, the Earth's climate has gone especially crazy and has been through a large number of big temperature swings ('ice ages') marked by very rapid changes; taking only a few decades to flip from one climate phase to another.
In addition there was probably often a more subtle (but difficult to detect) effect of the jockeying for position in ecosystems...for example a competitor species evolves a better strategy and wipes out the older species. Sometimes paleontologists think they see the effect of a new species or set of species evolving a trait and wiping out others as a result; but the cause-and-effect relationship is always difficult to establish, without having a time machine to go back and study the detailed ecology!
Probably both factors (environmental change, and evolution of predators, prey and competitors) contribute to such extinctions. Whatever its causes, random scattered extinction of species in the fossil record is called 'background extinction'.
It is probably in part the variability in the natural environment which is the reason for sexual reproduction, and why also every species has some maintained form of sexual reproduction instead of just cloning itself (asexual reproduction) - much easier to do in the short term, and less risky. Plants and animals must produce a range of gene combinations among their offspring to cope with these occasional unpredictable challenges, or the genetic line will eventually go extinct because it fails to meet some challenge. Even so, eventually on a timescale of millions of years it will probably fail to change fast enough, and it will go extinct...though in the meantime it may have given rise to other quite distinct species which will continue the genetic line.
This view of environmental change maintaining sexual reproduction and also causing background extinctions is called the 'Red Queen' view of the world, from the character in Alice in Wonderland who must 'always be running just to stand still'. The 'Red Queen Hypothesis' is the view that continuous environmental variability is the basis for maintaining sexual reproductive systems (and why most animals and plants don't just rely on cloning themselves), and why there is always a 'background rate' of extinction of species that don't quite manage to change their genetic profile fast enough to meet the challenges.
Mass Extinctions
Against this broad pattern of Earth history are phases when suddenly and more-or-less simultaneously, a large proportion of the animal species on Earth suddenly disappeared, leaving greatly depleted faunas that slowly recovered their diversity over millions of years as new forms evolved to exploit the gaps.
It is assumed that these mass extinctions, as they are called, were caused by environmental changes. But what sort of changes, and how suddenly did they occur?
The best known of the mass extinctions is the big end-Cretaceous mass extinction. The majority (about 70%) of animal families (on land and in the sea) and also many plants and zooplankton went extinct right at the same time 64 Myr ago. Almost nobody now disputes that this mass extinction was caused by a meteorite impact on the Earth, causing a range of effects; raining nitric acid (from combustion in the atmosphere of nitrogen and oxygen) onto the earth and into the ocean waters, cutting off the sun with a dust cloud for months or years. Soot from the fires that swept the Earth shows up everywhere in sediments of that age.
There was an even bigger mass extinction at the end of the Permian around 230 million years ago; 90% of the animal (vertebrate and invertebrate) familes went extinct. To wipe out 90% of the families, you need to eliminate nearly 100% of the species (since there are many species in each family). What caused it? No-one really knows. Some say a gradual catastrophe unfolded when plates closed in a ring to form a stagnent ocean basin, that released toxic hydrogen sulphide into the air in massive quantities (H2S is as toxic as cyanide gas). But just recently, geologists examaining sediments of end-Permian age found tell-tale cometary gas isotopes that suggest it was a comet that hit the Earth and caused the mass extinction.
Other smaller mass extinctions are also problematic. No-one really knows their causes, though rapid climate changes unrelated to meteorite impacts are a possibility. A mid-Tertiary mass extinction of most forms of land mammals seems to relate to a sudden brief warming event, after warm salty Equatorial water flooded the world's deep oceans.
Another problem with investigating most mass extinctions in the fossil record is that because the record is so 'bitty' going hundreds of millions of year back in time, that you can't tell if the various life forms died out in a single year, or over (say) five million years. The end-Cretaceous mass extinction is an exception because (being relatively recent in geological terms) the fossil record is fairly detailed and the dating is precise. Here at least, we can say with reasonable confidence from the speed with which species disappear from the fossil record that the mass extinction really was sudden (probably occurring within a few tens of thousands of years at most, possibly very much quicker like just a few years). For other mass extinctions, we really don't have much idea how sudden or not-sudden they were.
Mass extinctions 'free up' niches. As I mentioned in a previous lecture, mass extinctions 'free up' niches, allowing other species to evolve into them eventually. Mass extinction is thus of great importance in community structure: for example it is probably the only reason why mammals rather than dinosaurs are the dominant large herbivores and carnivores on land. Another example is how weird bivalve molluscs called rudists were taking over coral reefs in the late Cretaceous, slowly but relentlessly speciating into new niches and pushing out corals (the reef-building corals were probably mostly heading for extinction as they were losing their niches). Then rudists went extinct in the same mass extinction that killed the dinosaurs. Corals made a dramatic comeback and have been dominant in tropical reefs ever since.
Ice ages and extinction in the last few million years.
We get better clues to the relation between environmental change and extinction when we reach the last few milion years, because the fossil record is more detailed.
Coincidentally, the last 2 and a half million years have also been a time of frequent, dramatic fluctuations in the Earth's climate. This phase is known to geologists as the Quaternary. Every few thousand years, the Earth's climate has flipped from a colder drier state to a warmer, wetter state, or back again. The size of these jumps differs: the biggest ones would take the climate of Connecticut down to Georgia. It seems that many of these transitions have occurred over less than a single human lifetime: several decades, at the most.
What effects have these changes had on the species around the world? Have they driven many things extinct, as we might expect? Well, there are some signs of such effects:
As an example we look at patterns in temperate tree species richness, something I mentioned in a previous lecture with respect to spatial patterns between different regions. But here we'll look at it more from a temporal perspective.
A few million years before the start of biggest climate jumps, there were very diverse temperate forests throughout the Northern Hemisphere. This ancient flora is known as the 'Arcto-tertiary flora'. Forests in many places across Asia, Europe and North America were rich in things like giant redwoods, ginkgo, tulip trees, and other types of trees that now only have restricted distributions, as well as the still-common things like maples and oaks.
The start of the dramatic climate instability of the Quaternary eliminated many 'exotic' kinds of trees from Europe and North America. The diversity of the tree flora in each region declined. More genera of trees dropped out of European flora with each major glacial phase. European populations of giant redwoods lasted a long time, but finally succumbed about a million years ago. Their Californian relatives will still grow really well if planted in present climates in several parts of Europe: suggesting that it was perhaps 'bad luck' rather than any competitive inferiority that wiped them out. This is also true of many other old Arcto-Tertiary trees when re-introduced to Europe.
In contrast, East Asia (e.g. China) retained much of its original Arcto-tertiary diversity, at least in terms of genera (though probably many individual species were lost). A few things, like giant redwoods, were also lost from Asia however.
As I mentioned in my previous lecture, the European tree flora is now 'under-saturated' with species (if one compares the rising species richness curve of Europe with east Asia), presumably because it has repeatedly been hammered by cold, dry glaciations. Same for North America too, apparently.
The fossil record shows certain other effects of the last 2.4 milion years of climate instability on biodiversity. The onset of big sudden cold phases 2.4 Myr ago apparently resulted in a burst of extinctions in corals and other shelf sea marine life; around 75% of sea shell species in the Caribbean died out suddenly at that time (due to cooler sea temperatures?); the climate shift also seems to have caused widespread mammal extinctions on land at about the same time.
It's difficult to know what happened in other areas that are less well known. e.g. the tropics. If there were mass extinctions in rainforests from global cooling and onset of glacial phases, we have a pollen record that is too poor to see them.
The relative lack of extinction in the more recent fossil record
The closer one gets to the present-day, the better the fossil record becomes (because fewer of the fossil-containing sediments have been lost to erosion). When it is the last two million years or so, we have a great abundance of data on sea shells, mammals and temperate-zone trees. Getting closer to the present, for the last 50,000 years we even have a large amount of data on more fragile things such as insects in certain regions.
A surprising feature of most of the latter part of the last 2 million years (e.g. the last million years or so) is how few species went extinct as a result of those repeated big, severe climate changes. If one looks at the diverse sub-fossil beetle fauna known from the British Isles over the last 50,000 years or so (a detailed record only goes back that about far), of nearly 2000 insect species, all of them survive in the present-day world somewhere or other (though many of these types of beetles now survive only in other parts of western Eurasia...and one species is now confined to the Himalayas!).
Corals and sea shells also survived the last million years with almost no extinction, despite large and rapid shifts in sea level and sea temperature that must have seriously disrupted their ranges and left populations reduced and isolated (think how reefs would be drowned by a sea level rise of metres per century, carried on over thousands of years; that is what has happened many times over the last 2 million years).
One hint of what glaciations did or did not do to the tropical forests comes from the Queensland rainforest of northern Australia. One can compare the pollen flora from previous warm phases before this one (e.g. there was a very warm phase 130,000-100,000 years ago). Only 2 genera of trees have gone extinct during the upper Quaternary (last 200,000 years) despite severe reduction in forest areas. So maybe relatively little has been lost with each climate fluctuation - the forests remain very species-rich despite those Queensland rainforests being combined to very small scattered refuges in river gulleys or on localised flood plains during glacial maxima (though many species don't show up in the pollen record, it certainly suggests that each of the last few major glacials did not just kill off a great swathe of the species that had happened to survive the previous glacials).
And the Queensland forests were apparently hit relatively severely by arid phases. Maybe even fewer species (relatively speaking for size of the total flora) were lost from Amazonia or SE Asia with the glacials? (the fossil pollen record from these other tropical regions hasn't yet been looked at in enough detail to tell).
So in as much as we can tell, there were quite a few extinctions at the start of the big climate fluctuation phase 2.5 million years ago (the Quaternary), but generally speaking less extinction later on. All this makes you wonder; maybe the species we have here in the present-day world are natural 'survivors'? Maybe there is something about their biology, their ecology, or their reproductive systems that means they can survive the frequent rapid 'bottlenecks' that climate fluctuations bring about? The really sappy species were perhaps the ones that were lost at the beginning of the Quaternary.
If we believe this idea, what is different then about the survivors from the floras and faunas before 2.5 million years ago? It's very difficult to say. Maybe they have seeds or behavioural patterns that allow them disperse easily across open landscapes and recolonize quickly when the climate turns favourable. Maybe they aren't bound up in too many over-specialized symbioses with insects or fungi or food plants that might go wrong if the local population of their 'partner' goes extinct.
Maybe the 'survivors' have breeding systems that ensure wide cross-fertilization, compared to the broader set of species that were present before the start of the Quaternary? Rapid exchange and recombination of genes is a key to surviving a variable environment...with a wide enough repertoire of genes among your descendants, your lineage can duck and dodge any punches that the world can throw at it. Perhaps sometimes 'survivor' lineages have higher mutation rates, randomly generating new genes?
Despite having been knocked back to very local populations during cold, dry phases, many temperate tree populations in Europe show suprisingly high levels of genetic diversity within populations. Perhaps this is because the trees are frantically mutating and outcrossing their seeds on the offchance that a sudden climate change will come along and wipe almost everthing out, except the individuals that have some special chance combination of genes?
Really though, we don't know what is different or what has changed about the 'survivor species' of the late Quaternary that has helped them from going extinct!
Megafaunal extinctions
(this is a subject I mentioned in a previous lecture, but I'll reiterate it here)
But one group (large mammals) has suffered increased extinction towards the more recent part of the Quaternary. A wave of extinctions occurred toward the end of the Last Glacial phase. This began in Australia; between around 45,000 and 40,000 years ago, a diverse fauna of large marsupial mammals and birds dropped out of existence.
My Australian students managed to save this last specimen of the megafauna.
There was a similarly big extinction of large mammals in North & South America at the end of the Glacial between about 13,000 and 11,000 years ago. And a lesser burst of extinction occurred about the same time across northern Eurasia.
Was it climate change that caused extinction in these cases? Certainly there was climate variability in Australia between 40,000 and 25,000 years ago, but this was relatively minor compared to many other changes that had occurred previously during the last 2 million years. The large mammals apparently dropped out during a relatively 'favourable' stage for grazing mammals populations it would seem...one that was not expecially arid or cold, with lots of grassland around.
It was a similar story in North America. The megafauna (mammoths, American camels, American horses etc.) had just survived a glacial maximum (a really cold phase). Conditions warmed, they hung around for a while, and then they went extinct. The signs from the geological record are that North America had frantic climate variability during the Quaternary...these species had survived some real climatic shocks before. So why did they go extinct when they did?
In both Australia and the Americas, extinction seems to coincide with arrival of humans; around 45,000 years ago in Australia, a bit later (around 14,000 years ago) in the Americas. It seems almost beyond doubt that something the humans did - probably hunting - wiped out those species. Maybe it was the coincidence of human hunting pressure with some degree of environmental change (thus, a combination of factors) that pushed their populations over the brink. Some population modelling studies on mammoths (using modern elephants as the analogy for potential rate of natural increase) suggest that at a time when populations were reduced due to a reduction in food supply, a bit of hunting pressure could easily drive their numbers down to extinction.
In Europe/Eurasia humans had been around on-and-off for millions of years. So why did the mammals and woolly rhinos go extinct there too at the end of the Last Glacial? Possibly again a combination of reduced food supply related to climate warming, and hunting pressure from more advanced hunting techniques (human cultures were getting much more complex around the late Glacial).
These species and genera had survived many other similar changes over the last several hundred thousand years. The big difference this time was the presence of modern humans.
The megafaunal extinctions may be a foretaste of how in our modern world, with the 'squeeze' on each species provided by humans overhunting, overfishing and harvesting wild populations, can make them more likely to go extinct when climate change occurs.
The 'species-pump' hypothesis; hardship and near-extinction can create as well as destroy species.
If you break up ranges and shrink populations down, independant evolution of those isolated populations may eventually give more species overall, diverging from the 'parent' species. e.g. this may be the source of the incredible Eucalyptus species diversity in Australia; more than 500 species. So maybe range reduction, bottlenecks and fragmentation balances out in terms of effects of overall diversity (often cretating as many or more species than it destroys)? It is difficult to say, really.
Many evolutionary biologists suggest that all significant evolution occurs in small, isolated populations. But these usually exist at edges of a widespread species range anyway!
Some thoughts on extinction and the future 'greenhouse' world.
It seems unlikely a mass extinction will result from global warming over the coming centuries. Most species have weathered many drastic changes in the past (all those resistant late-Quaternary species).
The projected greenhouse warming seems a bit too much like the sort of patterns which been happening during Quaternary. But hunting or additional habitat reduction from humans might well push a lot more species over the edge. Especially animals. We also have to consider the interactions with direct CO2 effects on ecosystems as mentioned in the previous lecture; again, this combination of effects could be much more severe than climate change alone.
Summary.
1) Extinction occurs constantly in local populations, due to bad luck (stochasticity) and also probably inbreeding in small populations. This affects local patterns in species richness (e.g. on small islands, localised extinctions are more frequent than on larger islands).
2) There is 'background extinction' of individual species, scattered through the fossil record, and also 'mass extinction' when many species disappear at around the same time.
3) There is some evidence that climate change helped to cause some extinctions over the last 2 million years, but more recent species seem 'selected' to survive climate change quite well. Humans are a complicating factor in the last 50,000 years or so, apparently increasing the likelihood of extinction (especially when climate changes occurred).
..........
Deforestation
Throughout history, the fate of the world's forests has strongly reflected the pattern and intensity of land use by societies. Demand for agricultural land, timber, and other forest products, as well as technological change in agriculture, significantly impacts the mode and rate of transformation of forested areas. Biophysical triggers may also play a role, such as fire dynamics, which are linked to agricultural activities or natural phenomena such as ENSO droughts. These demands are often linked to present-day developing countries experiencing deforestation, which will be the focus of much of this chapter. It is worth noting, however, that technological changes in agriculture (e.g., development of sod-busting plows that opened up the American mid-West) contributed significantly to a "forest transition" in many European countries during the 19th and 20th centuries, in which net national forest cover stopped declining and began to increase (Mather 2001). Thus, technology cuts both ways, leading in some regions to declines in forest cover, and in others to increases.
Society's special interest in deforestation, as compared to other land use/land cover change issues, may be partly attributable to the stark nature of the transition from forest area to cleared land. Deforestation occurs relatively quickly, and in contrast to some other transitions (e.g., from crop land to pasture, or from productive land to degraded land), is easily observable by the human eye. Through the use of remote sensing technologies, large areas can be monitored, and estimates of deforestation can be obtained (see Section 3.1 for a short description of land cover change monitoring methodologies).
Deforestation has a number of repercussions, many of which are dealt with in separate chapters of this thematic guide:
Deforestation can lead to soil erosion or impoverishment, especially in tropical areas where soils tend to be thin and nutrient-poor.
Deforestation is linked to habitat loss, which is a leading cause of species endangerment and biodiversity loss, particularly in humid tropical forests.
Deforestation affects the hydrological cycle through changes in evapo-transpiration and run-off; and
Deforestation, and particularly forest burning, contributes to green-house gas emissions that bring about climate change.
Despite its apparent ease of detection, deforestation rates are still a matter of some debate. Section 3.1 addresses monitoring of land cover change, and the extent and rate of deforestation in temperate and tropical areas. Section 3.2 addresses the causes and processes of deforestation, drawing on a review of the most recent literature. Finally, Section 3.3 addresses potential policy interventions.
3.1 The Extent and Rate of Deforestation
Today, roughly 39 million square kilometers (29 percent) of the world's land surface is under forest cover (FAO 2000), and of that 28 million square kilometers is in so-called "closed forests" of 40 percent canopy cover or above (Singh et al. 2001). Since the end of the last ice age, approximately half the world's forest cover has been lost, most of it due to the expansion of human activities and settlements (Kapos 2000). In terms of primary forest, in contrast to secondary or other successional forests, much less remains. The World Resources Institute (1997) estimates that only one-fifth of the world's original forest cover remains, largely in blocks of undisturbed frontier forests in the Brazilian Amazon and boreal areas of Canada and Russia.
Measuring the extent and rate of deforestation is not as simple as it might at first appear (Singh et al. 2001). The first challenge is to define what is meant by a "forested area." In other words, what density of tree cover is required for an area to be considered a forest? Figure 3.1 shows a "continuous fields" tree cover map prepared by the Global Land Cover Facility (DeFries et al. 2000). This maps shows that far from being homogenous, land areas can vary from 10 to 100 percent forest cover and still be considered forests.
Once a threshold is defined, whether it be closed forests (i.e., trees with interlocking crowns and a canopy density of 40% or above) or open forest (i.e., 40% crown cover or less), the next challenge has to do with how forest cover change is monitored. For smaller areas, it may be possible to do a parcel-by-parcel inventory to determine rates of change. However, for large or inaccessible areas such as the Brazilian Amazon, the only realistic approach is to utilize remotely sensed imagery (generally from satellites, but also from airplanes). This requires, then, that the analyst has at least two sets of images, one set proceeding the deforestation event or events, and one set following.
The next step is image processing. Processing requires a classification of both sets of images (i.e., breaking the continuous field data into discrete categories such as forest, road, crops, pasture, etc.), and then a change matrix in which the analyst computes the change from one of the land use/land cover categories into other categories. In this way it is possible to obtain the percentage of land area that was forest and is now in one of several other types of land use. Note that an added difficulty, particularly with imagery from so-called "passive" sensors (sensors that rely on the sun's illumination), is that it is vital to obtain relatively cloud-free imagery, or else large areas may be obscured by cloud cover. This is a particular challenge in the humid tropics. Radar, or active sensors that bounce an energy pulse off the land surface, are being used in such zones with some success.
Because remote sensing imagery is expensive to acquire and to process, generally deforestation studies limit themselves to some sample area, say a sub-national administrative unit or a well defined geographic area. Thus, efforts must be made to obtain a random sample of forested areas, or else the estimates of deforestation will be biased.
The Global Forest Resources Assessment of the U.N. Food and Agriculture Organization (FAO) is a major assessment that has developed estimates of deforestation at the global, regional and national levels once every 10 years since 1980. For its 2000 assessment (also known as FRA 2000), the FAO utilized a relatively low threshold for forest cover of 10 percent minimum crown cover (FAO 2000). The assessment is based on a combination of reports by national authorities, and a 10 percent sample remote sensing survey for tropical areas.
Before presenting results of the FRA 2000, it is important to note that there is disagreement about the assessment's results, and even some recognition within the FRA report itself about the limitations of their methodology. A briefing paper by the World Resources Institute (WRI) identifies a number of potential problems with the FRA 2000 (Matthews 2000):
Methodological changes for each assessment since the first one in 1980 make comparisons to past assessments difficult (and therefore estimates of deforestation rates subject to uncertainty);
The use of self-reported data by countries is criticized on two grounds: (1) countries may have incentives to underestimate deforestation, and (2) data and monitoring systems in most countries are generally inadequate; and
The report relies on remote sensing surveys that cover randomly scattered plots in the world's forest areas. According to WRI, because deforestation is not randomly distributed, but tends to proceed outwards from transportation corridors, a 10 percent sampling rate is insufficient to identify how much forest is being lost.
There is a further concern that the 10 percent crown cover threshold includes lands that most non-specialists would consider to be tundra, wooded grassland, savanna or scrubland, not forest.
Interestingly, despite WRI's concerns that deforestation rates are being underestimated by the FRA methodology, the Tropical Ecosystem Environment Observation by Satellite (TREES) initiative has recently arrived at estimates of deforestation rates in the humid tropical domain (i.e. closed forests) that are 23 percent below the estimates developed by FRA 2000 for the same time period and type of forest (Achard et al. 2002). Furthermore, Steininger et al. (2001) found in their "wall-to-wall" remote sensing study of the Bolivian Amazon that the rate of deforestation is almost four times lower than that reported by the FRA 2000. According to FAO sample survey estimates from 1981-1990, annual forest loss in Bolivia was proceeding at the rate of 5,810 square kilometers per year, whereas the estimate based on wall-to-wall remote sensing coverage for the period 1987-1993 was only 1,529 square kilometers per year. Nevertheless, owing to different baseline figures, the FAO estimated a slightly larger remaining forested area (483,100 sq. km. in 1995) than did Steininger et al. (437,904 sq. km. in 1994).
Bearing in mind, then, some of these methodological issues and difficulties inherent in establishing firm deforestation rates, FRA 2000 results at global and regional levels are shown in Table 1. From the assessment, some interesting patterns are evident. The two most forested land areas are the European republics of the former Soviet Union (including Russian Siberia) and South America, each with just over 22 percent of global forest resources, and each with approximately half of their land areas under forest cover. The regions with least forest cover are Asia (due to land conversion for agriculture and large desert areas) and Africa (largely due to deserts). The highest changes in forested area were Africa and the Caribbean, each losing close to 1 percent of their forest cover over the decade. In contrast, most temperate and developed regions saw net growth in forested areas of between 0.1 and 0.3 percent.
Collectively, the Forest Resources Assessment, TREES and Global Land Cover 2000 (a recent initiative which has yet to publish deforestation statistics) contribute to our understanding of deforestation patterns and dynamics, and provide firmer basis for decision-making.
3.2 The Causes and Processes of Deforestation
This section is based upon on a recent study conducted by the Land Use and Land Cover Change (LUCC) project on the causes of tropical deforestation, which is the most complete examination of the topic to date (Geist and Lambin 2002, Geist and Lambin 2001). The study took the form of a meta-analysis - a statistical analysis of numerous case studies to examine patterns and processes of deforestation in many locations around the world. A phenomenon with as much local differentiation as land use and land cover change requires an over-arching analysis of individual case studies if we wish to generalize the findings and come up with policy recommendations.
In thinking about the processes of deforestation, it is useful to draw a distinction between the proximate causes and underlying driving forces. Proximate causes are human activities or immediate actions at the local level, such as agricultural expansion, that originate from intended land use and directly impact forest cover. For example, a proximate cause might be a farmer's decision to clear a plot of land for pasture. That decision, in turn, is embedded within a context, such as economic incentives and disincentives, government policies, access to markets, land tenure systems, and the socio-cultural environment in which the farmer lives. These constitute the driving forces - that is, the fundamental social processes that underpin the proximate causes, and that may operate at much broader scales.
The LUCC project meta-analysis examined 152 sub-national case studies - 78 from Latin America, 55 from Asia, and 19 from Africa - covering a time period from 1880 to 1996, with the majority of case studies falling in the fifty year period from 1940 to 1990. To be included, studies needed to quantify the rate of forest cover change, include quantitative data analysis or in-depth field investigations, consider clearly named factors as potential causes of deforestation, and be absent of obvious disciplinary biases. The study focused on four proximate causes: infrastructure extension, agricultural expansion, wood extraction, and other causes (e.g., predisposing environmental factors, biophysical factors, and social disruptions such as war and population displacements). These, in turn, were related to a number of underlying drivers which were subdivided into demographic, economic, technological, policy, institutional, and cultural factors (see Figure 1).
The study refuted two broad schools of thought that had hitherto dominated the debates about deforestation. One of them held that deforestation is the result of single-factor causation, such as shifting cultivation or population growth. The other school held that the causes behind deforestation are irreducibly complex. In other words, that correlations among deforestation and multiple causative factors are many and varied, revealing no distinct pattern.
What the meta-analysis revealed was that tropical deforestation is driven by identifiable regional patterns of causal factor synergies, of which the most prominent are economic factors, institutions, national policies and remote influences (at the underlying level) driving agricultural expansion, wood extraction, and infrastructure extension (at the proximate level).
3.2.1 Proximate Causes
In terms of immediate causation, tropical deforestation is best explained by multiple factors rather than single variables. Globally, the most prominent "triad" is agricultural expansion coupled with wood extraction and infrastructure expansion. These three factors combined were present in 25 percent of the 152 cases examined. Subsets (agriculture & wood, agriculture & infrastructure, and wood & infrastructure) were present in an additional 36 percent of cases. Agriculture leads the lists of causes. The expansion of cropped land and pastures is present, generally in combination with other causes, in 146 of 152 cases (or 96 percent).
Under these three broad categories - agriculture, wood extraction and infrastructure - it is possible to identify important subcategories. For example, within the category "extension of agricultural lands," permanent cultivation and cattle ranching were present in 48 and 46 percent of cases, respectively, whereas shifting cultivation was found in 40 percent. Under infrastructure, transportation extension (road building, railroads and water ways) was present in close to two-thirds of all cases. Settlement and market extension were less prominent, at just over a quarter of all cases. And, under wood extraction, commercial exploitation of forests outweighed fuel wood extraction almost two to one, with 52 percent and 28 percent of the cases respectively. Considering all the detailed categories, permanent cultivation, transport extension, and commercial wood extraction predominate, each being present in 50 percent or more of the cases.
There are some regional differences among the proximate causes. In Asia, agriculture-wood (22%) and agriculture-wood-infrastructure (38%) causes dominate, partly as a result of state enterprise forest exploitation and subsequent settlement of those areas by poor subsistence farmers. In Latin America, agriculture-infrastructure (32%) and agriculture-wood-infrastructure (19%) are predominate causes of forest loss. In Africa, all four factors (agriculture-wood-infrastructure-other) are found in 26 percent of all cases, with agriculture-other (16%) showing up also significant. The "other" in these cases includes civil wars and population displacements.
3.2.2 Driving Forces
At the aggregate level, it is striking that combinations of synergetic drivers rather than single drivers are associated with tropical deforestation. Eighty-eight percent of the cases are driven by multi-factor terms of causation, and the largest proportion of all cases (36%) includes some elements of each of the five major factors - economics, institutions, technology, culture, and demographic change.
Economic factors are present in 81 percent of all cases, and clearly dominate the underlying causes. Commercialization and the growth of mainly timber markets as well as market failures are frequently reported to drive deforestation. Low factor costs (for land, labor, fuel or timber), price increases for cash crops, and the "ecological footprint" of remote urban-industrial centers through the demand for raw materials underpin about one-third of the cases each. With few exceptions, factors related to economic development through a growing cash economy show little regional variation and, thus, constitute a strong underlying driving force of deforestation.
Institutional factors such as policies on land use and economic development (especially as related to colonization), transportation, or subsidies for land-based activities are found in 78 percent of the cases. Many of these policies directly or indirectly promoted the exploitation of resources in forest frontier areas. Lack of adequate governance structures, as manifested by corruption, lawlessness, cronyism, and mismanagement of the forestry sector, were found to be important institutional factors (42 percent of all cases). Land tenure and property rights issues, which are frequently highlighted in the literature on deforestation, showed up primarily in Asia (60% of Asian cases). Issues of open-access resources and squatting by landless farmers showed up in approximately one-fifth of all cases. So-called "land races," in which settlers clear forest in order to claim legal title to the land, were present in 13% of all cases, mostly in Latin America.
Technological factors in the wood and agriculture sectors, in combination with other driving forces, constitute the third most important driver, underlying 70 percent of all cases. Technological changes in the forestry sector in the form of chain saws and heavy equipment, and in wood processing, are associated with deforestation in 45 percent of all cases. Asia, in particular, was found to have a significant incidence of inappropriate logging technologies. Agro-technological factors were present in a similar proportion of cases, but the picture is complex and does not provide an easy-to-generalize pattern. Modification of farming systems through intensification (high-input, labor-intensive agriculture) and extensification (low-input, large area cultivation) was present in one-third of all cases; thus neither intensification nor extensification does a particularly good job of explaining deforestation in all cases.
Cultural factors were present in two-thirds of all cases. These include attitudes and perceptions such as unconcern for forests due to low morale and frontier mentalities, lack of stewardship values, and disregard for "nature." Such attitudes were more widespread in the Asian and Latin American cases. In parts of Asia (Thailand, Malaysia and Indonesia) and Latin America (Amazon lowlands, the Petýn region of Guatemala, and Costa Rica) forest colonization is or has been viewed as important for national land consolidation, security, unity and military defense. In a more limited number of cases in Latin America, forest frontiers were viewed as an important safety valve to forestall land reform in more populated areas. Household-level behavioral factors were present in over half of all cases with less regional variation. These include profit-orientation of actors (both local settlers and absentee landlords), traditional or inherited modes of cultivation or land-exploitation, and a commonly expressed sentiment that it is necessary to clear the land to establish an exclusive claim.
Finally, demographic factors such as natural increase or in-migration were explicitly mentioned in 61 percent of all cases. Most of its explanatory power tends to be derived from interlinkages with other underlying forces, especially in the full interplay of all five major drivers. Many cases did not specify beyond broad notions of population pressure and growth, but those that did tended to identify in-migration more frequently than natural increase. The authors also investigated the utility of the I=PAT (impact=population x affluence x technology) formulation used by Ehrlich and Ehrlich (1990) in explaining cases of deforestation. They found that in 46% of all cases P, A and T, broadly speaking, operate together in a synergetic driver combination. However, in 93% of theses cases, policy and institutional factors (which are left out of the I=PAT formulation) operated along with, or were even causative, of the PAT variables.
3.2.3 Conclusions
Deforestation is a complex, multiform process which cannot be represented by a mechanistic approach. Mechanistic models are built on the belief that we know the processes by which a system operates and that individual processes can be modeled using scientific laws, or rules, described by simple equations. Given the large number of interacting factors driving deforestation, and given interactions at different levels of causality (underlying forces, trigger events, mediating factors, proximate causes) only a system approach seems appropriate. System models are mathematical descriptions of several complex, interacting processes.
While the development of a "universal model" of deforestation is probably out of reach, a collection of specific models which represent the particular interactions between a reduced set of dominant driving forces for a given process of deforestation, specific to a geographic situation, is feasible. Some of the place-specific processes that could be modeled include subsistence agriculture, commercial agriculture, colonization activities, or logging, and some of the geographic situations in which different bundles of causal factors predominate include forest frontiers, roadside areas, and peri-urban areas.