- Oplysninger
- Spørg Stuwie
Species_species
Biologi
Københavns Universitet
Kommentarer
Forhåndsvisning af tekst
3 Species Pools
Definition. Data availability. Kinds of pools. A general model. Immigration and local extinction. Examples. Experiments. Invasive species. Measuring Pools.
the list of parts
A list of parts is the obvious beginning for any problem of assembly, be it the assembly of a bookcase or the assembly of mammals in a national park. The list of parts for community assembly is the pool, which is defined as the list of species which could occur in a particular eco- logical community (e., Diamond 1975, Wiens 1983, Keddy 1992, Pärtel et al. 1996). The size of this pool is determined by speciation and extinction rates (recall Figure 1) and, locally, to a lesser degree, by immigration. Since rates of speciation and extinction are generally slow relative to rates of ecological processes that are driven by filters, we may treat the pool for a habitat as a more or less stable quantity. This may frustrate those who think in terms of evolution, but the failure to distinguish among rates of different processes is perhaps one reason for the confusion in some areas of ecology today. We must of course acknowledge that the pool of species on Earth has changed in numbers and composition through time (e., Niklas et al. 1983, 1985, Levin and King 2017). However, we currently occupy a specific point in time, in which case the changes in the past, and the prospects for the future, can be set aside, allowing us to treat the pool as a constant. With this simplifying assumption, the numbers of species that occur in each pool provide a body of data that is worthy of study in its own right. We actually have a huge database of species pools for many kinds of organisms and many parts of the world, even if this database does not explicitly mention the word “pool.” This huge database is the
76
world collection of field guides and identification manuals describing the flora and fauna of different regions of Earth. To illustrate, consider Lanark County, where Paul lives. Species pools can be found in the following sources:
Birds: A Field Guide to the Birds (Peterson 1980) Reptiles and amphibians: Amphibians and Reptiles of the Great Lakes Region (Harding 2006) Plants: A Checklist of the Flora of Ontario: Vascular Plants (Morton and Venn 1990), which has been updated with new records to provide the Southern Ontario Vascular Plants Species List (Bradley 2013).
These books (Figure 3) are based upon vast numbers of obser- vations by people who are not community ecologists, by and large. They have been verified by large numbers of naturalists and biologists who have used these guides to identify and enumerate the species. Another piece of good news is that the work is already done. That is to say, we don’t have to start from scratch. We simply take the product off the shelf and use it. Now, if our budget is tight, we can buy a second-hand copy from a local used bookstore. Of course, there are some complications (there always are).
figure 3 Biologists have compiled guides to the plants, animals and fungi for many parts of the world, and these are a foundational source for information on species pools. These three guides are useful for Lanark County, Ontario, current home of the first author.
the list of parts 77
pool is half what we might gather from looking at this page in the guide. Here, again, the issue of scale matters: we need a list that includes only those species known to breed in this ecological region.
figure 3 The species pool provided by the Ontario flora includes plants that occupy very different habitats, ranging from deciduous forests in the south to tundra in the Hudson Bay lowlands. This inclusion of many habitats complicates the search for species pools in published sources.
the list of parts 79
In the realm of reptiles and amphibians (Harding 2006), we fare somewhat better on this criterion, since instead of all of eastern North America we need only to focus on those species that breed near the Great Lakes. This reduces the large number of species from the pool for all of eastern North America. However, the problem is not entirely solved, since eastern Ontario is at the northeast extreme, and many species that occur south of the Great Lakes are not found here: no Box Turtles, no Bog Turtles, no Glass Lizards, no Timber Rattlesnakes. Returning to plants, the pool for Ontario combines many eco- logical regions (Figure 3). We need smaller regions, preferably (but rarely) based on ecological boundaries. Many parts of the world have regional lists. Here, we have the Lanark County flora as an online database (White 2016). It informs us that there are 875 species of native plants in Lanark County. This is still a rather large number, but only about one-third of the pool in Ontario as a whole. We can also categorize them by ecological requirements, since another document (Oldham et al. 1995) assigns a wetness score for each species in the Ontario flora. For Lanark County, 390 species would be wetland plants (i., they have a wetness score of −3 to −5). So, for Lanark County, nearly half of the pool of plants occupies wetlands. Thus, as a first approximation, half of the flora occupies uplands, and half of the flora occupies flooded areas. We can also pull out individual groups, such as trees. There are 48 species of trees; of these, only 8 occur in wetlands. The published list for Lanark County is therefore quite close to the true pool of species that might be expected to occur in local communities of Lanark County. Of course, the issues of scale remain. Less than 100 km to the south are additional species of trees, including Pitch Pine (Pinus rigida) and Chinquapin Oak (Quercus muehlenber- gii). Perhaps we should consider these as candidates in the species pool. Actually, in this case it is unlikely, since extreme cold periods in the winter set a rather strict limit on trees in this part of the world. But the question needs to be asked: what about species not present, but nearby. And just how close is “nearby”? And, yet another question,
80 species pools
1975). At the largest scale lies the list of species in the ecological region, the upper limit set by evolution and dispersal. Ecologists call the length of this list “gamma diversity” (Magurran and McGill 2011). In between lie a range of nested and overlapping possibilities. Existing studies on species pools tend to fall into two areas of concern: some focus on predicting alpha diversity and gamma diversity, while others focus more upon predicting composition (Zobel 2016). Our book focuses mostly upon the latter problem, predicting composition of ecological communities. If you can predict composition, then richness is also known, even if that number was not the original objective. There is also a third type of diversity called “beta diversity.” This quantity describes how much turnover in community compos- ition exists within a landscape (or within your sample). Landscapes like the painting Rambling Rio Grande by Johnathan Harris, on the cover of this book, include riparian areas, arid woodlands, mesic mixed forests and alpine scree. This landscape exhibits higher beta diversity than a landscape with a single community. We mention this not because we will spend time pondering this number. Others have already done so in great depth (Anderson et al. 2011, Kraft et al. 2011, Myers et al. 2013). We mention this because understanding and pre- dicting the turnover in community composition across environmen- tal gradients is precisely the topic of this book. Some authors use the word pool in a slightly different sense: they consider the pool to be only that list of species that could poten- tially occupy a specific habitat type. This use of the word is much narrower, and hence the number of species in the pool will be much smaller. What we are calling simply the species pool, Zobel (2016) called the unfiltered pool, in contrast with the habitat-specific pool. Here is a practical example that makes the distinction clear: continu- ing for a while longer with Lanark County, does the species pool for wetland communities include adjoining terrestrial species? In our use of language in this chapter, thus far, we say yes to that question. Terrestrial species may well disperse into wetland habitats, but they will not establish, and if they do, they will not survive. This is part of
82 species pools
the process of filtering that we discussed in the preceding chapters. Hence, in this chapter, and in this book, when we use the word pool, we use it to mean the list of species that could potentially occupy the site before filtering begins. In contrast, in some published studies, the focus is upon the habitat-specific pool. In this case, continuing with the Lanark County wetland example, one would not only eliminate from the pool all terres- trial species, but one might actually create narrower subdivisions within the wetland pool, such as species that occur mostly in calcareous fens, or peat bogs, or shallow water. In such cases, the list of species in the habitat-specific pool is often produced by adding up the species found in quadrats from such habitats – which is a methodological issue we discuss further below. Some authors use the word regional species pool instead of the term unfiltered pool. “A regional species pool comprises all species available to colonize a focal site” (Cornell and Harrison 2014, p. 45). So, as you are reading the scientific literature on species pools, it is good to keep two distinctions in mind. Are people writing about unfiltered pools, or habitat-specific pools? Are they focused on rich- ness or composition? Some of the confusion in terminology arises out of these differences. Let us turn to an example from Estonia (Pärtel et al. 1996). The flora of Europe is well studied, in part because of the small number of species relative to the large number of botanists, and the tradition of botanical investigation going back to Linnaeus himself. Thus, we have accurate data on the regional species pool. There have also been some efforts to compile data on plant traits, including habitat requirements. Ellenberg and his co-workers (1991) compiled the list of habitat requirements for each species based upon the factors of, to name a few, light, soil moisture, pH and nitrogen, where each plant species has a score between 1 and 9 along each of these axes. For Europe, we know not only the pool, but also some general information on the ecological conditions in which each species in that pool is normally found. We can consider this list of Ellenberg scores to provide a kind of
types of pools 83
regional pool, and disperses enough spores widely enough to reach a single boulder within a gneiss landscape. To determine their values for the regional species pool (i., what we call the habitat-specific pool), it was necessary to extrapolate. They defined their regional pool as “the set of species, occurring in a certain region (here: Estonia) which are capable of coexisting in a target commu- nity” (p. 111). Pärtel et al. (1996) first set out to define the habitat, or environmental conditions of each community, by calculating the mean Ellenberg indicator value for the species that already occurred there. That is, these factors were not directly measured, but inferred from the exist- ing species and their indicator values. The regional pool was then con- structed by finding all other species in the Estonian species pool whose indicator values fit with the average Ellenberg value for that community. Four Ellenberg factors were used: light, soil moisture, pH and nitrogen. The existing species in a habitat, then, were used as a kind of bioassay for the conditions of that habitat, and this bioassay guided the selection of other candidate species from the Estonian pool. Pärtel and his team used only 1,073 species of the flora for P, excluding groups with large numbers of microspecies (e., Hieracium spp.) and excluding species recorded only once or twice in Estonia. This was another judgment call. Then, not unlike the methods used by Raunkaier a century earlier (recall Chapter 1), they chose possible communities at random using what we now call Monte Carlo methods. By accumulating a large number of possible random communities, they were able to test whether the pat- terns they found were statistically significant. Overall, they found a significant positive relationship between the actual species pool and the regional species pool (Figure 3). They also found a significant positive relationship between the num- ber of species in 1 m 2 quadrats and the actual species pool (Figure 3). Their work gives further insight into the problems of compiling data on species pools. When they first used the Ellenberg scores to create the species pool for each habitat, they concluded that “unreal- istically small” pools resulted. The size of the pool was increased by using an amplitude of 1 relative units around the habitat mean.
types of pools 85
Using amplitudes of 1 units, they obtained a four-dimensional hypervolume with a maximum distance from the centre of the hyper- volume being three units. The pool was constructed by “including from the regional flora (1) all indifferent species and (2) all other species for which the indicator values for light, soil moisture, pH and nitrogen content were within the interval of +/– three units around the average of that community in the four-dimensional hyper- space” (p. 113). This study illustrates work arising out of a well-studied situ- ation. The flora of Estonia is known. There is a long history of studying vegetation in Central Europe, with a rich array of other published studies. There is a compendium of data on the species pool, with key ecological requirements known for each species. The individual sites are small enough that they can be exhaustively surveyed. In most locations, we do not have such complete and reliable data. We started off this section with the intention of reviewing the use of words in describing pools, hoping to provide a summary and some sort of standard terminology. We soon realized that this was an
0 0 100
73 5
4 9
10 10
11 11
12 12
6 136
13
2
1
1
14
14
200 Regional species pool
Actual species pool
300 400
50
100
(a) 150 (b)
0 0 50 Actual species pool
Species richness (m
−2)
100 150
10
50
20
30
40
figure 3 Data from Estonian plant communities shows that the number of species in a community increases with the size of the regional species pool (a). At the smaller scale, the number of species in a quadrat increases with the number of species observed in the community (b). (After Pärtel et al. 1996.)
86 species pools
realm. Our usage is not entirely consistent with work published to date. For example, Pärtel et al. (1996) used the term regional pool, whereas we prefer the term habitat-specific pool to make it clear that this type of pool includes species that are all adapted to the same habitat type. Pärtel et al. (1996) compared the observed actual pool to the habitat-specific pool, whereas we prefer to conceptually distin- guish the community vector C from the species pool P: the local observed community is a subset of the unfiltered regional pool. There is one general lesson here: it is important to specify which kind of pool we are using in a study, what standard reference source has been used, and whether any specific rules have been used to expand or reduce the list provided in that standard reference.
some theory about pools and communities
We began with large sources of published information on species pools, such as regional plant or bird lists. Now let us go to a much smaller scale and think a little more about how a species pool relates to the composition of particular habitats (or even to specified sample units within those habitats). After all, most of us who do field work have the personal experience of enumerating species in particular sample units. Eriksson (1993) provides a useful start. He explores the relationship between the number of species in a quadrat, or sample unit, and the number in the pool. He begins by asking about the relationship between S, the number of species in “any arbitrary unit space on Earth,” and Q, the number of species in the pool. Our community vector C describing the biota of this arbitrary space on Earth is a set of species p 1 to pS (recall Figure 1). This vector could be the contents of a pitfall trap, the list of plants in an alvar community, the list of birds on an island or, to return to the examples of Chapter 1, the mammals in one “landscape” in Kruger National Park, or the plants comprising an Ecoelement in the Ontario land classification system. For simplicity, we shall use the general term community. We use the word community in the sense that it is rather larger than a single quadrat, but not as large as Kruger National Park.
88 species pools
Erikson invites us to remember that the number of species in a community is a balance between two other factors, the rate of input of new species (c, for colonization) and the rate of loss of existing species (e, local extinction). Some of this thinking will be familiar to those of you who have read The Theory of Island Biogeography, which asks, in an analogous way, how many species will occur on a particular island, given an adjacent mainland and its species pool. Thinking now about any community, Erikson observes that the number of species is determined by the following equation:
dS=dt ¼ c ðQ – SÞ – e S: (3)
That is, the colonization of species is proportional to the difference between the number of species already present (S), and those still available to arrive from the pool (Q). The extinction of species is proportional to the number of species already present (S), with some finite rate of local extinction. Setting the rate of change equal to zero, the equilibrium value of S, or the number of species in our commu- nity, is then
S ¼ Q c þc e
: (3)
Erikson describes how this simple expression yields some predictions.
- If the local extinction rate is very low, then S* will be close to Q. That is, nearly all the species in the pool will be found in a particular community or quadrat.
- If the local extinction rate nearly balances the rate of colonization of new species (that is, c = e), then the community will likely have about half the number of species in the pool, that is S*= ½Q.
- If the local extinction rate is higher, with e > c, then the community will have even a smaller proportion of the species found in the pool.
These equations give us a way of thinking about just how many species in a pool are likely to occur in a specific location. With rela- tively long-lived species, like clonal plants, rates of extinction may
some theory about pools and communities 89
inputs and outputs of species, and its own value of S. For most pur- poses, we select two scales out of this continuum, and focus on the relationship between the species pool P and the community C, as in Figure 1.
rates of arrival and departure in the community
The number of species in any community is only a subset of the pool. The lower number of species in the community is the result of two main causes. Some species may not have arrived from the pool yet – we call this an immigration effect, or by another name, dispersal limitation. Other species have arrived and then disap- peared, an extinction effect. Hence, the subset of species in the community is a result of equilibrium between these two pro- cesses. This is just classic island biogeography. One important assumption in the above models is this: the rate of extinction increases with the number of species present. This is a statistical given, in one sense. The more species there are, the more likely one of them is to disappear. This, too, is a key assumption dis- cussed in The Theory of Island Biogeography. This process is not, by the way, directly explained by a mechanism such as interspe- cific competition or disease – not at all: it is merely a statistical observation. Like the more pencils you have in your field bag, the more likely it is that one of them will be lost. It explicitly does not include the effects of biology. There are also real biological reasons why such a relationship might occur. For example, as the number of species in a community increases, rates of extinction may also increase from competition or from disease. That is an added factor. Similarly, the rate of extinction may also depend upon external factors such as the frequency or inten- sity of periods of fire, drought or hypoxia. And the traits that different species possess will determine their vulnerability to these factors. So we need to be cautious in assuming that the relationships between local extinction rates and S is simply a question of the number of species. The equation is a starting point that deliberately excludes
rates of arrival and departure in the community 91
most biology. Those of us who study plant and animal interactions are documenting possible causes for local extinctions. Here, however, we wish to look at the other side of the process: immigration. That is, inputs from the species pool into a community. We begin with the neutral model, that is, that the arrival of new propagules will indeed be simply a function of (Q − S), where Q is the number of species in the pool and S is the number in the commu- nity (recall Figure 1). This is the straight line showing that the number of new arrivals falls with the number of species present (Figure 3). But now we will add some biology. Consider traits for dispersal: some species in the pool will possess the kinds of traits that allow for more efficient dispersal. Overall, it is probable that species with traits evolved for enhanced dispersal will be the first to arrive, and that they will arrive rather quickly, shifting the line into a curvilinear form (Figure 3). This will tend to generally inflate values of S in the short term, and affect the kind of species found in C. With knowledge of traits, we know which species those are likely to be. And, simultan- eously, individuals without long-distance dispersal mechanisms will
Neutral
Trait based
Number of species present
New arrival rate
figure 3 The neutral model assumes a straight-line relationship between the arrival of new species and the number already present. The presence of species with traits for dispersal will shift the immigration curve in a predictable manner. The actual number of species found in a community (S) will normally be much smaller than Q.
92 species pools
In some cases, the presence of certain species on the site may actually enhance rates of establishment. This is known as facilitation. “Nurse plants” are a well-known example (Turner et al. 1966, Franco and Nobel 1989, Kellman and Kading 1992). So, facilitation may enhance survival and shift the curve higher. Soil mycorrhiza that are supported by existing plants may also be another source of facilitation (Wardle 2002, van der Heijden and Horton 2009, Keddy 2017). More generally, many kinds of mutualisms found in nature may increase local diversity in a community. In most other cases, it is likely that the presence of species already in the community will be an obstacle to new establish- ment. This is generally the result of competition, which, of course, has a huge scientific literature (Diamond and Case 1986, Grace and Tilman 1990, Keddy 2001). Competition for resources will likely reduce rates of establishment, particularly in commu- nities of sessile or territorial organisms where space is strictly limited (Yodzis 1978, 1986). Biomass may also be a surrogate for effects of competition. There are well-described cases in plant communities in which the presence of accumulated biomass of vegetation itself feeds back to control S and C. The principle effect of that biomass is to produce a canopy of existing plants that reduce light for new arrivals (Grime 1973a, 1973b, 1979). Let us look at a situation where, instead of putting S on the horizontal axis, we use biomass of existing species (Figure 3). Here, the trait differences among species is also import- ant – some plants have clonal growth and produce especially dense canopies and accumulations of leaf litter. These plants, when pre- sent, exclude many other species. In addition, early arrival by such species will give them an even greater local advantage. So, in this case, the probability of new species establishing is very much deter- mined by biomass, and particularly the biomass of such clonal spe- cies. Hence, we may expect the rate of establishment to be particularly low on the right-hand side of the figure, since shade is often fatal to the establishment of seedlings (Harper 1977). We might
94 species pools
therefore assume that establishment should be much higher at the low-biomass end of the gradient, where there is no canopy. But here it is possible that the low biomass is an indication that the site is unsuited to the growth of many potentially colonizing species. The competition from existing plants may indeed be low, but there is a reason for the lack of plants: lack of resources. So, in fact, the best conditions for establishment of new individuals may be at an inter- mediate position along this gradient, a region where neither physical stress nor competition predominate. This may partly explain why sites with intermediate levels of biomass have high diversity (Figure 3). That is, the corridor of diversity may reflect a situation where neither physical stress nor competition predominate. Grace (2001) has extended these ideas with a general exploratory model showing how biomass can regulate diversity in plant communities by controlling inputs from species pools. Fraser et al. (2015) have collected examples of this pattern at the global scale. Such work raises an interesting question: what might happen to the local com- munity if we experimentally supplement rates of invasion by sowing new individuals from the pool into the community? We will look at some such experiments in the next section. The main point here is that knowledge about the pool is a vital first step. And, general models exist that show us how immigration and local extinction can determine the likely number of species in a particular habitat, and the kinds of traits those species will possess. Once we have knowledge of species traits, and the environmental conditions, it is likely that real communities deviate significantly from neutral assumptions. In particular, species with high dispersal abilities are likely to be overrepresented. At least this sounds reason- able, until it is balanced by the fact that species able to dominate a site by clonal growth may eventually exclude these dispersal-adapted spe- cies. In which case, the value of S, and the kinds of species found in C, may be determined mostly by local conditions, particularly the rate at which competitive dominant species are removed to create gaps for new arrivals. This takes us into the realm of the role of natural
rates of arrival and departure in the community 95
Species_species
Kursus: Biologi
Universitet: Københavns Universitet
