Estimating Terrestrial Biodiversity Through Extrapolation
1994JONATHAN CODDINGTONEstimating terrestrial biodiversity through extrapolationROBERT K. COLWELL1 AND JONATHAN A. CODDINGTON2'Department of Ecology and Evolutionary Biology, University of Connecticut, Storrs, Connecticut 06269-3042, U.S.A.1Department of Entomology, National Museum of Natural History, Smithsonian Institution, Washington, D.C 20560, U.S.A.CONTENTSPAGE1. Introduction2. The organization of terrestrial biodiversity3. Richness and complementarity(a) Concepts(b) Optimizing complementarity in inventories4. Estimating local richness by sampling(a) Extrapolating species accumulation curves(b) Fitting parametric models of relative abundance to estimate richness(c) Non-parametric methods for estimating species richness from samples5. Complementarity(a) Measures of complementarity(b) Some examples of differing complementarity(c) Complementarity of samples6. Using ratios to estimate and extrapolate(a) Reference and comparison(b) Taxon 12113114114114116SUMMARYBoth the magnitude and the urgency of the task of assessing global biodiversity require that we make themost of what we know through the use of estimation and extrapolation. Likewise, future biodiversityinventories need to be designed around the use of effective sampling and estimation procedures,especially for 'hyperdiverse' groups of terrestrial organisms, such as arthropods, nematodes, fungi, andmicroorganisms. The challenge of estimating patterns of species richness from samples can be separatedinto (i) the problem of estimating local species richness, and (ii) the problem of estimating thedistinctness, or complementarity, of species assemblages. These concepts apply on a wide range ofspatial, temporal, and functional scales. Local richness can be estimated by extrapolating speciesaccumulation curves, fitting parametric distributions of relative abundance, or using non-parametrictechniques based on the distribution of individuals among species or of species among samples. Wepresent several of these methods and examine their effectiveness for an example data set. We present asimple measure of complementarity, with some biogeographic examples, and outline the difficultproblem of estimating complementarity from samples. Finally, we discuss the importance of using'reference' sites (or sub-sites) to assess the true richness and composition of species assemblages, tomeasure ecologically significant ratios between unrelated taxa, to measure taxon/sub-taxon (hierarchical) ratios, and to 'calibrate' standardized sampling methods. This information can then beapplied to the rapid, approximate assessment of species richness and faunal or floral composition at'comparative' sites.empirical estimation, lie at the heart of most research.in contemporary genetics and ecology. In systematics,although experimentation cannot play such a centralrole, phylogenetic hypotheses are increasingly basedon logical and quantitative criteria. Even in thesecases, however, reliable methods to interpolate andextrapolate, for instance, from the few speciesincluded in an analysis to the entire higher taxonthey exemplify, have been little assessed.1. INTRODUCTIONExtrapolating from the known to the unknown, fromthe past to the future, is a familiar and essentialprocess in those biological disciplines traditionallyinvolved in public policy, but seems rather alien tomany of the kinds of biologists whose expertise ispivotal to the scientific study of biodiversity. Experimentation and mechanistic hypothesis-testing, notPhil. Trans. R. Soc. Lond. B (1994) 345, 101-118Printed in Great Britain 1994 The Royal Society101
102 R. K. Colwell and J. A. CoddingtonEstimating terrestrial biodiversityThe urgent challenges of global climate change,massive habitat transformation, and the threat ofwidespread extinction, however, have made extrapolation and prediction a crucial component of manyresearch agendas in these fields. In the case ofterrestrial biodiversity (including freshwater habitats), a reasonably accurate picture for many groupsof vertebrate animals, most plants, and a very fewgroups of showy insects, can be developed byintegrating biogeographic information from faunisticand floristic surveys with the taxon-focused workof systematists (Groombridge 1992). This body ofknowledge has accumulated largely under its ownmomentum from thousands of independent sources.In contrast, our present state of taxonomic andbiogeographic knowledge for most other groups ofterrestrial organisms is sketchy at best, especially forthe 'hyperdiverse' terrestrial groups: insects, mitesand other arachnids, nematodes, fungi and microorganisms. Relying solely on traditional approaches,the current trajectory points to an adequate, worldwide picture for these groups no sooner than a fewcenturies from now (May 1990; Hawksworth 1991;Hammond 1992). (Of course, our ignorance of thetrue richness of these taxa makes any such projectionvery rough indeed.)Clearly, then, while aggressively building humanand institutional capacity in systematics (Gaston &May 1992; Anonymous 1993; Janzen 1993), approximate methods must be used to gain any useful sense ofthe richness, taxonomic diversity, and geographicpatterning of the hyperdiverse groups. In terms ofbiochemical diversity and the variety of potentiallyuseful 'evolutionary inventions' that natural selectionhas produced, the hyperdiverse groups present vastnumbers of unexploited opportunities for furtheringhuman welfare and solving environmental problems(Farnsworth 1988; Eisner 1990; Colwell 1992; Wilson1992; Reide(a/. 1993).Moreover, it seems only logical that themost diverse groups of organisms should play asignificant role in planning for the conservation andsustainable use of worldwide biodiversity (Brown1991; Hawksworth 1991; Kremerw at 1993), yetthey have so far been largely ignored. Reliance onlyon data from a few well-known taxa such as birds,mammals, trees, butterflies or ants (e.g. Raven &Wilson 1992) assumes that variation in diversity ofthese groups is closely concordant with the diversity ofunrepresented groups. If variation in importantproducer or decomposer diversity does not significantly correlate with bird diversity, for example,land-use decisions based on bird data may manage forbird diversity but against other taxa. From the pointof view of an invertebrate zoologist, mammals andbirds are fairly similar: mainly recent radiations oflarge, homeothermic heterotrophs. In contrast, thevast majority of other taxa have very different ages,histories, and lifestyles. Initial work on this questionsuggests that diversity patterns vary widely betweentaxa, and that relying on just a few groups would notoptimally preserve others (Prendergastef al. 19 93).More research on correlations between well-knownPhil. Trans. R. Soc. Lond. B (1994)but depauperate lineages and hyperdiverse groups isurgently needed before the 'indicator group' strategyis widely applied.In this paper, we will focus first on how terrestrialbiodiversity is organized, then on methods ofestimation and extrapolation. Some of the methodswe will discuss have been widely used to developquantitative estimates of terrestrial species richness,yet some promising quantitative techniques, such asnon-parametric estimators of local species richness,have been little used.As for actual numerical estimates of globalterrestrial species richness, we direct the reader tothe plethora of recent reviews and debates on thissubject (May 1988, 1990, 1992; Stork 1988, 1994;Gaston 1991; Hawksworth 1991; Hodkinson & Casson1991; Hammond 1992; Wilson 1992). Althoughestimating global species richness has attracted muchattention, further progress on this front awaits a betterunderstanding of the structure and variation ofbiodiversity on smaller scales, especially in landscapesor 'park-sized' units. Moreover, land-use decisions aremost often made at these levels and have great impacton the long-term future of biodiversity.2. THE ORGANIZATION OF TERRESTRIALBIODIVERSITYImagine a magnificent and omniscient GeographicInformation System(ois) for all the Earth's livingspecies, with the capacity to display any level of theLinnean hierarchy on any spatial scale, for any seasonof the year. To take an avian example that couldactually be approximated with present knowledge, wemight request that the distribution of the familyTrochilidae (hummingbirds) be superimposed on theworld map, indicating either absence of the family orthe presence of one or more of theca. 320 knownspecies of hummingbirds. Virtually all of the NewWorld continental land masses would light up(hummingbirds are strictly a New World group),from southern Alaska and central Canada to thetip of Tierra del Fuego, plus the Antilles and JuanFernandez archipelagos (Blake 1953; De Schauensee1970; Land 1970; Skutch 1973; Tyrrell & Tyrrell1985; Colwell 1989; Ridgely & Gwynne 1989; Stiles &Skutch 1989; Tyrrell & Tyrrell 1990).A species density map ('topographic' contoursshowing the number of hummingbird species at eachpoint on the map) would display a gradient from thelowland tropics, where the ranges of a dozen or morespecies often overlap, toward single species at thenorthern and southern ends of the family range(Skutch 1973; Feinsinger & Colwell 1978; Stiles1980). Zooming in on Central America, and then onCosta Rica would reveal further 'fine-structure' ofspecies density, from five species recorded from 3100 melevation at Cerro de la Muerte (Colwell 1973; Wolfet al. 1976), to 14 species at 1400m at Monteverde(Feinsinger 1976, 1978), to 25 species at La SelvaBiological Station in the Atlantic lowlands (Stiles1980; Karr et al. 1990). If we next request seasonalmaps, however, we would see that some of the species
Estimating terrestrial biodiversityat each site are year-round residents, whereas othersare seasonal migrants, dependent on seasonal nectarsources not only at those sites hut at other elevationsor latitudes. Some of the species are found at only oneof the three Costa Rican sites (among other places)and some are found at two of them. (None occurs atall three.) Finally, if we requested full geographicrange plots, species by species, for the hummingbirdsat these three sites, we would find that some areendemic to Costa Rica and Western Panama, someextend as far north as Arizona, and others as far southas the Amazonian basin.This complex mix of wide-ranging and narrowlyendemic species, of different patterns of seasonality,with broad latitudinal and elevational gradients oflocal species richness is absolutely characteristic ofterrestrial organisms: not only birds, but othervertebrates, insects, arachnids, plants, and no doubtfungi, protists, and bacteria as well. Moreover, thesame kinds of patterns are repeated in many formsand at many scales. Local assemblages of herbivorousinsects or mites are characteristically a mixture of hostplant specialists and generalists, and the same is truefor parasitic organisms in relation to their hosts(Futuyma & Moreno 1988). Pollinator assemblagesinclude everything from obligate, one-to-one relationships with plants (e.g. figs and fig wasps): to broadgeneralists that pollinate dozens or even hundreds ofplant species (Real 1983). Rainforest arboreal mitecommunities show the same kinds of complexgeographical patterning as the hummingbirds in theexample above, but also display striking faunaldifferences on a scale of meters, from forest floor, totree trunks, to leaves (Walteret al. 1994).3. RICHNESS AND COMPLEMENTARITY(a) ConceptsThe omniscient cis imagined above represents thetrue global pattern of biodiversity (from the specieslevel on up) that any estimation scheme should bedesigned to approximate. For the best-known groups,such as birds, mammals, or butterflies, species-byspecies patterns may be developed to estimate localspecies richness and patterns of biogeographicaloverlap, as in the hummingbird example. For thehyperdiverse groups, in contrast, exhaustive inventoryon a broad geographical scale is out of the question.Even the 'All Taxon Biological Inventories'(ATBis)now being discussed (Janzen & Hallwachs 1993;Yoon 1993) will require, at least, interpolationbetween sampled points along habitat gradients forthe smallest and most diverse organisms, and verylikely a variety of approximate methods for thesampling points themselves. For plants, records arestill sufficiently poor for some regions, especiallytropical forests, that we will need to rely on similarkinds of sampling and estimation for the foreseeablefuture (Raven 1988).As an idealized (and much-used) design for acomponent study in a regional biological inventory,imagine a series of local species inventories at 'points'Phil. Trans. R. Soc. Lond. B (1994)R. K. Colwell and J. A. Coddington103spaced along a gradient, or located randomly within ahabitat mosaic. For example, in a study of freshwaterfishes or algae, the points might be sampling stationsspaced along the gradient from the headwaters of ariver to its estuarine mouth. For plants or birdsthe gradient might be an elevational transect fromtemperate deciduous forest to alpine tundra, witha 4ha plot every 500 m elevation; or the tropicalequivalent. Or, the gradient might, instead, be aforest chronosequence, from early to late succession.As another temporal example, the 'points' might be aseries of malaise trap samples of flying insects taken inthe same trap over a 'gradient? from dry season to wetseason in a tropical deciduous forest. Alternatively,the 'points' might be tree species in the biochemicalmosaic of a rainforest, for a study of herbivorousinsects. On a global scale, each 'point' might be a50 000 ha ATBI site covering a range of macrohabitatgradients, as a component of a series of ATBIS placedwithin different phases of the worldwide mosaic ofmajor biomes (Solbrig 1991; di Castries al. 1992a,b ;Vernhes & Younes 1993; Yoon 1993).In each of these cases (and many more), theproblem of gaining an approximate description of thepattern of biodiversity for some taxon along agradient or among the phases of a mosaic can bebroken down into two parts: measuring or estimatingthe species richness of species assemblages locally,and measuring or estimating the complementarity the distinctness or dissimilarity - of these localinventories.The concept of complementarity is intended tocover distinctness in species composition over a broadspectrum of environmental scales, including smallscale ecological differences, such as the differencesbetween the mite faunas of the trunk versus the leavesof a single tree species (Walter et al. 1994); betweenhabitat and landscape-level differences alongenvironmental gradients ('beta diversity' or 'speciesturnover') (e.g. Shmida & Wilson 1985; Palmer &Dixon 1990); faunistic and floristic differencesbetween distant sites in the same biogeographicrealm; and (at the level of higher taxa) climaticallyanalogous sites on different continents or evenclimatically distinct sites in different biomes. Thisbroad use of the term 'complementarity' extendsVane-Wright's usage for comparing the biota ofpotential reserves (Vane-Wright et al 1991; Presseyet al 1993).We prefer a single, broad term to a series of morespecific, scale- or gradient-dependent concepts, toemphasize that the problem of characterizing differences in the species composition of componentassemblages is both universal and crucial to thesubject of estimating biodiversity, regardless ofcausal mechanism and of spatial or temporal scale.Using the concept of complementarity, whenappropriate and informative, in place of its logicalopposites, similarity or overlap, allows us to see bothlocal richness and biotic (floral or faunal) differencesas positive components of biodiversity. (Biotic similarity is negatively related to overall biodiversity.)The choice of complementarity over its statistical
104R. K. Colwell and J. A. Coddington Estimating terrestrial biodiversityequivalents, distinctness,' dissimilarity or distance, isstrictly a rhetorical preference, to capture the sensethat complementary faunas or floras form parts of awhole: a sense that distinctness (or its equivalent) doesnot convey.(b) Optimizing complementarity in inventoriesLocal richness and complementarity interact incomplex and vexing ways (as we will discussbelow), but treating them as separate componentsof biodiversity helps reveal common threads andcommon pitfalls in the methods that have been usedto estimate biodiversity, and may aid in designingefficient inventories (Longino 1994) and in developingstrategies for conservation (Presseye at. 1993).Measuring biodiversity in terms of the componentsdue to the species richness of local assemblages andthe complementarity between them does not requirethe world to follow any particular model of community or landscape structure, but it does meanmaking decisions about how to define the units to beinventoried and compared. As a first approximationfor this step, there is rarely any better strategy thanrelying on the informed intuition of experiencednaturalists. For a regional inventory of rainforesttrees, for example, perhaps over a 10 000 ha area,units might be defined by the intersection of factorsbased on life zones, major soil types, gap phases, slope,and elevation above sea level, with replicate plots ortransects placed within each inventory unit. In anyinventory, if preliminary data show that the speciescomposition of adjacent inventory units along atransect, or of the phases of a mosaic, are quitesimilar, the spatial or ecological scale might safely bemade coarser. On the other hand, if these units proveto have largely distinct species lists, the scale mighthave to be made finer to gain a reasonable picture ofthe full biota of the region for some taxon.The optimal spatial or ecological scale of inventoryunits clearly depends crucially on the biology of theorganisms to be sampled, as well as the size of theproject budget. Birds and beetles obviously respond todifferent environmental features on different scales,and so do hawks versus hummingbirds, and dungbeetles versus weevils. In addition to specifyingsampling or census methods, inventory protocolsneed to be specific about the scaling of inventoryunits. Often, scaling compromises will be made in theinterest of simplifying inventory protocols so that eachprotocol covers the broadest taxonomic spectrumfeasible. It is beyond the scope of this paper to makeeven a rough attempt to specify scales or protocols forparticular target taxa, or to review the enormoustaxon-specific literature on sampling methodologies.Although significant efforts have been made todevelop 'portable' inventory protocols that providereliable results among biomes and continents (e.g.Gadagkar et al. 1990; Hammond 1990; Coddington etai. 1991; Stork 1991; Heyera al. 1993), much remainsto be done, especially for the hyperdiverse taxa.When methods to estimate local richness andcomplementarity, including their confidence interPhil. Trans. R. Soe. Land. B (1994)vals, are more fully developed, integrated, and tested,the cost of inventorying should favour allocatingsampling effort as thinly and widely as possible,consistent with the degree of accuracy in thecomplementarity estimate, required. At present, it isunclear which groups scale geographically at similarrates. For large-scale inventories, each major taxon islikely to require a distinct 1 inventory strategy.4. ESTIMATING LOCAL RICHNESS BYSAMPLINGMeasurement of local richness by complete census isfeasible, in the terrestrial 'realm, only for plants andperhaps for conspicuous and highly philopatricmammals (e.g. territorial' primate troops). Even forthese groups, estimation by sampling may nonethelessbe the best option, but for virtually all others,measurement means sampling. Traditional collectionmethods employed in floral or faunal surveys byprofessional collectors for museums and herbaria mayintend to collect all species, but such a goal isnotori
Estimating terrestrial biodiversity through extrapolation . especially for 'hyperdiverse' groups of terrestrial organisms, such as arthropods, nematodes, fungi, and microorganisms. The challenge of estimating patterns of species richness from samples can be separated . exhaustive inventory on a broad geographical scale is out of the question.
Electrical Construction Estimating Introduction to Electrical Construction Estimating Estimating activites will use the North State Electric estimating procedures. Estimating and the Estimator Estimating is the science and the art by which a person or organization determines in advance of t
Sustaining Terrestrial Biodiversity: The Ecosystem Approach . There is no solution, I assure you, to save Earth’s biodiversity other . 1 or 2 species of same age . These two species found in tropical forests are part of the earth's biodiversity. On the left\ഠis an
Biodiversity and agriculture are strongly interrelated. Biodiversity is critical for agriculture whilst agriculture also contributes to conservation and sustainable management of biodiversity. Indeed, Integrated Farm Management both promotes and is enhanced by biodiversity. The protection, maintenance and enhancement of biodiversity is .
the central difference formula to the first derivative and Richardson’s Extrapolation to give an approximation of order O(h4). Recall N(h) f(x h) f(x h) 2h. Therefore, N(0.2) 22.414160. What do we evaluate next? N( ) Tim Chartier and Anne Greenbaum Richardson’s Extrapolation
change image dimensions/aspect ratios without changing the content of the original images. Extrapolation, however, is a much more challenging problem since there is much less in-formation available; while inpainting methods are given the entire boundary of the missing region, in image extrapolation we only know one border. This less constrained .
Sustaining Terrestrial Biodiversity: The Ecosystem Approach Chapter 10 . Core Case Study: Reintroducing Gray Wolves to Yellowstone . – Enhances invasion of plant species that cattle won’t eat . Concept 10-4 Sustaining biodiversity will
Biodiversity Revisited research and action agenda 4 REVISITING BIODIVERSITY RESEARCH AND ACTION Carina Wyborn, Jasper Montana, Nicole Kalas, Santiago Izquierdo Tort, Victoria Pilbeam This chapter examines the rationale and approach of the Biodiversity Revisited agenda. We first consider what biodiversity is and why it might need to be revisited.
Type A02 : Cable suffices This type comprises people with some limited interest in electronic technologies but who have neither the education nor income to become heavily engaged in using them. Many of this type are men who have recently retired or who are approaching retirement. A high proportion has access to cable television. Type A03 : Technology as fantasy This type contains many old .
