The spatial structure of Antarctic biodiversity

42 Pages·2014·20.25 MB·English

REVIEWS
EcologicalMonographs,84(2),2014,pp.203–244
(cid:2)2014bytheEcologicalSocietyofAmerica
The spatial structure of Antarctic biodiversity
PETERCONVEY,1,2,19 STEVEN L. CHOWN,3ANDREWCLARKE,1DAVIDK.A. BARNES,1STEFBOKHORST,4
VONDACUMMINGS,5HUGH W.DUCKLOW,6FRANCESCOFRATI,7 T. G.ALLANGREEN,8SHULAMITGORDON,9
HUWJ. GRIFFITHS,1CLIVEHOWARD-WILLIAMS,10 AD H. L. HUISKES,11,20 JOHANNALAYBOURN-PARRY,12
W.BERRY LYONS,13 ANDREWMCMINN,14SIMONA.MORLEY,1LLOYDS. PECK,1ANTONIOQUESADA,15
SHARON A.ROBINSON,16STEFANOSCHIAPARELLI,17ANDDIANAH. WALL18
1BritishAntarcticSurvey,NaturalEnvironmentResearchCouncil,HighCross,MadingleyRoad,
CambridgeCB30ETUnitedKingdom
2NationalAntarcticResearchCenter,IPSBuilding,UniversityMalaya,50603KualaLumpur,Malaysia
3SchoolofBiologicalSciences,MonashUniversity,Victoria3800Australia
4DepartmentofForestEcologyandManagement,SwedishUniversityofAgriculturalSciences,Umea˚SE90183Sweden
5NationalInstituteofWaterandAtmosphericResearch,PrivateBag14-901,Wellington6241NewZealand
6Lamont-DohertyEarthObservatory,Palisades,NewYork10964USA
7DepartmentofLifeSciences,UniversityofSiena,viaA.Moro2,53100Siena,Italy
8DepartmentofBiologicalSciences,UniversityofWaikato,PrivateBag3105,Hamilton,NewZealand
9AntarcticaNewZealand,PrivateBag4745,Christchurch8140NewZealand
10NationalInstituteofWaterandAtmosphericResearch,P.O.Box8502,Christchurch8440NewZealand
11NetherlandsInstituteofEcology(NIOO-KNAW),UnitforPolarEcology,P.O.Box140,4400ACYerseke,TheNetherlands
12BristolGlaciologyCentre,SchoolofGeographicalSciences,UniversityofBristol,BristolBS81SSUnitedKingdom
13ByrdPolarResearchCenter,SchoolofEarthSciences,TheOhioStateUniversity,1090CarmackRoad,
Columbus,Ohio43210-1002USA
14InstituteforMarineandAntarcticStudies,UniversityofTasmania,Hobart,Tasmania7001Australia
15DepartmentofBiology,UniversidadAutonomadeMadrid,28049Madrid,Spain
16InstituteforConservationBiology,TheUniversityofWollongong,NewSouthWales2522Australia
17DipartimentodiScienzedellaTerra,dell’AmbienteedellaVita(DISTAV),Universita`diGenova,C.soEuropa26,
GenovaI-16132Italy
18DepartmentofBiologyandNaturalResourceEcologyLaboratory,ColoradoStateUniversity,FortCollins,
Colorado80523-1499USA
Abstract. Patterns of environmental spatial structure lie at the heart of the most
fundamental and familiar patterns of diversity on Earth. Antarctica contains some of the
strongestenvironmentalgradientsontheplanetandthereforeprovidesanidealstudyground
totesthypothesesontherelevanceofenvironmentalvariabilityforbiodiversity.Toanswerthe
pivotal question, ‘‘How does spatial variation in physical and biological environmental
propertiesacrosstheAntarcticdrivebiodiversity?’’wehavesynthesizedcurrentknowledgeon
environmental variability across terrestrial, freshwater, and marine Antarctic biomes and
related this to the observed biotic patterns. The most important physical driver of Antarctic
terrestrial communities is the availability of liquid water, itself driven by solar irradiance
intensity. Patterns of biota distribution are further strongly influenced by the historical
development of any given location or region, and by geographical barriers. In freshwater
ecosystems,freewaterisalsocrucial,withfurtherimportantinfluencesfromsalinity,nutrient
availability, oxygenation, and characteristics of ice cover and extent. In the marine biome
theredoesnotappeartobeonemajordrivingforce,withtheexceptionoftheoceanographic
boundary of the Polar Front. At smaller spatial scales, ice cover, ice scour, and salinity
gradients are clearly important determinants of diversity at habitat and community level.
Stochastic and extreme events remain an important driving force in all environments,
particularly in the context of local extinction and colonization or recolonization, as well as
that of temporal environmental variability. Our synthesis demonstrates that the Antarctic
Manuscript received 20 December 2012; revised 17 June 2013; accepted 21 June 2013; final version received 20 July 2013.
CorrespondingEditor:B.J.Cardinale.
19E-mail:[email protected]
20Presentaddress:RoyalNetherlandsInstituteforSeaResearch,P.O.Box140,4400ACYerseke,TheNetherlands.
203
204 PETERCONVEYETAL. EcologicalMonographs
Vol.84,No.2
continent and surrounding oceans provide an ideal study ground to develop new
biogeographical models, including life history and physiological traits, and to address
questions regardingbiological responses to environmental variability andchange.
Keywords: adaptation;biogeography;environmentalgradients;historicalcontingency;marine;spatial
scaleandvariation;terrestrialenvironments.
INTRODUCTION fied by soils and water bodies in the McMurdo Dry
Valleys and the McMurdo Ice Shelf, respectively
Spatialstructure
[Ferna´ndez-Valiente et al. 2001, Barrett et al. 2007]), is
Spatial structure is one of the most fundamental
theterminusofmanyglobalgradients(UsherandBooth
characteristics of the planet. Early natural historians
1986, Peck et al. 2006), and its biota has formed the
and geographers realized that the environment varies
focusofmanybiogeographicinvestigations(reviewedin
systematically over a range of spatial scales, and
Convey et al. 2012a, Fraser et al. 2012, Terauds et al.
following the 19th-century voyages of discovery, wide-
2012).ExplicitattentiontospatialvariationinAntarctic
spreadrecognitiondevelopedthatthisvariationextends
diversity and its mechanistic underpinnings is growing
to global scales in the form of a diversity gradient. As
apace (e.g., Chown et al. 1998, Adams et al. 2006,
the fields of ecology and biogeography developed over
Chown and Convey 2007, Peat et al. 2007, Verleyen et
the 20th century, much progress was made in under-
al.2009,Griffiths2010,Cowanetal.2011,Mortimeret
standing the form of spatial variation and the likely
al. 2011a,b, Schiaparelli and Hopcroft 2011, Allcock
factors underlying it. Several key developments stand
and Strugnell 2012, Born et al. 2012). However, few
out from a biodiversity perspective. First is the early
description and analysis of gradients in the abiotic attempts have been made to draw together the
environment,andspeciesresponsestothem(Andrewar- conclusions from this work. Doing so would not only
tha and Birch 1954, Whittaker 1967). Such work provide substantial insight into the biodiversity impli-
remains important in modern investigations of species cations of spatial variation in a physically extreme
distributionandabundance,includingrecognitionofthe system, but would also indicate the extent to which
significance of direct, indirect, and resource gradients spatial frameworks developed elsewhere apply to the
S
W (Austin 1980). Second, geographers realized early on region.Inturn, this wouldtest their validity andreveal
E thatcloselylocatedfeaturesaremostlikelytobesimilar, the extent to which the variety encompassed by
VI
E a feature known as spatial autocorrelation. The signif- Antarctic systems can serve further to establish ecolog-
R icanceofautocorrelationforunderstandingvariationin icalgeneralities.
biodiversityremainsthesubjectofintenseinterest(e.g., As examples of the insights such an approach can
Beale et al. 2010, Hawkins 2012). Third, exploration of deliver, consider two key ecological predictions at
variationinfeaturesatdifferentspatialscaleshasledto different spatial scales. First, diversity at high latitudes
wide appreciation that different processes are likely to should typically show a decline as a consequence either
underlie patterns at each scale (Ricklefs 1987). Scale- ofdecliningenergy(Hawkinsetal.2003,butseeClarke
related thinking now permeates almost every area of
and Gaston 2006) or the change in the seasonal
ecology andevolutionaryphysiology.
availability of that energy (Archibald et al. 2010). The
Growing integration among these areas has led to
extent to which such declines are found varies among
recognitionthatenvironmentalspatialstructuremaylie
terrestrial and marine systems in the region (Peck et al.
at the heart of the most fundamental and familiar
2006)andamongtaxawithinthemarinesystem(Clarke
patternsofdiversity(Storchetal.2008).Moreover,itis
andJohnston2003).Inconsequence,considerablescope
now widely appreciated that to understand how
exists for testing unified theories to explain biodiversity
biodiversity changes through time, as a consequence of
variation (e.g., Price et al. 2012). Second, much
dispersal,extinction,speciation,andevolution,andhow
ecological theory suggests that predominantly abiotic
it may respond to environmental change driven by
factorsshouldinfluencethedistributionandabundance
humans, requires a spatially explicit approach (Black-
oforganismsinlow-diversitysystems(MacArthur1972,
burn and Gaston 1998, Sobero´n 2007, Gaston et al.
Southwood1988,Ricklefs2011).Ifthisisthecase,then
2008, 2009, McRae et al. 2008, Thuiller et al. 2008,
models seeking to explain variation in abundance and
Storferetal.2010,Ellneretal.2011,Bellardetal.2012).
In consequence, much attention is now being given to distribution, as well as in species richness, should be
the collection of spatially explicit data and the descrip- dominated by abiotic terms and by those describing
tionofvariationatseveralscalesasthestartingpointfor their spatial structure. In turn, the outcomes of
investigating biodiversity. mechanistic and environmental niche models (Kearney
Suchaperspectiveisalsobeingbroughttobearonthe andPorter2009)shouldcoincide,soassumptionsabout
Antarctic. Antarctica contains some of the strongest the extent to which each of these species distribution
environmental gradients on the planet (e.g., from modeling approaches deals with fundamental and
extreme hypersalinity to almost zero salinity, exempli- realized nichesshould besubstantiated.
May2014 ANTARCTICBIODIVERSITYSPATIALSTRUCTURE 205
Ouraimsherearethereforetoprovideanoverviewof the shelf area on the planet) (Arntz et al. 1994, Clarke
spatial variation in the Antarctic terrestrial, limnetic, and Johnston2003, Schiaparelli andHopcroft2011).
and marine environments, and to see what lessons for Unlike the Arctic, Antarctic marine and terrestrial
broader understanding of biodiversity variation gener- ecosystems are largely isolated, rather than forming a
ally have emerged from work in this area. We then continuumfromthoseatlowerlatitudes,aprocessthat
provideasynthesisofcurrentknowledgeintheseareas. commenced with the last stages of the breakup of
Because biodiversity has an explicitly historical context Gondwana, and was enhanced and then maintained by
(Ricklefs 1987, Clarke andCrame 1989), we commence the development of the atmospheric Polar Vortex and
with a brief discussion of the history of the region, oceanic Antarctic Polar Front (Clarke et al. 2005,
making reference to more comprehensive reviews for Barnes et al. 2006b, Bergstrom et al. 2006). In the deep
further information. Throughout, we draw on specific oceanssurroundingAntarctica,theisolatinginfluenceof
examples to illustrate our broader understanding, the Antarctic Circumpolar Current extends to about
recognizing that as a consequence, some significant 1000 m depth. Below this, connectivity with the other
areas of work have to remain less fully covered. The global ocean basins is more significant, exemplified by
latterincludetherolesofpolaroceanicfronts,mesoscale the northwards flow of cold Antarctic Bottom Water
features, and seabed topography in determining varia- that forms a major driver of the global overturning
tioninpelagicbiodiversity(see,e.g.,Murphyetal.2007, circulation or ‘‘ocean conveyor belt.’’ In marine pelagic
Tittensoretal.2010,Louzaoetal.2011,Wakefieldetal. orplanktonicecosystems,thedistribution oforganisms
2011, Ainley et al. 2012, Strugnell et al. 2012), the intimeandspaceisdeterminedbycomplexinteractions
complexities of diversity–environment interactions in among trophodynamics, population dynamics, physical
the many streams that are a feature of the terrestrial mixing, and circulation processes (Cullen et al. 2002,
Antarctic in the summer (e.g., Laybourn-Parry and Wakefield et al. 2011). At the mesoscale and larger
Pierce 2007), and detailed discussion of the functioning scales, distributions correspond to circulation features
of sub-Antarctic terrestrial systems (e.g., Chown and ranging from eddies and rings to basin-scale frontal
Froneman2008). boundariesseparatingmajor watermasses (Knox1994,
Longhurst1998,Ainleyetal.2012,Rogersetal.2012a).
The Antarctic environment Terrestrial ecosystem development is limited to areas
R
The‘‘Antarctic’’isdefinedhereinitswidestsense.For that are seasonally or permanently snow- and ice-free,
E
terrestrial and nonmarine aquatic environments, this andhaveappropriateenvironmentalconditions.Ice-free V
means the Antarctic continent and Peninsula, the groundiscurrentlylimitedto;0.34%oftheareaofthe IE
various archipelagos of the Scotia arc, and the sub- Antarctic continent, equating to ;45000 km2, while WS
Antarcticislands.Themarineenvironmentencompasses visible life is largely but not completely restricted to
the entire Southern Ocean, formed by southern prov- lower-altitude exposures in coastal regions (Convey et
incesoftheAtlantic,Indian,andPacificOceans(Fig.1), al. 2009, Convey 2013). In contrast with much of the
and is delineated to the north by the mean position of marine environment, typically island-like terrestrial
theAntarcticPolarFront.Antarcticecosystemsvaryon ecosystems are isolated from each other across a range
land from polar deserts, including the continent’s ice of scales, from meters to many hundreds of kilometers.
itself, freshwater to hypersaline lakes and their ice The ice of Antarctica is also not devoid of life.
covers,tolushgrasslandsandeutrophicponds(Thomas Considerable biomass is associated with snow and ice
et al. 2008) (Fig. 2). Marine ecosystems range from algalcommunitiesthatdevelopinsummer,especiallyin
shallow coastal regions to abyssal depths of the open coastal regions, although these have received relatively
ocean, ice-free to permanently ice-covered areas, and littleresearchattention(Bagshawetal.2007,Stibaletal.
highlydiverseto verysimple,featureless habitats. 2012; see Hodson et al. 2008 for discussion of Arctic
parallels). The existence of subglacial microbial com-
The Southern Ocean is dominated by the deep sea
munities, increasingly recognized in alpine regions, is
(depth .3000 m). However, very little is known about
now being examined in Antarctica (Tranter et al. 2005,
Antarctic deep-sea biodiversity (Kaiser and Barnes
Lanoil et al. 2009), while much attention is focused on
2008), other than that it appears to be rich in some
thepotentiallyexceptionalbiotatobefoundinthemany
groups and undescribed species (Brandt et al. 2007,
lakesnowknowntoliebeneaththecontinent’sicesheets
Rogers et al. 2012b), and that this richness is patchy
orin itspermafrost (Skidmore 2011).
across spatial scales (Kaiser et al. 2007, Griffiths et al.
2009,Conveyetal.2012a).Thecontinentalslope(shelf
HISTORY
break to 3000 m) is also poorly sampled and known,
Cenozoic climate changeandglaciations
although it would appear to be central to a cline from
species-rich and abundant shelf faunas to generally Understanding current biological processes, particu-
poorer abyssal depths. The majority of sampling, larly those relating to diversity, requires sound knowl-
recorded diversity, and knowledge of biological struc- edgeoftheregionalhistoryofthebiota.Onthelongest
ture concerns the fauna of the continental shelf depths timescales,Antarcticahasshiftedfrombeingwarm,ice-
(generallyshallowerthan1000m,accountingfor8%of free, and broadly connected to other land masses, to
206 PETERCONVEYETAL. EcologicalMonographs
Vol.84,No.2
S
W
E
VI
E
R
FIG.1. (A,B)OverviewmapsofAntarctica,theSouthernOcean,andadjacentregionsoftheSouthernHemisphere,indicating
locationsmentionedthroughoutthetext.
cold, glaciated, and isolated (Bertler and Barrett 2010). appear to be enhancements of existing features rather
The fossil history of the region is limited and does not thannoveladaptations(Convey1996),whilecontinental
permit description of changes in detail, at least not for cooling and extensive glaciation led to more extensive
thepurposesrequiredhere.Nonetheless,muchworkhas extinctions thanseenin the marineenvironment.
been undertaken on the fossil history of the Antarctic
Climatesince the Last Glacial Maximum
(seeforinstanceHaywoodetal.2009,StilwellandLong
2011,Barrett2013).ThecoolingofAntarcticwaterswas The last 15–20 kyr have seen considerable climatic
keyindrivingchangestothemarinefauna,leading,for change (Steig et al. 2000, Fountain and Lyons 2003,
example, to the evolution of antifreeze in teleost fishes Mulvaney et al. 2012). The Last Glacial Maximum
(DeVries1988,Chenetal.2008),thelossofaheatshock (LGM) ended at 15 kyr BP with a very rapid warming
response (Clark and Peck 2009), and of invertebrate (Bølling-AllerødEvent),whenmeanannualtemperature
groups such as many decapods (Clarke and Johnston increased by almost 128C over 1 kyr and the snow
2003), and evolutionary radiations in other groups. On accumulation rate increased. The Younger Dryas
land, physiological responses to desiccation and cold cooling event from 13–11 kyr BP then temporarily
May2014 ANTARCTICBIODIVERSITYSPATIALSTRUCTURE 207
R
E
V
IE
W
S
FIG.1. Continued.
dropped temperatures by ;48C. After 10 kyr BP there 12 kyr BP (Webster et al. 1996). As the Ross Ice Shelf
was a gradual cooling of ;58C, with sharp cooling thenretreated,thesedrained,leavingaseriesofsmaller
events 9.5 and 6.5 kyr BP. Finally, over the last 1000 lakes. Cooling uptoabout1 kyragowasaccompanied
years, there has been a warming of ;28C. A similar by lake evaporation to a minimum volume (Wilson
climatic history is evident on the sub-Antarctic islands, 1964, Lyons et al. 1998), and the lakes have since
although it varies with their spatial location (e.g., Hall partially refilled as temperatures have warmed. During
2002, Fraseretal. 2012). the high lake stands, previously deposited salts were
This historical background has left traces detectable solubilized, now underlying structurally important
inmodernterrestrialandaquatichabitats.Forexample, salinity gradients (Vincent et al. 1981), and providing
molecular signals are present in populations of Mc- nutrientsforchlorophyllmaximaatdepth(Priscu1995).
MurdoDryValleyspringtailsthatindicatetheinfluence The ecological history (or legacy) of a given area is
of ancient shorelines (Nolan et al. 2006). Lakes can important in the understanding of inland ice-free
preserveclearrecordsofchangeoverlongtimescales.In ecosystemssuchastheMcMurdoDryValleys,Vestfold
VictoriaLand,lakesreachedtheirmaximumextent10– Hills (Pickard 1986, Zwart et al. 1998, Moorhead et al.
208 PETERCONVEYETAL. EcologicalMonographs
Vol.84,No.2
1999,Lyonsetal.2001,Hodgsonetal.2004),Larseman Refugia andisolation
Hills (Burgess et al. 1994), and Schirmacher Oasis
Signals of spatial regionalization and temporal
(Bormann and Fritzsche 1995). Soils in the McMurdo
isolation are apparent in both terrestrial and marine
DryValleyregionhavebeenproducedbythesuccessive
environments of Antarctica (Griffiths et al. 2009,
movementsoftheEastAntarcticandWestAntarcticIce
Teraudsetal.2012).Inthesea,theoverwhelmingcause
Sheets through the Valleys as climate has fluctuated
ofisolationisthatoftheAntarcticCircumpolarCurrent
(Hall et al. 2000, Hendy 2000). Thus the soils in the
(ACC) and Polar Front (Clarke et al. 2005). This
easternendoftheTaylorValleyareyoung(;24kyrand
eastwards flowing current, formed between 24 and 41
less), while those in the western part of the valley are
mya,createsasteeptemperaturegradientof38–48Cover
much older (;75–130 kyr), indicating that even at the
a distance of tens of kilometers, and forms a strong
last glacial maximum the Antarctic terrestrial environ-
biogeographic discontinuity. Subsequent evolution in
ment was not completely covered by ice. The younger
the cold Antarctic marine environment has selected for
soilshavemuchlowerN:Pratios(Barrettetal.2007),as
stenothermal (Somero and DeVries 1967, Peck et al.
they have more soluble P (low weathering loss), while
2010b)andeurybathictaxa(Breyetal.1996),andledto
theoldersoilshavehighertotalnitrogencontentdueto high species-level endemism (50–70%) (Griffiths et al.
longerexposuretoatmosphericnitrateinput.Theoldest 2009, Convey et al. 2012a).At glacial maxima, most of
soil surfaces have higher conductivities because they the continental shelf was covered by ice, restricting
have accumulated atmospheric aerosols, leading to fauna to isolated refugia or forcing them into deeper
lower soil invertebrate abundance (Virginia and Wall water(Thatjeetal.2005,Conveyetal.2009).Cyclesof
1999). Some of the soil organic carbon in the Taylor contraction to refugia followed by re-expansion are
Valleyisofearlierlacustrineorigin,andisnowamajor likelytohavebeenamajorinfluenceonevolutioninthe
energysourcefortheterrestrialinvertebratecommunity Antarctic marine fauna (Clarke and Crame 1989, 1992,
(Burkins et al. 2000). The soil stoichiometry gradient is 1997,Fraseretal.2012),andincreasinglyitappearsalso
alsoreflectedinthelakeswithinthebasins(Priscu1995). on the Antarctic and sub-Antarctic terrestrial biota,
Older soil surfaces also exist in this area (mostly at which also shows substantial endemism (Greve et al.
higher elevations) and have been exposed for up to 2005,StevensandHogg2006,Stevensetal.2006,Pugh
several million years, having an order of magnitude and Convey 2008, McGaughran et al. 2010, Grobler et
S
W higherN:Pratioagain(Barrettetal.2007).Barrettetal. al.2011, Mortimer et al.2011a,b).
E (2007)hypothesizedthat,ingeneral,theinputelemental
VI composition (i.e., C, N, P) not only constrains produc- SPATIALVARIATION INTERRESTRIALSYSTEMS
E
R tivity within the ecosystem, but is modified by the Terrestrial Antarctica is characterized by limited and
biological processes occurring there by changing the insularexposureofice-freegroundaswellaspatchyand
stoichiometry ‘‘downstream.’’ As an example, the N:P discontinuous substrates. Thus, the potential to link
ratiochangesfrom21:1forglacialsnow/ice,to15:1for biological features with ‘‘continuous’’ physical environ-
cryoconite water, to 12:1 for stream water, to 25:1 for mental gradients such as temperature or water avail-
lake water. Over longer time periods, as lake levels ability is limited. The sub- and maritime Antarctic
fluctuatewithclimatechange,soilandlakenutrientscan islands experience less extreme daily and seasonal
be redistributed between the terrestrial and aquatic temperature variations than do locations within the
systems, providing legacy subsidies for both systems mainbodyofthecontinent,butshowmorevariationin
(Moorhead et al. 1999, Lyons et al. 2000). The other variables such as precipitation (Convey 2013). In
drawdownshaveconcentratednutrientsinthehypolim- consequence, investigations of biodiversity at a wider
niaofolderlakes(Priscu1995).Thiscryoconcentration range of spatial scales are possible (e.g., Terauds et al.
of nutrients in the deeper, saline portions also creates 2011).A spatial approach,basedon latitude,elevation,
important nutrient gradients within the lakes them- and distance from the coast (see Table 1), was first
selves. The diffusion of nitrate, ammonium, and proposed as a means to investigate variation in the
phosphate across the chemoclines drives production in Antarctic terrestrial fauna by Janetschek (1970). In the
the deep chlorophyll maxima of the lakes in the lastdecade interest in this approach hasincreased both
McMurdo DryValleys. for fundamental theoretical reasons (e.g., le Roux and
In the Windmill Islands region the best-developed McGeoch 2008a) and to address complex questions
vegetation communities are found in sites where related to climate change impacts in the polar regions
nutrientsweredepositedinpenguincoloniesabandoned (Chownet al.2012b).
thousandsofyearsago(Goodwin1993).Thevegetation Antarctic terrestrial organisms typically have patchy
changes from extensive moss beds in the lower-lying localdistributions(UsherandBooth1984,1986,Caruso
regions to lichen-dominated communities toward the and Bargagli 2007, Caruso et al. 2007, 2009, 2012a), as
summits (Melick and Seppelt 1997). Opposing resource do the microbiota (Caruso et al. 2011, Chong et al.
gradients accompanythe vegetationchange,withwater 2011). Various biotic factors, such as productivity and
decreasing and plant nitrogen content increasing with chlorophyll a content (Sinclair and Sjursen 2001),
elevation (Wasley et al.2006a, 2012). macroscopic vegetation (Sinclair 2001), and food
May2014 ANTARCTICBIODIVERSITYSPATIALSTRUCTURE 209
R
E
V
IE
W
S
FIG.2. TerrestrialandmarineenvironmentsintheAntarcticregion.(A)AmummifiedsealinMcKelveyValley,McMurdoDry
Valleys.(B)OnyxRiver,WrightValley,McMurdoDryValleys.(C)ThecushionplantAzorellaselago(Apiaceae)alongsideahigh-
altitudelakeonsub-AntarcticMarionIsland.(D)WanderingAlbatross(Diomedeaexulans)nests,lowlandtussockgrassland,and
mirevegetationonsub-AntarcticPrinceEdwardIsland.(E)Yeticrabs(Kiwasp.)ataScotiaarchydrothermalventsite.2500min
depth. (F) The impact of ice scour on the rich benthic diversity near Rothera Research Station, Adelaide Island, Antarctic
Peninsula.Photocredits:A–D,S.L.Chown;E,CourtesyofNERCChEsSoConsortium;F,K.Brown,BritishAntarcticSurvey.
preference(Kennedy1999)influencedistributionsofthe al. 1998, Bargagli et al. 1999, Porazinska et al. 2002b,
fauna (see also Hogg et 2006, Caruso et al. 2012b). Sinclair 2002, Poage et al. 2008) (see Table 2). Similar
Notwithstanding these factors, Antarctic biodiversity patterns are manifest in the sub-Antarctic, though here
variation and ecosystem functioning are predominantly interspecific (i.e., biotic) interactions may become more
driven by abiotic factors (Convey 1996), including soil significant(e.g.,Vogel1985,Frenot1987,Chown1994,
structure (Bo¨lter et al. 1997, Wall and Virginia 1998), BergstromandSelkirk1997,Smithetal.2001,Davieset
chemistry (Porazinska et al. 2002a), and, in particular, al. 2011, Lebouvier et al. 2011, Terauds et al. 2011).
wateravailability(Kennedy1993,Block1996,Powerset Thus, any discussion of spatial variation in diversity
210 PETERCONVEYETAL. EcologicalMonographs
Vol.84,No.2
TABLE 1. Dependence matrix of terrestrial and marine mustcommencewithatreatmentofthedirecteffectsof
physical factors vs. latitude, altitude/depth, and distance abiotic environmental variation. We do so here, then
from the coast in Antarctica (from Peterson and Howard-
Williams2001). discuss direct gradients, such as resource gradients
(including soils and productivity), followed by indirect
Terrestrialand Altitude/ Distance gradients, most commonly investigated with latitude or
marinefactors Latitude depth fromcoast
altitude as the independent factor. Although this
Terrestrialphysicalfactor approach follows, in broad outline, Austin’s (1980)
Totalsolarradiation 3 1 0 classification, the strong influence of solar radiation on
Temperature 3 2 1
water availability means that we accord it initial
Humidity 2 1 2
Wind 1 1 1 attention. Usually,or atleast in the caseof plants, it is
Precipitation(amount) 2 1 1 consideredaresourcegradient.Throughoutweconsider
Precipitation(type) 2 1 1
both strictly terrestrial and limnetic environments,
Ablation(amount) 2 2 2
Ablation(type) 2 2 2 which may include a variety of streams, lakes, and
Terrestrialgeochemistry 1 1 1 smaller water bodies such as cryoconite holes (see
Seasonality 3 1 0
Bergstromet al.2006).
Marinephysicalfactor
Polynya 2 0 1 Solarradiation
Current 2 2 2
Salinity 1 2 2 The two physical attributes that have the greatest
Seaicecover 2 0 1 influence on Antarctic ecosystems are the presence of
Iceshelfcover 3 0 0
liquid-phase water on land and of ice cover in marine
Notes: Key: 0, no dependence; 1, weak dependence; 2, and freshwater systems. These are both influenced by
medium dependence; 3, strong dependence. The original solar radiation and hence show a latitudinal response.
measurements for latitude were in degrees, for altitude/depth
were per 100 m, and for distance from the coast were in Other important factors include extreme seasonality
kilometers. combined with steep latitudinal gradients of light–dark
periodicity (Fig. 3) and exposure to high incident
radiationandtodamagingwavelengthsoftheradiation
S spectrum.
W
E
VI
E TABLE2. EnvironmentalstressorsforandresponsepatternsbyAntarcticterrestrialbiota.
R
Factor Protectionprovidedby Negativeeffectsofprotection
Radiation(PARandUV) screeningpigments loweredlightresponseofphotosynthesis;
costofproduction
shadedhabitat,refraction/reflectancedueto loweredPARandtemperature
plantformand/oroverlyingsnow
Temperature(includingfreeze– insulationbysnow loweredPAR;increasedrespirationcost;
thawconsequences) potentiallyshortenedgrowthseason
northerngrowthaspect increaseddesiccation
wetground(extracellularfreezingofwater surfaceicelayercanleadtoincreased
protectsactivemossesfromsubzero soilinvertebratemortality
temperatures(continent))
buttonform,lowstature potentialtoreducePARandthus
photosynthesis
darkpigmentation(enhancessolarheating) accelerateddesiccation
Wind groundbetweenrocks,soilcrusts,endo-and reducedlightavailability
chasmo-lithichabitats(stronglinkwith
wateravailability),leelocation
Desiccation(notestronglinkage south-facingaspect(continent) lowertemperatures
withwind,radiationand
endo-andchasmo-lithichabitat,soilcrusts loweredPAR;physicallimittoorganism
temperature)
size
sublithichabitat loweredPAR
permanentsnowbank/drift loweredPARandtemperature;increased
respirationcost;atcontinentalsites
insulationpreventswarminginearly
summer
flowingmeltwater substrateinstability;abrasion;burial
Notes:Protectionagainstthemainenvironmentalstressors(showninthemiddlecolumn)forAntarcticterrestrialbiotaalso
resultsinnegativeconsequencesforgrowthanddevelopmentasshownintherightcolumn.
May2014 ANTARCTICBIODIVERSITYSPATIALSTRUCTURE 211
FIG.3. Solarradiationatthetopoftheatmosphereasafunctionoflatitudeandtimeofyearforthezone608–908S(adapted
fromVincent1988).Valuesfollowingthelatitudevaluesaretotalannualradiation(MJ/yr).
The existence of gradients in exposure to solar nominal full sunlight, 2000 lmol photons(cid:2)m(cid:3)2(cid:2)s(cid:3)1
radiation is implicit in many studies of terrestrial (Pannewitz etal. 2003a,Schroeter et al.2011).
vegetation distribution, ecophysiology, and photosyn- Littleevidenceexiststhatnaturallyoccurringlevelsof
thetic biochemistry. These include photosynthetically PAR or UV radiation are major limiting factors for
active radiation (PAR), ultraviolet radiation (UV-A) autotrophs in Antarctica (Kappen et al. 1998a, Clarke
and seasonally anthropogenically enhanced levels of andRobinson2008,Schroeteretal.2012).UVradiation R
biologically harmful UV-B radiation. On an annual generally is lower in Antarctica than in tropical and E
V
basis there is a latitudinal decline in total radiation temperate areas; the mean UV index for January is ;3 IE
incident on the Earth’s atmosphere, from a monthly atRossIslandand;4atAnversIsland,comparedto7– W
S
average of about 22 MJ/m2 in the northern Antarctic 12inNewZealand.IncreasedUV-Bradiationasaresult
Peninsula to around 9 MJ/m2 at the South Pole. PAR of the ozone hole occurs between September and early
comprisesonlypartoftheradiationspectrum(400–700 December,sohaslittleeffectwithinthemainAntarctic
nm,typically43%ofthetotalradiationenergy),andthe continent, as vegetation activity occurs mainly within
fraction that reaches ecosystems at Earth’s surface is theperiodDecembertoearlyFebruary.Enhancedlevels
influencedbylarge-scalepatternsincloudinessaswellas do occur on the Antarctic Peninsula, but there is
reflectance through albedo, with the result that the evidence that the bryophytes at least can rapidly
radiation gradient experienced at ground level can be acclimate to changes in UV radiation (Newsham 2003,
reversed. Green et al. 2005, Newsham and Robinson 2009,
Insummertheintegrateddailyradiationisashigh,or Turnbull et al. 2009). Exposure does result in damage
even higher, in continental Antarctica than in many to biochemical constituents of cells (e.g., DNA and
temperateareas.Geography,however,canalsostrongly photosystems[Greenetal.2000,Georgeetal.2002,Lud
influence local weather conditions, consequently affect- etal.2002,TurnbullandRobinson2009,Robinsonand
ingtheproportionofradiationreachingecosystems.For Waterman 2013]), but these effects are transient. The
example, as a result of different cloud cover and path dynamicbehavioroftheprotectivepigments,whichare
length,islandsofftheAntarcticPeninsulareceiveabout typicallylostiftheplantsareshaded,suggeststhatthere
10% of radiation incident on the Earth’s atmosphere, is a metabolic cost to their presence. However, their
while on the polar plateau this can reach 90%. The productioncostsaresmall,estimatedat2%ofdailynet
almost continuous cloud in the northwestern Antarctic photosynthesis (Snell et al. 2009). While protective
Peninsulameansthatradiationlevelsatgroundlevelare ‘‘sunscreen’’ pigments are found across many mosses,
low and even, and temperatures relatively stable lichens, and microbes, and there is clearly greater
compared to elsewhere in Antarctica. As a result the production in more exposed locations (Post 1990, Post
active times of lichens are mainly confined to the two and Vesk 1992, Lovelock and Robinson 2002, Clarke
summermonthsatBotanyBay(778S)comparedtothe andRobinson2008),studiesquantifyingtheenergeticor
wholeyearatLivingstonIsland(628S)(Schroeteretal. resource investment in these strategies across exposure
2010). However, photosynthesizing organisms in the gradients are still required. Dynamic avoidance strate-
former region can experience incident PAR exceeding gies such as vertical migrations to avoid high UV have
212 PETERCONVEYETAL. EcologicalMonographs
Vol.84,No.2
also been demonstrated in Antarctic microbial mats hours(Pecketal.2006).Similarly,freshwaterpoolscan
(Nadeau et al. 1999). Such investment is arguably vary by .508C annually and by 208C daily. Small,
sufficient to generate a significant selection pressure, chemically stratified ponds in ice-free areas can have
suggesting that spatial variation in exposure will be summer temperature differences between surface and
mirrored in biological responses that affect organism bottom waters (;40–70 cm depth) of .108C (Healy et
energybudgetsandlifehistorytrade-offs(Convey2011). al. 2006). At broader scales, pronounced variation is
The interaction between insolation and desiccation is clear, as might be expected. For example, Howard-
important. The mosses and lichens that dominate the Williamsetal.(2010)foundasignificantlineardeclinein
limited continental Antarctic flora and survive in summer temperatures across Victoria Land. As an
extremely xeric environments, such as rock surfaces, integrating measure of ‘‘probability of melt,’’ they also
can avoid light stress by drying out rapidly, becoming calculated degree-days above freezing, finding a wide
inactiveandincreasinglytolerantofradiationincluding variation at latitudes above 808 S, with values rapidly
UV-B (Schlensog and Schroeter 2000, Kappen and droppingtowardzero athigherlatitudes.
Valladares 2007, Turnbull et al. 2009). Interspecific Variabilityinabioticconditionsisrecognizedasakey
differencesintheresponseofmossestodesiccationand feature affecting organismal responses (e.g., Tufto
UVradiation (Smith 1999, Robinsonet al.2003, Dunn 2000), and possibly also biodiversity variation (e.g.,
and Robinson 2006, Turnbull and Robinson 2009, Archibald et al. 2010). However, the effects of
Wasley et al. 2012), determine both their spatial temperature variation and its predictability per se
distribution on small scales, and their likely different (ratherthanmeanannualtemperatures),havenotbeen
future response to changes in the pattern and/or as widely studied in Antarctic organisms as elsewhere.
magnitude of stresses experienced. On exposed surfaces Most of the work in this area is concerned with the
that dry quickly, lichen development is limited to physiological implications of thermal variation at
crevices,interstitialspaces,orsouth-facingsurfacesthat several spatial scales both on the continent and on the
offer shade but also where a reliable water supply is sub-Antarcticislands(e.g.,Daveyetal.1992,Deereand
present. One result of this is the development of Chown 2006, Rinehart et al. 2006, Teets et al. 2011).
endolithic and chasmolithic communities, which are Nonetheless, adaptation to large diurnal temperature
completelyconfinedtothenorthfacesofsuitablerocks variation means that the terrestrial biota may not
S
W (Friedmann 1982). Mosses and lichens that grow in respond clearly to small local temperature gradients,
E mesicareaswheremeltwateristhemainsupplymayface especiallyifconfoundedbyotherenvironmentalfactors
VI veryhighlightlevelswhenactive.However,thespecific (Smith 2003). High natural variability in temperature
E
R conditionswhenbiotaareactiveatagivenlocationare may also underlie the comparatively small effects of
important. Temperatures, for example, are practically artificial warming often reported in climate manipula-
identicalforactivelichensinthemaritimeAntarcticand tion experiments (Sinclair 2002, Bokhorst et al. 2007b,
continentaldryvalleys,buttheirperiodsofactivityvary 2008) compared with changes in response to moisture
drastically (Schroeter et al. 2010). Interstitial, sublithic, availability and freeze–thaw cycles (Wasley et al.
and soil crust communities become relatively more 2006a,b, 2012, Yergeau and Kowalchuk 2008, Lenne´
important in drier areas, with translucent rock and soil et al. 2010).
particles allowing light transmission while protecting
Wateravailability
from desiccation (Hughes and Lawley 2003, Cowan et
al.2010). At a large geographic scale, a strong desiccation
As a result of these activity patterns, there is no (water availability) gradient exists from the sub- and
relationship at continental scale between light levels or maritime Antarctic into the continent (Walton 1984),
interceptedquantityoflightandgrowthrate.Lichensin thelatterincludingsomeofthedriestsitesonEarth.The
themaritimeAntarctichavesomeofthehighestgrowth availability of liquid water to organisms broadly
rates in the world for this group, while those in the depends on the balance between annual precipitation
continent can be one (Cape Hallett) or two orders of andlossesbyevaporation,sublimation,andfreezing.In
magnitude (McMurdo Dry Valleys) slower (Sancho et locations such as the McMurdo Dry Valleys, most or
al. 2007, Green et al. 2012). The duration of annual even all precipitation falling as snow may sublime, and
metabolic activity, controlled by water availability, thusnotbecomeavailabletobiota.Precipitationoccurs
appears to be correlated with richness (Green et al. mainly as rain in the warmer sub-Antarctic zones (and
2011a, Clarkeet al.2012). increasingly in the maritime Antarctic during summer),
andsnowonthecontinent.Additional,thoughnotwell
Temperature
quantified,sourcesaredew(Bu¨deletal.2008)andcloud
Antarctica contains among the most and the least ormist(especiallyforridgelinebiotasinmontaneareas).
variable thermal environments, with the former being Thisbalancecanbemodifiedbysnowremovalbywind
mostly terrestrial and the latter marine. Temperatures scour or addition by redistribution, or by release of
on open rock surfaces can vary by .758C through the concentrated precipitation during local melt of ice or
year and by at least 108C and even up to 508C within snow. The concentration of precipitation is particularly