Biodiversity patterns and their conservation in wetlands of the...

Records of the Western Australian lvluseum Supplement No. 67: 337-364

Biodiversity patterns and their conservation in wetlandsof the Western Australian wheatbelt

S.A. Halse l, M.N. Lyons!, A.M. Pinder l and R.J. ShieF

'Department of Conservation and Land Management, Science Division,P.O. Box 51 Wanneroo, Western Australia 6946, Australia

2Department of Environmental Biology, University of Adelaide, South Australia 5005, Australia

Abstract A total of 197 wetlands were sampled between 1997 and 2000 in asurvey designed to record wetland biodiversity across the wheatbelt andsouth coast of Western Australia, an area of 205000 km2 Altogether, 986wetland-associated plant, 844 aquatic invertebrate and 57 waterbird specieswere recorded, with an average of 73 and a range of 10-174 species perwetland. Thirty-four per cent of species were found at only one wetland.

Sixteen types of wetland were recognized, based on their biologicalassemblages. Overall, salinity was the factor most responsible fordifferentiation between wetlands in terms of biodiversitv but differencesbetween some freshwater wetland types, such as sedge sw~mps and graniterock pools, were largely attributable to other abiotic factors Among salinewetlands, the biota of naturally saline (and usually very salty) seasonal playaswas distinct from that of wetlands with longer penods of inundation. It wasunclear whether differences related to inundation or salinitv.

Usmg cluster analysis, 22 assemblages of co-occur;ing species wereidentified and the distributions of 18 of them were modelled. Between 33 and86'/'0 of the species richness of each assemblage at a wetland was explained bytwo to four abiotic variables. The assemblages that modelled most poorlyconsisted of species that were widespread and had broad ecologicaltolerances, with ranges extending beyond the wheatbelt, so that the surveywas unlikely to have circumscribed their ecological requirements. Mostassemblages consisted of a mix of plant, invertebrate and (fewer) waterbirdspecies. Factors affecting the distribution of plants and animals within a cooccurring assemblage often appeared to differ, especially for plants growingon the bank of a wetland. Riparian plants are probably exposed to differentenvironmental factors, especially salinity patterns, than those influencinganimals using the waterbody itself.

The dramatic increase in secondary salinisation that has been observed inthe wheatbelt and south coast over the past 100 years, with associated loss offreshwater habitat and changes to naturally saline playas, is likely to lead tosignificant loss of biodiversity. Most assemblages (and species) wereassociated with particular salinity ranges and there was an inverserelationship between overall community richness and salimty, especiallywithin the waterbody. Many species typical of naturally saline playas wererarely found at secondarilv saline wetlands.

Th'e predictable occurr~nce of some species assemblages, such as thosecharacteristic of sedge swamps and saline playas of the north-easternwheatbelt, makes it possible to identify wet lands that, if protected fromsecondary salinisation, will conserve large proportions of them. Some otherassemblages occur at many wetlands and their conservation is assured byalmost any strategy, even if they are unpredictably distributed. However,protection of rare species and assemblages that occur infrequently andsomewhat randomly within the wheatbelt poses a significant challenge towetland managers.

INTRODUCTIONSouth-west Western Australia is an old, deeply

weathered landscape with flat topography andlarge expanses of nutrient deficient soils (Mulcahy,1967; Wyrwoll, 1988). The steep rainfall gradientwith distance from the coast (Figure 1) means the

extreme south-west supports wet sclerophyll forestand permanent wetlands while more inland areascontain open woodland or shrub lands andepisodically flooded waterbodies (Hopper, 1979;Beard, 1990).

This paper reports results of a biological survey

338

Paynes Find

Albany

S. A. Halse, M. N. lyons, A. M. Pinder, R. J. Shiel

Kalgoorlie

ESP

.~,/~~

o 200

Kilometres

400j

Figure 1 Wheatbelt and south coast of south-west Western Australia showing the wetlands surveyed, rainfall isohyets,IBRA regions (Thackway and Cresswell 1995), and the Meckering line (Mulcahy 1967). 0, site on publicland; e, site on private land; GSP, Geraldton Sandplain; SCP, Swan Coastal Plain; AWB, Avon Whealtbelt;]F, ]arrah Forest; WR, Warren; ESP, Esperance Sandplain; MR, Mallee Region.

(WBS) of wetlands of the wheatbelt and south coastof Western Australia, with small incursions intoadjacent forested or open woodland areas. Manywetlands in the surveyed area were saline. Salt hasbeen accumulating in soil profiles andgroundwaters of the wheatbelt for hundreds ofthousands of years (Commander et al., 1994) as theresult of a higher rate of deposition of marineaerosol salt on the landscape than salt discharge viarivers and groundwater flow (Hingston andGailitis, 1976). Consequently, most playa lakesformed by groundwater discharge were naturallysaline (e.g. Salama, 1994), as were some riversystems on the south coast. Even most 'freshwater'wetlands had relatively high salt levels by globalstandards prior to clearing of native vegetation(Schofield et aI., 1988). However, nowadays manyof the saline wetlands in the surveyed area are saltybecause of secondary salinisation (Mulcahy, 1978;

George et aI., 1995). The clearing of perennial nativevegetation and its replacement with annual cropsand pastures caused run-off and recharge toincrease. As a consequence, groundwater levelshave risen and salt previously stored in the soilprofile, as well as in groundwater, has beenmobilised and percolated to the surface (Clarke etal., 2002). It has been estimated that about 6% ofland in the surveyed area is currently salinised(groundwater within 2 m of surface) and that thiswill eventually increase to 33% (Short andMcConnell, 2001; George et al., 2002). Theproportion of wetlands affected is much higherbecause of their low position in the landscape(Halse et al., 1993b, 2000a).

Salinity is a major environmental gradientstructuring aquatic communities (Hammer, 1986)and the historical prevalence of salt in wetlands ofthe surveyed area has resulted in an aquatic fauna

Aquatic biodiversity

that is relatively salt-tolerant (Williams et a/' 1991;Halse et a/, 2000a; Kay et aI, 2001; Pinder et aI,2002). Likewise, there has been considerableradiation of salt-tolerant plant species in, andadjacent to, wheatbelt wetlands (Short, 1982;Wilson, 1984; Lyons et ai, 2004). Nevertheless,secondary salinisation has the potential to wreakdevastating changes on the biodiversity of thewheatbelt because of loss of freshwater wetlandsand the likely changes to temporal and spatialpatterns of salinity in naturally saline wetlands(Williams, 1999; Cramer and I-Iobbs, 2002; Halse etai, 2003).

The impact of salinisation in south-west WesternAustralia is made more acute by the region's highconservation values. The extensive radiation ofvascular plant groups, especially the Myrtaceae,Proteaceae, Papilionaceae and Mimosaceae, haslong been recognized (Diels, 1906; Beard et a/' 2000)and the south-west was listed by Myers et al (2000)as one of 25 global hotspots for biodiversity, basedon a combination of terrestrial (mainly plant)species richness and the extent of land clearing.Many high-conservation value, endemic plantspecies occur in high-rainfall areas (WardellJohnson and Horwitz, 1996) but the number isgreater in what Hopper (1979) termed the'intermediate rainfall zone' between 300-800 mmannual rainfall. This is the area covered by the WBS.Hopper (1979) attributed plant richness in the studyarea to the existence of a mosaic of soil types andthe isolation brought about by climatic variability,mostly since the Quaternary.

The importance of south-west Western Australiafor aquatic invertebrates is less well documentedthan for plants but it appears to be a region ofsignificant richness and endemism for groups withdrought-resistant eggs, especially crustaceans (Frey,1991; Maly and Bayly, 1991; Thomsen, 1999). Thedistribution of crustaceans has parallels with plants:while many endemic species occur in areas of highrainfall (Storey et a/' 1993; Wardell-Johnson and!-!orwitz, 1996), much of the crustacean endemismoccurs in intermediate and low rainfall zones (Halseand McRae, 2001; Remigio et ai, 2001; Halse, 2002;Timms, 2002).

The WBS was a direct response by theCovernment of Western Australia to the threat ofbroad-scale loss of biodiversity because ofsecondary salinisation (Anonymous, 1996b). It hadthree objectives: (1) to document patterns ofbiodiversity in wetlands of the wheatbelt and southcoast, (2) to investigate the role of salinity and otherenvironmental factors in structuring communities,and (3) to select a set of Natural Diversity RecoveryCatchments as a focus for government andcommunity actions to ameliorate the impact ofsalinity on biodiversity and thus conserverepresentative wetland communities. Existing

339

information relevant to the three objectives isreviewed below.

Prior to the survey, there had been littleinvestigation of biodiversity patterns in wetlands ofthe surveyed area other than for waterbirds (Jaenschet ai, 1988; Halse et a/, 1993b, 1995). Studies ofinvertebrates were limited to a few wet lands (Halse,1981; Williams et al, 1991; Doupe and Horwitz,1995; Halse et a/' 2000a) or focus sed on particulartaxonomic groups (Ceddes et al., 1981; Brock andShiel, 1983). The comprehensive survey of rivers byKay et al. (2001) identified invertebrates only tofamily. Studies of submerged plants were few(Brock and Lane, 1983; Brock and Shiel,1983) and,other than the broadscale survey by Halse et al.(1993a), studies of emergent and riparian vegetationwere restricted to particular wetlands (Froend et al.,1987; Froend and McComb, 1991; Froend and vander Moezel, 1994). Substantially more informationon plants and invertebrates, as well as diatoms, isavailable as a result of the WBS (Pinder et al, 2000,2002, 2004, 2005; Blinn et ai, 2004; Lyons et al.,2004).

The general role of salinity in structuring plantand animal communities within a waterbody isreasonably well understood. While there is a trendfor community richness to be inversely related tosalinity, the affinity of different higher-leveltaxonomic groups for saline conditions varies(Hammer, 1986). These patterns were confirmed inthe WBS with Pinder et al. (2005) showing a tightnegative relationship between invertebrate richnessand salinity at values >4 g L-l across bothsecondarily and naturally saline wetlands.Waterbird and diatom richness also showednegative relationships with salinity (Blinn et a!.,2004; Cale et a!', 2004). The role of salinity instructuring riparian plant communities, especiallythose of naturally saline systems, is less clear (seeHart et al, 1991). Halse et al. (1993a) found norelationship between plant species richness andnatural salinity of wetlands, although secondarysalinisation reduced richness. Lyons et al. (2004)reached similar conclusions, after comparing soilsalinity and plant richness of small quadrats invarious vegetation associations around wetlandsduring the WBS. Richness of riparian plants waspoorly correlated with wet land salinity probablybecause the salt content of riparian soils reflectsmicroscale topography and leaching rather than thesalinity value of wetland water (Cramer and Hobbs,2002).

The practice of reserving or managing areas toprotect the biota within them, and thus the floraand fauna of the region, has a significantinternational history (Margules and Pressey, 2000).The first national park in Western Australia (JohnForrest) was gazetted in 1900 but it was another 52years before the first nature reserve was proclaimed

340

(Tammin Railway Dam). Between then and 1980,many wet lands within the surveyed area werereserved to provide habitat for waterfO\vl, as wellas being places where ducks could be shot (Lane,1985). Other wetlands were included in naturereserves where the primary purpose was protectionof terrestrial flora and fauna. Early reservation wasan ad hoc process but, in recent years, a number ofrelatively sophisticated mathematical methods havebeen developed in Australia, and elsewhere, tomaximise the number of species from the regionalspecies pool that are protected by reserving a givenarea. The focus of these methods has been terrestrialecosystems (e.g. McKenzie et ai, 1989; Nicholls,1989; Justus and Sarkar, 2002; Scotts and Drielsma,2003) but many of the principles apply to wetlandreservation. In conjunction with the development ofanalytical methods, there has been a program ofregional surveys in Western Australia to providethe biological data on which to base decisions about

S. A. Halse, M. N. Lyons, A. M. Pinder, R. J. Shiel

reservation (Biological Survey Committee, 1984;McKenzie and Robinson, 1987; McKenzie et ai,1991, 2000), with aquatic ecosystems being includedin these surveys since the mid-1990s (Halse et al,2000b).

Selection of Natural Diversity RecoveryCatchments (BRCs) to ameliorate impact of salinityon biodiversitv uses the same information as areserve selection process. There are currently sixBRCs in south-west Western Australia (Figure 2).Each BRC is a sub-catchment of 50 000 - 120 000 hawith several or many wetlands at low points withinthe catchment. All contain a small number of naturereserves, and some remnant vegetation on freeholdland, but much of each catchment is clearedagricultural land. Public money is being spentwithin these catchments on salinity control,revegetation and management of uncleared landwith the objective of maintaining existing levels ofbiodiversity. At the Toolibin BRC, groundwater is

Albany

Kilometres

Figure 2 Wheatbelt and south coast of south-west Western Australia showing existing Natural Diversity RecoveryCatchments (hatched), major nature reserves and the Potential Recovery Wetlands selected as a result of thesurvey.• PRW.


being pumped from under Toolibin Lake to lowerthe water-table and maintain health of the lakevegetation (Froend et ai, 1997; Dogramaci, 2003). Itis intended that an additional 10-20 BRCs will beselected and that most of the species assemblagesidentified in the WBS will be protected in thesecatchments (see Anonymous, 1996a; Keighery,2001).

STUDY AREA

The area covered by the WBS is shown in Figure1. Nearly all wet lands were located between the 600and 300 mm isohvets. The area has a Mediterraneanclimate with hot dry summers and predominantlywinter rainfall, although the proportion of summerrain increases with distance to the north and east(Gentilli, 1972). Annual evaporation varies from1320 to 2750 mm. Rainfall was relatively light, andwetland levels low, during 1997 and 1998 (Figure3). Extensive late summer and au tumn rains in 1999meant that wetlands remained extensively floodedthroughout that year. Rainfall patterns within thesurveyed area during 2000 were similar to 1997 and1998, although there was extensive rain farther east.

The different landforrns and vegetationassociations of the surveyed area are reflected inthe IBRA regions it covers. The boundaries of theseregions largely reflect geology and vegetationformations (Thackway and CresswelL 1995). Mostof the wetlands surveyed are in the AvonWheatbelt, Esperance Sandplains and GeraldtonSandplains regions, with a small number in theMallee and Jarrah Forest. Lake Cronin (SPS003) isin the Coolgardie region. In broad terms, the naturalvegetation of the Geraldton and EsperanceSandplains consists of shrub lands, the AvonWheatbelt contains open eucalypt woodlands andthe Mallee contains eucalypt mallee formations(Gibson et aI., 2004). However, surrounding plantformations have relatively little influence on plantspecies occurring within the regularly inundatedparts of wetlands. It must also be recognizedapproximately 90'10 of the Avon-Wheatbelt, 80% ofthe Mallee, 73% of the Geraldton Sandplains and55'i{, of the Esperance Sandplains have been clearedof their original vegetation during the past 100years (Shepherd et al. 2001), with broad-acrecropping now the main activity.

Physiognomy and patterns of inundation ofwetlands in the surveyed area are highly variable.No wetland is permanently filled, although some ofthe deeper wetlands in the most south-westerly partof the surveyed area, around Lake Muir, rarely dry.West of the Meckering line (Figure 1) and along thesouth coast, rivers have defined channels, flowseasonally for several months and usually dryduring summer. Farther inland, drainage linesconsist of broad palaeochannels that are rarely

341

connected to active rivers and contain mosaics offlat channels and small playas (Mulcahy. 1967;Beard. 1999). They contain water only briefly afterlarge rainfall events. Lentic waterbodies includesemi-permanent or seasonal basin wetlands, oftenwith rivers flowing into and out of them that mayconnect a series of wetlands, such as L.ake Toolibinon the northern Arthur River (Froend et ai, 1997).Most basin wetlands are in the western andsouthern parts of the surveyed area. In higherrainfall zones, the bed is usually covered by sedgeswhile lower rainfall results in trees growing acrossthe bed unless the centre of the wetland remainswater-logged for prolonged periods (as oftenoccurs), when it will be open (Halse et al. 1993a).Shrubs such as Muehlenbaeckia spp. sometimes occurinstead of trees in wetlands of low rainfall zones.

Other lentic wetland types include seasonally orepisodically filled playa lakes. These are morecommon in the eastern part of the surveyed areaand vary in size from a few 100 m 2 to several 100km 2

. Many are naturally saline and support lunetteson the downwind side (Bowler, 1983) withvegetation in low-lying areas around the wetlandmargin consisting mostly of chenopods. Playa lakesare maintained as landscape features bygroundwater discharge (Salama, 1994; Harper andGilkes, 2003) but most fill only after surface or subsurface inflow.

A variety of vernal pools, claypans and pools ongranite outcrops also occur. Vernal pools areperched above regional water-tables, usually onimpervious clay sediments, and seasonally containshallow water for a few months. They are fed bysurface run-off from their immediate catchment andsupport dense stands of sedges and small shrubsacross the wetland. They form only in areas ofmoderately high rainfall and occur mostly on thewestern edge of the surveyed area. Claypans areanalogous to vernal pools, occurring in low rainfallareas in the eastern part of the surveyed area. Whenflooded they contain either open water (often veryturbid) or emergent plants, such as Tecticorniaverrucosa, Muehlellbaeckia j70nilenta and Eragrostisaustralasica. Granite rock pools have beencomparatively well studied (Hopper et aI., 1997;Bayly, 1999; Pinder et ai, 2000). They occur in smalldepressions on inselbergs and share manysimilarities with vernal pools although, because oftheir rock substrate, they usually do not supportemergent vegetation.

The surveyed area extended to both the south andwest coasts (Figure 1) and, near the coastlines,saline lakes of marine origin are present. These areeither part of old estuary systems or shallow lakesin inter-dunal swales (Hodgkin and Hesp, 1998).They usually support extensive chenopd marsheson their margins and sometimes stands of salttolerant lvlelalellca cllticularis.

342 S. A. Halse, M. N. lyons, A. M. Pinder, R. J. Shiel

(a)

(c)

"''''-5P'di'".::::: :-"?'".........~

\b,lU~

(b)

I Iight:si on Record

VerY much above AverageDecile 10 '

Above AverageDeeile 8 - <)

AverageDecife 4 - 7

Below Avera"eDecile ~ - 3"

Very much below averageDecile 1

Figure 3 Rainfall deciles in the surveyed area during 1997-2000, and rainfall over all Western Australia during 1999.(a) 1997, (b) 1998, (c) 1999, (d) 2000, (e) whole State in 1999 (data from Bureau of Meteorology).


Secondary salinisation has already affected manywetlands of the surveyed area. The most visibleeffects are increased inundation, increased salinity,and death of lake-bed and riparian vegetation(Cramer and Hobbs, 2002; Halse et aI., 2003)

METHODS

This paper is based data from 197 wetlands(Figure 1, Appendix 1), which are a subset of thosesurveyed for plants by Lyons et al. (2004) andaquatic invertebrates by Pinder et al. (2004). Dams,reservoirs and wetlands receiving hypersalinedischarge water were excluded from our analysis,which dealt only with naturally occurring wetlands.Most of these were on Crown land (nature, water orrecreation reserves or unvested) but 65 were onfreehold land, which had usually been cleared foragriculture except for a narrow belt of riparianvegetation, and seven were on pastoral leases.Many wetlands in reserves also had little unclearedland around them and in some cases the reservedarea did not encompass the whole wetland (seeHalse et ai, 1993a).

The WBS examined patterns in composition anddistribution of the overall biological communitiesin wetlands, as represented by three elements withdifferent life history traits: namely waterbirds,aquatic invertebrates, and vascular plants withinthe wetland and riparian zone. We define theriparian zone as extending to the high-water markof major, regular (though not necessarily annual)flood events. Aquatic invertebrates and waterbirdsin wet lands of the central third of the study areawere surveyed in spring 1997, the southern thirdin spring 1998 and the northern third in spring1999 (Pinder et ai, 2004). Four additional wetlandswere surveyed in 2000. Wetlands of southern andsome central parts of the study area were surveyedfor plants in 1998 and those in northern and theremaining central parts were sampled in 1999 withspring and summer visits to each wetland tocoincide with periods of maximal floweringactivity (Lyons et al., 2004). Five wetlands weresurveyed in 2000.

There was temporal disparity between theanimal and plant data-sets for many wetlands,including all those sampled for fauna in 1997.Conditions did not change much between animaland plant sampling if both occurred in 1997 and1998, because these years had similar rainfallpatterns (Figure 2), but differences weresometimes marked if plants were sampled in 1999and animals during the preceding two years. Wehave been unable to compensate for temporaldiscrepancies except for seven wetlands in thedata-set of Pinder et al (2004) that were sampledtwice, in different years. For these, faunal datawere used from the year when wetland conditions

343

were most similar to those when plant data werecollected.

A species list of vascular plants, invertebrates andwaterbirds was compiled for each wetland basedon scoring all plant species within a variablenumber of 100 m2 quadrats at the wetland (Lyons etai, 2004), taking two 50 m invertebrate sweeps(Pinder et al., 2004), and surveying either the wholewetland, or a large section of it, for waterbirds. Weuse the term "species" loosely to refer to the animaland plant units used in analysis. Identificationswere usually at species level but existing taxonomickeys enabled some invertebrate groups to bedistinguished only at high taxonomic levels (e.g.Nematoda). Even when keys were adequate, it wassometimes impossible to identify all animals orplants of a genus to species level because specimenswere immature or sterile and then it was necessaryto lump identifications at genus level for analysis.

AnalysisSingleton species were excluded from the data-set

for all multivariate analyses. Wetlands wereclassified into types according to the similarity oftheir biota using the PATN analysis package(Belbin, 1993) and presence/absence species data.Czekanowski's coefficient was used as a measure ofdissimilarity and under-estimated dissimilarities(>0.95) were re-calculated using the Shortest Pathoption. The Unweighted Pair-Group Mean Average(UPGMA) fusion method, with ~=-0.1, was used togroup wetlands (Sneath and Sokal, 1973). Thediscreteness of wetland types identified byclassification was examined by ordination usingSemi-Strong Hybrid Multidimensional Scaling(SSH) (Belbin, 1991).

Species were classified into assemblages ofspecies with similar patterns of occurrence usingthe Two-Step coefficient of dissimilarity (Austinand Belbin, 1982) and UPGMA. The degree ofnestedness in each assemblage was calculated usingNESTED (Atmar and Patterson, 1993, 1995).Nestedness was further examined by checking thatrichness of each assemblage was unimodallydistributed against I-dimensional ordination scoresof community structure, derived by SSH withoutmasking singletons. Whether richness values ofeach assemblage fitted a Poisson distribution waschecked by visual inspection after plotting them.

For each assemblage that appeared to beecologically meaningful, the relationship betweenspecies richness of the assemblage andenvironmental attributes across the 197 wetlandswas modelled using a generalised linear model(Poisson regression) in the STATISTICA analysispackage (StatSoft, 2001). The environmentalattributes available for modelling related togeography, climate, water physico-chemistry andtwo particular habitats (Table 1). Environmental

344 S. A. Halse, M. N. Lyons, A. M. Pinder, R. J. Shiel

Table 1 Environmental attributes measured at, or derived for, each wetland and used in modelling. Ionic ratioscalculated using milliequivalent values.

Code Attribute Code Attribute

Mg LI calcium carbonateDissolved persulphate nitrogen (mg L I)Dissolved persulphate phosphorus (mg LI)Chlorophyll a,b,c, Phaeophytin (mg L 1

)

GeographicLat Latitude (OS)Long Longitude (OE)Alt Altitude (m)ClimaticTann Annual average temperature (QC)Pann Annual average precipitation (mm)Pdry Driest quarter precipitation (mm)Pcv Coefficient of variation precipitationEvap Annual average pan evaporation (mm)Physico-chemicalSaP Total dissolved solids (mg L-I)pI-PAlkTNTPChi

I Used as either a continuous or categorical variable

TurbColSilNa'Ca2'

M g2'

K'Mn2•

ClHCO,SO/DMCCSCCHabitatRockFlow

Turbidity (NTU)Colour (TCU)Silica (mg LI)Sodium (% meq)Calcium (% meq)Magnesium (% meq)Potassium (% meq)Manganese (% meq)Chloride (% meq)Bicarbonate (% meq)Sulphate (Ya meq)Mgl.+ Cal' : ClCa2

': SO/Cal' : HC0

3+CO/-

Waterbody on granite outcropFlowing water when sampled

attributes were screened before constructing eachmodel and attributes that were not significantlyrelated to richness of the assemblage were excludedfrom the regression analysis. If several attributeswere strongly inter-correlated, only the one withmost obvious biological meaning was included inanalysis. A regression equation was constructedwith two to four environmental attributes, using theBest Subsets routine. Significance of the equationwas assessed using the Wald statistic after checkingthat outliers were not disproportionatelyinfluencing coefficients. The amount of variationexplained was calculated by an Rl-value adjustedfor number of cases and environmental variablesused (Tabachnik and Fiddell, 1983).

In an attempt to define the habitat preferences ofdifferent assemblages, their fidelity to each wetlandtype was calculated as:

F 4i = PAi • w/w i • (l/LpAi' w/w)1-n

where F4i

is the fidelity of assemblage A to wetlandtype i, p Ai is the proportion of assemblage A inwetland type i, w i is the number of wetlands in typei and w· is the average number of wetlands per type(see Boesch, 1977).

The WBS results were used to identify somePotential Recovery Wetlands (PRWs) that couldform the core of future BRCs. The PRWs representthe surveyed wetlands containing the largestproportion(s) of the assemblages that appeared toneed active protection to persist in the surveyedarea in the face of further salinisation. Theproportion of each assemblage that would beconserved if the PRWs and existing BRCs wereprotected while the remainder of the surveyed area

became salinised was calculated to provide notionalindication of the adequacy of the proposed PRWs.The calculation assumed eventual loss of all speciesoutside the PRWs and BRCs, although we recognizethe assumption is unrealistic (see Halse et aI., 2003).

RESULTS

Altogether, 986 wetland or wetland-associatedplant, 844 aquatic invertebrate and 57 waterbirdspecies were recorded, with an average of 73 (range10-174) species per wetland (Appendix 2). Fifteenper cent of plant species were naturalised aliens(weeds). Species occurrence was very patchy with34% of species occurring at a single wetland, 15% attwo wetlands and only 1.1 % at more than one-thirdof sites. No species was present at all wetlands andfive of the 10 more common 'species' were higherlevel taxa. The more common true species were thechironomid midge Procladius paludicola (58% ofwetlands), introduced herbs Hypochaeris glabra(54%) and Sonchus oleraceus (43%), introduced grassParapholis incurva (49%) and grey teal Anasgibberifrons (48%).

Species assemblagesAfter deleting singleton species from the data-set,

22 species assemblages were recognized but threeof these appeared to be artefacts driven by thespecies array at individual sites (assemblage 15 byLake Pleasant View SPM024, 19 by Arro SwampSPS183, 22 by Goonaping Swamp SPS023)(Appendix 3). Information about assemblages issummarized in Table 2. The assemblages were:Assemblage 1. A group of 27 species mostly

Table 2 matrix ofmember was recorded,

wetbnds of each

,lS:il'rnbla\,:l'S (rows) to wetland types (columns), N represents the number of wetlands of each type or the number of times an "c;,;prnhh"p

and S represent the mean number of species per wetland and the number of species in each assemblage, respectively, Mean TOS (g Land for occurrences of members of each assemblai!e, and mean oH at wetlands of each tvoe arc also shown. Sce text for method of calculatin"

;;...0r:;l:J

Cl

c;-C'0.

<'re....'",~

53 16Y

74 516 22±06

27 127 2.6±1

8Y 854 30±OA

84 244 3.3±1

51 232 43±1

64 2381 4Y±03

62 191 7.

70 88Y ryI

106 577

61 183 24±3

35 2124 24t I

31 452 26±2

30 43Y 26±2

63 270 32±4

68 2236 44±1

37 802 59±2

44 156 64±4

100 603 90±3

""...(>1

0.000 0000

0002 0002

0.000 o 000

0.001 0001

0004 (W05

0008 0000

0.004 0006

0005 0007

0007 0008

0021 0005

0.036 0032

0.033 0.027

0.095 0.036

0043 0.010

0.013

0.302 0.204

Wetland type

VI I VIII IX II VI IV V III XI X XII XV XVI XIII XIVN 13 13 .'1 12 18 11 12 12 10 4 2.'1 11 22 12

100±Y 6Y±13 108±5 78±3 74±5 8th5 68±7 103±4 54±3 5Y±5 73±4 2~±,S 52±6 50±605±O2 1 ±OA 4.0±0.Y 38±]0 43±lJ6 45±11 58±1.1 16±3 39±6 48±16 5l±8 116±27 128±21 146±35

76±OA 7 lJ±lJ 2 8.lJ±O.2 7.6±0.2 85±0.2 84±0.2 8.2±0.2 7.6±0.3 8.8±lJ.2 8.6±lJ.2 8.Y±0.2 47±lJ7 8.0±0.1 51 S N TDS

21 • 0006 0.102 0.000 0.036 0.007 0024 0018 0007 OODO 0000 0000 0.003 O.OOD

20 • • 0.171 W.il;. 0.081 0.066 0.02Y OODY 0020 0050 0.005 0000 0012 0.010 0000

0(133 0015 O(m 0.01 4 0025 0042 0021 0.000 0.000 0.0l1

17 0.054 0058 0082 0016 0.093 0.104 0.078 005Y 0.023 0010 0.012 0.003 0002

8 0.055 0007 0.000 0.003 0013 0028 0008 0019 0048 0000 0.000 0002 0.000

16 - 0088 0183 0.025 0.021 0019 0.030 0.018 0.068 0011 0.037 0004 0.000

5 0.081 0.149 0.076 0,069 0.110 0.079 0.095 0.096 0.133 0036 0.044 0.009 0.0l 2 0001

4 0.051 0051 0024 .. OOlO .. 0.146 0035 0006 0.000 0005 0.000 (W05 0008

8 0.069 0.103 0.051 0.082 0036 0.128 0.168 0.124 0.135 0.142 0.046 002] 0.002 DO I(]

2 0.027 0.088 0025 0002 0.097 0.193 0.132 0.134 0.144 0026 0.061 0013 0.023 001

3 0011 0007 002Y 0036 0.011 0.082 0.127 0020 0048 .. 0.023 0.199 0000 0.006

6 0.03Y 0.077 0.025 0033 0.069 0.059 0.093 0.073 0.110 0.067 0.156 0.059 0.056 0.025

7* 0.119 0.142 004Y 0045 0.107 0.032 0.027 0016 0.079 0.061 0.068 0.044 0.061 0017

9 0008 0.007 0022 0.065 O.OIY 0028 0.152 0022 0.078 l1li 0.060 0.180 0017 0019

12 0.126 0043 0.154 0004 ... 0.037 OOOY 0009 0.056 0.000 0.021 0.019 0.179 0.006

13 0.018 0.030 0010 0004 0.056 0.012 0.028 0.031 0.090 0.050 __ 0.087 0.117 0.039

14 0.022 0004 OOlY 0.000 0.109 0.000 0003 0006 0.048 0005 0.085 0036. 0.02Y

11 0000 0.000 0.000 0.000 0.052 0.000 0000 0.010 0.023 0000 0038 0.Cll1 ..' 0.015

10 0.002 0003 0.012 0.000 0.030 ()(lOO 0007 0005 0.002 0.005 0.120 0.125 0.100 0.085

346 S. A. Halse, M. N. Lyons, A. M. Finder, R. J. Shiel

•.~ ..' ..

I .'

:: .- .•

;, ••

•• • 0 o.0

•

....- .. \

, .00

• 0

.0

,.

•

Assemblage 4

o'

I .•...

o

..t .. ,.-~

• Assemblage 6

••. ..,, .f·· •.\• •tt. • •;., .. \ ...-..

• ••# •I. ...... ... ",. .~.., ..

• • • •.r. ...: ". I : 0

:- 0 A I·:· ,.tl',tII6.· ..•• •lit: o· ~.j:

•....•

I _f

••

• ',0

••o

Assemblage 5

..• 0

0......

,.•• '. ~."!. ,··Iliro•o 0

~~,,/,

• •

••• • 0

••• •• • ••. '-:. .

• o •• '• •..'..: ••e o •

• •....,. ...

••. '.,. .~f ....•

Figure 4 Distribution of species assemblages at wetlands surveyed in the wheatbelt. Assemblages 1-6.

l_________


associated with fresh water (mean 2.6±1.0 g LI)that represented 0.9(~, of species occurrences inthe survey. The assemblage contained 11invertebrate, 1 waterbird and 15 plant species.Members were absent from most wet lands andthe assemblage tended to occur in the easterncentral and southern wheatbelt (Figure 4). Itwas best represented at Lake Cronin (SPS003,15 species).

Assemblage 2. A group of 106 species found mostlyin association with subsaline water (mean10.6±1.1 g Ll). The assemblage contained 54invertebrate, 10 waterbird and 42 plant speciesand represented 4.3% of all species occurrences.Members of the assemblage were distributedthroughout the wheatbelt, with few at anyonesite (Figure 4). The assemblage was bestrepresented at Murchison River (SPSI98, 29species).

Assemblage 3. A group of 61 species found mostlyin or adjacent to saline rivers and wetlands ofthe south coast (mean 23.5±1.1 g L I) (Figure 4).The assemblage contained 29 invertebrate, 1waterbird and 31 plant species and represented1.4% of all species occurrences. The assemblagewas best represented at the Old field River(SPSI23, 12 species).

Assemblage 4. A group of 62 species in orassociated with fresh to subsaline, flowingwater (mean 7.1±2.5 g Ll). The assemblagecontained 38 invertebrate and 24 plant speciesand represented 1.4% of all species occurrences.Members were concentrated at stream sites inwandoo woodland east of Perth (Figure 4) andthe assemblage was best represented at]imperding Brook (SPS004, 20 species).

Assemblage 5. A group of 64 species widelydistributed in the wheatbelt (Figure 4) inassociation with fresh to subsaline water (mean4.9±0.3 g Ll). The assemblage contained 56invertebrate, 1 waterbird and 7 plant species,representing 17.7(),;, of all species occurrences,and was best represented at Range Road yateswamp (SPS033, 39 species).

Assemblage 6. A group of 35 widespread (Figure4), salt-tolerant species (mean 23.7±0.9 g LI).The assemblage contained 12 invertebrate, 10waterbird and 13 plant species, representing15.8'Yo of all species occurrences, and was bestrepresented at Dulbinning Lake (SPS007, 25species).

Assemblage 7. A group of 31 species found over thefull range of salinities (mean 26.1±2.3 g LI)(Table 2). The assemblage, which contained 10invertebrate, 1 waterbird and 20 plant species,was patchily distributed across the wholewheatbelt and represented 3.4°!c, of all speciesoccurrences. Plant species were nearly all exoticweeds, without conservation value. The

347

assemblage was best represented at ArroSwamp (SPSI83, 14 species).

Assemblage 8. A group of 70 species foundmostly in fresh to subsaline or weakly salinewater (mean 7.2±0.7 g L I). The assemblagecontained 47 invertebrate, 2 waterbird and 21plant species, representing 6.6% of all speciesoccurrences. It was patchily distributedthroughout the wheatbelt but with a tendencyfor greater frequency of occurrence on thesouth coast (Figure 5). It was best representedat Boyacup Bridge swamp (SPS111, 20species).

Assemblage 9. A group of 30 halobiont species(mean 26.1±2.0 g LI) that occurred mostfrequently along the south coast and southwestern part of the study area (Figure 5). Theassemblage contained 11 invertebrate, 1waterbird and 18 plant species, representing3.3% of all species occurrences. It was bestrepresented at Oldfield River (SPSI23, 15species).

Assemblage 10. A group of 100 species associatedwith saline and hypersaline water (mean 90±3g Ll) outside high rainfall and, usually, coastalareas (Figure 5). The assemblage wasdominated by plants, containing 22invertebrate, 7 waterbird and 71 plant species,which represented 4.9% of all speciesoccurrences. The site with most species of theassemblage was Isthmus Lake (SPS058, 24species).

Assemblage 11. A group of 44 species associatedwith hypersaline water (mean 64±4 g LI) of thenorth-eastern wheatbelt (Figure 5). Theassemblage comprised almost exclusively plantswith 2 invertebrate, 1 waterbird and 41 plantspecies, which represented 1.2% of all speciesoccurrences. It was best represented at LakeMoore (SPS148, 19 species).

Assemblage 12. A group of 63 species associatedwith some of the freshest and some of the mosthypersaline wet lands (mean 31.8±3.6 g L I) inthe north-eastern wheatbelt (Figure 5). Theassemblage was dominated by 47 plants specieswhich, together with 16 invertebrate species,represented 2.0% of all species occurrences.There was considerable discrepancy betweenplants and invertebrates in terms of thesalinities with which they were associated(Table 3). The assemblage was best representedat Yarra Yarra Lake (SPSI62, 12 species).

Assemblage 13. A group of 68 species mostcommonly found in or near secondarily salinewetlands (mean 44±1 g Cl). The assemblage wasWidespread in the wheatbelt (Figure 5) andaccounted for 16.6% of all species occurrences.It comprised 27 invertebrate, 4 waterbird and 37plant species and was best represented at Lakes


\

Perth '~ir

Assemblage 9Percentage of asseI';lblage sp.x'\es at

.5010<75.3010<50

•.

r<algcorlil';

•• •...

.... .

:...•

\perth~

j

\\

~.

\ .\

\

,.

";. Kalgoor!le

"

..

•

..

Assemblage 8Percentage ot assemblage s;;aX;;E:'s at each sile

• :5:0<30• 5 to<15• 0 \0<5

.'.........•

...-.· ...•

f·"r.\.\\

\

r~ ..\.

Getaldtotl ~\ •

\, .

•

Assemblage 11Percentage of assemblage species at s<lch site

.30 to <50e15tc<30• 510<15

010<5o

.,

Kalgooriie...•

.~

•••• •

•••

•

.... ,

••• ..• • •••

••

'.

...

'.'

....,fI'

•

.'

· ..~•

; .·.\

\

Albany

~.... ..·(16)

•

Assemblage 13Percentage of assemblage species at eac!", sIte

.30 te <50815to<30• 510<15

Cto<$o

e, Ka-lgoofhe

· .•'.

Assemblage 12Percentage of assemblage species at each site

81510<30

• 510<15010<5o

• • ..•.' ..

•

..

..,..,• •• •

Geraldton

Figure 5 Distribution of species assemblages at wetlands surveyed in the wheatbelt. Assemblages 8-13,

Aquatic biodiversity 349

(4\

.50

.30

• to <30·

•

Assemblage 16P"'M""" '" '"e~"'..,p speCle-s at e8{:h

••..

...••

~.

. ",. .,

•• ••.' . •.•

.. .-.- ..•

".,...• • 1.~. .,

• ,..

Albany

Gen31dton

• Assemblage 18of "",mblage spems

~.) Kalgoor1:e

• .'.\ ."

• • ,.'• . ;.., • •

..

•

1 •

Albany

.30 la <50.,5·Assemblage 21

Percentage

••

•

Assemblage 20Percentage of assemolage spec,t'c'S at eaCh

••

•

• •

....

• •

•..

•

••. . • • • •

• •·

Figure 6 Distribution of as'3ernblai~esat wetlands ~111r\f{"",rl in the wheatbelt. AsseJrntJla);;es 14-21.

350

Table 3 Mean (±SE) salinity of occurrence (g LI) ofplant, invertebrate and waterbird species ineach assemblage. SE shown only if there were~5 occurrences in a category

Assemblage Plants Invertebrates Birds

21 0.9±O.7 0.4±0.120 3.7±1.6 1.3±0.21 3.2±1.6 1.0±O.217 4.3±O.8 2.3±O.4 1.4±O.518 4.1±2.2 2.4±O.916 7.6±2.3 1.8±O.65 12.2±2.3 4.2±O.2 6.7±2.14 14.6±6.7 2.7±0.58 11.8±2.0 5.1±0.6 10.7±3.02 20.3±3.1 5.6±0.5 5.6±0.73 25.3±4.8 21.6±3.2 22.06 32.1±1.9 20.6±1.3 15.8±1.27 30.4±3.1 14.5±2.5 23.8±16.69 28.6±2.8 22.6±2.9 12.2±5.112 38.1±4.3 10.7±5.113 50.3±1.8 34.1±1.4 43.1±4.514 58.4±2.2 68.0±12.611 62.5±4.3 67.0 108.0±24.110 88.7±3.9 91.9±5.7 84.9±9.8

Biddy and Coyrecup (SPS067 and SPM004, 33species).

Assemblage 14. A group of 37 species typical ofnaturally hypersaline wetlands (mean 59±2 g L-l)in the northern wheatbelt, although morewidely distributed (Figure 6). The assemblagecomprised almost exclusively plants (36 species)which, together with 1 waterbird species,represented 6% of all species occurrences. Theassemblage was best represented at WeelhambyLake and Mongers samphire pan (SPS169 andSPS167, 25 species).

Assemblage 16. A group of 51 species occurringpredominantly around fresh water althoughmean salinity was 4.3±1.1 g L-I because ofisolated plant occurrences around hypersalinewetlands. The assemblage was more commonin high rainfall areas and the south coast (Figure6). It consisted of 30 invertebrate and 21 plantspecies, representing 1.7% of all speciesoccurrences and was best represented atYoulabup Swamp (SPS124, 26 species).

Assemblage 17. A group of 89 species associatedwith fresh to subsaline water (mean 3.O±OA g L,l)predominantly on the western margin of thewheatbelt (Figure 6). The assemblage contained54 invertebrate, 1 waterbird and 34 plantspecies, representing 6.4% of all speciesoccurrences. It was best represented at NalyerinLake (SPS031, 49 species).

Assemblage 18. A group of 84 species almostcompletely restricted to sedge swamps, withgreatest occurrence in swamps of the southwestern part of the study area (Figure 6). Whilethe majority of occurrences were in fresh water,


mean salinity of occurrence was 3.3±1.3 g L,I asa result of a few plant records aroundhypersaline wetlands. The assemblagecontained 35 invertebrate and 49 plant species,representing 1.8% of all species occurrences. Itwas best represented at Poorginup Swamp(SPS103, 39 species).

Assemblage 20. A group of 74 species patchilydistributed though the wheatbelt (Figure 6) andoccurring predominantly in fresh water (mean2.2±O.6 g L,I). The assemblage contained 41invertebrate, 1 waterbird and 32 plant species,representing 3.8% of all species occurrences. Itwas best represented at Punjerwerry Claypanand Weelawadji Lake (SPS197 and 177, 34species).

Assemblage 21. A group of 53 species typical ofpools on granite outcrops (mean O.6±O.3 g L,l,Figure 6). The assemblage contained 27invertebrate and 26 plant species, representing1.3% of all species occurrences. It was bestrepresented at Wannara Rock (SPS168, 21species),

Several trends were noticeable amongassemblages. Firstly, waterbirds rarely formed asignificant component of the species richness of anassemblage: even in the most bird-dominatedassemblage 6, waterbirds comprised only 29% oftaxa. Plants comprised most of the biota of naturallysaline assemblages. The ratio of plants:plants+invertebrates averaged >0.9 in theassemblages most typical of naturally hypersalinesites (10, 11 and 14) compared with <0.5 in otherassemblages

Secondly, many assemblages occurred at allwetland types, albeit often at very low frequency(Table 2). This was largely attributable to manyplant species that grew on dunes or rises around awetland not closely reflecting conditions in thewaterbody itself and being widely distributedacross wetland types. For most assemblages, theaverage wetland salinity associated with plantrecords was higher than for invertebrate records.The discrepancy increased as the assemblagebecame more halophilic, except for assemblagesoccurring predominantly at hypersaline wetlands(Table 3).

Thirdly, few assemblages were tightly definedspatially and the occurrence of assemblages wasoften patchy. For example, wetlands containingsignificant numbers of assemblage 4 species werelocated near Geraldton, Perth and Esperance withonly scattered, low-frequency occurrence atwetlands between (Figure 4). The somewhatstochastic occurrence of assemblages was reflectedin the relatively low number of assemblage speciesat individual sites. Even the site best representingan assemblage contained, on average, only 45% ofassemblage species (range 20-71%).


Table 4 EquatIOns predicting the logarithm of species richness at each site for each species assemblage and thedegree of nestedness of assemblages. The adjusted R2 IS a measure of how much variation in assemblagerichness is explained by each equation; rand '/0 fill are measures of noise and the frequency of occurrence ofspecies In the assemblage, and P is the probabilitv that an assemblage is not nested (Atmar and Patterson1995). See Table 1 for abbreviations and footnote for categories of salinity' and pH2.

Assemblage Equation for Log,S Adj R2 T % fill P

1 2.3644 - 0.0457 Pcv + 0.0405 HCOJ

- 0.1034 sa, -0.0374 Sal' 62 16 11 le'S2 -124464 0.0176 Sal + 0.0024 Alk + 0.0285 Pcv + 0.0995 Long 33 8 4 4e,223 -11.7119 + 1.3273 flow + 0.3931 latitude - 0.0068 elevation - 53 11.3 5 2e'

0.1572 Ca:S04meq4 -4.8503 + 1.3979 Flow + 0.0050 Pann + 0.0184 Pcv + 0.0495 Mg 47 11 6 3e J

5 7.3355 - 0.0532 Sal - 0.0362 Long 70 33 23 8e"56 2.6369 - 0.0040 Sal- 0.7251 pHI - 0.1397 pH2 - 108641 Mn 33 38 31 2e-"8 -.0358 + 1.0679 Sail + 1.5586 Sal2 + 0.6770 Sal3 + 00367 Mg 55 22 9 3e-2

"

00405 CC9 -9.6860 + 0.3569 Lat - 0.0062 Alt - 1.9000 Sail 0.0681 Sal2 + 59 22 12 Se_I;'

0.4372 Sal310 3.0831 - 0.0180Pcv -3.2756 Sail - 2.8694 Sal2 - 06227 Sal3 57 10 6 3e_3S

11 -5.3949 + 0.6050 Tann -0.0235 Pann - 10.6212 SaIl 2.8034 Sal2 - 75 16 8 2e-40.6572 Sal3 + 0.3606 pH

12 3.8148 + 0.0108 Alt - 0.0658 Mg - 0.0030 Sal - 0.0026 Pann 51 12 5 2e-413 3.6323 0.0020 Pann - 1.2524 Sail - 0.8087 Sal2 + 0.2018 Sal3 - 53 26 18 2e_97

0.7926 pHI - 0.2944 pH214 - 24369 + 0.0093 Evap - 24704 Sail - 1.4907 Sal2 00581 Mg 65 23 19 2e_4916 -98974 + 04045 Lat 0.0383 Sal 0.3998 pH + 01901 K 43 13 6 5e-~

17 10.2881 + 02703 Lat + 0.0038 Pann + 0.0268 Pcv 00709 Sal 76 9 8 2e-7'

18 -14.5642 + 3.9927 logPann - 0.6591 Tann + 1.3916 pH2 - 0.0005 TN 84 13 8 2e"020 3.7039 + 0.6485 Rock 0.0930 Sal 0.0044 Alk - 0.0254 Cl 633 12 8 4e_36

21 -2.7073 + 2.9708 Rock + 0.1348 S04 - 0.8150 CC 86 15 11 6e-9

------, Sail < 150 mg L', Sal2 < 8000 mg L', Sal3 < 50 000 mg L', Sal4 ;:: 50 000 mg L'2 pHI < 6.0, pH2 ;:: 6.0

drop SPS025, SPS177 R2 = 70%

Modelling assemblage occurrenceSpecies richness values of all modelled

assemblages were unimodally distributed acrossordination scores, suggesting the assemblages werenested (see also P-values in Table 4). Richnessvalues approximated Poisson distributions for allassemblages other than 5 and 13. Equationsdescribing the distribution of species richness wereproduced for 18 assemblages (no satisfactory modelcould be constructed for assemblage 7 and noattempt was made to model the artefactualassemblages 15, 19 and 22). Adjusted R'-values ofthe equations varied from 33-86% (Table 4).

The low amount of variation (33%) explained byequations for assemblages 2 and 6 appeared toreflect biological reality rather than violation ofmodel assumptions. Both assemblages containedfreshwater species that were salt-tolerant and foundthroughout the wheatbelt (Figure 4) and beyond, sothat the WBS did not circumscribe their ecologicallimits efficiently. A quarter of the species belongingto assemblage 6 were mobile waterbirds, half theinvertebrates were higher taxa (usually families orphyla) and many of the plants were weeds. Morethan half the invertebrates in assemblage 2 had poordispersal ability and their occurrence may bedetermined largely by historical events (see Maly et

aI., 1997 for a discussion of copepod distributions)that are stochastic with regard to the environmentalparameters measured in the WBS.

The variables used in models of species richnessdo not necessarily reflect the main environmentaldrivers of species distributions or assemblagerichness, even when R'-values are high. However,salinity was an explanatory variable for 14assemblages (Table 4) and for other assemblagesthere were significant correlations between salinityand one of the variables used in the model. Thisstrongly suggests salinity is a major driver ofcommunity patterns. Other variables that appearedcommonly in models were annual rainfall, pH,latitude, alkalinity and magnesium ionconcentration, while the presence of a graniteoutcrop was a strong predictor of the occurrence ofassemblage 21 and flowing water was a predictor ofassemblages 3 and 4.

Types of wetlandSixteen types of wetland were recognized in the

surveyed area, based on their plant and animalassemblages. The major split in the classificationrelated to the separation of primarily and secondarilysaline sites from fresh or subsaline ones (Figure 7,AppendiX 4). The major wetland types were:


Type Ill. Biodiverse subsaline wetlands. Fresh tosubsaline water (mean 5.8±1.1 g L 1

). Wetlandsin this group had little in common with eachother. Lake Logue (SPM002) on the coastsouth of Geraldton was in the final stages ofdrying and had become saline as it evapoconcentrated, the secondarily saline LakeDulbinning (SPS007) in the central wheatbeltfilled a month or so prior to sampling andwas fresher than usual, Capamouro Swamp(SPS163) in the northern wheatbelt usuallycontains hypersaline water (if water present)but was subsaline when sampled because of alarge flood event, and Peenebup Creek(SPSl15) is a subsaline river flowing towardsthe south coast. Mean richness was 100±3,with Boyacup Bridge swamp (SPS111, 113species) the richest site. Most of theassemblages found at wetlands ofintermediate salinities occurred at lowfrequency in type III wetlands; invertebratesformed a greater percentage of the biota (52%of species) than plants (39%).

Type IV. Semi-permanent subsaline wetlands. Freshto subsaline water (mean 4.3±0.6 g Ll). Mostlyon south coast, usually with extensive fringingtrees. Mean richness was 84±5, which wasalmost 20% lower than type Ill. Theassemblages occurring in the two wetland typeswere similar, with type IV wetlandsdifferentiated largely by a greater proportion ofspecies in assemblages 3 and 9. Invertebratesdominated the biota (57% of species) but morewaterbird species were present (13%) than atany other wetland type except X. The richestsite was Meeking Lake (SPS038, 102 species).

Type V. Disturbed subsaline wetlands. Fresh tosubsaline water (mean 3.6±0.7 g L·l, excludingSPS199 where sampling included areas muchfresher than the recorded salinity). Occurredthroughout the wheatbelt, other than the southcoast. Mean richness was 67±6, which was morethan 30% lower than type III and the number ofassemblages making use of the wetlands wasfewer than for III and IV (Table 2). Invertebrateswere very much the dominant element of thebiota (71% of species) with the lower speciesrichness in type V wetlands largely theconsequence of a depauperate flora. The richestsite was Qualeup Lake (SPS032, 126 species),which had almost twice as many species as anyother wetland in the group.

Type VI. Westward flowing rivers. Fresh tosubsaline water (mean 3.8±1.0 g L·l). Meanrichness was 70±4, with Skelton Gully (SPS193,97 species) being the most species-rich site. Halfthe records of assemblage 4 occurred in thiswetland type and several other assemblages ofintermediate salinities also occurred at

II

I

I

I

IType XV, saline playas (22)

Type XVI, species-poor acidichypersaline playas (8)

Type X, secondarily salinewetlands (25)

Type XI, saline sou1h coastrivers (4)

Type XII, coastal saline lakes(11)Type XIII, alkaline hypersalinelakes (12)

Type XIV, species-rich acidichypersaline playas (9)

Type VII, granite rock andvernal pools (13)

Type VIII, daypans andshallow yate swamps(5)

Type IX, sedge swamps (12)

Type 11, disturbed northernswamps (18)

Type Ill, biodiverse subsalinewetlands (10)

Type IV, semi-permanentsubsaline wetlands (12)

Type V, disturbed subsalinewetlands (12)

Type VI, westward flowingrivers (11)

Figure 7 UPGMA classification of wetlands into typesbased on their biota. Number of wetlands ineach type shown.

Type I, undisturbed freshwaterswamps (9)

Type I. Minimally disturbed freshwater swamps.Fresh water (mean 0.5±0.1 g Lt) with eitheremergent trees, shrubs or grasses on the lakebed. Supported a high number of ubiquitousspecies. Mean richness was 95±8, with ArroSwamp (SPS183, 151 species) and WeelawadjiLake (SPS177, 135 species) having the mostspecies-rich communities. Assemblage 1occurred predominantly in this wetland type(Table 2) and invertebrates (61% of species)dominated the biota.

Type II. Disturbed northern swamps. Fresh tosubsaline water (mean 4.0±0.9 g L'l). Allmember wetlands were in the northernwheatbelt and were disturbed to some extent;most contained emergent vegetation. Meanrichness was 76±3, with Tardun CBC swamp(SPS187, 103 species) being the richest. Type IIwetlands were significant habitat (>20% ofrecords, adjusted for group size) for species inassemblage 12 and invertebrates (55% ofspecies) dominated the biota.


moderate frequency; invertebrates dominatedthe biota (70(>;) of species),

Type VII. eremite rock pools and vernal pools,With their very small, mostly hard rock,catchments these sites were the least saline ofany wetland type (mean OA±0,2 g LI), Therewere two obvious sub-groups: pools on graniteoutcrops and three shallow, clay-based vernalpools plus a shallow creek high in thelandscape, Mean richness was 82±5, withWanarra Rock (SPS176, 117 species) being themost species-rich site, Type VII wetlands werethe dominant habitat for assemblage 21, withassemblages 20 and 16 also occurring insignificant proportions, Slightly moreinvertebrate species (55'1'0) were present thanplants (45°1<»),

Type VIII. Claypans and shallow yate swamps, Twogroups of wetlands occurred within this type:turbid claypans and clear-water shallow yateswamps, Both were fresh (mean 05±0,2 g LI).Mean richness was 62±12, with PunjerwerryClaypan (SPS197, 100 species) being the richestsite. Assemblages 20, 1 and 16 occurred insignificant proportions; the biota wasdominated by invertebrates (70'1<) of species),

Type IX. Sedge swamps. Fresh water (1 ,1±OA g L,I),Mean richness was 98±5, with Nalyerin Lake(SPS031, 130 species) being the richest site,These wetlands provided the main habitat forassemblage 18, and significant habitat forassemblages 17, 16 and 4, The biota wasdominated by invertebrates (63% of species),

Type X. Secondarily saline wetlands. Almost allthese wetlands contained fresh or subsalinewater prior to land clearing but are now moresaline (Bennetts Lake spson may be a naturallysaline exception), Mean salinity was 38.7±65 gLI and mean richness 54±3, with Stennetts Lake(SPS075, 71 species) the richest site. Thewetlands provided significant habitat forassemblage 13 and most assemblages typical ofsaline water were present at low frequencies.,Half the species were plants, with waterbirdscomprised 13(;<) of the community to make typeX (together with IV) the wetland typesupporting most waterbird species.

Type XL Saline south coast rivers. Three south coastrivers and an adjacent swamp that appeared tocluster with rivers on the basis of plantcomposition, Saline water (16.0±6.5 g LI) andmean richness 68±6, with Oldfield River(SPSI23, 68 species) the richest of the river sites.The sites appeared to be important habitat forspecies of assemblages 3 and 9, although smallgroup size means this conclusion should betreated cautiously, Invertebrates comprised thelargest proportion of the biota (58% of species).

Type XII. Coastal saline lakes. Saline water

353

(47.9±165 g Ll). l'he two subsaline wetlands inthis group (SPS121, SPS122) were saline butfreshened after rain shortly before sampling(Pinder et ai, 2004a). The wetlands were locatedon the coastal plain of the south and west coastsand had mean richness 56±4, with Mullet Lake(SPSI41, 82 species) the richest site,Assemblages 3 and 9 occurred in type XIIwetIands in significant proportions; just overhalf the biota (53%) comprised plant species.

Type XIII. Alkaline hypersaline playas. Hypersaline(128±21 g LI), alkaline (8.0±0.1) water. Primarilyhypersaline playa lakes, mean richness 50±6,with Anderson Lake (SPS106, 94 species) beingthe richest site. Assemblage 10 occurred insignificant proportions and other assemblagesassociated with high salinities were present inlow frequencies. Plants dominated the biota(76% of species).

Type XIV. Species-rich acidic hypersaline playas.Hypersaline (146±35 g L'!), acidic water (4.6±0.6,excluding the alkaline Kondinin samphiremarsh SPS017 which had a salinity of 10 g LIand pH of 9.2). Primarily hypersaline playalakes, mean richness 49±6, with Crook's wetland(SPS076, 58 species) being the richest acidic site.Assemblage 10 occurred in significantproportions and other assemblages associatedwith hypersaline conditions were present in lowfrequencies. Plants dominated the biota (80% ofspecies).

Type XV. Saline playas. Saline water (51±8 g L I).Primarily saline and alkaline, mean richness71±3, with Mongers samphire pan (SPS167, 114species) being the richest site. Assemblages 11and 14 occurred in significant proportions andmost other assemblages typical of salineconditions were present. Plants dominated thebiota (66% of species).

Type XVI. Species-poor acidic hypersaline playas.Hypersaline (116±26 g LI), acidic water(4.7±0.71, including two sites with pH -8).Disturbed primarily hypersaline playa lakes,usually with increased inundation as a result ofsecondary salinisation, these wet lands weredepauperate with mean richness 21±4. Wet landswith most species were Masters saline lakeSPS097 and Kondinin salt marsh lake (43 and 34species, respectively) but neither was typical ofthe type. Only assemblage 10 had more than 5(/0of its occurrences in this wetland type. Plantsdominated the biota (67% of species).

The different wetland groups were comparativelywell separated in ordination space (Figure 8),providing evidence that the wetland typesidentified in Figure 7 represented distinctcommunities. Sedge swamps and granite rock pools(types IX and VII) were the wetland types in freshwater with most clearly distinguished communities.


20

, 5

10

o Fresh• Saline

10

05

00

.()5

-1.0

•o

20 '5 10 05 00 .() 5 ·'0 ., 5 .20 .25

., 0

00

'()5

05

(b)

(a)

The ordination of freshwater wetlands supported adivision within type VII between the vernal pools(Goonaping SPS023, Youlabup SPS124 andNgopitchup SPS102) and granite rock pools,although we did not recognize this in the wetlandclassification. Claypans and shallow yate swamps(type VIII) formed a heterogeneous group withPunjerwerry Claypan (SPS197) and Frank HahnNational Park claypan (SPS087) having differentcommunities from each other and other type VIIIwetlands. Among the saline wetland types, therewas considerable heterogeneity within species-pooracidic hypersaline playas (type XVI), which resultedin some overlap with other hypersaline wetlands(types XIII-XV) (Figure 8). Species-rich acidichypersaline playas and saline playas (types XIV andXV) also showed some overlap with each other.Jurien coastal lake (SPS178) and Hutt Lagoon(SPS189), the only northern wetlands among typeXII (predominantly southern coastal saline lakes),were outliers but grouped together.

os

Figure 8

• X• XIo XIIo XIIIt:. XIV<>xvAXVl

<>1t:. 11o IIIo IV

• VA VI

• VII• VIIIo IX

10

os

10

00

00

-os

-1.0

•

o 0o

00 ·10

·2.0

os

-1.S

1.0

Ordination of wetland groups in thewheatbelt as defined by UPGMA clustering.(a) all groups, (b) freshwater groups I-IX, (c)saline groups X-XVI. Stress = 0.19.

·10

1 S

-os

00

-os

O.S

10

1S

00

O.S

10

(c)

Singleton speciesMore than one-third of species were recorded at

only one wetland and were excluded frommultivariate analyses for mathematical reasons(Belbin, 1993). Plants comprised 55% of allsingletons, invertebrates 43% and waterbirds 2%.Within the three elements the percentage singletonswas similar for plants and invertebrates (36% and33%, Figure 9a,b) but lower for waterbirds (19%).Individual wetlands contributed disproportionatelyto the list of singletons, with 6% of sites accountingfor one-quarter of all singletons. Most of these siteswere either on the wetter, western side of thewheatbelt or in the north (Figure 10) and it is likelythat many of the singletons in these wetlands werespecies with core ranges outside the wheatbelt,either in the south-west forest or the northern aridzone.

The number of singleton species was significantlycorrelated with overall species richness at each site(r = 0.43) and negatively correlated with salinity (r =

-0.23) but the relationships had little explanatorypower. Most wetland types had few singletons,with sedge swamps (type IX) the most obviousexception (Figure 9c). Individual sites with largenumbers of singletons occurred in type I (ArroSwamp SPS183, 23 species), type VI (MurchisonRiver SPS198, 13 species), type VII (GoonapingSwamp SPS023, 24 species) and type IX (PoorginupSwamp SPS103, 22 species). Hypersaline wetlandssupported relatively few singletons (Figure 9c),with the greatest number occurring in Kondinin saltmarsh lake (SPS016, 7 species). However, theproportion of the flora at individual hypersalinewetlands that were singletons was much higherthan for the invertebrate fauna (Figure 9d).Beaumont Nature Reserve (SPS130, 4 species) stood


out for its invertebrate singletons. Secondarilysaline wetlands (type X) and disturbed subsalinewetlands (type V) supported few singleton species.

DISCUSSION

Waterbird distributions in the wheatbelt andsouth coast, and the factors controlling them, werewell understood prior to the WBS as a result ofsurveys in the early 1980s (Jaensch et aI., 1988; Halseet aI., 1993b) but the WBS considerably improvedknowledge of invertebrates and vascular plants. Forexample, it was estimated by Anonymous (1996a)that there were only 200 aquatic invertebratespecies in the wheatbelt compared with nearly 1000now known from the WBS and related monitoringprograms (Cale et ai, 2004; Pinder et ai, 2004).While the vascular flora was better known thanaquatic invertebrates, the WBS represented the firstsystematic floristic survey of wheatbelt wetlands. In

particular, knowledge of the naturally salinesystems within inland palaeo-channels has beengreatly improved. Eight new species werediscovered, as well as numerous populations ofspecies previously found in few localities andconsidered restricted (Lyons et aI., 2004). One of themajor outcomes of the WBS has been showing thatwetland communities of the wheatbelt and southcoast are substantially richer and more complexthan previously recognized. In addition, thesurveyed area has a moderately rich diatom fauna(Blinn et aI., 2004). Protection of these biologicalcommunities should be a conservation priority.

In terms of localised biodiversity patterns, nearlyall wetlands in the Wheatbelt are significant.Average instantaneous numbers of plant,invertebrate and waterbird species associated withwetlands in types I (freshwater), III (biodiversesubsaline) and IX (sedge) were 2:: 100. Only typeXVI wetlands (species-poor acidic hypersaline)

3>

Wetland type

>>....................

(d) 14

:: 12';;.. 10~ 8<:><: 6"6ll,~ 4'0 2"I- O-+--.......,............~.......,...-L....L.,,...........,.-...,..

Figure 9 Frequency of occurrence of singletons in wetlands. (a) invertebrates, (b) plants, (c) percentage of all plant orinvertebrate singletons in each wetland type, (d) average percentage singletons in the plant or invertebratespecies list at wetlands of each type. Invertebrates, solid bars; plants open bars.


Biodiversity patterns and surrogate taxaMany approaches to reservation, based on

mapping plant associations or environmentalsurrogates of biodiversity, implicitly assume thatthe different taxonomic elements of biologicalcommunities respond in the same way to the sameenvironmental parameters and, thus, are similarlydistributed (see Margules and Pressey, 2000).Recent work, including the WBS, has shown this isnot true at the scale of a wetland or terrestrial site(McKenzie et al., 2000; Davis et al., 2001; Fleishmanet al., 2002). In part, this is the result of variousorganisms being distributed at different scales.Most manageable conservation units (such as awetland and associated vegetation) are composedof mosaics of microhabitats, with the distribution ofthe biota being controlled within thesemicrohabitats. For example, the relationshipbetween wetland salinity and invertebrates isdifferent from that between wetland salinity andplants, even for members of the same assemblage(Table 3), because most plants grow on the bankrather than in the water column. Wetland salinityvalues are often poorly correlated with soil salinityon the bank and, in fact, there is considerable

were depauperate (average of 23 species). However,some wetland types (e.g. X secondarily saline) weredominated by widespread species and theirirreplaceability at a regional scale is low (seePressey et al., 1994). Protection of individualwetlands containing widespread species is usuallynot critical to maintaining the regional species poolbut may be very important for maintenance of theoverall populations of some waterbird species(Halse et al., 1995 and references therein).

The most commonly occurring assemblages were5 (widespread, fresh to subsaline), 8 (fresh toweakly saline, mostly south coast) and 13(secondarily saline), which accounted for about halfthe species records (Table 2). The abundance of twosalt-tolerant freshwater assemblages reflects thegreater richness of the freshwater biota, comparedwith that of highly saline systems, and theimportance of salt tolerance if species are to occurcommonly. It should be emphasized, however, thatthe relative frequency of different assemblages, asperceived by survey analysis, was partly a reflectionof sampling design. An attempt was made tosample the full range of natural wetland types inthe surveyed area (see Lyons et al., 2004; Pinder etal., 2004) but not necessarily in proportion to theirabundance. For example, assemblage 21 (granitespecialists) comprised 1% of species records. Theassemblage was more-or-less restricted to graniteoutcrop pools, which represented 5% of thewetlands sampled. Its frequency of occurrencewould have increased if more granite pools hadbeen sampled.

".'

••

•··

•

....

, .

'. KargoOfl!€

1nvertebratesOf !;mqlf>IOlY& at each

PlantsNumber of slr'lg'etons at each site

.>~O

.>5

•

:

,

.•...•

.,. • j'cO~--i).

\"~~ ~1"'-'-"~--'~-,"'~~

Albar:y

.,

\Perth~' •

1 •..~

\

f~,--I

\~~

1 •. • ~

'01 '.\. .~ ..\

\

\ ,

\r\.

(""",Idlon .\\

':- .\~ .

\.Geraldtor. '\,

Figure 10 Distribution of singleton species at wetlandssurveyed in the wheatbelt.


variation in soil salinity on the b,mk allording tosmall-slale topographiGll variation (Lvons t't ai,2(04). The ability of waterbirds to move betweenwetlands means their distribution may not be fullydetermined by the charalteristils of the wetland atwhich they were recorded. Thev mav feed at aproductive saline microhabitat, flv to a freshwaterseep in another wetland to drink and roost insuitable microhabitat of a third wetland (Norman,1983; see also Roshier et al., 2(01).

The varied responses of different taxa to theirenvironment means that it is rare for one group oforganisms to be an effilient surrogate for theOlcurrence of others at a site and, ideallv, reserveselection should be based on survevs of all the biota(see Gaston, 2(00). This was prevented bv logisticalconstraints in the WBS, as is always likely to Ollurin broad-scale surveys. Nevertheless, the necessityof procedures for reserve (or in this case BRC)selection being explicitly based on a range of taxawas recognized and we surveyed aquaticinvertebrates, waterbirds and plants. Theseelements operate at different spatial sca havedifferent life histories and represent both thewaterbody and its surroundings. In terms ofconservation and public interest, the riparian zone(or immediate surroundings of a vvetland) is asimportant a habitat as the waterbodv itself. Theprotection of terrestrial fauna species, includingfrogs, using the riparian zone was addressed in theWBS by McKenzie et al. (2004).

Reliability of survey patternsOne of the most significant issues with respect to

the WBS is how well single surveys characterisedwetlands and, therefore, whether the wetland typesand species assemblages identified in this paperreally reflect patterns of the wheatbelt and southcoast. Reliability issues fall into two classes: (1) howwell sampling reflected instantaneous communitypatterns, and (2) whether temporal variation inclimate is likely to have affected patterns obtained.We use the term 'instantaneous' principally to referto community composition during the season ofsampling, although our testing of the invertebrateprotocol also examined how well communitycomposition at the actual time of sampling wasdocumented.

Halse et al (2002) examined the adequaly ofinvertebrate WBS sampling and analyticaltechniques to document patterns of instantaneousinvertebrate communities. Thev were able todiscriminate between communities of five basinwetlands along a salinity gradient. Given the staticnature of vaslular plants and the considerablesurvey effort used, it is likely that instantaneouscommunity patterns of plants were also sufficientlywell documented to characterise a wetland (Lyonset ai, 2(04), although some species of submerged

357

plants found in fresh water at the end of summer,some late-flowering herbs around claypans andRlIppia spp. (identification difficulties) were underrepresented in the dataset. We undertook furtheranalysis of some of Cale et al.'s (2004) waterbirdand wetland data to show that, although waterbirdsare highly mobile, instantaneous composition ofwaterbird communities can be characterised wellenough by single surveys during spring, such asused in the WBS, to distinguish broad categories ofwetland (Figure 11; see also Halse et aI., 2000a).

Large temporal fluctuations in wetland depth andsalinity (both increases and decreases) sometimescause substantial shifts in wetland communities. Forexample, the waterbird community at Lake Logue(SPM002) showed substantial differences betweenthe dry year of 1997 when the lake was 0.36 m deepin spring and the wet years of 1999 and 2000 whenit was ca 3 m (Figure 11, Table 6). On the otherhand, Halse et al (2000b) found the relationshipbetween sites based on invertebrate samplesremained constant across annual rainfall events inthe southern Carnarvon Basin, and invertebrate andwaterbird data collected in south-west WesternAustralia according to WBS protocols have shown aconsistent relationship between wetlands acrossyears when wet land conditions have been similar(Cale et aI., 2004; Figure 11). Similar results arelikely for plants, with extreme flooding and verylow water levels affecting communitiessubstantially, but the communities remaining stableat intermediate water depths, except for minorrainfall-induced variation in the speciescomposition of annual plants. Even annuals haveadaptations to recruit at varying elevationsdepending on water depth in a given season, sothat overall plant composition is relativelyindependent of the degree of flooding. The WBSplant sampling was designed to capture all zones ofrecruitment and minimise the effect of year onsurvey results (Lyons et al., 2(04).

Given that the northern wetlands were floodedbeyond their usual boundaries and were less salinethan usual when sampled in 1999, while wetlandssurveyed in 1997 and 1998 in central and southernareas were experiencing dry, saline conditions (seeFigure 2), it is possible that some long-termrelationships between wetlands of the wheatbeltand south coast were obscured by rainfall patternsduring the WBS by the kind of responses observedat Lake Logue (Figure 11). Extensive rainfallthroughout the northern parts of Western Australiain 1999, at the same time as the northern wheatbeltwas flooded, also allowed some tropicalinvertebrates to extend their ranges southwards tothe wheatbelt (see Pinder et ai, 2004a). Somewaterbird species (e.g. Hima/ltoplls hima/ltopus)made the reciprocal movements and more-or-lessdisappeared from the wheatbelt (S.A, Halse and

358 S. A. Halse, M. N. Lyons, A. M. Finder, R. J. Shiel

0.5 * *<>*

0OA + • •• + Bryde wet0.0 ... <> Bryde++ <> Bryde dry• Logue wet

-0.5 A.0 Logue dry+ Towerinning

OAA CoyrecupA Coyrecup dry

-1.0 * Wheatfield0 Noobijup• Bennett's wet0 Bennett's dry

Coomelberrup dry

Figure 11 Ordination of waterbird communities at a series of wetlands surveyed in spring of different years (see Caleet al. 2004 for survey methods and Table 6 for information on wetland depths and salinities). Stress = 0.18.

G.B. Pearson unpublished data). Nevertheless,conditions at most northern wetlands in 1999 wereprobably not extreme enough to have had majorimpact on community characterisation and wesuggest most of the WBS patterns are reliable.

Potential Recovery WetlandsOne of the major objectives of the WBS was to

identify wetlands that might form parts of BRCs.The approach we adopted to identify PRWs issimilar to that used by McKenzie et al. (1989, 2000)to identify conservation reserve requirements ofterrestrial systems in the Nullarbor and southernCarnarvon Basin regions of Western Australia. Theymodelled the occurrence of each assemblage in theregion of interest to provide the basis for selectingconservation areas in parts of the region that hadnot been surveyed. However, wetland assemblagesare often poorly predicted by models (see Table 4)and only amenable to mapping as species richnessisoclines when the assemblage is almost ubiquitous(see Figures LH'». McKenzie et al. (2000) attemptedto overcome these difficulties by basingconservation recommendations in the southernCarnarvon Basin on a wetland classification ratherthan on assemblages.

-------

In the WBS, species assemblages were adopted asthe units on which conservation measures would bebased but a combination of modelling, fidelityanalysis of species assemblages to wetland types,proportion of assemblages at species-rich sites andassemblage mapping have been used to form thebasis for recommendations about conservation ofassemblages and the wheatbelt biota. Summaries foreach assemblage in terms of types of wetlands andareas where the assemblage is best represented,need for conservation, and the proportion of thewhole assemblage likely to occur at anyone site,are presented in Table 5. Where there is a need forconservation, we have suggested 1-3 surveyedwetlands or complexes that might be protected toconserve the assemblage. These PRWs should beregarded as indicative of the kinds of wetlands inneed of protection; in many cases further surveymay reveal wetlands supporting a higherproportion of the relevant assemblage. The notionalpercentage of each assemblage that wouldconserved by protecting the 19 PRWs, in additionto existing BRCs, in an otherwise salinisedlandscape varied between 42 and 100% (Table 5).This represents 72% of the waterbirds recordedduring the WBS, 59% of the aquatic invertebrates


Table 5 Suggested focus for conservation of the various wheatbelt and south coast species assemblages. Preferredwetland types, part of the surveyed area where assemblage most common, sensitivity to salinisation, potentialrecovery wet lands (PRWs) and the maximum proportion of the assemblage present at individual wetlandsare listed. The proportion of each assemblage that would be conserved (PC) by protecting the PWRs andexisting Natural Diversity Recoverv Catchments from salinisation is also shown.

Assemblage Comments PC

Type I and VIII wetlands, mostly ll1. south-eastern wheatbelt and south coast with restricted 0.67occurrence, sensitive to salinisation, extends beyond wheatbelt, PRW Lake Cronin,proportion of assemblage at individual sites >0.5

2 Types I, II, III, IV, V, VI wetlands and rivers, Widespread occurrence, some tolerance of 0.42salinisation, no special conservation requirement, proportion of assemblage at individualsites <0.3

3 Types IV, XI and XII rivers and wet lands, mostly on south coast with restricted occurrence, 0.44some tolerance of salinlsation, PRW Lake Warden BRC, Oldfield River, proportion ofassemblage at individual sites <0.3

4 Type VI rivers, mostly in wandoo woodland east of Perth, sensitive to salinisation, PRW 0.47Jimperding Brook, proportion of assemblage at individual sites <0.3

5 All fresh or sub-saline wetland and river types, Widespread occurrence, sensitive to 0.97salinisation, no special habitat requirement, proportion of assemblage at individual sites >0.5

6 All wetland types except some very fresh and some hypersaline ones, widespread occurrence, 1.0tolerant of salinisation, no special habitat requirement, proportion of assemblage atindividual sites >0.5

7 A plant-dominated assemblage, most of which arc weeds, no special habitat requirement 0.87

8 Most fresh or sub-saline wetland and rIver types, mostly on south coast and higher rainfall 0.81parts of wheatbelt, sensitive to salinisation, no special habitat requirement, proportion ofassemblage at individual sites <0.3

9 Types IV, XI and XII rivers and wetlands, mostly on south coast, moderately restricted 0.97occurrence, some tolerance of salinisation, PRW Lake Warden BRC, Oldfield River, proportionof assemblage at individual sites 0.3-D.5

10 Types XIII and XIV wetlands, mostly in eastern wheatbelt, some tolerance of salinisation, PRW 0.64Isthmus Lake, Lake Altham, Kondinin Samphire Marsh, proportion of assemblage atindividual sites <0.3

11 Type XV wetlands, in north-eastern wheatbelt, sensitive to salinisation, PRW Lake Moore, 0.80Weelhamby Lake, proportion of assemblage at individual sites 0.3-D.5

12 Stronger affinity to northern and eastern wheatbelt than to wetland type, probably sensitive 0.48to salinisation, no special conservation requirement other than wetland in north-easternwheatbelt (PRW Wannara Claypan), proportion of assemblage at individual sites <0.3

13 Type X and other saline/hypersaline wet lands, Widespread occurrence, tolerant of salinisation, 0.91no special habitat requirement, proportion of assemblage at individual sites 0.3-D.5

14 Type XV wetlands, mostly in north-eastern wheatbelt, sensitive to salinisation, PRW 0.97Buntine-Marchagee BRC, Weelhamby Lake, proportion of assemblage at individual sites >0.5

16 Types VII and IX wetlands, mostly south-east of wheatbelt, sensitive to salinisation, PRW 0.73Youlabup Swamp, Ngopitchup Swamp), proportion of assemblage at individual sites <0.3

17 Type IX wetlands, western edge of wheatbelt, sensitive to salinisation, PRW Nalyerin Lake, 0.93Lake Muir BRC, proportion of assemblage at individual sites <0.3

18 Type IX wetlands, south-western edge of wheatbelt, sensitive to salinisation, PRW Lake 0.95Muir BRC, Nalyerin Lake, proportion of assemblage at individual sites 0.3-D.5

20 Types VII and VIII wetlands, patchy occurrence throughout wheatbelt, sensitive to salinisation, 0.82PRW Punjerwerry Claypan, proportion of assemblage at individual sites 0.3-D.5

21 Type VII wetlands, granite rock outcrops in eastern wheatbelt, protected from salinisation, PRW 0.66Wannara Rock, Dunn Rock, proportion of assemblage at individual sites 0.3-D.5

and 55% of the plants based on single surveys.When additional data from some PRWs and BRCSthat had been sampled on multiple occasions wereincluded in the analysis, 93% of waterbirds weresometimes found in the protected wetlands (seeCall.' et aI., 2004).

The above two estimates of waterbird protection

afforded by PRWs demonstrate that levels ofprotection estimated in Table 5 need qualification.Firstly, on the positive side, some assemblages areunlikely to be threatened by salinisation. This ismost obviously true for assemblages 6 and 21. Theformer consists of Widespread salt-tolerant speciesthat occur everywhere, including salinised

360

Table 6 October depth Zind sallnltv (electricalconductivitv) 111 different vears at eightwet lands in the Wheatbelt where waterbirdcommunities were monitored (Call' ct al. 2004).

Year Depth Salinity(m) (mS cm'!)

Lake Bryde 1997 1.74 25401999 0.05 537002001 0.48 9510

Bennett's Lake 1998 0.6 18652000 2.43 38102001 2.95 2770

Lake Coomelberrup 1998 03 328002000 N/A 913002002 0.05 160900

Lake Coyrecup 1997 0.9 561001999 1.09 568002001 0.35 125000

Lake Logue 1997 0.36 120401999 3.65 19292000 2.52 33902002 0.28 35700

Noobijup Swamp 1998 0.7 38002000 N/A 38102002 0.8 5700

Lake Towerinning 1997 3.19 90601999 3.34 93002001 2.37 16880

Lake Wheat field 1997 2.0 109001999 2.5 90102001 1.92 9080

wetlands, and it was the only assemblage fullyprotected by the suggested 19 PRWs. Assemblage21 is the least salt-tolerant assemblage but occurs ina habitat (granite rock outcrops) that will notbecome salinised (Pinder et ai, 2000).

Secondly, BRCs will nearly always contain agreater array of wetlands than we have sampledand, consequently, will support a greaterproportion of some assemblages than the PRWswith which they are associated. For example, theLake Muir BRC contains ca 100 wetlands, of whichwe sampled three. Twenty-seven Muir wetlandswere sampled two or three times each in 1996-97and yielded 488 invertebrate species (A.W. Storey,R.J. Shiel and S.A. Halse unpublished data)compared with 228 from the WBS sampling.Similarly, the Lake Bryde and Buntine-MarchageeRecovery Catchments support some freshwaterwetlands we did not sample, as well as a largenumber of small salt pans in braided palaeochannels, and are likely to support more speciesthan WBS data suggest. In addition to extrawetlands providing different habitat and extraspecies, annual variation in wetland conditionsusually causes the long-term species list of awetland to be greater than instantaneous lists (e.g.Halse et al, 1993b; Timms, 1998). Provided sourcesof colonization exist, or propagules remain at the

---------


wetland, PRWs may help achieve conservi1tion ofmore pli1nt i1nd invertebri1te species than theani1lyses herein suggest.

A third, somewhat negative, consideration is theachievability of complete protection in recoverycatchments. Some PRWs are likely to becomedegraded despite attempts to control salinity(Clarke et ai, 2002) and assemblage membersrelying on these wetlands may be lost (Haise et ai,2003). However, the figures presented here onfuture protection of the biota relate only to thesurveyed area and many species deemed to be atrisk i1lso occur in regions without a salinity threat.For example, the blue-billed duck OXYllra allstralisoccurs on the Swan Coastal Plain (as well as ineastern Australia), the copepod Calamoecia attCll11ataand mayfly Tasmallocoellis tilyardi are widespreadthrough southern Western Australia, the sedgeBaumea articulata is widespread in coastalfreshwater wetlands across southern Australia, andthe chenopod Tecticornia verrucosa grows whereversuitable habitat occurs throughout WesternAustralia. Loss of these species from the wheatbeltwill not cause extinction.

Rare speciesOur analysis of the level of protection afforded to

the aquatic biota by PRWs ignores singleton species,which comprised 34% of the biota. While somesingletons have considerable conservationsignificance, many are not localised, rare species.The occurrence of many singletons on the edge ofthe surveyed area (Figure 10) suggests theirapparent rarity is a sampling artefact. The steeprainfall gradients to the west and south of the studyarea cause considerable turnover in the biota overshort distances (Hopper, 1979), which can beexpected to lead to frequent occurrence ofsingletons in surveys. The invertebrate singletons inthe northern wheatbelt were mostly tropical speciesthat had extended south in a wet year (Pinder et al.,2004).

Even within the main body of the study area,where the highest proportion of singletonsrequiring special conservation focus would beexpected to occur, most singeletons do not warrantprotection. Many plant singletons were species thatoccurred commonly at higher elevations interrestrial habitats around wetlands. Only 5% werePriority Taxa (see Coates and Atkins, 2001) that,along with the single Specially Protected waterbird(Botaurlls poiciloptilllS), will require specialconservation focus. Some of the 20% of singletoninvertebrates that were new species or appeared tobe localised (Pinder et al., 2004a) will also requireconservation focus, although many are likely to befound more widely after further survey.

Many of the rare plants, and some rareinvertebrates, occurred in naturally saline wetlands.


r('glll1t'S in

Australia:11.;11'111

COllllllandl'r, D.l), litil'ld. I.K. Thorpe. PAL. Dalie, R.E.,Bird, I.R, and Turner, I,V. (1994) Chlurinl'-.16 andCclrbon-14 llleasurenwnts on hl'fwrs'llitll'g 1'0 und w a te I' in [v I·t ia 1'1' p,lleoc h,1 nne I.s Iwa r

Western Australia. I)wll'ssior\al l\l\,erS5.1·60, ot Wl'stern ;\ustralia, Perth,

lkard. 1.5. (1990). /'111111 o!\V,'.;lcrll /\II.;lmllll. KdngarooPress, Svdnev.

Beard, IS (1999). Evolution 01 thc river svstems of tIll'south-wcst drainage division, Western AustraliaIOllllllllolllic (,;ol/ol 01 \Vc.;lcrll /\II.;lm{lo 82: 1,+7164.

BC'lI'd. 1.5. Ch,lpman, I\.R. ,md Cioia, ['. (2()()O). Speciesrichness and endemism in thc Western Australianflora. 10111'1101 III Bioxcoxmplll/27: 1257-126x.

[klbin. L. (1991). Semi-strong hvbrid scaling: ,1 ncwordination algorithm. IOllrJlIII 01 S"iclI"c 2:

491-496.

Belbin, L. (1993). /'11 IN: pllllcm 1IIIIIIy.;i.; CSl RO,Canberra.

Biological Survcv Committec. (19x4). Thc biologicalsurvcv of the Eastern Coldficlds of Western Austra[ia.Part I. Introduction and methods. I';ccol'd.; 01 Illc

\Vc.;lcrll IIw;11'II1i1l ;\;II1';CIIIII SlIpp{clIICl/l18: 1-19.

Blinn, D.W" Halse, S,A" Pinder, A,M,. Shic[. R.I, andMcRae, I,M. (20(H). Diatom and zooplanktoncommunities and environmental determinants in theWcslcrn Australian wheatbelt: a rcsponse tosecundarv salinisation. III/dro[lIoloXIII 528: 229· 24x,

Boesch, D.F, (1977). Application of numcricalclassification in ecologicll investigations of waterpollution, EPA-600/.1-77-0.13, United StatesEnvironmental Protection Agencv, Washington D.e.

Bowler, I.M, (19x3), Lunettes as ind ices of hyd rologicchange: A review of Australian evidence, Proceedillg.;ollhe Royal Sociely ot' Victoria 95: ]47-68.

Brock, M.A, and Lane, I,A,K, (19x.1). The aquaticmacrophyte flora of saline wetlands in WesternAustralia in relation to salinity and permanence,Ilydrohioloxia 105: 6.1-76,

Brock, M,A, and Shiel, RJ (19:-:.1), The composition ofaquatic communities in saline waters in \NesternAustralia, Ifydrohio!ogia 105: 77-:-:4,

Cale, DJ, Halse, S.A. and Walker, CD. (20lB), WetlandIllonitoring in the Wheatbelt of south-west WesternAustralia: site descriptions. waterbird, aquaticinvertebrate and groundwater data. COII.;er('aliollSciellce Wc.;lcl'II A 11'; I1'111 ia 5: 20-·B5,

CLuke, CL Ceorge, R,I" Bell, R.W. and J latton, TJ(2002), Drvland salinity in south-western Australia:its origins, remedies, and future' research dil'l'ctions,AII,ll'II/lIlIlII'lIl'IIo! (It'Soill~''.-;eilnh40: en-I 1.1,

Coates, DJ and Atkins, K.A, (2001), Prioritv sl'lting ,1Ildthe conSl'l'\'ation of Western Australia's diverse andhighlv endemic flora, C'II-;C['('IIIIOII 97: 251263.

Cra nw r, \ A ,m d 1lob bs, RI. (2002). Lcol og iCillcOllseqlll'nCl's ot ,1Itere'd hvdrologiLaltragmented ecos\'stl'ms in soutlwrn

and m,magellll'nt ~lI',lfI'" ""-

27: 5,16- 564.

1\1\i.;. I.A" llals.., S.A. ,md hOl'IlL!, I~.II. (2001), l.aLtors

Austin. :vU) dnd [klbill. 1 . (] 9:-:2). A new a[Jprodch tothe species cLlSsiticdtion problcms in floristic ,1I1,)I\SIS

11'11'11111111 10111'11011'1 7: 75-:-:9.

I.I\.E. (1999). 1';1 ,,,A: ologc.;.luart llouse. Perth.

REFERENCES

ACKNOWLEDG EMENTS

Anonvmous. (1996a). 11 .;illlOlullI .,IIIIl'IIII'1I1\Vc.;ICrJl /11I.;ll'IIlill. Covernment ot Western Australia,Perth.

Anonvmous. (1996b).\Vc.;lcrll 11 11.;1 I'IIlillll ';01111111/ ollillll

1'1011. C;overnment ot \Vestern Australia. Perth.

;\tmeH. Wand Patlerson, BD. (199.1). TIll' medsure ofordl'l' ,md disordl'l' in the distribution ot species III

fragmenll'd hdbit,lt. ( 96: .17.1.1:-:2

Atlllar. W. dnd [)dtll'I'son, B.D. (1995). fill' 1I1',lc,{lIl'';,

11'IIII'CI'IIIIII'C "0"'111111,'1': II ('1'11111 />il'll pl'l'gl'llIlI,2lJ.j 1IIIIII'uc,. A[C5 ResceHch [nc,

We wish to thank M.R. Williams for statisticaladvice, P. Goia for climatic data from theANUCUM model, and J.c. Cocking and J.M.McRae for producing the figures. Funding for thisstudv was provided as part of the State SalinitvStrategy.

Integrating wetland and upland protectionThis paper has focussed on protecting

biodiversity associated with wetlands which,because they arc low in the landscape, arcdisproportionately threatened by salinisation (Halseet 11/., 2(03). However, protection of uplandbiodiversity is equally important and BRCs cannotbe selected on wetland values alone. Results fromterrestrial components of the WBS (McKenzie et a/.,2004) should be combined with recommendationsabout PRWs (Table 5) and species-rich examples ofthe identified wetland types (Figure 7) to identifycatchments of 50,000-]()O,OOO ha with particularlyhigh biodiversity value. Irreplaceability analysis isone method of combining datasets of disparatebiodiversity values and Walshe et 11/. (2004) haveused the survey data to identify potential SRCs thatencompass much of the biodiversity of thewheatbelt and south coast and provide theframework for landscape-scale conservation efforts.

In most G1SeS, whether on private or public land,these wetlands were threatened with increasedinundation because of salinisation. Whensalinisation is the principal threat, rare speciesusually cannot be protected through small-scalemanagenwnt actions aimed at the organisms andtheir immediate habitat (see Coates and Atkins,2(01). Catchment-scale solutions (C1arke cl al. 2(02)or very expensive localised engineering works(Froend et aI., 1997) arc required.

362

influt'ncing biodiversity in coastal plain wt'tlands ofsouthwestern Australia. III B. Gopa!' W.J. Junk andJ.A. Davis (t'dsJ, III (('et/lll1ds: IIssessII/e11 I.lilllclloll IIl1d collscrullllolI: 89-100. Backhuys, Leiden.

Diels. L. (1906). Die UOII We';I~Auslrallell

,;udllch de,; Wellderkrel,;es. Engelmann. Leipzig.

Dogramaci, S., George, R., Mauger, G and Ruprecht, J.(2003). Waler /Jillallce alld salllllly Irelld. TooJi/JIIIcalchll/elll. Weslerll Au,;lralla. Department ofConservation and Land Management, Perth.

Doupe, R.G. and Horwitz, 1'. (1995). The value ofmacroinvertebrate assemblages for determiningpriorities in wetland rehabilitation: a case study fromLake Toolibin, Western Australia. jourJIlIl of Ihe RoyalSoclely of WeslerJI Au,;lralia 78: 33-38.

Fleishman, E., Betrus, CJ., Blair, RB., MacNally, R. andMurphy, 0.0. (2002). Nestedness analysis andconservation planning: the importance of place,environment, and life history across taxonomicgroups. Oecologlll 133: 78-89.

Frey, D.G. (1991). The species of Pleuro.lus and threerelated genera (Anomopda, Chydoridae) in southernAustralia and New Zealand. Records oflhe AuslrallallMuseul1l 43: 291-372.

Froend, IU-I., Halse, S.A. and Storey, A.W. (1997).Planning for the recovery of Lake Toolibin, WesternAustralia. Wellalld,; Ecology alld Mallagel1lelll 5: 73-85.

Froend, R.H., Heddle, E.M., Bell, 0.1'. and McComb, AJ.(1987). Effects of salinity and waterlogging on thevegetation of Toolibin Lake, Western Australia.Auslraliall jourJIal of Ecology 12: 281-298.

Froend, RlI. and McComb, AJ. (1991). An account of thedecline of Lake Towerrinning, a wheatbelt wetland.Journal of Ihe Royal Socletl/ of WeslerJI Auslralia 73: 1238.

Froend, R.H. and van del' Moeze!, P.G. (1994) The impactof prolonged flooding on the vegetation ofCoomalbidgup Swamp, Western Australia. jourJIal ofIhe Royal SoclL'ly of WeslerJI Au,;lralia 77: 15-22.

Gaston, K.J. (2000). Global patterns in biodiversity. Nalure405: 220-227.

Geddes, M.C, De Deckker, 1'., Williams, W.D., Morton,D.W. and Topping, M. (198I). On the chemistry andbiota of some saline lakes in Western Australia.Hydro/Jlologll/ 82: 20I-222.

Gentilli, J. (1972) Auslrallall clll1Iale pal/erns. Nelson,Melbourne.

George, R.J., Clarke, CJ. and Hatton, T.J. (2002).Computer modelled ground water response torecharge management for dryland salinity control inWestern Australia. Ad1'allces III Ell1'lrolll1lelllalMOllllorlllg alld A10delilllg 2: 3-35.

Ceorge. I\.J., McFarlane, DJ and Speed, RJ. (1995). Theconsequences of a changing hydrologic environmentfor native vegetation in southwestern Australia. IIID.A Saunders, J.L. Craig and E.M. Mattiske (eds).Nalure collscrUllllolI 4: Ihe role of lIelworks: 9-22. SurreyBeatty & Sons, Sydney.

Cibson, N., Keighery, GJ., Lyons, M.N. and Webb, A.(2004). Terrt'strial flora and vegetation of the WesternAustralian wheatbelt. I\ecor,(s of Ihe Weslem AuslralialllviuseUlI1 Supplelllelll 67: 139-189.

7 _

S. A. Halse, M. N. Lyons, A. M. Pinder, R J. Shiel

Halse, S.A. (1981). Faunal assemblagt's of some salinelakes near Marchagee, Western Australia./\uslrallllll Journlll of A11/rllle Illld Fresllil'lller Resellrdl32: 133-142.

lIaIse, S.A. (2002). Diversity of Ostracoda (Crustacea) ininland waters of Western Australia. VcrhalldlulIgellIlIlerllallollllle VerellllgulIg fur Iheorelische ulldallgczl'lllldle Lilllllologle 28: 9 I 4-91 8.

Halse, S.A., Call', D.J., Jasinska, E.J. and Shie!, R.J. (2002).Monitoring change in aquatic invt'rtebratebiodiversity: sample sizt', faunal elements andanalytical methods. Aquallc Ecology 36: 395-410.

Halse, S.A., McRae, J.M. (2001). Calal1loecla Iri/o/JI/la n sp(Copepoda: Calanoida) from salt lakes in south~

westt'rn Australia. Joumlll of Ihe ROl/al Society ofWeslem Au,;lralia 84: 5-1 I.

liaIse, S.A., Pearson, C.B., McRae, J.M. and Shiel,R.J.(2000a). Monitoring aquatic invertebrates andwaterbirds at Toolibin and Walbyring Lakes in theWt'stern Australian wheatbelt. joumal of Ihe RoyalSoclely of WeslerJI Auslralill 83: 17-28.

Halse, S.A., Pearson, C.B. and I'atrick, S. (1993a)Vegetation of dt'pth~gauged wetlands in naturereserves of south~west Western Australia. TechnicalReport 30. Department of Conservation and LandManagement, Perth.

Halse, S.A., Pearson, GB., Vervest, R.M. and Yung, F.H.(1995). Annual waterfowl counts in south~west

Western Australia 1991/92. CALMSclellce 2: 1-24.

Halse, S.A., l\uprecht, J.K. and I'inder, A.M. (2003).Salinisation and prospects for biodiversity in riversand wetlands of south~west Western Australia.Auslralillll jourJIlIl of Holalll/ 51: 673-688.

Halse, S.A., Shiel, R.J., Storey, AW., Edward, D.H.D.,Lansbury, I., Cale, D.J. and Harvey, MS. (2000b).Aquatic invertebrates and waterbirds of wetlands andrivers of the southern Carnarvon Basin, WesternAustralia. Records of Ihe Weslem Auslralillll Museul1lSupplell/elll 61: 217-267.

Halse, S.A., Williams, M.R., Jaensch, R.I'. and Lane,J.A.K. (1993b). Wetland characteristics and waterbirduse of wetlands in south~western Australia. WildliFeResearch 20: 103-126.

Hammer, U.T. (1986). Salille lake ecosl/slell/s of Ihe world.Junk, Dordrecht.

Harper, RJ. and Cilkes, R.J. (2003). Aeolian influences onthe soils and land forms of the southern Yigarn Cratonof semi~arid, southwestern Australia. Geol1lorphology59: 215-235.

Hart, B.T., Bailey, P., Edwards, R., Hortle, K., James, K.,McMahon, A., Meredith, C and Swadling, K. (1991).A review of the salt sensitivity of the Australianfreshwater biota. Hydro/Jlologla 210: 105-44.

Hingston, F.J. and Gailitis, V. (1976). The geographicvariation of salt precipitated over Western Australia.Au,;lralillll Journal of 5011 Resellrch 14: 319-335.

Hopper, S.D. (1979). Biogeographical rates of speciationin the southwest Australian flora. All/l/IIII l\e1'lew of

alld 10: 399--422.

Hopper, S.D., Brown, A.I) and Marchant, N.G (1997).Plants of Western Australian granite outcrops. jourJIalof Ihe Royal of Wes/ern A 1I,;lralill 80: 141-158.


UJI1sl'l'vation planll1l1g; all

tauna spatial distributiunset l/ I";,' li',i 1/011

C,rholil'ld, N.I., Ruprl'cht, IK and Loh, Le. (19~~). fhl'

impatt of agricultural dl'vl'lopml'nt on till' salinitl· 01

SUrfarl' wall'r l'l";'Jurrl'S 01 south,wl'st \Vl'Sll'1'Il

/\ustralia. W5 27. Watl'r 01 Wl'sternl\ustra!i,l, Pl'rth.

5rotts, D. and Drielsm'l, \:1 (201l.1)

Irame\\orks tor

tor

82F 254.

5hl'phnd, D.I'., l\cestoll, (;.R. and Ilol'kins, I\.I.M.(21l1l1). Natl\l' \egdatiul1 in Wl'skrn Australia.

l'inder, I\.M., lIaise, 5.1\., Mc!\,1l', I.M. and 5hil'!, RI

(200'i). Influl'nn' ot salinitv on thl' occurrencl' of

whl'atbelt aqualic invertebrates. lIydl'o/Jiologill (in

I'rl'ss).

Pindl'r, A.M., lIaise, S.A., Shil'l, R.I., Cale, D.I. and

McRal', I.M. (2002). I Jalophilic invertl'brates uf salinl'wl'tlands in thl' whl'atbl'lt of south-wl'stl'rn AustralIa.

,'1'/11/110/111/<.'1'1/ del' 1IIII'I'I/ll1lOIlllle

I'indl'r, A.M., Halsl', 5.1\., Md\,ll', l.:vl and shil'l, R.J.

(2004). Aquatir invl'rtl'bratl' assemblagl's of wl'tlands,llld rivers in the wl1l'atlll'lt rl'giun ot Wl'stl'rn

I\UstLllia PI'IIIIt!, 01 lill' WI'''II'I'I/ 1I"lmllllll :\111."'1/11167: 7-37

N, \1 Jt ll' rnll' il'r, 1\. A., \:1 i tll' rnw in, e.C, d a

hll1 ll'S,l , CA.B. and Kl'I1!.s, I (2000). Biodi\l'rsitvho!.sl'o!.s lor Ul\lSl'rvation I'rioritil's. ,'JIIIIIII' 403: S'i.1S~S.

Nicholls, 1\.0. (19S9). Ilow to m,lkJ' """"1-'.'"'' surn'vs goturtlll'r with ised Illll'ar mudl'!.s.

C'OIl"I'I('IIIIOII 50: 51·· 75.

Norman, 1.1. (19~3). Cn'\' Tl'a!, Chl'stnut TI',ll, ,1l1d Pacific

Black Duck at ,I s,lline habitat in Vittoria. EIIIII 83:

262270.

28: 16~71694.

Pindl'r, ;\\:1, Halsc, 5.A., shil'l, KI. and \:1cRae, I.M.(2000). (;rallitc outcrop puols in south-westl'rn

Australia: foci for divnsification and retugia for

aqu,ltic invl'rkbratl's. jOlll'l/1I1 of 1111' 1\111/111 of

Wl'sll'lII AI/slmlill 83: 117-129.

l'n'ssl'v, R.L., lohnson, I.R. and Wilson, P.D. (1994).

5hadl's of irrl'placl'ability: towards a ml'asurl' of thl'contribution of sites to a rl'sl'rvation goal. /)111,117 ""'-'1,11/

COlIsel'mlioll 3: 242-262.

Rl'migio, LA., Hl'bl'rt, PD.N. and savagl', A. (2001).

Phvlogenl'tic relationships and rl'mark,lbll' radiation

in I'lImrll'lllill (Crustacl'a: Anostrara), thl' l'ndl'mic

brinl' shrimp of Australia: l'vidl'ncl' from

mitochondrial DNA sequl'ncl's. 10111'1/111 of I!II'I.i,lll['i/II 74: 59--71.

Roshil'r, D.A., Robl'rtson, A.I, Kingsford, R. r. ,1l1d Grl'en,

D.G. (200 I). Continl'nt,ll-scale inkrilctions with

temporary resourcl'S may l'xplain thl' paradox uf

large pupulatiulls ot dl'sl'rt watl'rbirds in Austral id.l.lIl1d'lll/1[' 16: 547-556.

5alama, R.B. (199,+). Thl' l'\olution of salinl' lakl's ill till'

rl'lilit drain,lgl' ot thl' Yilg,nn Rin'r, Weskrll

Austrdlid. III R. Rl'llaut ,1Ild W. last (l'ds),

IIl1d of 1/I0dl'I'I/ IIl1d 111111"111

1'11111111111011 :i(}: IWJI99.luls,l,Okldhoma.

(1967). lakrites ,1l1d soils 111

suuthwestnn Australia. III I.N knllings alld I. \.

Mabbutt ,llIdl'" 1I"lllllllllllldC;IIIIII'II: 211·2.10. l\ustralIan Natiollal Universitv,Callbl'rra.

\1.1. (I 5al1l1isat!on in the southwest utWestl'rn Austr,llia. SI'i!/(!1 9: 269 72.

McKell/ie, N.I., Belbin, L., M'lrgules, CR. and K,'igherv,

C;·I· Selecting representative resel'\'e systems

in remote area: a case study in the Nullarbor region,Western Australia. COlhelTllllll1I 50: 239,261.

\:lcKen/ie, N.L., Halsl', S.A. and C;ibson, N. (2000). Some

g,lpS in the reserve system of the southern Carnarvon

Basin, Western Australia. I\ecords of Iltl' We"lcrn

A IISIIllI ill 11 M liSI'll III 511 ppIn 11 I'll I 61: 547-567.

McKell/il', N.!., lohnston, R.B. ,lnd Kendrick, I'C;. (Eds)

(1991) AII"lmlll1. surrev Beattv,Svdnev.

i'vlcKen/ie, N.I. ,1Ild Robinsun, I\e. (l'ds) (19~7). ASIIITel/ ,,/ Ill,' Nllllllr/,or SOIlIIt IIlId

ivV,'"I,'rll Irlllll/ III South l\ustr,llian

Departnll'nt ot Fnvironn1l'nt and I'ldnning, Adelaide.

\:kKen/ie, N.L., C;ibson, N., Keighel'\, dnd Rolfe, IK

(2004) 1',ltterns in the blodiversitv of tnrestridlenvironnll'nts in tIll' Western Austr,llian Wlw,ltbelt.

/\"«1nl" of Ihe tVe"IClII AII"IIIIIIIIII ;\111'[1'1111167: 293-335,

1,H'nslh, R 1', Vc'nes!, 1\1\1. ,1nd I kWISh, \:11 (I '!~~)

vV,lterbird surveys ot wt'tL1l1d n,ltun' ITSl'rves Insouth,western I\ustrali,): 19~1-~'i. Report 30. I\OV,11

Austr,llasian Ornithologists Union, \:klbuurne.

lustus, I. ,1nll sarkar, S. (2002). Ihe princq,I,' ot

compkmentaritv in the ut reserle nt'twurks tu

cunSl'r\'t' biudiversitv: ,1 l,relimin,Hv histurv. 1011111111

,11 BI'''II''1I11 27: 421-4.15.

K,n, W.R, Ilalsl" s.A., scanlun, \:1.1). and Smith, \:11.

(20l1l). Distribution and el1\'irunmenLl1 tulerdnces uf

aquatic ma(Tuinvertebrdte families in tIll' agriniltur,ll

/une of suuthwl'stern Australia. 1011111111 of Iltl' ['Jorllt

AIIICllmll BOlllloloSlll11 20: 1~2-199.

Keigl1l'rv, C (2001). Whl',ltbelt wonders under threat.

011/1'11."101'<' 16 .1742.

1,1111', I.I\.K. (19~5). Important aspl'cts ut duck hunting in

Australia, with particuldr reterence tu Western

l\ustrali,l. III 1.1. Bunning (l'd.L of lite

<1/,'"1'0""1/11 Oil /!inl" IIlId IIll11l11Sl'llIelll, 1011l11lIle"'lIlrs 1983:2~ 1-307. Witwatersrdnd Bird Club, luhannesburg.

I.vuns, M.N., Cibsun, N., Keiglll'ry, Cl. and Ivons, SD.

(20()4). The \vetland flora and vcgl'tation uf the suuthwcst ,lgricultural /une of VVeskrn !\ustralia Pelord,

of Ille 1I>llIIlllIiI :\/111,,'11111 SUPI,!L'mcnt Nu 67:()( 10-000

\:blv, E.J. and B'lvlv, I.A.E. (1991). F,lCtors influencing

patterns uf Australasi,lll centropagidcupepods. 1011111111 of 18: 455-61.

tvlaly, E.I., Halse, S.A. and Maly, M.I) (I Distribution

and incidence patterns uf BOl'lkcllll, C1I1111110e1'll1, and

f[ellllZ,ol'ckelll1 (Cupepuda: Calanuida) in Western

Australia. ;\;1111'111e I1l1d Frcs/1Z1'I11C1 [\I'''l'lIr1'1t 48: 615-621.

Margules, CR. and Pressey, R.L. (2000). Review article:

systematic conservation planning. Nalllre 405: 243-')r-,..... :)~) .

364

Technical Report 249. Department of Agriculture,Perth.

Short, P.S. (1982). Breeding systems and distributionpatterns of some arid Australian genera of thesubtribe Gnaphaliinae (Compositae:lnuleae). In WRBarker and PJM Greenslade (eds), El'olution of the j10raand fauna of arid Australia. Peacock Publications,Adelaide.

Short, Rand McConnell, C. (2001). Extent and impactsof dryland salinity. Resource Management TechnicalReport 202. Agriculture Western Australia, Perth.

Sneath, P.H.A. and Soka\, R.R. (1973). NumericalTaxonomy: the principles and practice of numericalclassification. Freeman, San Francisco.

StatSoft. (2001). Statistica system reference. StatSoft, Tulsa,Oklahoma.

Storey, A.W., Halse, S.A. and Shiel, RJ. (1993). Aquaticinvertebrate fauna of the Two Peoples Bay area,southwestern Australia. Journal of the Royal Society ofWestern Australia 76: 25-32.

Thackway, R. and Cresswell, LD. (eds) (1995). An InterimBiogeographic regitmalisation for Australia: a frameworkfi)/' esta/llishing the national system of reserl'es, l'ersion4.0. Australian Nature Conservation Agency,Canberra.

Thomsen, J.B. (1999). Looking for the hotspots. WorldConserl'atioll 2/99: 6-7.

S. A. HaIse, M. N. Lyons, A. M. Pinder, R. J. ShieI

Timms, B.V. (1998). Further studies on the saline lakes ofthe eastern Paroo, inland New South Wales,Australia. fhldrobiologia 381: 31-42.

Timms, B.V. (2002). The fairy shrimp genus BranchinellaSayee (Crustacea: Anostraca: Thamnocephalidae) inWestern Australia, including a description of fournew species. Hydrobiologia 486: 71-89.

Walshe, T.V., Halse, S.A., McKenzie, N. L. and Gibson,N. (2004). Toward identification of an efficient set ofconservation recovery catchments for WesternAustralian wheatbelt biodiversity. Records of theWestern Australian Museum SupplemCllt 67: 365-384.

Wardell-Johnson, G. and Horwitz, P. (1996). Conservingbiodiversity and the recognition of heterogeneity inancient landscapes: a case study from south-westernAustralia. Forest Ecology alld Management 85: 219-38.

Williams, W.D. (1999). Salinisation: a major threat towater resources in the arid and semi-arid regions ofthe world. Lakes (7' Reserl'oirs: Research and Management4: 85-91.

Williams, W.D., Taaffe, R.G. and Boulton, A.J. (1991).Longitudinal distribution of macroinvertebrates intwo rivers subject to salinisation. Hydrobiologia 210:151-60.

Wilson, P.G. (1984). Chenopodiaceae. III AS. George (ed),Flora of Australia, vof. 4. Phytolaccaceae toChellopodiaceae: 81-235. A u stra Iian Governmen t,Canberra.

Wyrwoll, K.-It (1988). Time in the geomorphology ofWestern Australia. Progress ill Physical Geography 12:237-263.

Electronic appendices are on CD inside the back cover

Biodiversity patterns and their conservation in wetlands of the...

Documents

Transcript of Biodiversity patterns and their conservation in wetlands of the...