DOI: 10.18195/issn.0313-122x.78(2).2015.485-513 Records of the Western Australian Museum, Supplement 78: 485–513 (2015). The riparian fl ora and plant communities of the Pilbara region of Western Australia

M.N. Lyons

Department of Parks and Wildlife, Science and Conservation Division, Kieran McNamara Conservation Science Centre, Locked Bag 104, Bentley Delivery Centre, Western Australia 6983, Australia. Email: [email protected]

Abstract – A survey of riparian fl ora and plant communities was undertaken at 98 wetlands and rivers in the Pilbara region of Western Australia. Sampling was quadrat-based, with fl oristics, surface soils and wetland attributes recorded. Selected sites captured the full range of Pilbara wetland types including springs, river pools, claypans, salt marshes and rock pools. A total of 455 taxa was recorded from the survey sites, representing ca. 25% of the known fl ora of the Pilbara bioregion. The fl ora is dominated by taxa with Eremaean and tropical affi nities, with only six taxa endemic in the region. Of recorded taxa known from four or fewer bioregions, most are shared with the adjacent Carnarvon and Gascoyne bioregions rather than the adjoining internally draining deserts. Sixteen taxa of conservation signifi cance were documented, with claypans, the Fortescue Marsh, and Millstream and Karijini National Park sites dominating occurrences of rare species. Eight major groups were defi ned by classifying wetlands in terms of species presence/absence data. Floristic patterning was strongly aligned with the major wetland types (geomorphic/hydrological) used in the primary sampling stratifi cation. A combination of wetland morphology/hydrological setting, site edaphic attributes and distance to the coast were dominant variables related to riparian fl oristic composition. Primary compositional separation was observed between riverine and non-riverine sites, with lowland turbid riverine sites with fi ne- textured soils compositionally related to claypans and clay fl ats. Limited biogeographic patterning was evident except where individual IBRA subregions and drainage basins were dominated by few wetland types.

Keywords – Pilbara, wetland, riparian, floristic composition, botanical survey, vegetation, rivers

INTRODUCTION coupled with a rocky middle and upper landscape, produces river fl ows of high energy and volume The Pilbara IBRA region of Western Australia and relatively short duration. Extensive aquifer 2 is a region of approximately 179,000 km that systems discharge at the surface to form springs corresponds broadly with the Pilbara Craton, and contribute perennial fl ows to rivers and creeks. a major geological block of Archaean origin In lowlands, major rivers have broad channels (Geological Survey of Western Australia 1990). The with fl ood fl ows forming numerous side channels. region is recognised as biogeographically distinct Clay fl ats, claypans and lowland creeks with fi ne under the scheme developed by Thackway and sediments are typically highly turbid. Creswell (1994), based on the original Fortescue Land use in the region is overwhelmingly Botanical District defined by Beard (1990). A dominated by mining and extensive pastoral use. detailed description of the soils, climate and These uses place pressures on Pilbara wetlands landforms of the region is provided by McKenzie et through hydrological changes associated with al. (2011). mine dewatering and discharges, and the impact The region’s climate is broadly arid, with highly of grazing cattle. In the Pilbara, stock watering at seasonal, typically intense, summer rainfall that natural water points is uncontrolled, concentrating may include cyclonic systems (Leighton 2004). This, trampling and grazing in riparian zones, leading to 486 M.N. Lyons

Plate 1 A highly turbid ephemeral claypan at Yarraloola (site PSW074A) in lowlands dominated by Eriachne benthamii. Trampling by cattle is evident (M.N. Lyons).

bank erosion and waterbody siltation. Aggregation the current study, coupled with studies of aquatic of stock, particularly late in the dry season, invertebrates by Pinder et al. (2011), will provide the leads to the eutrophication of waterbodies from basis for spatially explicit modelling of the biotic accumulated faeces. composition of Pilbara wetlands and rivers. Protection of wetland biological values across the Pilbara is contingent on a greater understanding STUDY AREA AND METHODS of their biogeographic patterning, for a number The Pilbara region as defi ned in this study is of their biotic components. The existing reserve network in the Pilbara consists of four reserves bounded to the south by the Ashburton River and that, while capturing some major wetlands (e.g. the east by the De Grey–Oakover River system. Karijini National Park – gorges, Millstream A small number of wetlands were also sampled National Park – springs), do not include the full in the Rudall River area to the east of the main diversity of Pilbara wetlands. The inclusion of study area (Figure 1). The study area includes wetland fl oristic data to inform formal conservation the entire Pilbara IBRA region (Thackway and planning is a relatively recent feature of broad- Cresswell 1994) and additional small areas of scale biodiversity survey in Western Australia the adjoining Gascoyne and Little Sandy Desert (Keighery et al. 2000; Lyons et al. 2004; Walshe et al. regions, representing a total area of approximately 2 2004). This paper aims to document the riparian 225,000 km . flora of the wetlands and rivers of the Pilbara Precipitation across the region is dominated and examine how the major geographic and site by summer rainfall associated with tropical low- environments correlate with fl oristic composition. pressure systems producing thunderstorms and The data sets and understanding developed in cyclones. Annual average rainfall is 290 mm. Pilbara riparian fl ora and plant communities 487

Figure 1 Maps of the Pilbara study area showing the 98 wetland sites sampled (note A and B sites at the same wetland are shown as a single point). Figure 1A shows IBRA6.1 subregions as shaded areas with full names; the Pilbara Bioregion comprises four subregions: Roebourne Plains (Eastern & Western), Chichester, Fortescue and Hamersley. Drainage basin boundaries are outlined in black and labelled with abbreviated names as follows: ASH, Ashburton; OC, Onslow Coast; FOR, Fortescue; PHC, Port Hedland Coast; DGO, De Grey-Oakover; GSD, Great Sandy Desert. Figure 1B shows the main rivers in blue and two major wetlands (shaded grey). The sampled wetland sites are numbered as per Table 1. 488 M.N. Lyons

Inter-annually and spatially, rainfall variability is high, driven largely by the occurrence and path of cyclones. Winter rainfall from cold fronts originating in the south contributes signifi cantly in the western coastal and central uplands of the region, and attenuates markedly towards the east. Two resulting bioclimatic zones were identifi ed by Beard (1990), with the western coastal area characterised as semi-desert tropical, and the remainder as desert.

Site selection and sampling Ninety-eight wetlands were selected to capture the diversity of wetland types present in each of the fi ve major drainage basins within the Pilbara (Figure 1A, Table 1). The selected wetlands formed the basis of an integrated wetland survey of the Pilbara that covered a variety of biotic groups of the water column, including the riparian fl ora and vegetation reported here (see McKenzie et al. 2009, Pinder et al. 2010). Here, riparian is used to describe environments at the edge of waterbodies; i.e. under the hydrological infl uence of the wetland but not inundated at the time of sampling. This includes river banks, the waterlogged shallow soil aprons of rock pools, and the margins of claypans. Some riverine communities, such as Melaleuca glomerata on sandy islands, were rarely sampled since they were often not adjacent to waterbodies. The primary stratifi cation employed in the survey was the a priori designation of a number of wetland types based on hydrological and morphological Plate 2 Munreemya Billabong, a seasonal/episodic wetland on Yarrie Station (wetland site attributes. These closely follow the scheme adopted PSW036A). The riparian zone dominated for the region by Masini (1998): rock pools; springs by Eucalyptus camaldulensis and E. victrix and their outfl ow creeks; river pools; turbid linear over Eriachne benthamii and Dichanthium pools; claypans and fl ooded clay fl ats; ephemeral fecundum sampled using a linear quadrat creeks; salt marshes; and gorges. Wetlands were (M.N. Lyons). selected by examining topographic maps and consulting local experts, followed by aerial and ground-based reconnaissance. Within major river 1) were sampled at each of fi ve large wetlands systems, sampling sites were distributed along the (PSW002, PSW003, PSW004, PSW035, and PSW080), length of the main channel, and included the full to capture the major wetland zonation. At some range of stream orders from ephemeral headwater creeks to near-coastal river pools (Figure 1B). small wetlands (typically small claypans and ephemeral streams), discontiguous subquadrats Sampling was undertaken within a representative were surveyed in order to cumulatively sample 200 section within each selected wetland. At most m2 and to avoid upland vegetation. Quadrats were wetlands (93), a single linear quadrat (200 m2) was marked with steel pegs and their locations recorded established parallel to the waterbody margin and positioned to capture the typical riparian setting using handheld GPS. For wetlands where the 2 of the wetland. Typically, quadrat confi gurations entire riparian zone was less than 200 m , such as at each wetland were a single 4 m x 50 m or 5 m x very small claypans and rock pools with fringing 40 m rectangle, depending on riparian zone width. shallow soil aprons, the whole wetland fringe Within each quadrat, all vascular plants were was sampled (Table 1). These different quadrat scored within eight contiguous 25 m2 subquadrats confi gurations are herein referred to as sites. Sites to evaluate the adequacy of quadrat size. Two were scored twice, in autumn and spring, between 200 m2 quadrats (designated A and B, see Table 2004 and 2006. Pilbara riparian fl ora and plant communities 489

Plate 3 Ephemeral claypan on Roy Hill Station (site PSW039A). Sampling the grass- and herb-rich margin downslope of upland vegetation dominated by Triodia (D. Mickle).

Species accumulation curves were produced in versus non-riverine sites, and altitude (Table 3). A EstimateS (using Mao Tau) for eight sites to examine suite of climatic variables (Table 3) was derived the adequacy of the quadrat design (Colwell 2009). from the BIOCLIM module of ANUCLIM (Houlder Input data were compiled as a presence/absence et al. 2000). 2 matrix of species at the eight subquadrats (25 m ) To provide a broader perspective on the riparian 2 within each 200 m quadrat. fl ora, the proportion it represents of the known Within sites, surface soil samples (at 5–15 cm Pilbara fl ora, based on records held at the Western depth) were collected from 30 spaced points Australian Herbarium (S. Dillon, unpublished and bulked in the fi eld to yield approximately data), was determined. Additionally, the broader 2.5 kg samples. Particle size and soil chemistry biogeographic relationships of the riparian fl ora were analysed by the Chemistry Centre of were explored by examining the IBRA regional Western Australia. Parameters analysed included occurrences of the taxa across Western Australia pH, electrical conductivity, organic carbon, based on voucher specimen geocodes. Due to phosphorous, nitrogen, potassium, calcium, differences in taxonomic resolution across the data magnesium, sodium, and percent silt, clay and sand, and three gravel fractions (Table 2). A number sets, some taxa were amalgamated at the specifi c of additional variables were derived for the sites by level to yield a fi nal dataset of 430 taxa across the Pinder et al. (2011) and are used here. They include State’s IBRA regions. Data on the occurrence of Strahler stream order (Strahler 1952), four inferred Pilbara riparian taxa were also compiled for the permanence/hydroperiod classes (1 = ephemeral, botanical provinces (Northern, Eremaean and 2 = seasonal episodic, 3 = near permanent, 4 = South-West) of Beard (1990), treating the Coolgardie permanent), straight line distance to coast, riverine IBRA region as Eremaean.

490 M.N. Lyons

Quadrat dimensions Quadrat

cation; cation; PrecColQ

PrecSeas annual range (°C); annual range (°C); AnnPreci

ecColQ – precipitation ecColQ – precipitation

– Strahler stream order order – Strahler stream MeTemWeQ TempAnRa

eTemp – annual mean temperature – annual mean temperature eTemp us 4 x 37.5 + 4 x 12.5; E – 10 x 20; F – us 4 x 37.5 + 12.5; E – 10 20; F MaxTeCoP

MedDiuRa AnMeTemp

24.8 15.625.2 40.2 14.9 32.325.2 40.1 31 14.9 30.6 40.1 31.2 279 30.6 303 88 31.2 93 303 42 40 93 A 23.4 40 13.4 A 24.6 37.9 13.5 29.4 A 38.7 29.6 371 2924.8 30.7 14.3 8825.6 39.3 357 53 14.1 29.7 9625.6 38.7 30.8 14.1 27.9 A 350 4825.7 38.8 31.1 14.2 88 352 28 A 25.6 38.8 100 31.1 59 14.2 27.9 38.8 31.1 354 48 27.9 A 345 100 31.1 A 99 48 346 48 98 A 48 A B 24.7 15.324.8 40.1 15.6 32.1 40.2 31 32.3 275 31 87 279 44 88 42 A A 25.7 13.925.7 39.7 13.9 29.225.7 39.7 31.4 13.8 29.2 345 39.6 31.4 29.1 96 345 31.4 44 96 34924.9 44 14.2 A 97 39.8 44 31.1 A 31 A 299 84 51 A

DisCoast

Riverine Stream order Stream

guration) ; Site group – group derived from UPGMA classifi UPGMA derived from – group guration) ; Site group

Perm class Perm Site group Site 1200267 2200228 2200228 1431216 1441177 1411169 145195 141197 145192 141195 1341286 1200267 2200180 2200180 2221174 7451365 Longitude (°E) Latitude (°S) er (see Quadrat dimensions for sampling confi ora. Fortescue Marsh East (Moojari)Fort Marsh West (Moojari)Fort Marsh West Gnalka Gnoona Claypan salt marsh salt marshGnalka Gnoona Claypan salt marshKoondepindawarrina PoolFortescue Falls claypanHamersley Gorge claypan 22.3162 22.5085 pool turbid creek Bobswim Pool 119.1499 22.3132 119.7796Palm Spring on Caves Creek 119.1530 5. 5. 22.1203Palm Pool at Millstream 29.1530 5. 118.3952 DeltaMillstream 118.4920 spring 29.1546 (spring fed) creek Gorge (spring fed) creek Gorge 22.4774 Pool Gorge Gregory 118.4751 22.2587 118.5513 117.9859Palm Spring at Millstream river pool river pool 6. 6. spring 22.2674 spring river pool 117.0383 21.5698 23.0585 117.0553 6. 118.0903 6. 21.5519 21.5721 21.5839 116.9686 116.9615 117.0677 6. 6. 6. Perm class – Inferred permanence class (1 – ephemeral, 2 – seasonal/episodic, 3 – near permanent, 4 – permanent); Stream order order permanence class (1 – ephemeral, 2 seasonal/episodic, 3 near permanent, 4 permanent); Stream Perm class – Inferred AnM 1952); Riverine/non-riverine ( 0 – non-riverine, 1 riverine); DisCoast straight line distance to coast (km); (Strahler, – temperature coldest period (°C); TempAnRa – maximum temperature diurnal range (°C); MaxTeCoP (°C); MedDiuRa- mean temperature seasonality; Pr – precipitation (mm); PrecSeas – annual precipitation AnnPrec wettest quarter (°C); – mean temperature MeTemWeQ – 5 x 40; B 4 50 ; C discontiguous 30 + 10; D discontiguo (A coldest quarter (mm); Quadrat dimensions in metres discontiguous 5 x15 + x 25; G – 20 +5 20; H sampled out. Sites sampled for riparian fl PSW001A Coondiner PoolPSW002B PSW003A PSW003B PSW004A pool turbid creek PSW004B PSW005A 22.7256PSW006A 119.6566 PSW007A PSW008A PSW009A PSW010A PSW011A PSW012A PSW013A PSW002A Fortescue Marsh East salt marsh 22.5091 119.7769 5. Site code Site name type Wetland Table 1 Table Codes: Site code – unique sample site identifi

Pilbara riparian fl ora and plant communities 491

Quadrat dimensions Quadrat

PrecColQ

PrecSeas

AnnPreci

MeTemWeQ

TempAnRa

MaxTeCoP MedDiuRa

14.5 39.1 31.2 30.6 297 88 44 A 15.1 39.8 29.3 31.5 314 102 37 A AnMeTemp

26.2 14.825.9 39.5 14.8 28.1 39.8 31.5 307 2925.9 31.6 13.1 91 36.4 351 57 24.5 9226.9 30.6 14.6 A 304 58 40.5 101 28.8 A 25.9 31.8 41 15.5 329 40.4 A 104 30.6 31.4 37 33024.8 A 15.7 99 40.2 42 32 31.1 C 301 91 39 A 25.9 15 40.1 29.3 31.5 321 94 61 A 25.9 14.2 37.9 26.3 30.9 291 91 5427.4 14.2 A 40.5 27.826.3 32.1 15.6 319 40.7 107 30.726.3 31.8 34 15.3 30226.1 40.1 A 15.6 99 30 40.4 31.5 37 30.9 31.7 286 A 272 9425.1 15.1 93 3824.9 40 36 15.1 31.9 A 25.1 39.9 31.2 15.2 31.9 A 25.7 40.1 31.1 230 15.4 32.1 231 8825.3 40.8 31.3 13.1 30.6 87 224 34 37.3 31.8 35 24.9 88 277 A 25.7 30.3 33 13.2 A 87 30426.6 36.2 65 13.8 24.8 92 B 26.5 30.3 38 69 13.6 A 308 25.926.5 37.7 104 31.2 13.6 25.5 A 37.7 31.1 309 39 25.5 305 103 31.1 A 103 35 305 35 103 A 35 A A

DisCoast

Riverine

Stream order Stream

Perm class Perm Site group Site 144162 143199 14518 1421104 1431141 1431226 142129124 232190 322121 735111426 836185 7451161 7321164 8461192 3321342 3331337 7221362 3121128 220014 322118 330038 120024 120024 Longitude (°E) Latitude (°S) Myanore Creek Pool Creek Myanore Chalyarn PoolNyeetbury Spring PoolWodgina Munda Homestead Pool pool turbid creek StationYarrie at De Grey Coppin Gap Pool 21.4435 river pool spring 115.8625 Pelican Pool river poolGlen Herring Pool river pool river pool RockholeTarquin Carawine Gorge 21.7551 river pool 20.5220Bamboo Spring 116.0365 21.8593 118.0618 20.6800 Spring Wolli Weeli 116.5147 21.4255 6. river pool 120.2020 Queen Desert Spring river pool 6. 118.4725 6. river pool Pool Creek Watrara 20.8840 PoolPoonamerrala Creek river pool 120.1180 PoolMoreton spring 21.3424 spring 6. 21.3321Cane River Claypan river pool 119.6136 21.1906 120.3744 river poolAmy Well Pool near Mt. Creek 121.0838 river pool 6. 21.4774Cooliarin Pool 121.0279 river poolParadise Pool 22.4674 22.1489 ClaypanDeGrey 22.9156 claypan 122.2583 119.6793 claypan 22.6186 119.2050 ClaypanDeGrey 22.5038 122.3586 122.0768 6. 6. 22.1945 115.9556 pool turbid creek 21.8057 6. claypan 22.6765 claypan 115.1048 20.5051 115.7161 claypan 118.6219 20.3777 20.2894 119.4204 119.4310 20.2886 119.4334 PSW014A PSW015A PSW016A PSW017A PSW018A PSW019A PSW020A PSW021A PSW022A PSW023A PSW024A PSW025A PSW026A PSW027A PSW028A PSW029A PSW030A PSW031A PSW032A PSW033A PSW034A PSW035A PSW035B Site code Site name type Wetland

492 M.N. Lyons

Quadrat dimensions Quadrat

PrecColQ

PrecSeas

AnnPreci

MeTemWeQ

TempAnRa

MaxTeCoP MedDiuRa

14.9 39.8 28.7 31.5 330 92 61 A AnMeTemp

25.8 15.2 41.5 32.4 32.2 234 81 47 A 27.2 14.226.9 40.4 13.9 27.826.8 39.9 31.9 14.6 27.3 32324.7 39.9 31.5 106 15.5 28.1 33325.4 40.1 31.5 35 105 14.5 32.3 32124.4 38 A 103 40 31 15.6 30.324.4 39.7 272 35 B 31.3 15.5 32.6 8622.8 39.7 30.8 309 B 14.9 32.6 257 43 9422.8 38.6 30.8 14.9 33.2 86 257 40 D 23.8 38.6 29.6 36 14.8 33.2 85 223 A 24.8 39.6 29.6 37 14.4 32.8 A 73 223 40.1 30.6 37 31.6 A 73 23625.8 31.2 37 14.6 70 E 26925.7 39.4 45 15.4 28.2 78 E 25.5 40.9 31.3 49 30.8 A 318 1525.7 31.8 40.1 A 98 274 15 29.425.8 40.6 31.4 63 15.1 87 30.525.7 41.2 288 31.8 64 15.1 31.6 A 88 41.4 332 32 32.5 A 70 8524.5 32.2 275 13.8 224 61 A 8225.1 38.8 13.9 29.5 78 52 A 26.2 38.3 30.5 47 13.4 28.1 341 A 26.7 37.4 30.6 13.9 25.4 A 88 372 38.6 105 31 51 26.4 31.5 304 48 A 319 103 A 103 41 37 A A

DisCoast

Riverine

Stream order Stream

Perm class Perm Site group Site 8371257 324184 320073 410092 4100285 3200225 1100312 4100317 2200436 2200436 7451369 2311289 73417426 835165 8451132 847195 7451148 2411215 2200288 3221185 7421105 745112 846143 Longitude (°E) Latitude (°S) Ashburton R.sh Pool on river pool 23.3667 116.9596 Munreemya BillabongMunreemya Coppin Pool ClaypanSweet Well Roy Hill Claypan claypanMulga Downs Outcamp Claypan ClaypanEthel Creek claypan ClaypanJackson Bore claypan claypan Salt LakeWeelarrana claypan 20.6696 Salt LakeWeelarrana 120.2269 PoolYandibiddy claypan 22.3619 claypanFork Spring 118.9763 20.6521 salt marsh PoolRed Hill Creek 20.5729 120.6327 salt marsh 120.2604 House Pool on Cane River 22.6750 119.8111 Henry River Pool 22.8234 river poolAshburton R.Whiskey Pool on 22.8502 120.2587 24.0661 river pool 120.2356 Pool on Duck CreekWallarook river pool 119.8781 24.0706 river poolCheela Spring spring 119.8734 river pool 5. Berringarra Claypan 23.9022 5. river poolCatfi 118.7123 22.0960 Pool Creek Wackilina 21.9623 115.6209 22.5296 116.0477 Errawallana Spring 115.2832 22.4791 23.2836Pool Spring on Sherlock River claypan 116.4684 118.1020 22.7273 springCarleecarleethong Pool 115.6993 6. pool turbid creek river pool spring 22.3828 23.4922 river pool 117.6140 22.9573 117.1981 20.8124 116.8403 117.5969 6. 21.6265 20.4655 117.7752 119.5295 PSW036A PSW037A PSW038A PSW039A PSW040A PSW041A PSW042A PSW043A PSW043B PSW044A PSW045A PSW046A PSW047A PSW048A PSW049A PSW050A PSW051A PSW052A PSW053A PSW054A PSW055A PSW056A PSW057A Site code Site name type Wetland

Pilbara riparian fl ora and plant communities 493

Quadrat dimensions Quadrat

PrecColQ

PrecSeas

AnnPreci

MeTemWeQ

TempAnRa

MaxTeCoP MedDiuRa

15.3 40 30.3 31.3 285 91 39 A 13.3 37.215.4 25.1 41.1 30.9 30.9 269 32 93 280 49 87 H 60 A AnMeTemp

25.1 14.225.9 38.9 13.7 28.727.1 37.8 30.8 14.2 26.3 36627.3 40.2 31 14.4 27.6 95 40.5 31.8 323 57 28.2 324 11125.4 105 32 15.7 A 4024.3 40.2 318 35 16.1 31.7 A 10625.2 39.8 31.3 A 15.9 32.6 34 26523.4 40.2 30.6 14.6 31.9 89 A 31923.3 38.4 31.1 37 15.1 31.7 89 30124.8 39.6 30 42 14.4 33.6 A 9124.7 30.4 291 40 41 14.7 A 226 31.5 7925.9 40.2 31.1 15.6 32.2 A 68 45 40.3 31.3 275 43 31.3 245 A 79 31.5 A 74 263 4925.1 47 13.1 91 A 25.8 37.2 36 13.6 24.9 A 25.9 37.3 30.1 12.5 25.2 A 30026.1 36.3 30.7 14.6 23.7 90 29225.9 38.3 30.6 72 12.5 26.9 92 27426.1 36.4 31.1 57 13.7 23.7 94 B 30326.1 37.9 30.6 43 13.5 26.4 90 B 27225.8 37.2 31 61 14.3 25.6 94 F 37.9 30.9 319 43 A 27.1 307 107 30.7 102 H 41 309 41 102 G 39 A H

DisCoast

Riverine

Stream order Stream

Perm class Perm Site group Site 3331125 142153 131177 724194 112117826 1411230 7431217 3331193 7341336 8441418 7411288 2451320 1411218 41002226 821114226 410020 310011 72110.9 145135 73111 411138 324124 420062 Longitude (°E) Latitude (°S) Kumina CreekThe Springs on Jones RiverChinaman SpringMiningarra Creek spring Range CreekGregory Skull Springs river poolBonnie Pool spring PoolCookes Creek ephemeral creek ephemeral creekKalgan Pool on Creek 21.1039 Spring 117.2483Wannagunna 21.8626 20.7492 SpringFlatrocks 116.9090 river pool 21.2538 6. spring 120.3067 121.0783Horrigans Pool 20.6105 river poolRunning Waters river pool 120.3448 6. Karratha Granite Rock river pool 6. 23.1871Curara Claypan 119.6976 ClaypanMindaroo spring 21.6806 21.8635 21.9778 ClaypanYarraloola 120.4813 121.0078 river pool 119.9912 pools rock 24.3006 Peninsula CreekBurrup spring 6. 118.8731 Pool on Lower Fortescue River Peninsula RockholeBurrup claypan claypan 23.2533 river pool ClaypanCroyden 23.5552 ephemeral creek 20.9100 claypan 118.1514 118.2586 Pool Peawah Creek West 116.7353 ephemeral creek 21.6858 on Indee StationRedrock 6. 121.1253 20.5656 116.8247 22.6647 21.3324 21.9146 6. 20.5736 pool turbid creek 116.0702 116.1547 114.9090 116.8075 21.4285 claypan pool rock 115.6847 6. 20.7641 118.0150 20.8755 21.0383 118.5885 117.6534 PSW058A PSW059A PSW060A PSW061A PSW062A PSW063A PSW064A PSW065A PSW066A PSW067A PSW068A PSW069A PSW070A PSW071A PSW072A PSW073A PSW074A PSW075A PSW076A PSW077A PSW078A PSW079A PSW080A Site code Site name type Wetland

494 M.N. Lyons

Quadrat dimensions Quadrat

PrecColQ

PrecSeas

AnnPreci

MeTemWeQ

TempAnRa

MaxTeCoP MedDiuRa

13.2 37.1 24.9 30.8 269 93 49 A AnMeTemp

25.8 14.325.2 37.9 15.8 27.125.1 40.2 30.7 15.6 309 3225.1 102 40 31.2 14.8 32.124.6 40.7 265 39 31.1 32.5 15 8924.4 31.6 247 H 40.7 14.9 230 37 33.1 8824.6 40.5 31.5 15.6 33.2 72 34 A 23.6 39.7 30.9 221 47 32.4 223 15 A 6825.8 30.9 38.8 D 67 230 50 16 32.526.1 40.8 30.3 52 15.7 87 31.2 A 25.7 40.3 255 31.5 32 13.6 30.2 B 8125.3 37.7 31.5 319 13.8 26.4 A 336 39 9925.9 37.8 30.9 103 15.2 329 38 27 B 23.1 40.3 42 108 30.6 14.5 30 A 24.2 38.1 344 43 A 31.5 13.3 31.6 10525.5 38.4 29.7 313 A 14.1 29.1 47 296 10026.1 38.7 30.3 13.7 27.9 79 37 A 37526.2 37.9 31 46 13.9 26.4 91 A 38.1 356 31 51 26.5 A 102 31.2 319 B 48 316 107 107 41 A 41 A A

DisCoast

Riverine

Stream order Stream

Perm class Perm Site group Site 410062 7321238 3350266 3101337 1261378 3231386 3431308 3341370 7351150 1411115 745141 742170 1341147 7221335 3241201 322198 121140 311143 31002426 Longitude (°E) Latitude (°S) at at 21.0638 117.7091 20.8759 116.7298 Redrock on Indee StationRedrock PoolBillan Ballan Creek Pool Creek Tongololo Claypan Creek Turee pool rock Junction PoolGorge river poolGlen Ross Creek river pool PoolCarrowina claypan 20.8746Innawally Pool 118.5895 Rocky Island Pool 21.9690 river pool 120.8556 Panorama Spring 22.1653Kangan Pool on Sherlock River ephemeral creek 121.0803 River PoolHarding 23.6525 river pool river pool 118.0988 25.0842 Ck.Cangan Pool on Cockerega 24.1790 river pool river pool 118.0548Homestead Creek 118.0298 river pool spring CreekJoffre 6. 22.5784 N.P. in Millstream Un-named Creek river pool 21.0977 ephemeral creek 121.0343 Claypan Well Benmore 23.3621 117.6271 21.5481 120.1904 Crabhole FlatCroydon 119.3435 21.6953 ephemeral creek 21.5799Karratha Crabhole Flat 118.6296 21.2018 117.1029 21.3377 119.3205 claypan 6. 117.0697 23.2485 ephemeral creek 6. clay fl 119.5992 clay fl 22.3918 118.2583 21.0444 117.6636 PSW080B PSW081A PSW082A PSW083A PSW084A PSW085A PSW086A PSW087A PSW088A PSW089A PSW090A PSW091A PSW092A PSW093A PSW094A PSW095A PSW096A PSW097A PSW098A Site code Site name type Wetland

Pilbara riparian fl ora and plant communities 495

K (exch) me% (exch) K

Na (exch) me% (exch) Na assium milliequivalent assium milliequivalent me% (exch) Mg ch) me% – exchangeable ch) me% – exchangeable

ly Black); P (total) mg/kg – (total) mg/kg – ly Black); P

Ca (exch) me% (exch) Ca

P (HCO3) mg/kg (HCO3) P

Gravel 2–4.8mm % 2–4.8mm Gravel

Gravel 4.8–8mm % 4.8–8mm Gravel . (2009). % 8–16mm Gravel

et al

Gravel >16mm % >16mm Gravel

Clay % Clay

Silt % Silt

Sand % Sand

P (total) mg/kg (total) P

OrgC (W/B) % (W/B) OrgC

pH (H2O) pH

mS/m EC (1:5) (1:5) EC total Phosphorus; P (HCO3) mg/kg – extractable Phosphorus; Ca (exch) me% – exchangeable Calcium milliequivalent percent; Mg (ex Ca (exch) me% – exchangeable Calcium milliequivalent percent; (HCO3) mg/kg – extractable Phosphorus; P total Phosphorus; K (exch) me% – exchangeable Pot Na (exch) me% – exchangeable Sodium milliequivalent percent; Magnesium milliequivalent percent; percent. Soil variables for each riparian site sampled. For detail on soils analyses see Meissner carbon (Walke organic (W/B) % – percent Codes: EC (1:5) mS/m – electrical conductivity; pH (H20) in distilled water; OrgC Table 2 Table Site code PSW001APSW002APSW002B 8PSW003A 140PSW003B 6.4 150 7.4PSW004A 290PSW004B 7.9 1.55 770 0.37 7.9PSW005A 8 640 0.52PSW006A 450 8.2 8 0.49PSW007A 56.5 370 11 6.5 34 0.78PSW008A 350 70 12 6.5 63PSW009A 240 0.9 6.3 32 19 49PSW010A 8.3 0.9 31.5 12 15.5 49.5 480 1PSW011A 34 8.4 32.5 200 0 21.5 2.02 480PSW012A 31 8.6 44 83 18.5 0.76 8.4 520 0PSW013A 0 480 410 44 0 19.5 0.66PSW014A 19.5 0 8.8 270 85 36 0.86 83.5 8.3 0 19.5 0 190 0 420 36.5 83.5 0 0.41 220 8.3 10.5 0 22 5 3.7 36.5 90.5 8.3 3.9 0 8 170 0 0 74.5 6 1.28 0 3.9 42 5.5 290 0 6.3 1.02 9.8 77 11.5 0 270 8.5 0 0 15.7 9.8 55.5 0 4 50 110 1.24 13.3 14 0 8 74.5 19.6 15.4 29.5 34 13.3 0 24 380 75 18.1 6.3 0 4.3 11.1 11.5 10.2 19 17.1 15 15 18.1 7.11 24 23 5.48 17 2.75 8 15.9 14 8.3 12.9 19.2 0 23 15.02 0 3.13 24.1 2.12 0.33 29 10.76 30.1 8 28.9 5.3 6 3.02 9 9.32 3.41 19.5 10.76 4.64 0 2.27 1.18 16.5 0 3 47 1.64 33 6.54 4.64 0 3.56 15.5 4.12 0.46 24 2.34 3 54 4.12 3.32 0 0.46 0 5.59 16.4 5 6.9 2.02 0 40.5 5 5.46 10.18 2.02 4.14 2 1.5 0 0 14.4 4.47 4.42 14.23 0 0.41 4 0.69 4.83 1.32 2.99 0.63 0 9 0.55 0 7.16 0.22 1.2 2.03 7.24 0 12.1 2.07 0.17 3 5.14 15.58 0.89 9 1.78 11.89 11.57 0.9 5 9.7 15.54 5.73 7.62 0.18 1.6 1.16

496 M.N. Lyons

K (exch) me% (exch) K

Na (exch) me% (exch) Na

Mg (exch) me% (exch) Mg

Ca (exch) me% (exch) Ca

P (HCO3) mg/kg (HCO3) P

Gravel 2–4.8mm % 2–4.8mm Gravel

Gravel 4.8–8mm % 4.8–8mm Gravel

Gravel 8–16mm % 8–16mm Gravel

Gravel >16mm % >16mm Gravel

Clay % Clay

Silt % Silt

Sand % Sand

P (total) mg/kg (total) P

OrgC (W/B) % (W/B) OrgC

pH (H2O) pH

mS/m EC (1:5) (1:5) EC Site code PSW015APSW016APSW017A 67PSW018A 50PSW019A 9 30PSW020A 8.4 38PSW021A 9 0.92 23 1PSW022A 8.4 48 330PSW023A 8 0.5 33 0.56 280PSW024A 9.7 77.5 110 190PSW025A 9.4 0.54 150 39 88 14 0.38 9.6PSW026A 100 87.5 190 81 0.4PSW027A 9 100 8 48 8.5 0.74 5.5 8PSW028A 88 40 150 7 91.5PSW029A 200 0 8.4 1.56 4 53 7 2.41PSW030A 5 93 8.5 5 84 89 260 12 1.22PSW031A 0 7.1 230 310 15.7 26.1 0.89 4.5PSW032A 7 8 81.5 380 3.5 6 17 9 16.6 1.02 7.9 74.5PSW033A 13.6 0 300 10 2.5 12 81.5 12PSW034A 0 6.6 7.8 0.67 87 14 350 5 6.8 11.7 0.38 88PSW035A 7.4 8.5 0 0 4 6.5 200 10.7 1.68 16.7 8.9 16.9PSW035B 110 93.5 16.5 11.5 0 5 8.1 7 0.28 10 0 7.9 90 380 0 3 9 11.2 7 7.4 3.5 4 93.5 0 1.17 14.3 190 0 6.9 4 5 6.4 12 27 25 2.5 250 5.5 0.17 3 0 0 6.52 4.35 7.3 25.1 6.34 0 73.5 0.81 0 12.3 93 57.5 6.1 8.37 34 4 2 4.5 6 6.8 6.86 19 1.91 0.14 10 0 0 0 240 0 19.1 29.5 2.02 0.18 3.11 94 0 1.67 10 39 1 130 0.9 1.72 4 6.4 16.5 86 30.5 0 0 13 0.84 5 140 1.8 0 0.26 52 3.52 2 6.1 0 3.6 0 0 12.1 0.44 9 0.48 2.96 7 0 73 0 16 3.18 1.88 17 9 1.86 27.4 5.3 4 4.54 0 0 0 10.06 0.26 7 1.48 8 6.25 0 0.36 3 0 12.19 4.75 12.2 6.33 39 0 0 1.66 0 0 4.91 0.38 5.63 0 19 0.19 5 1.06 0 4.58 4 0 0.13 2.72 1.74 0 0 0.28 9 0 4.63 0 0 2.31 0 0.41 1.19 0.45 0 0.83 0 38 2.11 2.22 15 0 1.93 0 13 0.53 0 1.18 2.48 9.22 1.01 0 2.21 0 0 1.73 0.18 1.38 5.09 0.35 1.27 0 5 16.03 0.22 0 0.64 28 0.85 4.2 5 2.61 0.98 0.5 4.56 7 1.49 0.27 4.24 1.87 1.32 0.14 5.5 0.25 2.34 0.18 1.67 0.42 0.35 0.86 0.39

Pilbara riparian fl ora and plant communities 497

K (exch) me% (exch) K

Na (exch) me% (exch) Na

Mg (exch) me% (exch) Mg

Ca (exch) me% (exch) Ca

P (HCO3) mg/kg (HCO3) P

Gravel 2–4.8mm % 2–4.8mm Gravel

Gravel 4.8–8mm % 4.8–8mm Gravel

Gravel 8–16mm % 8–16mm Gravel

Gravel >16mm % >16mm Gravel

Clay % Clay

Silt % Silt

Sand % Sand

P (total) mg/kg (total) P

OrgC (W/B) % (W/B) OrgC

pH (H2O) pH

mS/m EC (1:5) (1:5) EC Site code PSW036APSW037APSW038A 6PSW039A 2PSW040A 9 6.6PSW041A 5 6.5PSW042A 0.86 39 7.7PSW043A 0.16 8 7.8 280PSW043B 0.08 7.3 17 83PSW044A 56.5 0.16 300 7.4 78 0.21PSW045A 7.7 97.5 17 530 93 8PSW046A 0.27 460 93 32 0.5 0.11PSW047A 8.2 26.5 95 130 840 58.5 0.18PSW048A 1 8.7 220 48 2 0 0.14 8.5PSW049A 53.5 1 13 50 280 55 0.71PSW050A 200 8.4 6 21 15.5 0 2.93 0 86PSW051A 28.5 7.7 4 180 20 9 94 0.59 31PSW052A 240 8.5 0 54 0 0 6 84.5 0 0.57PSW053A 8.3 390 0 280 2 88 36 0 0.44PSW054A 8.4 6 200 0 25 0 0 88.5 8 0.42 0 9.1PSW055A 160 0 7.5 71 0 4 0 79 0.46PSW056A 7.5 5.5 230 9.5 8 0 0 0.45 0 91 0 14 4.5 8 280 49 0 0 0 11 80.5 0 0.36 6 0 92 56 6 0 4.5 0 4.5 6 9.61 81 0 8.6 9 0.44 260 10 0 0 0 52 0 26.5 0 5.22 8.9 4.5 10.3 0.96 5 7.5 210 12 83 0.66 1.72 0 0 10.5 19.5 4 0.27 30 0 0 9.9 0.29 360 2.9 19 0 82 0.25 240 11.5 5.43 0.76 8 0 11.1 0.64 0 120 0 18 26.4 43 0.73 3.8 0.08 0 13 5.52 0 82.5 2.41 9 0.41 0 10 9 0 93 12 0.45 0.1 7 1.74 8 0 16 3.1 1.42 0 0.55 6.66 0 12 9 9 3 0.49 0.81 2 6.63 2.7 3.1 0 0.29 3.79 4.8 41 9.84 9.5 0 0 8.01 0 0.46 2.57 4.02 9.5 0 4.78 1.69 11.1 1.06 9 5 0 0 0 1.25 0 2.13 4 2.16 3.22 0.77 14 0 0.66 0 1.42 2.7 0 0.97 0 11 0 0.9 2.54 0 1.11 4.98 22.4 1.2 5 0.36 0 1.5 6.58 0.31 0 2.5 0 14 0 4 5.74 0.25 1.77 0 0.33 3.02 0 0 0 3.84 0.22 2.94 0.23 0.34 1.36 12 1 0 3.45 0.36 15 10 0.38 5.72 7.28 0.5 9.61 4 6.4 0.49 2.2 0.87 5.21 6.48 3.6 1.02 0.37 1.82 0.43 2.29 1.2 0.33 1.39 0.29

498 M.N. Lyons

K (exch) me% (exch) K

Na (exch) me% (exch) Na

Mg (exch) me% (exch) Mg

Ca (exch) me% (exch) Ca

P (HCO3) mg/kg (HCO3) P

Gravel 2–4.8mm % 2–4.8mm Gravel

Gravel 4.8–8mm % 4.8–8mm Gravel

Gravel 8–16mm % 8–16mm Gravel

Gravel >16mm % >16mm Gravel

Clay % Clay

Silt % Silt

Sand % Sand

P (total) mg/kg (total) P

OrgC (W/B) % (W/B) OrgC

pH (H2O) pH

mS/m EC (1:5) (1:5) EC Site code PSW057APSW058APSW059A 6PSW060A 20PSW061A 38 8.3PSW062A 7.4 100PSW063A 0.07 8.5 31 0.38 6.8PSW064A 32 110 1.27PSW065A 9.2 440 150 4.22PSW066A 97 8.8 140 190 93.5 0.37 8.4PSW067A 320 11 84 1.44 0.5 8.9PSW068A 3 190 24 86.5 0.45PSW069A 7 210 36 2.5 1.37 8 96 8PSW070A 350 8.4 3.5 140 84PSW071A 480 8.6 0 0.38 24 95 1.5 8 0.62 8.5 16.4PSW072A 5.5 37 69 9.5 110 0.45PSW073A 8.4 330 2.5 0 2 22.9 15 1.1 0 0PSW074A 8.3 90.5 180 6.5 13.5 9 18.8 92.5 0.3 8.1PSW075A 6.2 200 0 6 3 5.5 17.5 0 88.5 0 0.21 0PSW076A 19.1 3.5 12 5.2 78 5.4 91.5 0.78PSW077A 0 5.5 270 4 190 0 9.2 0 5 0 0 4PSW078A 0 1.15 7.4 5 300 6 95.5 290 95 6 8.2 0.06 0 0 7 5 410 0 2.47 2 71.5 0.28 0 8.2 0 20.6 0 3.5 35 23 78 1.68 1 5.4 1.58 31.5 14 290 0 22.1 0 5.9 2.85 0.84 0 2.5 0 15 240 2 0 4 7.6 93 0.52 23.5 9.9 4 12.5 57.5 0.91 280 14.5 0.76 0 0 0 82.5 0 9.21 1.17 0.16 45 0 8.01 0.5 19.1 12 6 3.06 12.7 0.32 0 270 85 3.1 10.5 7.64 110 32 12.76 0 0 4 1.32 0 21 6 0.11 6.5 30.5 63 0 5.63 6 7 7.9 0 1.27 0.89 82.5 1.34 7.77 0 0 3 0 2 3.37 5.56 0 2.14 8 0.85 0.17 9.9 0 4 9 0 0.11 0 3.82 2.75 1.38 2.4 0 0 5.26 0 1.5 24.2 0 29 10.03 7 13.5 0.73 23.4 0 0.7 0 1.39 3.32 2 0.46 1.11 0 0 0 0 1 0.33 18.4 0 4.43 0.15 0.94 0 22 0.37 2.01 12.8 0 0 4.02 0 0.49 0 1.51 25 0.12 4.17 0 1.07 10.6 2.72 0.7 10 2 6.5 0 0 3.03 0.42 11 1.02 8 0.3 0.5 5.5 0.25 5.12 1.05 0 7.34 0 0.15 7.81 2.59 0.39 0.31 0.66 6.35 4 22 3.62 0.71 0.89 1.2 1.78 2.42 1.18 4.19 1.12 0.55 0.22 0.41 2.39 0.2 0.25 0.85 0.81 0.31

Pilbara riparian fl ora and plant communities 499

K (exch) me% (exch) K

Na (exch) me% (exch) Na

Mg (exch) me% (exch) Mg

Ca (exch) me% (exch) Ca

P (HCO3) mg/kg (HCO3) P

Gravel 2–4.8mm % 2–4.8mm Gravel

Gravel 4.8–8mm % 4.8–8mm Gravel

Gravel 8–16mm % 8–16mm Gravel

Gravel >16mm % >16mm Gravel

Clay % Clay

Silt % Silt

Sand % Sand

P (total) mg/kg (total) P

OrgC (W/B) % (W/B) OrgC

pH (H2O) pH

mS/m EC (1:5) (1:5) EC Site code PSW079APSW080APSW080B 6PSW081A 4PSW082A 6.4 20PSW083A 110 6PSW084A 1.02 5.3 14 9.3PSW085A 17 0.48 130 0.42PSW086A 7.3 64 1.95 120PSW087A 85.5 6.7 480 160 1.06PSW088A 390 7.6 35 81.5 6 0.24 86 7.5PSW089A 280 14 77 1.09 10.5PSW090A 7.7 190 53 8.5 0.5 7 77.5PSW091A 7.6 260 8 13 190 78.5 1.13PSW092A 9.4 9.5 0 390 19 7 71 0.45 8.2PSW093A 10 5 290 2.4 94 84 0.54 13PSW094A 8.6 120 0 41 1.94 13 80 22.1 0PSW095A 3.2 8.8 180 16.5 8 7 5.8 89.5 0.64PSW096A 220 9.6 14 0 5 16 7.5 0 7.1 91.5 1.72PSW097A 0 4 140 10.1 11 8.5 87 9 0.71PSW098A 10.4 4 26.5 390 12.5 19.4 2 0 6.4 0 6.9 92 0 0.37 8 250 6.5 8 9 17.3 8.1 0 6.2 3 77.5 1.6 4.5 12 8.6 6.6 18 240 3.5 92.5 0 96 0 0 12 1.01 7.7 5 9.6 8 0 400 1.44 24.8 92.5 11 0.21 8.6 3 4.5 4.34 1.17 270 0 16.6 10.5 0 0 29 0.94 80.5 20.2 3.5 0 0.38 190 1.34 0 0.42 8.11 7.6 0.73 60 4.5 7 15.8 0.17 10.5 20.1 180 13 4.62 79.5 0 0.15 4 24.7 120 2.14 0.52 0 3.2 10.5 0 7.4 19.5 0.18 9 19.8 2.32 9.46 46 6.5 0.41 11 0.39 2.58 53.5 0.28 0 11.8 5.9 4 20.5 7 0 14.6 7.91 3.04 10 0.57 14 2.75 14 3.6 16.5 4.93 26.5 0 15 12.3 3 1.02 0.57 14.7 0.77 2.69 37.5 0 6.71 4.18 17.1 0 32.5 16.5 30.2 0.81 6 6 0 0.58 0.9 0 1.22 2.46 3.19 11.8 0 21.6 8 5 0 0.38 0.36 1.22 23.9 2.77 5.53 5.05 6 0 0 0 4.42 3.81 1.06 0 0.66 4 0.26 10.52 2.16 5.53 0 0 8.37 1.95 4.58 0.19 0 0.66 0 4.51 1.98 3.17 3.29 0.31 0.24 4 0 2.12 0 0.2 0.24 0.28 8 0.14 11.48 8 0.27 5 1.59 3.23 0.33 19.53 0.93 0.27 16.12 2.82 4.99 0.09 0.79 0.39 0.89 0.28 0.7 0.65 500 M.N. Lyons °C); TempAnRa – °C); TempAnRa – precipitation seasonality; – precipitation (exch) me% – exchangeable ium milliequivalent percent; ium milliequivalent percent; (W/B) % – percent organic carbon organic (W/B) % – percent c, 3 – near permanent, 4 permanent); ight line distance to coast (km); AnMeTemp AnMeTemp ight line distance to coast (km); sampling Pilbara riparian vegetation. ned from 12345.15.26.16.278 Mean SE Mean SE Mean SE Mean SE Mean SE Mean SE Mean SE Mean SE Mean SE Mean SE Stream order – Strahler stream order (Strahler, 1952); Riverine/non-riverine ( 0 – non-riverine, 1 riverine); DisCoast stra (Strahler, order – Strahler stream order Stream coldest period ( – maximum temperature diurnal range (°C); MaxTeCoP (°C); MedDiuRa – mean temperature – annual mean temperature temperature annual range (°C); MeTemWeQ – mean temperature wettest quarter (°C); AnnPrec – annual precipitation (mm); PrecSeas (mm); PrecSeas – annual precipitation AnnPrec wettest quarter (°C); – mean temperature annual range (°C); MeTemWeQ temperature coldest quarter (mm); EC (1:5) mS/m – electrical conductivity; pH (H20) in distilled water; OrgC – precipitation PrecColQ Ca (exch) me% – exchangeable Calc (HCO3) mg/kg – extractable Phosphorus; P (total) mg/kg – total Phosphorus; Black); P (Walkely K Na (exch) me% – exchangeable Sodium milliequivalent percent; Mg (exch) me% – exchangeable Magnesium milliequivalent percent; Potassium milliequivalent percent. defi variables for the site groups (SE) for species richness and environmental error Mean values ± standard Site Group Species richnessPerm class orderStream 13.8Riverine/lacustrine 2.2DisCoast 22.0 0.4 2.0Climatic variables 1.0 4.6AnMeTemp 0.2 0.2 41.8 0.8 0.2 2.0MedDiuRa 3.0 0.4MaxTeCoP 39.4 0.2 0.0 137.2 0.4 66.2TempAnRa 0.7 2.2 3.6 2.1 25.6 167.2 23.5MeTemWeQ 0.1 0.2 43.9 0.4 0.3 14.4 166.8 7.5AnnPreci 0.1 1.1 25.6 0.1 27.3 0.4 10.0 38.6PrecSeas 119.4 0.1 0.1 0.1 14.0 28.4 0.9 0.5 0.1 47.9PrecColQ 0.0 2.0 25.4 31.0 267.0 0.3 18.4 39.5 0.0 1.4Soil variables 0.2 0.0 0.0 0.0 0.0 14.4 29.0 1.0 0.7 0.0Gravel >16mm (%) 25.6 332.0 292.2 31.3 0.0 2.0 0.2 9.8 39.0 1.2 0.0 60.0 10.2 0.2Gravel 8–16mm (%) 0.3 14.3 97.2 29.1 0.3 143.6 0.0 0.0 313.4 2.6 24.8 0.0 0.0 31.0 18.2 38.2 22.5 0.4 38.6 4.0 0.6 1.0 3.7 26.9 228.5 2.8 292.1 0.0 0.0 0.1 15.6 1.6 91.8 28.1 0.4 0.0 51.1 1.6 9.7 0.0 0.2 24.0 30.9 49.6 0.3 173.6 0.0 40.2 0.0 3.4 2.2 1.2 19.9 292.4 1.0 3.5 0.6 27.7 0.1 2.3 14.9 4.7 6.1 92.2 32.3 0.0 8.7 2.7 0.8 158.8 25.6 0.0 0.3 31.0 44.1 0.0 39.4 2.1 279.0 0.0 39.2 0.9 2.4 1.4 0.1 1.0 3.2 0.0 0.0 14.6 1.8 95.1 31.9 0.4 3.3 2.6 25.3 0.8 263.0 30.4 45.1 0.0 0.2 0.2 39.3 3.1 0.7 0.3 1.3 20.0 0.3 0.3 0.4 1.0 3.4 14.8 4.5 88.0 29.2 0.2 318.6 5.2 0.5 25.3 31.2 42.0 0.3 0.1 40.4 7.1 0.0 0.4 0.0 0.2 0.2 0.6 0.5 0.1 0.0 14.6 0.0 277.5 83.0 31.2 0.3 25.8 7.4 0.0 31.5 38.5 0.2 39.2 5.0 0.2 0.0 0.3 300.0 0.2 14.9 0.8 94.9 29.4 0.3 0.0 0.0 8.0 31.0 45.5 0.2 40.2 0.0 1.8 0.6 281.1 0.0 0.1 1.6 82.0 30.0 0.3 10.6 0.0 8.6 31.6 52.3 1.9 0.7 9.0 1.8 0.2 1.3 92.5 1.6 1.1 43.8 2.4 2.6 1.7 90.2 1.1 50.4 2.4 7.0 3.3 6.3 4.2 2.3 2.0 0.9 1.5 0.6 1.1 Table 3 Table permanence class (1 – ephemeral, 2 seasonal/episodi Codes: Species richness – number of taxa per site; Perm class Inferred Pilbara riparian fl ora and plant communities 501 12345.15.26.16.278 Mean SE Mean SE Mean SE Mean SE Mean SE Mean SE Mean SE Mean SE Mean SE Mean SE Site Group Gravel 4.8–8mm (%)Gravel 2–4.8mm (%) 0.0Sand % 0.0 0.0Silt % 8.1 0.0Clay % 11.7EC (1:5) mS/m 3.1 4.6 1.6pH (H2O) 62.9 3.3 (W/B) %OrgC 1.0 5.0 1.6 5.8 1.0 (total) mg/kg 10.2P 56.1 26.9 2.8 1.4 (HCO3) mg/kgP 1.1 1.0 8.8 4.1 15.8 0.5 2.7 12.4Ca (exch) me% 0.0 6.8 71.6 28.1 0.0 388.0 2.7 3.0Mg (exch) me% 0.2 4.1 0.0 0.2 17.8 6.1 131.9 10.8 23.9 386.0 0.7 0.0Na (exch) me% 81.6 0.0 6.8 17.7 7.5 63.0 1.5 7.3 0.0 3.8K (exch) me% 5.1 251.0 25.8 0.1 2.7 0.0 0.2 5.7 11.3 2.7 48.5 24.9 0.7 0.0 12.6 7.3 10.8 7.2 0.9 142.7 2.7 1.9 7.3 0.7 13.3 10.6 7.4 0.7 4.0 2.2 30.2 0.1 0.1 145.0 23.8 69.6 410.0 3.3 2.3 27.8 1.7 2.5 0.3 0.3 2.5 1.1 7.5 1.8 4.1 20.0 10.3 6.6 7.6 3.1 0.6 472.5 6.5 0.7 267.5 2.3 82.6 0.1 17.9 0.4 0.4 98.4 12.5 27.7 2.7 6.2 2.7 4.5 0.1 0.4 1.9 1.4 7.7 1.0 7.0 114.3 254.3 3.3 9.4 29.0 0.6 23.9 73.3 15.5 2.3 0.3 9.1 1.4 0.0 0.3 0.1 8.3 196.0 165.0 2.5 1.4 2.3 9.6 1.8 0.1 0.4 0.8 8.1 54.1 0.8 1.2 37.3 11.5 0.9 2.9 85.7 69.0 245.7 0.8 6.3 0.5 17.3 1.4 0.1 0.1 0.1 9.5 2.4 22.8 2.0 14.7 2.6 2.3 6.1 0.2 1.2 222.2 0.4 8.5 0.4 5.8 37.3 79.1 3.5 6.5 29.3 2.8 10.5 0.3 10.2 0.2 0.1 0.1 6.3 7.7 0.8 3.0 1.3 0.9 0.3 1.2 3.0 8.7 7.5 1.4 6.2 9.1 1.8 1.1 0.6 0.3 0.2 11.8 2.4 6.4 0.6 2.2 0.5 0.9 2.9 8.6 9.4 4.2 3.2 2.5 0.8 0.1 0.9 0.2 1.9 7.4 1.1 0.5 0.7 0.8 7.8 10.7 4.8 3.8 2.0 2.3 0.2 0.2 0.3 3.1 0.4 1.4 1.3 5.9 2.2 0.5 0.5 2.8 0.9 0.6 0.4 0.6 0.6 0.1 0.2 0.4 0.1 502 M.N. Lyons

Data analysis PATN (v3.12) to classify sites according to Species presence/absence data were compiled similarities in their species composition (Bray- for sites by combining seasonal samples. Site Curtis measure of dissimilarity), and species resemblance matrices were generated using according to their co-occurrence at sites (the two- the Bray-Curtis similarity measure in Primer step algorithm, Belbin 1980). Both classifi cations (Primer-E Ltd 2008). The infl uence of singleton used the fl exible UPGMA (unweighted pair-group species (species with a single site occurrence) was mean average) sorting strategy with the beta assessed by comparing resemblance matrices, value set to -0.1 (Belbin 1995; Sneath and Sokal including and excluding singletons, using Primer’s 1973). Indicator species showing high fi delity and/ RELATE routine (999 permutations). Analysis of or specifi city to site groups were identifi ed using similarity (anosim) routines (Clarke and Green, the ‘Indicspecies package’ in the R statistical 1998) were used to examine the signifi cance of computing language (De Caceres and Legendre relationships between the floristic composition 2009, R development core team 2009). Monte Carlo of sites with a priori wetland types, IBRA sub- randomisation was used to test the signifi cance region, permanence, riverine versus lacustrine of taxa indicative of both individual and multiple waterbodies, and catchment basin occurrence. groups of sites (P <0.05). For interpretation, categorical variables were overlain on non-metric multidimensional scaling RESULTS ordinations of sites (nMDS, 50 starts; Primer-E Ltd. 2008). Two-dimensional ordinations were used as Sampling adequacy they showed similar stress to 3D outputs. Sites included a range of species richness scores The BEST/BIOENV procedure in PRIMER (12–52 taxa) and wetland types. Comparison of was used to explore the relationship between the observed richness for each site and estimators site compositional similarity and environmental of total richness (Figure 2 – four examples shown) attributes. This module uses Spearman’s rank– revealed that the quadrat design employed had order correlation to match distances in a site captured, on average, 89% (79%–100%) of the total association matrix to Euclidean distances among richness in the sites (ICE estimator, see Colwell, each of its environmental attributes. Continuous 2009). variables used in the analysis included chemical and textural attributes of the soil, BIOCLIM climatic Riparian plant diversity variables, distance to coast and altitude. Categorical A total of 455 taxa was identifi ed from the 98 variables included ordinal permanence class, wetlands and rivers sampled during the survey Strahler stream order and the two-state categorical (Table 1, Appendix 2). These taxa were distributed variable riverine/non riverine. For pairs of variables among 62 families and 202 genera. Dominant plant that were highly intercorrelated (P > 0.9), a single families were Poaceae (18.5% of taxa), Fabaceae variable was retained. In the fi nal data set, distance (11.4%), Asteraceae (10.8%), and Cyperaceae (8.1%). to coast was retained in preference to annual range Dominant genera included Acacia (4.4%), Eragrostis of temperature, and percentage clay was retained (2.9%), Cyperus (2.9%), Ptilotus (2.0%), Eriachne (2.0%) in preference to percentage sand. Highly skewed and Fimbristylis (2.0%). variables were log-transformed and the resulting The taxa represent approximately 25% of the variable matrix normalised. known fl ora of the Pilbara bioregion (S. Dillon Some sites were excluded from the analysis unpublished data), and are dominated by species of environmental correlates. Soil data were not that occur elsewhere in the Eremaean (398 taxa) available for PSW004B, and PSW003A was an and Northern (295 taxa) Botanical Provinces (after extreme compositional outlier. Six saline sites Beard 1990). Pilbara populations of twenty taxa (e.g. sampling the Fortescue Mar sh and Weelarana Salt Atalaya hemiglauca, Lobelia arnhemiaca and Terminalia Lake were also excluded from the BEST/BIOENV canescens) occur in the Pilbara as populations procedure since they were compositionally and disjunct from the Kimberley. A group of 143 taxa environmentally very distinct from the remainder extend into the South-West Botanical Province. of the data set. Of these, the majority occur at the periphery of this province, i.e. within the Avon-Wheatbelt (105 Site and species classifi cation taxa) and Geraldton Sandplains (also 105 taxa) The data matrix of species presence/absence bioregions. at sites (excluding singletons) was re-ordered Of those taxa with more restricted distributions to produce a two-way table for interpretation (known from six or fewer IBRA regions – 88 taxa), (Appendix 1) (Belbin 1995) by using the program only the cosmopolitan Pteris vittata occurs in the Pilbara riparian fl ora and plant communities 503

Figure 2 Expected species richness accumulation curves based on the Mao Tau function for four sites with a range of total observed richness. Error bars equal one standard deviation. temperate south-west of Western Australia. Taxa Koondepindawarrina Pool (PSW005). Tecticornia known from four or fewer bioregions (64 taxa) medusa and T. globulifera, formerly known only largely extend into adjacent bioregions. Of these, from the Fortescue Marsh, were also recorded from the Carnarvon (29 taxa) and Gascoyne (29 taxa) Weelaranna Salt Lake (PSW043). Eleocharis papillosa, share a greater number of taxa with the Pilbara previously known from scattered localities across than Dampierland (10 taxa) and the Murchison central Australia and near Onslow at the edge of (nine taxa), with a small component occurring in the Pilbara, was recorded from two sites on the the adjacent Great Sandy Desert (seven taxa) and Fortescue Marsh (PSW002, PSW003). The new Little Sandy Desert (nine taxa) bioregions. record of Myriocephalus scalpellus from near Onslow Naturalised taxa represented 6.6 % of the fl ora, (PSW073) represents the only known population noting that the weed status of some species (e.g. other than the type locality at Coondiner Pool Flaveria trimera) remains contentious (see Keighery (PSW001). Eremophila spongiocarpa, endemic in the 2010). The most frequently recorded weeds included Pilbara and restricted to the Fortescue Marsh, was Cenchrus ciliaris (46% of sites), Cynodon dactylon recorded within its known range (PSW003). Other (22%), and Cenchrus setiger (17%). Notable weed notable priority taxa include Cladium procerum records included Conyza parva collected from which occurs in the Pilbara as a major outlier from Chinaman Spring (PSW060), on the northern edge of the study area, a major range extension from populations in eastern Australia. Known from near-coastal areas of south-west Western Australia. both Millstream and Karijini National Park, it was Previously known anecdotally (Keighery 2010), recorded at a known locality (at PSW006). Also Gnaphalium polycaulon is confi rmed as occurring recorded from a known population at Fortescue in the Pilbara, with a record from Coondiner Pool Falls, Adiantum capillus-veneris has a scattering of (PSW001). localised populations across much of Australia. Sixteen taxa on the Department of Parks Eragrostis surreyana is endemic in the Pilbara IBRA and Wildlife’s Priority Flora List (Smith 2013) region, and was recorded from a known population were recorded (see Appendix 2). A number are on the Burrup Peninsula. Although not listed as associated with wetlands within and bordering the priority taxa, the Pilbara endemics Bergia perennis Fortescue Marsh Land System as defi ned by Van subsp. perennis, Eriachne tenuiculmis, Sida arsiniata Vreeswyk et al. (2004). Both Helichrysum oligochaetum and Stemodia kingii were also recorded during the and Rhodanthe ascendens were recorded from survey. 504 M.N. Lyons

Major range extensions and permanent sites the lowest (20.8 ± 2.4). This is The grasses Imperata cylindrica and Pseudoraphis related largely to the relative proportion of different spinescens have distributions that extend from wetland types within each class. Ephemeral sites the Kimberley region across northern Australia included a number of rock pools (PSW071A, and down the eastern Australian coast to South PSW080A) with herb-rich aprons, and claypa ns Australia, with numerous inland occurrences. and clay fl ats with rich assemblages of annual taxa. Not previously known from the Pilbara, Imperata Permanent sites included river pools (15 sites), gorge cylindrica occurred at three sites: Fortescue Falls creeks and the majority of species-poor springs (13 (PSW006), Palm Spring on Cave Creek (PSW009) of 15). Seasonal and near-permanent sites showed and Millstream Delta (PSW011). Prior to the current little difference in richness and were dominated by survey, the only Pilbara record of Pseudoraphis river pools. Little difference in richness was evident spinescens was a collection by A. Forrest on the for riverine sites in relation to Strahler stream order; Fortescue River in 1878 (Mueller 1881a, 1881b). with only order-one versus order-three showing During the survey it was collected at four sites: signifi cant pairwise differences (P < 0.05). Moreton Pool (PSW030), Munreemya Billabong Richness of rare taxa, including singletons and (PSW036), Curara Claypan (PSW072), and West species occurring at fewer than six sites, was Peawah Creek Pool (PSW079). Nymphoides indica, examined for geographic patterns (Figure 3). recorded at De Grey claypan (PSW035), is the fi rst Concentration of rare taxa was evident within collection for the Pilbara bioregion. Its known the Fortescue Marsh and on the periphery of distribution includes much of the world’s tropics the study area at a small number of sites which and extends from the Kimberley across the north included Weelarrana Salt Marsh (PSW043B), and down the east coast of Australia. claypans, granite rocks and a single seasonal creek (PSW085) at the southern edge of the region. Some Species frequency and richness concentration of rare taxa along the northern edge Of the 450 taxa analysed, two-thirds (66.4%) of the Hamersley IBRA subregion is associated occurred at fewer than fi ve sites (see Appendix with rarely recorded taxa from the gorges and 1), and one-third (33.2%) occurred at a single upland ephemeral creeks of the Hamersley Range. site (Appendix 3). Three taxa, Cyperus vaginatus Notably, few singleton and rare taxa were observed (perennial sedge), Eragrostis tenellula (annual grass) in the Chichester IBRA subregion (Figure 3). High and Centipeda minima (annual herb), occurred numbers of rare taxa were associated with both at more than half the sites. Species richness wetland-type and water permanence. Overall, ranged from three (PSW032) to 87 taxa per site ephemeral and seasonal wetland types such as salt (PSW085) and averaged 26.3 ± 1.4. Richness showed marshes, rock pools and claypans showed highest significant differences between wetland types numbers of rare taxa, and their non-uniform (Kruskal-Wallace 1-way ANOVA, P < 0.005), with distribution across the study area explains much of gorge creeks (20.0 ± 2), springs (18.3 ± 1.4) and salt the observed geographical patterning. marshes (14.5 ± 3.5), showing the lowest richness. Gorge creeks were characterised by massive rock Site classifi cation and provided limited habitats for annual taxa, Comparison of the full and reduced (singletons instead being dominated by dense perennial sedges excluded) species presence/absence matrices and shrubs in soil pockets that are perennially revealed that the two were highly correlated (Rho moist. Springs were commonly dominated by dense = 0.99, p < 0.001). In preference, the reduced matrix patches of perennial sedges (typically Cyperus was used. vaginatus), often interspersed with bare areas of The primary division in the classification calcrete. Salt marsh sites had a limited suite of largely separated non-turbid riverine sites from halophytic taxa. predominantly non-riverine wetlands and turbid Rock pools (42.5 ± 2.5), clay flats (41.0 ± 6), riverine sites with fi ne-textured soils. Within the clay pans (29.7 ± 3.2) and ephemeral creeks (28.7 last major group, the small number of saline sites ± 4.4) had the highest mean richness with rich formed a distinct group at the next division in the assemblages of herbs and annual and perennial classifi cation. The site classifi cation was ultimately grasses. examined at the 8-group level (Figure 4). The ten- Richness also showed significant differences group classification and a priori wetland types between inferred permanence classes (Kruskal- overlain on 2D nMDS ordinations (Figure 5 A Wallace 1-way ANOVA, P < 0.005). Ephemeral sites and B), revealed a strong correspondence between had the highest mean species richness (39.2 ± 3.6) fl oristic groupings and wetland type. Pilbara riparian fl ora and plant communities 505

Figure 3 Number of rare taxa (single site record) or uncommon taxa (≤ 5 site records) recorded per site, with symbol-size proportional to richness. (A) Richness within sites containing rare taxa (maximum 15 taxa). (B) Richness within sites containing rare taxa as a proportion of total site richness (maximum 0.27). (C) Richness within sites containing uncommon taxa (maximum 31). (D) Richness within sites containing uncommon taxa as a proportion of total site richness (0.87). Sites with no rare taxa are solid dots.

Dissimilarity Large claypans 1 & turbid creek pools with broad margins 2 Claypans, clay flats, rock pools 3 and riverine sites with fine 4 textured soils

5.1 Salt marshes 5.2 Springs & river pools 6.1 with high water 6.2 permanence

River 7 pools & creeks 8 0.86 1.13 1.40

Figure 4 Classifi cation of 103 survey sites (UPGMA) based on species presence/absence. Classifi cation is shown at the eight group level with descriptions summarising divisions in the classifi cation. 506 M.N. Lyons

Plate 4 Large seasonal/episodic claypan on Mulga Downs Station (site PSW040A). Stands of Acacia stenophylla occur across the inundated bed and within the riparian zone.

Large claypans and turbid linear pools with broad margins Group 2. A group of fi ve sites at four wetlands including three claypans and the margin of a linear, Group 1. A group of fi ve claypan sites at four turbid pool. This group also featured Eucalyptus wetlands including large claypans and a single victrix as a structural dominant over a relatively linear turbid claypan (Coondiner Pool), all typifi ed species-rich herb layer. Signifi cant indicator species by very broad, low-gradient margins. Structural included Sporobolus mitchellii, Bergia trimera, B. dominants within the group included scattered perennis, Myriocephalus oldfieldii, M. rudallii and Eucalyptus victrix over Eriachne benthamii with Muehlenbeckia fl orulenta. The sites showed similar Marsilea spp. forming dense swards. Signifi cant soil chemistry attributes to group 1 (Table 2). indicator species included Schoenoplectus laevis, Elytrophorus spicatus, Eriocaulon cinereum, Claypans, clay fl ats, rock pools, and riverine sites with Nymphoides indica, Glossostigma diandrum and Rotala fi ne-textured soils diandra. Sites sampled showed a high percentage of Group 3. A large group of 23 sites with the clay content (26.9 ± 4.1) and the lowest EC (5.8 ± 1.1) highest species richness of all site groups (41.8 and pH (6.8 ± 1.1) of all site groups. These sites, with taxa per site). Sites sampled the riparian zones fi ne-textured soils, also had high total phosphorus of linear turbid creeks, river pools, ephemeral (388.0 ± 131.9 mg/kg). The few wetlands in the creeks, claypans and clay fl ats (Table 1). Although group were scattered from the north-west coastal including a number of wetland morphologies, margin to the inland eastern edge of the study area sites within the group shared fi ne-textured soils in and had the lowest species richness of groups 1–4 both riverine and non-riverine settings (mean silt (13.8 taxa per site). 10.8 ± 1.5 % and clay 17.7± 2.7%). Seventeen sites Pilbara riparian fl ora and plant communities 507 included Eucalyptus victrix or E. camaldulensis subsp. Springs and river pools with high water permanence refulgens as overstorey trees within the riparian Group 6.1. A subgroup of 23 sites from river zone. Analysis revealed a single taxon, Fimbristylis pools and 11 of the 15 springs sampled. Sites littoralis, as the signifi cant indicator species for the had a mean species richness of 18.4 taxa, the group. The group shared a number of taxa with group 4 (see below) with Fimbristylis microcarya, signifi cant indicator species being Schoenoplectus Wahlenbergia tumidifructa, Lipocarpha microcephala, subulatus and Fimbristylis sieberiana. Structural Oldenlandia galioides and Drosera indica being dominants included Melaleuca argentea, Eucalyptus signifi cant indicator species for both site groups camaldulensis subsp. refulgens, Acacia ampliceps, taken together. Typha domingensis and Cyperus vaginatus. The group Group 4. A group of seven sites on the margins also included all the sampled springs and gorges of claypans and the aprons of rock pools. As a within Karijini and Millstream National Parks. group, the sites were close to the coast relative The group had the highest mean permanence score to all other groups (119.4 ± 47.7 km). Claypans (3.7) of all site groups. This was coupled with the in the group were small, highly ephemeral, and highest mean annual rainfall (318.6 mm/yr). The occurred within Triodia-dominated fl ats and dune soils of the sites contained the highest percentage systems. The edges of pans were low sandy rises gravel (all fractions) and were otherwise typically with a narrow band of annual taxa recruiting after sandy. Exchangeable calcium was the highest of rainfall and wetland-fi lling. The group showed all groups (6.2 me%) and is likely to refl ect the similarly high species richness to group 3 (39.4 taxa contribution of calcium-rich groundwater to the per site). Sixteen herbs and annual grasses were riparian zone substrate. signifi cant indicator species for the group with Synaptantha tillaeacea var. tillaeacea, Eriachne aristidea, Bulbostylis barbata, Calandrinia ptychosperma, Cyperus pulchellus and Gomphrena leptoclada subsp. leptoclada signifi cant at P < 0.05.

Salt marshes Group 5.1. A subgroup of two sites within the upper fringe of the Fortescue Marsh, characterised by a group of salt-tolerant indicator species including Cressa australis, Muellerolimon salicorniaceum, Cyperus bulbosus and Eleocharis papillosa. The sites showed high EC (145 ± 2.5) and phosphorus (total P, 410 ± 20) and high clay content (27.8 ± 3.1). The current survey, however, did not attempt to capture the full fl oristic variation evident within and bordering the Fortescue Marsh. These two sites sample an extensive zone within the Fortescue Marsh that is dominated structurally by Muellerolimon salicorniaceum. Group 5.2. A subgroup comprising four sites at two wetlands, sampling the edges of the main basins of the Fortescue Marsh and Weelarrana Salt Lake. They were characterised by low shrublands dominated by Tecticornia spp. Signifi cant indicator species for the subgroup included Tecticornia halocnemoides sens. lat., T. indica, T. medusa, T. globulifera and Triglochin hexagona. The sites showed the highest soil EC (472.5 ± 98.4) of all groups, and Plate 5 Areas of bedrock and calcrete dominate the high total P (267.5 ±27.7 mg/kg). Species richness margins of Palm Spring (site PSW013A). Melaleuca glomerata, juvenile Livistona alfredii was low with 10 taxa per site. The sites are likely to (Millstream fan-palm) and sedges occur in experience signifi cant periods of inundation during the shallow soil pockets along the spring-fed major fl ooding events. creek (M.N. Lyons). 508 M.N. Lyons

Group 6.2. A subgroup of four species-poor sites group. Eucalyptus camaldulensis subsp. refulgens and (9.8 taxa per site) comprising two river pools and Cyperus vaginatus are structural dominants. The two small springs. Structural dominants included separation of groups 6.1 and 7 is due largely to the Acacia ampliceps and Cyperus vaginatus. Sites showed very low frequency of taxa dependent on high/ significant cattle disturbance or flood damage, permanent moisture, such as Melaleuca argentea, which probably accounted for their low species Schoenoplectus subulatus and Typha domingensis richness scores. Stylobasium spathulatum and the within group 7, and the presence of a group of naturalised taxon Bidens bipinnata were signifi cant widespread annual taxa including Eragrostis indicator species for the group. tenellula, Wahlenbergia tumidifructa, and Centipeda minima that are largely absent from group 6.1. Soil River pools and creeks texture attributes of groups 6.1 and 7 were very Group 7. A group of 21 riverine sites including similar (Table 2). 15 clear river pools, four ephemeral creeks and Group 8. A group of nine sites on permanent and two springs. The group shows compositional near-permanent higher-order river pools structurally similarities to subgroup 6.1 yet showed higher dominated by Melaleuca argentea, Eucalyptus mean species richness (26.9 cf. 18.4 taxa per site). camaldulensis subsp. refulgens, Cyperus vaginatus and The group had a mean permanence score of 3.1 (cf. Cynodon dactylon. Average species richness was 3.7 for group 6.1), refl ecting the small number of lower than the other group dominated by river pools springs and inclusion of ephemeral creeks within (19.9 taxa per site, cf. 26.9 in group 7). Soils within the the group. Indicator species analysis did not reveal group were more fi nely textured than groups 6.1 and any taxon to be individually signifi cant for the 7, with lower gravel fractions (Table 2).

Plate 6 Pool in ephemeral creek (site PSW062A) at the base of the Gregory Range. The riparian zone is dominated by Eucalyptus victrix and Cyperus vaginatus, with Schoenoplectus subulatus in the waterbody (N. Gibson). Pilbara riparian fl ora and plant communities 509

Figure 5 Two dimensional. non-metric multi-dimensional scaling ordination (2D nMDS) of 103 sites based on presence / absence of plant species overlain with (A) fl oristic groups and (B) a priori wetland types. Also, 2D nMDS ordination excluding salt marshes (fl oristic subgroups 5.1 and 5.2) with sites coded (C) as riverine/non-riverine and (D) according to permanence class (1, ephemeral; 2, seasonal episodic; 3, near permanent; 4, permanent).

Environmental variables and fl oristic The BEST procedure revealed that a combination composition of fi ve variables [riverine/non-riverine, DisCoast, Significant compositional differences were clay%, EC (1:5), pH] provided the highest rank observed for riverine versus non-riverine sites correlation with the site dissimilarity matrix (Figure 5C) (Rho = 0.525), and permanence class (Spearman’s rho = 0.53, P < 0.01). The inclusion (Figure 5D) (Rho = 0.383). Within riverine sites, of the categorical variable riverine/non-riverine excluding the very small number of ephemeral confirms the major compositional separation of sites, permanent sites were compositionally these two major wetland classes. The soil variables different from seasonal and near-permanent sites relate to a number of compositional gradients. High (Rho = 0.25), although there was no signifi cant clay content corresponds to the separation of clear difference between stream orders (Rho = 0.022). riverine sites with relatively coarse sediments, from Geographic patterns were examined by testing claypans and turbid riverine sites typifi ed by fi nely compositional differences against IBRA subregions textured soils. The latter were compositionally and major drainage basins. IBRA sub-regions distinct with relatively rich assemblages of annual showed only moderate compositional differences taxa. The inclusion of soil electrical conductivity and (Rho = 0.241), with pairwise differences between pH also correlates with the compositional separation both the Fortescue and Roebourne Plain against of riverine and non-riverine sites. The inclusion of all other sub-regions accounting for the bulk of DisCoast refl ects the greater occurrence of a number the patterning. Little support for compositional of compositionally distinct groups along the subdued differences between drainage basins was observed landscapes of the coastal plains of the study area, (Rho = 0.073). including claypans and turbid creek pools. 510 M.N. Lyons

DISCUSSION drainage with few major riverine habitats. Where these occur they are commonly broad, sometimes Flora and diversity saline, drainage lines flowing into basins, or Prior to the current study, the only regional highly ephemeral creek systems arising in isolated survey of the Pilbara riparian fl ora was undertaken ranges. The saline coastal areas of the Pilbara and by Masini (1998) and Masini and Walker (1989) as the Fortescue Marsh share taxa with these desert two separate studies sampling subsections of the systems (e.g. Tecticornia auriculata). The deserts, like region. Although interrupted by cyclonic activity the Pilbara, contain a suite of claypans and clay fl ats and subsequent sampling late in the dry season, that are episodically inundated and elements of the authors recorded ca. 300 taxa. Many localised their fl ora are shared. surveys assessing impacts of proposed mining The proportion of uncommon taxa recorded developments have also been completed (e.g. Biota for the Pilbara – singletons (33.2%) and those Environmental Sciences, 2012). Other studies have from fewer than fi ve sites (66.4%) – is of similar been undertaken to assess the impacts on Pilbara magnitude to other similar studies. Gibson et al. wetlands of water abstraction and mine dewatering (2000) recorded that 55% of taxa occurred at a single and its disposal (e.g. Johnson and Wright, 2001; site (singletons) in the southern Carnarvon Basin. Braimbridge et al. 2010). The only other regional- A more comprehensive survey of temperate zone scale wetland floristic survey undertaken in wetlands in Western Australia showed very similar Western Australia’s arid zone was completed by proportions (31% and 59.3% respectively) to the Gibson et al. (2000), in the southern half of the current study (Lyons et al. 2004). The occurrence of Carnarvon Basin bioregion to the south of the this uncommon component of the fl ora (at similar Pilbara. They recorded 263 taxa from 57 wetlands levels) is also a consistent feature of quadrat-based (albeit with smaller quadrat size) representing floristic surveys carried out across the wider ca.12% of the region’s fl ora (Keighery et al. 2000). Western Australian landscape (Gibson et al. 2004; Gibson et al. (2000) observed that the wetland fl ora Keighery et al. 2000; Lyons et al. 2014), and poses of the southern Carnarvon Basin, while dominated problems for designing reserve systems that aim to by taxa of the arid zone (Eremaean), had affi nities capture major fl oristic communities. with both temperate and tropical fl oras. In part, Species richness of sites was related to both this is true of the Pilbara, although the temperate the a priori wetland types and the site groups component of the flora is smaller and largely derived from the classifi cation based on fl oristic composed of very widespread taxa rather than those composition. Maximum richness occurred in less with temperate distributions for which the Pilbara permanent wetlands, with concomitant higher is their northern range limit. The Pilbara riparian richness in non-riverine sites. A similar pattern flora also includes a larger component of the was observed for the proportion of uncommon taxa tropical fl ora, from both the rocky landscapes of the across the range of wetland types. Coarsely textured northern Kimberley and the more topographically soils that limit moisture retention, as well as major subdued central Kimberley, Ord–Victoria Plain and disturbance associated with the high energy/ Victoria–Bonaparte bioregions. A subset of these volume fl ows following fl ood events, are likely to does not occur elsewhere in the Eremaean areas of limit the suite of taxa that can persist along many Western Australia (e.g. Atalaya hemiglauca). Notable riverine riparian zones in the Pilbara. The plant elements of the tropical riparian fl ora are also near communities of these larger riverine sites are often their southern limit in the Pilbara (e.g. Melaleuca simple, with groundwater-dependent overstorey argentea and Sesbania formosa). trees (Melaleuca argentea, Eucalyptus camaldulensis While sharing broadly similar climates (Leighton, subsp. refulgens) (O’Grady 2006) and simple 2004; Wyrwoll et al. 2000), the strong affi nities of understoreys of shrubs, rhizomatous grasses and the Pilbara fl ora with the adjoining Carnarvon and sedges that are more resistant to fl ood waters (e.g. Gascoyne bioregions are also attributable to the Melaleuca linophylla, Cyperus vaginatus). In contrast, regions possessing a similar suite of wetland types. more ephemeral sites, those with lower fl ow-energy All share a combination of riverine and non-riverine (e.g. turbid linear creeks), as well as claypans, clay water bodies that represent similar hydrological flats and rock pools, experience much-reduced settings. Indeed some of the major structural fl ood disturbance and their soils are more fi nely dominants in riverine riparian zones are common textured. In many of these shallow wetlands, the to the three regions (e.g. Eucalyptus camaldulensis wet/dry ecotone is broad, providing habitats for subsp. refulgens and Acacia coriacea subsp. pendens). a larger range of taxa with differing responses to In contrast, the sandy deserts to the east and north- wetting and drying (Brock and Casanova 1997). The east of the Pilbara are characterised by internal development of species-rich seed banks that persist Pilbara riparian fl ora and plant communities 511 for long periods between fi ll events has also been similarities to claypans. Turbid sites are a feature shown to contribute to the richness of these sites of creeks and smaller rivers in Pilbara lowlands. (Brock et al. 2006). Like claypans, they have fi ne soils, have generally The extensive aquifer systems of the Pilbara short hydroperiods and are not subject to major (Waters and Rivers Commission 1996) discharge fl ood fl ows. This is refl ected in the analysis with groundwater to form permanently fl owing creeks, soil texture emerging as a signifi cant term in the typically in rocky sites, or contribute flows to BEST analysis. Masini (1988) also highlighted permanent pools within river channels. Springs the importance of sediment texture and water form a group that is relatively poor in species (18.3 permanence in structuring Pilbara wetland plant taxa per site) in the Pilbara, and are characterised by communities, finding strong correspondence taxa requiring permanently damp or wet soil that between morphological/hydrological (subjective) physically dominated sites, e.g. Cyperus vaginatus, and floristic (objective) classification of sites. Melaleuca argentea, Typha domingensis. Studies of Importantly, both the current study and that of springs elsewhere in Australia confi rm that springs Masini (1988) are very site-specifi c. Dryland rivers of the Pilbara are fl oristically distinct from the show considerable heterogeneity in terms of remainder of the continent. In the Kimberley and channel morphology and sediment attributes along their length, particularly where over-bank fl ows tropical Northern Territory, rainforest floristic fi ll secondary channels (Thoms et al. 2006). As a elements are a major component of spring fl oras consequence, even a small stretch of river (< 5 km) (Kenneally et al. 1991; Russell-Smith 1991). Detailed may capture a number of fl oristic elements. studies of the springs of the Great Artesian Basin (Fensham and Fairfax 2003; Fensham and Price Biogeographic patterns in the riparian fl ora across 2004), while very different fl oristically from the the Pilbara are explained by gross differences in Pilbara, highlight the importance of springs in the the distribution of major wetland types across its conservation of endemic and relictual taxa. In the four IBRA subregions. This is not surprising, given Pilbara, springs and spring-fed river pools and that, relative to the broader landscape, riparian associated gorge habitats provide important mesic zones are somewhat buffered from the dominant refugia for taxa such as Livistona alfredii, Stylidium climatic gradient that occurs from west to east weeliwolli and Fimbristylis sieberiana, including major in the region, particularly in terms of moisture distributional outliers such as Cladium procerum, availability (Leighton, 2004). In particular, at the Phragmites karka, Imperata cylindrica and Adiantum regional scale, spring sites are largely decoupled capillus-veneris. In arid Central Australia, permanent from the prevailing climate. The topographically water bodies, including springs and seepages, have subdued, largely depositional landscapes of the also been shown to be important refugia for relict Fortescue and Roebourne Plains IBRA sub-regions and endemic species (White et al. 2000; Brim Box et show signifi cant differences from the remaining al. 2008). sub-regions.

Floristic composition Conclusions Patterns in the fl oristic composition of Pilbara The Pilbara riparian fl ora represents 25% of the wetlands and rivers correspond to major Pilbara fl ora, and is dominated by Eremaean and hydrological and substrate attributes of the sites. tropical taxa, with strongest affi nities to adjoining These environmental differences are largely externally draining bioregions. embodied by the initial a priori site groups that Biogeographic patterning is limited, with infl uenced the sampling design and are refl ected in sub-regional differences being related to the the site groups I derived. relative proportion of wetland types within each The vegetation of the Fortescue Marsh is subregion. Riverine sites, in particular, show compositionally distinct from the remainder of the limited compositional difference across the region. site groups, even though sampling of the Marsh The Fortescue valley, coastal plains and other was limited relative to its complexity and size. lowland areas make important contributions to the While I did not sample tidal environments along diversity of non-riverine wetlands. Claypans and the Pilbara coast that are likely to share some taxa clay fl ats capture a large component of the Pilbara with the Marsh (e.g. Tecticornia indica) it contains a wetland fl ora, and their scattered occurrence across fl oristically unique element of the Pilbara wetland the lowlands of the region poses diffi culties in fl ora. capturing their diversity in the reserve system. Among the remaining rivers and wetlands a Within the region, the Fortescue Marsh, along with major compositional separation is seen between spring sites in Karijini and Millstream National riverine and non-riverine (wetland) sites. Turbid Parks, support unique elements of the Pilbara creek pools, while riverine, show clear fl oristic riparian fl ora. 512 M.N. Lyons

ACKNOWLEDGEMENTS Colwell, R.K. (2009). EstimateS: Statistical estimation of species richness and shared species from samples. Version I thank David Mickle for his unwavering 8.2. User's Guide and application, published at assistance and companionship in the field and http://purl.oclc.org/estimates offi ce. Field work was assisted by Neil Gibson, De Caceres, M. and Legendre, P. (2009). Associations Natalia Huang and Simon Lyons. Adrian Pinder led between species and groups of sites: indices and statistical inference. Ecology 90: 3566–3574. the site selection with valuable initial input from Fensham, R.J. and Fairfax, R.J. (2003). Spring wetlands Peter Kendrick and Stephen van Leeuwen. Pilbara of the Great Artesian Basin, Queensland, Australia. DEC staff provided welcome accommodation and Wetlands Ecology and Management 11: 343–362. delivered much-needed parts for broken field Fensham, R.J., Fairfax, R.J. and Sharpe, P.R. (2004). equipment. Numerous pastoralists and indigenous Spring wetlands in seasonally arid Queensland, communities provided local knowledge, hospitality environmental relations, classification and conservation values. Australian Journal of Botany 52: and access to their lands. Adrian Pinder drafted 583–595. Figure 1. Steve Dillon provided his Pilbara species Geological Survey of Western Australia (1990). Geology list and Ben Richardson provided data on the and mineral resources of Western Australia. Memoir species IBRA occurrences. Lesley Gibson assisted 3. Western Australian Geological Survey: Perth, with indicator species analysis. Norm McKenzie Australia. along with two anonymous referees provided Gibson, N.G., Keighery, G.J. and Lyons, M.N. (2000). The valuable comments on the draft manuscript. fl ora and vegetation of the seasonal and perennial wetlands of the southern Carnarvon Basin, Western Australia. Records of the Western Australian Museum, REFERENCES Supplement 61: 175–216. Gibson, N., Keighery, G.J., Lyons, M.N. and Webb, A. Anderson, M.J., Gorley, R.N. and Clarke, K.R. (2008). (2004). Terrestrial fl ora and vegetation of the Western Plymouth: Primer-E; 2008. PERMANOVA+ for Australian wheatbelt. Records of the Western Australian PRIMER: Guide to software and statistical methods. Museum, Supplement 67: 139–189. Beard, J.S. (1990). Plant life of Western Australia. Kangaroo Houlder, D.J., Hutchinson, M.F., Nix, H.A. and McMahon, Press: Kenthurst, Australia. J.P. (2000). ANUCLIM User Guide, Version 5.1. Centre Belbin, L. (1980). TWOSTEP: A program incorporating for Resource and Environmental Studies, Australian asymmetric comparisons that uses two steps to produce National University: Canberra, Australia. a dissimilarity matrix. Technical Memorandum 80/9, Johnson, S.L. and Wright, A.H. (2001). Central Pilbara CSIRO Division of Land Use Research: Canberra, groundwater study. Hydrological Record Series, Report Australia. HG 8. Waters and Rivers Commission: Perth, Belbin, L. (1995). PATN technical reference. CSIRO Division Australia. of Wildlife and Ecology: Canberra, Australia. Keighery, G.J. (2010). The naturalised vascular plants of Biota Environmental Sciences (2012). A vegetation and fl ora the Pilbara region, Western Australia. Records of the survey of the Koodaideri study area. Unpublished report Western Australian Museum, Supplement 78: 299–311. prepared for Rio Tinto: Perth, Australia. (online: Keighery, G.J., Gibson, N., Lyons M.N. and Burbidge, http://www.epa.wa.gov.au/EIA/referralofProp- A.H. (2000). Flora and vegetation of the southern schemes/Lists/Proposal/Attachments/178/ Carnarvon Basin. Records of the Western Australian Appendix%204_A%20Vegetation%20and%20 Museum, Supplement 61: 77–154. Flora%20Survey.pdf) Kenneally, K.F., Keighery, G.J. and Hyland, B.P.M. Braimbridge, M., Antao, M. and Loomes, R. (2010). (1991). Floristics and phytogeography of Kimberley Groundwater dependent ecosystems for Millstream: rainforests (pp 91–131). In: McKenzie, N.L., Johnston, ecological values and issues, Environmental water R.B. and Kendrick, P.G. (eds), Kimberley rainforests of report series, Report no. 13, Department of Water, Australia. Surrey Beatty and Sons: Chipping-Norton, Government of Western Australia. Australia. Brim Box, J., Duguid, A., Read, R.E., Kimber, R.G., Leighton, K.A. (2004). Climate (pp 19–38). In: van Knapton, A., Davis, J. and Bowland, A.E. (2008). Vreeswyk, A.M.E., Payne, A.L., Leighton, K.A. and Central Australian waterbodies: The importance of Hennig, P. (eds) (2004). An inventory and condition permanence in a desert landscape. Journal of Arid survey of the Pilbara region, Western Australia. Technical Environments 72: 1395–1413. Bulletin No. 92. Western Australian Department of Brock, M., Capon S.J. and Porter, J.L. (2006). Disturbance Agriculture: Perth, Australia. of plant communities of desert rivers (pp 100–132). In: Lyons, M.N., Gibson, N., Keighery, G.J. and Lyons, S.D. Kingsford, R. (ed.), Ecology of desert rivers. Cambridge (2004). Wetland fl ora and vegetation of the Western University Press: Cambridge, UK. Australian wheatbelt. Records of the Western Australian Brock, M. and Casanova, M.T. (1997). Plant life at the Museum, Supplement 67: 39–89. edge of wetlands: ecological responses to wetting and Lyons, M.N., Keighery, G.J., Gibson, L.A. and Handasyde, drying (pp. 181–192). In: Klomp, N. and Lunt, I. (eds), T. (2014). Flora and vegetation communities of Frontiers in ecology: Building the links. Elsevier Science: selected islands off the Kimberley coast of Western Oxford, UK. Australia. Records of the Western Australian Museum, Clarke, K.R. and Green, R.H. (1988). Statistical design and Supplement 81: 205–243. analysis for a ‘biological effects’ study. Marine Ecology McKenzie, N.L., van Leeuwen, S. and Pinder, A.M. (2009). Progress Series 92: 205–219. Introduction to the Pilbara Biodiversity Survey, Pilbara riparian fl ora and plant communities 513

2002–2007. Records of the Western Australian Museum, Freeman: San Francisco, USA. Supplement 78: 3–89. Strahler, A.N. (1952). Hypsometric (area-altitude) analysis Masini, R.J. (1988). Inland waters of the Pilbara Western of erosional topography. Bulletin of the Geological Australia (Part 1), A report of a fi eld study carried out in Society of America 63: 1117–1142. March–April 1983. Technical Series 10, Environmental Thackway, R. and Cresswell, I.D. (1995). An interim Protection Authority Perth, Western Australia. biogeographic regionalisation of Australia. Australian Masini, R.J. and Walker, B.A. (1989). Inland waters of the Nature Conservation Agency: Canberra, Australia. Pilbara Western Australia (Part 2), A report of a fi eld Thoms, M.C., Beyer, P.J. and Rogers, K.H. (2006). study carried out in October–November 1984. Technical Series 24, Environmental Protection Authority: Perth, Variability, complexity, and diversity: the Western Australia. geomorphology of river ecosystems in dryland regions Meissner, R., Owen, G. and Bayliss, B. (2009). Flora (pp 47–75). In: Kingsford, R.T. (ed), Ecology of desert and vegetation of the banded iron formation of the rivers. Cambridge University Press: Cambridge, UK. Yilgarn Craton: Robinson Range and Mount Gould. van Vreeswyk, A.M.E., Payne, A.L., Leighton, K.A. and Conservation Science Western Australia 7: 363–376. Hennig, P. (eds) (2004). An inventory and condition Mueller, F. (1881a). Plants of north-western Australia. Votes survey of the Pilbara region, Western Australia. Technical and Proceedings of the Legislative Council during the Bulletin No. 92. Western Australian Department of fi rst session of 1881, no. 1. Government Printer: Perth, Agriculture: Perth, Australia. Australia. Walshe, T.V., Halse, S.A., McKenzie N.L. and Gibson, N. Mueller, F. (1881b). A catalogue of plants collected during (2004). Towards identifi cation of an effi cient set of Mr. Alexander Forrest's geographical exploration of natural diversity recovery catchments in the Western North-west Australia in 1879. Journal and Proceedings Australian Wheatbelt. Records of the Western Australian of the Royal Society of New South Wales 14: 81–95. Museum, Supplement 67: 365–384. Pinder, A.M., Halse, S.A., Shiel, R.J. and McRae, J.M. Waters and Rivers Commission (1996). Pilbara region water (2010). An arid zone awash with diversity: patterns in resources: review and development plan, summary report 4. the distribution of aquatic invertebrates in the Pilbara Western Australian Waters and Rivers Commission: region of Western Australia. Records of the Western Perth, Australia. Australian Museum, Supplement 78: 205–246. R Development Core Team (2009). R: A language and White, M., Albrecht, A., Duguid, A., Latz, A., Hamilton, environment for statistical computing. R Foundation for M., Latz, P. and White, M. (2000). Plant species and sites Statistical Computing: Vienna, Austria. of botanical signifi cance in the southern bioregions of the Russell-Smith, J. (1991). Classifi cation, species richness, Northern Territory; vol. 1: Signifi cant vascular plants. and environmental relations of monsoon rainforest Report to the Australian Heritage Commission from in northern Australia. Journal of Vegetation Science 2: the Arid Lands Environment Centre, and the Parks 259–278. and Wildlife Commission of the Northern Territory: Smith, M.G. (2013). Threatened and priority fl ora list for Alice Springs, Australia. Western Australia. Department of Parks and Wildlife: Wyrwoll, K.-H., Courtney, J. and Sandercock, P. (2000). Kensington, Western Australia. The climatic environment of the Carnarvon Basin, Sneath, P.H.A. and Sokal, R.R. (1973). Numerical taxonomy: Western Australia. Records of the Western Australian the principles and practice of numerical classifi cation. Museum, Supplement 61: 13–27.

APPENDIX 1 [ELECTRONIC] Site by species data matrix, reordered according to their site and species classifi cations. The fl oristic groups are indicated. Taxa recorded from one site are excluded (see Electronic Appendix 3).

APPENDIX 2 [ELECTRONIC] List of taxa recorded from 103 sites sampled during the survey (superscript 1 denotes pairs of infraspecifi c taxa analysed at specifi c rank, superscript 2 denotes closely related taxa that could not be reliably discriminated and were amalgamated for analysis). Numbers following taxa are Department of Parks and Wildlife priority fl ora codes (see Smith 2013).

APPENDIX 3 [ELECTRONIC] Site by singletons species data matrix.

See CD inside the back cover or visit http://www.museum.wa.gov.au/research/records-supplements/