S.E. Connor et al. Forgotten land-use impacts in NW Australia
1 Manuscript for Journal of Vegetation Science (special issue edited by Thomas Giesecke, Triin Reitalu
2 and Petr Kuneš).
3
4 Title:
5 Forgotten impacts of European land-use on riparian and savanna vegetation in North-Western
6 Australia
7
8 Author names and addresses:
9 Simon E. Connor, Larissa Schneider, Jessica Trezise, Susan Rule, Russell L. Barrett, Atun Zawadzki &
10 Simon G. Haberle
11
12 Connor, S.E. (corresponding author, [email protected]) 1,2
13 Schneider, L. ([email protected]) 3
14 Trezise, J. ([email protected]) 3
15 Rule, S. ([email protected]) 3
16 Barrett, R.L. ([email protected]) 4,5
17 Zawadzki, A. ([email protected]) 6
18 Haberle, S.G. ([email protected]) 3
19 1School of Geography, 221 Bouverie Street, University of Melbourne, 3010 VIC, Australia;
20 2CIMA-FCT, University of the Algarve, Campus de Gambelas, 8005-139 Faro, Portugal;
1
S.E. Connor et al. Forgotten land-use impacts in NW Australia
21 3Centre of Excellence in Australian Biodiversity & Heritage, and Department of Archaeology and
22 Natural History, Australian National University, Fellows Rd, Canberra, 2600 ACT, Australia;
23 4National Herbarium of New South Wales, Royal Botanic Gardens and Domain Trust, Mrs Macquaries
24 Road, Sydney NSW 2000, Australia;
25 5College of Medicine, Biology and Environment, Research School of Biology, Australian National
26 University, Canberra, 2601 ACT, Australia;
27 6Institute for Environmental Research, Australian Nuclear Science and Technology Organisation, PMB
28 1 Menai, 2234 NSW, Australia.
29
30 Printed journal page estimate:
31 8597 words including references (10.7 pages), figures 1.3 pages, total 12 pages.
32
33 Abstract
34 Questions
35 Fire and livestock grazing are regarded as current threats to biodiversity and landscape integrity in
36 Northern Australia, yet it remains unclear what biodiversity losses and habitat changes occurred in
37 the 19–20th centuries as livestock and novel fire regimes were introduced by Europeans. What
38 baseline is appropriate for assessing current and future environmental change?
39 Location
40 Australia's Kimberley region is internationally recognised for its unique biodiversity and cultural
41 heritage. The region is home to some of the world's most extensive and ancient rock art galleries,
42 created by Aboriginal peoples since their arrival on the continent 65,000 years ago. The Kimberley is
2
S.E. Connor et al. Forgotten land-use impacts in NW Australia
43 considered one of Australia's most intact landscapes and its assumed natural vegetation has been
44 mapped in detail.
45 Methods
46 Interpretations are based on a continuous sediment record obtained from a waterhole on the
47 Mitchell River floodplain. Sediments were analysed for geochemical and palynological proxies of
48 environmental change and dated using Lead-210 and Carbon-14 techniques.
49 Results
50 We show that the present-day vegetation in and around the waterhole is very different to its pre-
51 European counterpart. Pre-European riparian vegetation was dominated by Antidesma ghaesembilla
52 and Banksia dentata, both of which declined rapidly at the beginning of the 20th century. Soon after,
53 savanna density around the site declined and grasses became more prevalent. These vegetation
54 shifts were accompanied by geochemical and biological evidence for increased grazing, local
55 burning, erosion and eutrophication.
56 Conclusions
57 We suggest that the Kimberley region's vegetation, while maintaining a 'natural' appearance, has
58 been altered dramatically during the last hundred years through grazing and fire regime changes.
59 Landscape management should consider whether the current (impacted) vegetation is a desirable or
60 realistic baseline target for biodiversity conservation.
61
62 Keywords geochemistry, palaeoecology, human impact, riverine vegetation, savanna, fire regimes,
63 grazing, Australian Monsoon Tropics, Potential Natural Vegetation
64
65 Nomenclature Western Australian Herbarium (2017); Mucina & Daniel (2013).
3
S.E. Connor et al. Forgotten land-use impacts in NW Australia
66
67 Running head Forgotten land-use impacts in NW Australia
68
69 Introduction
70 Human impacts associated with European colonisation beginning in the 15th century
71 profoundly altered ecosystems around the globe. Impacts included the introduction of exotic
72 species, manipulation of fire regimes, deforestation, major land-use changes, displacement of
73 indigenous land management practices and alterations to biogeochemical cycles (Kirkpatrick 1999;
74 Ireland and Booth 2012; Johnson et al. 2017). As these impacts preceded a widespread
75 understanding of ecosystem ecology and research, in many parts of the world it is difficult to
76 understand how ecosystems functioned prior to European colonisation. This raises questions about
77 how to best manage altered ecologies in an era of rapid environmental change (Johnson et al. 2017).
78 A common baseline for large-scale ecosystem management is the concept of Potential
79 Natural Vegetation (PNV). PNV takes information from areas of assumed natural vegetation (late-
80 successional) and extrapolates to other (impacted) areas with similar environmental conditions
81 (Loidi & Fernández-González 2012). The PNV approach has been criticised because its predictions
82 sometimes conflict with the fossil record (Carrión & Fernández 2009; Rull 2015; Abraham et al.
83 2015). PNV also fails to account for ecosystem dynamics (Chiarucci et al. 2010) and is unable to
84 simulate ecological structure in cultural landscapes (Strona et al. 2016). Despite these constraints,
85 PNV remains an accessible baseline for conservation and management decisions.
86 Northern Australia is regarded as one of the largest areas of intact tropical savanna
87 worldwide (Woinarski et al. 2007; Bowman et al. 2010; Ziembicki et al. 2015). If this is true, PNV
88 maps for the region should constitute a valuable baseline (Beard et al. 2013; Mucina & Daniel 2013).
89 Yet the ecosystems of Northern Australia have been extensively modified since European
4
S.E. Connor et al. Forgotten land-use impacts in NW Australia
90 colonisation through livestock grazing, disruption of Indigenous land curation and the introduction of
91 invasive species, among other impacts (Hnatiuk & Kenneally 1981; Lonsdale 1994; Vigilante &
92 Bowman 2004; Douglas et al. 2015; Radford et al. 2015). Impacts are linked to extinctions, declining
93 native mammal populations, altered stream ecology, eutrophication, erosion and catchment
94 instability (Payne et al. 2004; Woinarski et al. 2007; Ziembicki et al. 2015).
95 Taking a different view, recent research links the recent decline of fire-sensitive species
96 across Northern Australia to interactions between fire and flammable grasses (Trauernicht et al.
97 2013; Bowman et al. 2014). Uncertainties surrounding the causes of recent species losses remain
98 obscure in the absence of reliable historical evidence for 19th and early-20th century environmental
99 change in this sparsely populated region. Such evidence could provide a critical test of ecosystem
100 integrity and reveal long-term drivers of species decline.
101 Northern Australia is home to high levels of biodiversity, much of which remains
102 undocumented (Barrett & Barrett 2011, 2015; Barrett 2013, 2015; Maslin et al. 2013; Moritz et al.
103 2013). Phylogenetic research reveals extraordinary levels of genetic diversity and endemism,
104 perhaps equal to the recognised biodiversity hotspots of Eastern Australia (Moritz et al. 2013). The
105 Kimberley region in NW Australia is internationally recognised for its extensive rock art galleries,
106 which are some of the most ancient examples of human artistic expression globally (Aubert 2012)
107 and are situated in an ancient cultural landscape created by Indigenous Australians (Rangan et al.
108 2015; Hiscock et al. 2016). There is a pressing need to understand and document the region’s
109 biological and cultural heritage as economic development pressures on the Northern Australian
110 environment intensify.
111 In this paper, we confront the assumed natural vegetation of the Kimberley region with
112 fossil evidence, aiming to assess the ecological integrity of the region’s tropical savanna and riparian
113 vegetation types. Our data pertain to the savanna woodlands and riparian thickets (Mucina & Daniel
114 2013) of the Mitchell Plateau, one of the last areas of the Australian continent to experience
5
S.E. Connor et al. Forgotten land-use impacts in NW Australia
115 European colonisation (McGonigal 1990). We compare pollen data with independent geochemical
116 data to understand the timing, direction and drivers of two centuries of environmental change.
117
118 Methods
119 Study area
120 The Kimberley Region is situated in NW Australia and constitutes a biogeographically and
121 geologically distinct entity within the Australian Monsoonal Tropics (Pepper & Keogh 2014). The
122 region is characterised by complex and ancient geology, which has combined with long periods of
123 subaerial exposure and weathering to create unique landforms and topography (Pillans 2007; Tyler
124 2016). The Kimberley region is broadly divided into three geological basins: the extensive Kimberley
125 Basin in the north, the Ord Basin to the south and east, and the Canning Basin to the south and west.
126 These basins are separated by two orogenic belts of metamorphosed and intrusive rocks. The
127 western portion of the Kimberley Basin is dominated by the King Leopold Sandstone (erosion-
128 resistant arenites) and Carson Volcanics (weathered basalts), while the Warton and Pentecost
129 Sandstones dominate the eastern portion. These rocks have been subjected to only minor faulting
130 and deformation since their deposition some 1840–1800 million years ago and thick, erosion-
131 resistant laterites developed over the basalts 70–50 million years ago (Tyler 2016). Geology has a
132 defining influence on the distribution of soil types and vegetation units, as well as biogeographical
133 and phylogenetic patterns (Hnatiuk & Kenneally 1981; Pepper and Keogh 2014; Mucina & Daniel
134 2013; Moritz et al. 2013; Barrett 2015; Barrett & Barrett 2015).
135 On the Mitchell Plateau, savanna woodlands on basalt tend to have a ground layer of
136 Allopteropsis semialata, Chrysopogon fallax, Heteropogon contortus, Sehima nervosum, Sorghum
137 plumosum and many other grasses, with a tree layer of eucalypts (Eucalyptus bigalerita, E. tectifera,
138 Corymbia greeniana, C. bella, C. disjuncta), ironwood trees (Erythrophloemum aff. chlorostachys) and
6
S.E. Connor et al. Forgotten land-use impacts in NW Australia
139 terminalia (Terminalia canescens, T. fitzgeraldii) depending on topography (Mucina & Daniel 2013).
140 Savanna woodlands on sandstone-derived soils have dominant Sorghum and Triodia grasses and a
141 dominant tree layer of eucalypts (Eucalyptus tetradonta, E. miniata, E. apodophylla, E. houseana,
142 Corymbia latifolia, C. nesophila, C. torta) and palms (Livistona eastonii) (Hnatiuk & Kenneally 1981;
143 Mucina & Daniel 2013). Sandstone outcrops have an altogether different flora, having a high
144 proportion of species considered to have Gondwanan heritage, as well as many endemics (Dunlop &
145 Webb 1991). Key species of these scrubs/heaths include Acacia gonocarpa, A. kelleri, A. arida, A.
146 tumida, Bossiaea arenitensis, Calytrix brownii, C. exstipulata, C. achaeta, Corymbia polycarpa,
147 Gardenia pyriformis and Grevillea agrifolia (Mucina & Daniel 2013). Rainforest remnants (vine
148 thickets) occur in fire-protected locations where moisture is available all year (Hnatiuk & Kenneally
149 1981; Ondei et al. 2017). Albizia lebbeck, Atalaya variifolia, Cochlospermum fraseri, Sersalisia sericea
150 and Wrightia pubescens are important canopy species in these rainforests (Beard 1976). Wetlands
151 and riparian zones are dominated by Pandanus and Melaleuca species (P. aquaticus, P. spiralis, M.
152 leucadendron, M. nervosa, M. viridiflora). Nauclea orientalis, Timonius timon and the deciduous
153 shrub Antidesma ghaesembilla are also common in these areas.
154 The Kimberley region has a dry tropical monsoonal climate, with rainfall seasonality driven
155 by the Indo-Australian Summer Monsoon. Doongan Station, near our study site, receives an average
156 of 1205 mm of annual precipitation, 95% of which falls from November to April (Bureau of
157 Meteorology, http://www.bom.gov.au/climate/data/ accessed 5 June 2017). Rainfall events usually
158 occur in the form of prolonged tropical storms and/or cyclones during the humid ‘wet’ season,
159 flooding the region’s watercourses (Mucina & Daniel 2013). A steep rainfall gradient exists across the
160 Kimberley region – on average, over 1400 mm of precipitation falls annually on the seaward slopes
161 of the Mitchell Plateau, while the southern parts of the Kimberley adjoining the Great Sandy Desert
162 receive less than 400 mm. Maximum daily temperatures average over 30°C in most parts of the
163 Kimberley Region throughout the year (Bureau of Meteorology,
164 http://www.bom.gov.au/climate/data/ accessed 5 June 2017).
7
S.E. Connor et al. Forgotten land-use impacts in NW Australia
165
166 Sample collection and dating
167 A sediment core was collected from a small (approx. 100 m2) elliptical waterhole on the
168 floodplain of the Mitchell River (15°10’30.5”S, 125°53’29.3”E; Fig. 1) using a mud–water interface
169 corer operated from a floating platform. The site (MP11A) is 150 m from the main river channel and
170 forms part of a complex of lakes and swamps where the floodplain widens as it follows the
171 geological contact between basalt and sandstone. The waterhole is fringed by Melaleuca
172 leucadendron to the south and surrounded by savanna woodland on all sides. Nymphaea violacea
173 and Eleocharis grow in the water, which was approximately 50 cm deep at the time of core
174 collection.
175 The core was dated using a combination of 210Pb and 14C techniques (Wright et al. 2017).
176 Eight 210Pb samples were selected from the upper 35 cm of the core. Macroscopic plant remains
177 were removed before the samples were freeze dried, ground and weighed. Subsamples of >1 g were
178 dated at the Australian Nuclear Science and Technology Organisation (ANSTO) using alpha
179 spectrometry. Radiocarbon (14C) dating was performed on three samples >5 g using Accelerator
180 Mass Spectrometry at DirectAMS laboratories (Seattle, USA). Macroscopic plant material was
181 removed prior to further pretreatment. An age–depth model was constructed using cubic spline
182 regression in Clam 2.2 (Blaauw 2010).
183 Surface sediment samples are routinely collected in palaeoecological research to calibrate
184 pollen–vegetation relationships (Morris et al. 2017). In the Kimberley Region, surface samples have
185 been analysed to aid the interpretation of pollen records (Proske et al. 2014; Field et al. 2017), but
186 multi-site calibration has not been attempted. We collected surface sediment samples from ten
187 small lakes and wetlands across the region (Fig. 1). Multiple representative stands of the
188 surrounding vegetation were surveyed using a Haglöf HEC-R electronic clinometer and a GRS
189 Densitometer to estimate canopy height, tree volume and density. Composition was measured on
8
S.E. Connor et al. Forgotten land-use impacts in NW Australia
190 the DAFOR scale (Kent 2012). Satellite imagery was used to estimate the proportion of each
191 vegetation type within a 100 m radius of the coring site. Stand-based estimates were weighted using
192 these proportions to obtain vegetation estimates for comparison with pollen assemblages.
193
194 Pollen and charcoal analysis
195 Seasonally variable environments often have less suitable conditions for pollen preservation
196 than more stable environments (Head & Fullager 1992), requiring relatively large sediment samples.
197 Samples of 5 ml of sediment were treated according to standard palynological techniques (Jones &
198 Rowe 1999), including removal of clays through settling and decantation, sieving to remove large
199 particles (>120 μm), treatment with 10% KOH at 90°C for 20 minutes to remove humic material,
200 heavy liquid separation (s.g. 2.0) to remove minerogenic material, acetolysis (a 9:1 mixture of
201 C4H6O3to H2SO4), carbonate removal with 10% HCl, HF (50%) to remove fine silica, dehydration in
202 100% ethanol, and mounting in glycerol on glass slides. Lycopodium spore tablets were added early
203 in pretreatment to enable calculation of pollen concentrations and accumulation rates. Pollen,
204 spores and selected fungal and algal remains were identified at 400 x magnification, using the
205 extensive Kimberley collection in the Australasian Pollen and Spore Atlas (Rowe 2006) and published
206 guides (Macphail & Stevenson 2004; van Geel & Aptroot 2006). Identifications were agreed upon by
207 three analysts (JT, SR and SH).
208 Charcoal particles are an indicator of fire occurrence, with macroscopic particles considered
209 indicative of local fires and microscopic particles indicative of regional burning (Whitlock & Larsen
210 2001). Microscopic charcoal was quantified on pollen slides by counting opaque black particles 10–
211 120 μm. Macroscopic charcoal was sieved from a known volume of sediment using a 120 μm mesh
212 and particles identified using a binocular microscope (Stevenson & Haberle 2005). Results were
213 converted to accumulation rates for comparison with pollen data.
9
S.E. Connor et al. Forgotten land-use impacts in NW Australia
214 The pollen sequence was divided into relatively homogenous timespans using zonation
215 (CONISS), implemented in Tilia (version 2.0.41, Eric Grimm, http://www.TiliaIT.com). Pollen–
216 vegetation relationships were assessed by comparing the abundance of each pollen type to the
217 abundance of pollen-producing plants within a 100-m radius (the most commonly-used sampling
218 radius in comparable studies: Bunting et al. 2013).
219
220 Geochemical analysis
221 To examine the relationship between past vegetation and changes in the floodplain
222 environment, we analysed elemental concentrations of 34 geochemical elements. We used the
223 alkaline elements sodium (Na), potassium (K), magnesium (Mg), calcium (Ca) and barium (Ba) to
224 track erosion. These elements are relatively water-soluble and their mobility during erosion
225 processes results in elevated concentrations in water-body sediments (Davies et al. 2015). The
226 lithogenic element titanium (Ti) was used as a control for the mobile elements as it is geochemically
227 stable and hosted by resistant minerals (Davies et al. 2015). The element phosphorus (P) was used to
228 track livestock grazing. Cattle directly affect the local P cycle because the phosphorous they
229 consume is returned to the environment as dung (Parham et al. 2002). This increases P
230 concentrations in the soil and ultimately contributes to P enrichment in water bodies and sediments.
231 Sediment samples (33 from the core and 10 surface samples) of 0.2 g each were digested
232 using 1 ml of HNO3 (Aristar, BDH, Australia) and 2 ml HCl (Merck Suprapur, Germany) and analysed
233 for metals using an inductively coupled plasma mass spectrometer (ICP-MS) (Perkin-Elmer; model
234 DRC-e) with an AS-90 auto-sampler (Telford et al. 2008). Certified reference materials (NRC-BCSS1,
235 NRC-Mess1, NIST-1646, NRC-PCAS1) and blanks were used as to check the quality and traceability of
236 metals. Principal Components Analysis (PCA) was used to link geochemical variables to pollen
237 assemblages.
10
S.E. Connor et al. Forgotten land-use impacts in NW Australia
238
239
240 Results
241 Figure 2 shows the adopted age–depth model for the MP11A sediment record with dated
242 levels and associated errors. Two 14C dates were excluded from the model (30.5 and 40.5 cm depth).
243 These samples yielded ages that were the same or older than the lowermost 14C age in the core.
244 Redeposition of older carbon by floods is a common problem when dating recent floodplain deposits
245 (Ely et al. 1992). The chronology established by 210Pb has low errors and indicates relatively constant
246 sedimentation rates. The apparent increase in sedimentation toward the top of the core is likely to
247 reflect the relatively uncompacted nature of the more recent sediments.
248 Figure 3 shows key changes in the vegetation of the MP11A site and its surrounds through the
249 last 200 years (see Supporting Information for complete pollen record). Four main phases were
250 defined by zonation, described below in relation to changes in charcoal and geochemistry:
251 1. Phase 1 (1823–1907 AD) is characterised by relatively high proportions of Corymbia, other
252 Myrtaceae, Callitris, Terminalia, Banksia and Antidesma ghaesembilla pollen. Phase 1 pollen
253 assemblages are very different to those of surface samples collected in modern waterholes
254 (see PCA, Supporting Information). Microscopic charcoal accumulation rates were relatively
255 high in Phase 1 compared with later phases, whilst macroscopic charcoal accumulation rates
256 were low (average 4% total charcoal). Geochemical indicators show a dip in concentrations
257 of weathering resistant Ti later in this phase and an increase in silicon (Si).
258 2. Phase 2 (1907–1948 AD) exhibits a steep decline in Antidesma ghaesembilla and Banksia
259 pollen. Pollen of the canopy dominants fluctuated, with Corymbia and Callitris attaining
260 their highest proportions in the entire sequence before rapidly decreasing toward the end
261 of the phase. Botryococcus decreased substantially at the same time and Sordaria fungal
11
S.E. Connor et al. Forgotten land-use impacts in NW Australia
262 spores increased in abundance. Charcoal declined substantially toward the end of this
263 phase, reaching its lowest concentrations in the entire record. In contrast, the mobile
264 elements (Na, K and Ba) increased while Ti concentrations declined. The end of this phase is
265 marked by a sharp increase in P, reaching levels over 10 times background concentrations.
266 3. Phase 3 (1948–1993 AD) begins with major increases in Poaceae and Pandanus pollen.
267 These changes were followed by a major increase in both macroscopic and microscopic
268 charcoal, reaching the highest concentrations in the record. Tree pollen decreased to an
269 average of 34% (compared to 50% in the previous zone). Phase 3 pollen assemblages are
270 similar to those of surface samples from modern waterholes (see PCA, Supporting
271 Information). Geochemical data show a temporary decline in concentrations of Na, K and Ba
272 and an increase in Ti, Mg and Ca.
273 4. Phase 4 (1993–2016 AD) is differentiated from the previous phase by a further decline in
274 tree pollen (average: 30%) and the highest levels of Poaceae, Acacia and Dodonaea pollen
275 for the entire record. Macroscopic charcoal is abundant compared to microscopic charcoal
276 (average 30% total charcoal). Sordaria remained abundant through this phase. Geochemical
277 indicators show relatively high concentrations for all elements except P, which decreased.
278
279 Figure 4 shows the relationships between major pollen taxa and vegetation parameters for
280 surface samples (data in Supporting Information). Corymbia pollen correlates strongly with area-
281 weighted plant abundance (r2: 0.87), savanna basal area (r2: 0.74) and savanna canopy cover (r2:
282 0.48). Pandanus pollen was less strongly related to abundance (r2: 0.37) and had no relationship to
283 savanna basal area (i.e. tree density). Poaceae produced a short abundance gradient (being
284 dominant in vegetation surveys), but displayed an inverse relationship with savanna density (r2:
285 0.42) and canopy cover (r2: 0.27). Corymbia:Poaceae ratios correlated positively with both Corymbia
286 abundance (r2: 0.64) and savanna density (r2: 0.67).
12
S.E. Connor et al. Forgotten land-use impacts in NW Australia
287
288 Discussion
289 Our multi-proxy results provide a glimpse of historical changes that occurred in the savannas
290 and inland floodplains of the Kimberley Region. A longer record of palaeovegetation change comes
291 from Black Springs, 75 km SW of site MP11A (15°38’S; 126°23’23”E), showing Pandanus expansion
292 during Holocene wet phases and aridification cycles linked to weakening monsoons (Field et al.
293 2017). This record, however, has low temporal resolution (centennial-scale) and uncertain dating for
294 recent centuries. The MP11A record provides decadal resolution, revealing a sequence of
295 environmental changes in the upper Mitchell River area since European colonisation. The 210Pb
296 chronology shows no signs of flood disturbance (Arnaud et al. 2002). Nevertheless, the assigned ages
297 should be viewed as approximate.
298 The sequence of events is summarised in Figure 3. An initial decline of the riparian shrub
299 Antidesma ghaesembilla began around 1900 (Phase 1), followed by Banksia decline a decade later.
300 Corymbia, indicative of savanna density (Fig. 4), increased around 1912 and Callitris followed in the
301 1920s (Phase 2). At the same time, regional fires decreased (indicated by lower levels of microscopic
302 charcoal) and grazing intensified (indicated a rapid increase in dung fungal spores, such as Sordaria).
303 The 1930s brought greater grazing intensity, increasing erosion (indicated by rising Ba and Na
304 concentrations) and a major decline in savanna trees (Corymbia and Callitris) and regional burning.
305 Eutrophication followed in the 1940s, with a pronounced peak in P lasting until the 1970s (Phase 3).
306 During this interval, grass abundance increased and Pandanus replaced A. ghaesembilla as the
307 dominant riparian element. Fire occurrence rose substantially in the late 1970s, particularly the local
308 fires indicated by macroscopic charcoal and Neurospora fungal spores (Whitlock & Larsen 2001; van
309 Geel & Aptroot 2006). This increase in local burning was accompanied by evidence of greater grazing
310 intensity (dung fungi), peaking around 1990. The last two decades have seen further expansion of
13
S.E. Connor et al. Forgotten land-use impacts in NW Australia
311 grasses and a decline in savanna density, as well as the disappearance of Pandanus from the
312 waterhole margins (Phase 4).
313 Observed changes in savanna fire regime, savanna vegetation structure and riparian zone
314 vegetation are discussed in the following sections in relation to major drivers of environmental
315 change.
316
317 Changes in savanna fire regime
318 On a global scale, savannas are thought to exist as an alternative stable state to forests,
319 maintained by frequent fire (Bond et al. 2005). Changing fire regimes are implicated in major
320 vegetation shifts across the Northern Australian savanna zone (O’Neill et al. 1993; Russell-Smith et
321 al.2003; Bowman 2003), including the simplification of vegetation structure (Craig 1997),
322 compositional change (Trauernicht et al. 2013; Bowman et al. 2014), shifting forest–savanna
323 boundaries (Hnatiuk & Kenneally 1981; Banfai & Bowman 2005) and negative impacts on native
324 mammals (Radford et al. 2015; Ziembicki et al. 2015). The effects of fire regime change in Australia
325 are not completely understood, given a lack of information on historical fire regimes and fire ecology
326 (Silcock et al. 2016). Fire regime change is often linked to the dispossession of Aboriginal peoples of
327 their land and consequent loss of traditional burning practices, followed by the different burning
328 practices of European colonists (Russell-Smith et al. 2003; Ritchie 2009). Alternatively, fire regime
329 change may stem from the invasion of fire-promoting grass species, creating a grass–fire positive
330 feedback cycle (Bowman et al. 2014).
331 Our data demonstrate profound changes in fire regime around the MP11A site over 200
332 years. Phase 1 samples are dominated by microscopic charcoal. This suggests that regional savanna
333 fires were common, while local fires in the riparian zone were apparently rare. This agrees with
334 ethnographic evidence (Russell-Smith et al. 1997; Vigilante 2001). The major reduction in regional
14
S.E. Connor et al. Forgotten land-use impacts in NW Australia
335 burning in the 1930s and 1940s (Phase 2) coincides with the decline of continuous Aboriginal
336 occupation of the Mitchell Plateau, ending in the 1950s (Wilson 1981). Data from the 1960s onwards
337 (Phases 3–4) show that local fires became more prevalent, impacting both savanna and riparian
338 zones (Fig. 3). Historically, Aboriginal fire management is associated with light/early-season fires
339 compared to unmanaged areas, where severe dry-season fires are more frequent (O’Neill et al.
340 1993).
341 The nature of the MP11A site limits the conclusions that may be drawn from charcoal
342 analysis. Fluvial transport is a major source of primary charcoal (direct products of fire) and
343 secondary charcoal (redeposited by erosion : Patterson et al. 1987; Whitlock & Larsen 2001).
344 Overlapping radiocarbon dates at different sediment depths could indicate contamination by older
345 (secondary) charcoal (Fig. 2). However, the macroscopic charcoal record is validated by a similar
346 trend in Neurospora, a biological proxy for local fire (van Geel & Aptroot 2006). Charcoal abundance
347 is not clearly related to erosion proxies (Fig. 3) and there is no indication of a major shift in sediment
348 source that explains the dramatic increase in macroscopic charcoal since the 1960s. Hence, the
349 MP11A charcoal record is cautiously interpreted as a reflection of fire occurrence in the riparian
350 zone (macroscopic charcoal) and surrounding savanna (microscopic charcoal). The changes are
351 consistent with a transition from Aboriginal to European fire management (Haberle et al. 2006), but
352 remain to be corroborated at other sites.
353
354 Changes in savanna structure
355 A critical question in understanding savanna vegetation dynamics is how global-scale factors
356 interact with the local environment (Case & Staver 2017). Previous research has debated the role of
357 fire versus local edaphic factors in determining vegetation structure in Northern Australia (e.g.
358 Bowman 1988; Craig 1997; Douglas et al. 2015; Ondei et al. 2017). These studies often used space-
359 for-time substitution or repeat aerial photography to analyse vegetation changes. Space-for-time
15
S.E. Connor et al. Forgotten land-use impacts in NW Australia
360 substitution is limited by confounding variables such as differences in soil type, climate, species
361 pools and land use at different sampling locations. Aerial photography provides long-term data on a
362 broad geographic scale, but has limited temporal scope (usually late 20th century) and provides little
363 detail on changes in understorey vegetation beneath the canopy where the impacts of savanna fire
364 are presumably greatest. Palaeoecological data from small basins, such as the MP11A record, reveal
365 vegetation changes at a small spatial scale over long timescales and are ideally placed to inform
366 debates on vegetation structure (Mariani et al. 2017; Morris et al. 2017).
367 Our results point to a long-term decline in savanna tree density. The decline is registered in
368 decreasing Corymbia pollen abundances and Corymbia:Poaceae ratios, proxies strongly associated
369 with savanna basal area (Fig. 4). Tree density decline was particularly acute in the 1930s. This
370 evidence, albeit from a single site, contrasts with observations of 20th-century ‘woody thickening’
371 across Northern Australia (e.g. Bowman et al. 2010; Tng et al. 2011, cf. Murphy et al. 2014).
372 Observed changes in vegetation density are dependent on the baseline used. The earliest aerial
373 photography for most of Northern Australia dates to the 1940s and 50s. Savanna density change at
374 MP11A since the 1940s has been negligible (Fig. 3). Compared to a baseline of the 19th or early 20th
375 century, however, loss has been substantial.
376 Around MP11A, grass cover increased as savanna density declined (Fig. 3). Historically,
377 livestock have caused major changes in grassland species composition in the Kimberley Region
378 (Petherham et al. 1986). Cattle, sheep, pigs, horses and donkeys were introduced by European
379 graziers from the 1880s (McGonigal 1990). Livestock numbers increased rapidly from less than
380 100,000 cattle in the 1890s to more than 600,000 in 1914 (Bolton 1953), and were linked to
381 environmental degradation from the 1930s (Payne et al. 2004). Cattle grazing can trigger a transition
382 from perennial to annual grasses in Australian savannas (Ash et al. 1997; Fensham & Skull 1999). The
383 same type of transition is linked to increasing fire frequency (Bowman et al. 2014). The histories of
16
S.E. Connor et al. Forgotten land-use impacts in NW Australia
384 grazing and fire at MP11A are so long and intertwined that a single cause of recent savanna
385 structural change is unlikely to emerge.
386 Perennial Triodia grasses, if left unburnt for long periods, can gain significant size and
387 density, providing habitat to a range of vertebrates and invertebrates. Old tussocks, often exceeding
388 two metres in height and five metres in breadth, were frequent along the Fitzroy Crossing Road until
389 the mid-1980s. Recent changes in fire regime have greatly reduced the extent of unburnt Triodia
390 tussocks, a change that may be linked to the decline of small mammals in the region. Pollen data
391 cannot be used to interrogate the history of specific grass species, because different species produce
392 pollen that is morphologically indistinguishable. Analysis of phytoliths or ancient DNA may help
393 address this shortfall.
394 Callitris intratropica has been the target of considerable research into fire impacts in
395 Northern Australian savannas (e.g. Bowman & Panton 1993; Bowman et al. 2014). Callitris pollen is
396 morphologically distinct and the MP11A pollen record provides insight into the recent history of this
397 fire-sensitive conifer. Callitris apparently experienced a major decline in the 1930s and has since
398 recovered slightly. C. intratropica trees were not sufficiently abundant in our vegetation surveys to
399 permit a quantitative test of the grass–fire cycle hypothesis (Bowman et al. 2014). Theoretically, if
400 the expansion of grasses is linked to Callitris decline, the ratio of Callitris:Poaceae pollen should
401 decrease over time. Instead there is an increase since the mid-20th century at MP11A (Fig. 3). The
402 purported link between fire and grass cover finds little support in our data, with charcoal and
403 Poaceae pollen following different trajectories. Our data cannot answer questions about grazing–
404 fire–grass interactions on a regional scale, but point to a profound legacy of 20th-century grazing and
405 fire impacts in the Mitchell Plateau savannas.
406
407 Changes in the riparian zone
17
S.E. Connor et al. Forgotten land-use impacts in NW Australia
408 Data from the MP11A waterhole provide strong evidence for the 20th-century deterioration
409 of a riparian shrub layer comprising Banksia and Antidesma. Banksia pollen represents B. dentata,
410 the only Banksia species extant in the Kimberley region. This non-serotinous species occurs in
411 seasonally inundated areas and swamp margins on sandstone bedrock, where it grows alongside
412 Eucalyptus apodophylla (Hnatiuk & Kenneally 1981). Extensive aerial surveys across the Kimberley
413 confirm this almost exclusive habitat restriction (R. Barrett, pers. obs.). Our data suggest that
414 Banksia dentata also occupied riparian zones in the past (i.e. Melaleuca leucadendron alliance of
415 Hnatiuk & Kenneally, 1981; Riparian Thickets group of Mucina & Daniel 2013).
416 Antidesma ghaesembilla is a fire-sensitive riparian shrub that bears edible fruits and is
417 known to expand rapidly in areas of unburnt woodland (Bowman et al. 1988; Russell-Smith et al.
418 2003). It occurs frequently along the Mitchell River floodplain today. The MP11A pollen record
419 suggests A. ghaesembilla was dominant around this site until the early 20th century. Its decline
420 follows a peak in macroscopic charcoal, implicating local fires as the cause. According to
421 ethnographic sources, Aboriginal people avoided burning riparian vegetation (Russell-Smith et al.
422 1997; Vigilante 2001). Post-fire recovery may have been impeded by grazing, as Antidesma species
423 are more palatable to cattle than browse- and burning-resistant Pandanus.
424 Riparian and wetland zones are where cattle and feral animal impacts in the Kimberley are
425 most acute (Legge et al. 2011). Grazing and trampling have had significant detrimental impacts on
426 riparian ‘bamboo’ (Phragmites karka), once common in the Kimberley, but now listed as a species of
427 conservation concern in Western Australia. Enhanced erosion is attested geochemically by increased
428 concentrations of mobile versus immobile elements (Bierman et al. 1998). At MP11A, mobile Na, K
429 and Ba increased as relatively immobile Ti decreased during the 20th century. Erosion is further
430 indicated by Pseudoschizaea palynomorphs (Lopez-Marino et al. 2010), which peak in Phases 3–4.
431 Grazing may cause P enrichment from faeces and losses of Ca and Mg from urine patches (Early et al.
432 1993; Parham et al. 2002). In the MP11A record, the early–mid 20th century is characterised by an
18
S.E. Connor et al. Forgotten land-use impacts in NW Australia
433 increase in P (Fig. 3) and a decrease in Ca and Mg (Supporting Information). With the additional
434 support of proxies for livestock dung (Fig. 3), these geochemical changes may be linked to grazing
435 impacts.
436 Some of the changes witnessed by the MP11A record could be explained by other
437 environmental shifts occurring in the Mitchell River catchment . Larsen et al. (2016) demonstrated
438 how the geomorphic process of knick-point migration caused riparian forest loss in Australia’s
439 Northern Territory. Sand slugs released by upstream erosion cause long-lasting impacts on
440 downstream ecosystems in many Australian rivers (Prosser et al. 2001). These explanations seem
441 inadequate in the case of MP11A – aerial photographs of the study area in 1949 and 2017 show no
442 obvious change in riparian forest and floodplain geomorphology (see Supporting Information).
443 There are no major flooding events corresponding to the time of peak impact at MP11A
444 (Gillieson et al. 1991; Wohl et al. 1994). Meteorological records indicate some higher-than-average
445 rainfall years in the 1940s and 1950s, but much higher values in recent decades (Bureau of
446 Meteorology, http://www.bom.gov.au/climate/data/ accessed 5 June 2017), part of a wetting trend
447 across Northern Australia (Bowman et al. 2010). The decades-long environmental transformation
448 evident at MP11A cannot be explained by a single disturbance event. Agreement between
449 geochemical and biological proxies for grazing and fire, alongside clear evidence for the loss of
450 savanna tree density and the shrub stratum within the riparian zone, suggest geomorphic processes
451 are an unlikely cause of historic vegetation changes in the study area.
452
453 Implications for conservation and vegetation mapping
454 Our evidence casts doubt on the notion that Northern Australia’s landscapes and
455 ecosystems are intact and unmodified (Woinarski et al. 2007; Bowman et al. 2010; Ziembicki et al.
456 2015). Just as the intact canopy of ‘Australian’ eucalypts disguises a savanna flora with strong SE
19
S.E. Connor et al. Forgotten land-use impacts in NW Australia
457 Asian affinities (Haynes et al. 1991), the same canopy may disguise the pervasive ecological
458 transformations that have occurred over the last 100 years. The forgotten impacts of early 20th-
459 century fire and grazing have profoundly altered savanna and riparian ecosystems on the Mitchell
460 Plateau and perhaps more widely.
461 Our data call for caution in adopting mid-20th-century ecological baselines for assessing
462 ecological change, given that the impacts of European colonisation often manifested themselves
463 much earlier. Potential Natural Vegetation (PNV) maps have similar limitations. Because PNV maps
464 are populated from data collected in the current (impacted) environment, there are likely to be
465 knowledge shortfalls in reconstructing pre-European vegetation (e.g. Beard et al. 2013). Given the
466 duration and magnitude of historical impact, it could be asked whether restoration to pre-European
467 conditions is desirable or realistic.
468 There is a clear need for replication beyond the single palaeoecological record presented
469 here to better understand spatial and temporal variation in the Kimberley region’s ecosystems. It is
470 acknowledged that obtaining palaeoecological data for vegetation types existing under more arid
471 conditions may be difficult (Head & Fullager 1992). Hence PNV provides a useful ‘null model’
472 (Somodi et al. 2012) that can only be improved as better data and new sources of information come
473 to hand.
474 Failure to recognise the Kimberley Region’s historically-altered ecological state could place
475 the remaining biodiversity and ecosystems at greater risk than could be expected under truly ‘intact’
476 ecological conditions.
477
478 Acknowledgements
479 We acknowledge the traditional owners of the land on which our research was conducted. Sincere
480 thanks to Cecelia Myers, Susan Bradley and Doongan Station staff for facilitating and supporting our
20
S.E. Connor et al. Forgotten land-use impacts in NW Australia
481 research. Thanks to Janelle Stevenson, Peter Kershaw and an anonymous reviewer for constructive
482 reviews. Funding was provided by the Kimberley Foundation of Australia and the ARC
483 (DP140103591) project ‘Environmental Transformations linked to Early Human Occupation in
484 Northern Australia’. Research is also supported by the ARC Centre of Excellence for Australian
485 Biodiversity and Heritage (CE170100015).
486
487 Author contributions
488 SEC led the writing and conducted fieldwork
489 LS led the geochemical and statistical analyses and contributed to the manuscript
490 JT analysed the palynological data and led the literature review
491 SR analysed additional palynological data and oversaw pollen taxonomy
492 RLB contributed to the manuscript, particularly its introduction and discussion sections
493 AZ provided Pb-210 dating and expertise
494 SGH coordinated the research, conducted fieldwork and contributed to the manuscript
495
496 References
497 Abraham, V., Kuneš, P., Petr, L., Svitavská-Svobodová, H., Jamrichová, E., Švarcová, M.G. & Pokorný,
498 P. 2016. A pollen-based quantitative reconstruction of the Holocene vegetation updates a
499 perspective on the natural vegetation in the Czech Republic and Slovakia. Preslia 88: 409-
500 434.
21
S.E. Connor et al. Forgotten land-use impacts in NW Australia
501 Arnaud, F., Lignier, V., Revel, M., Desmet, M., Beck, C., Pourchet, M., Charlet, F., Trentesaux, A. &
502 Tribovillard, N. 2002. Flood and earthquake disturbance of 210Pb geochronology (Lake
503 Anterne, NW Alps). Terra Nova 14: 225-232.
504 Ash, A.J., McIvor, J.G., Mott, J.J. & Andrew, M.H. 1997. Building grass castles: integrating ecology and
505 management of Australia's tropical tallgrass rangelands. The Rangeland Journal 19: 123-144.
506 Aubert, M. 2012. A review of rock art dating in the Kimberley, Western Australia. Journal of
507 Archaeological Science 39: 573–577.
508 Banfai, D.S. & Bowman, D.M.J.S. 2005. Dynamics of a savanna-forest mosaic in the Australian
509 monsoon tropics inferred from stand structures and historical aerial photography. Australian
510 Journal of Botany 53: 185-194.
511 Barrett, R.L. & Barrett, M.D. 2011. Two new species of Triodia (Poaceae: Triodieae) from the
512 Kimberley region of Western Australia. Telopea 13: 57–67.
513 Barrett, R.L. & Barrett, M.D. 2015. Twenty-seven new species of vascular plants from Western
514 Australia. Nuytsia 26: 21–87.
515 Barrett, R.L. 2013. Solanum zoeae (Solanaceae), a new species of bush tomato from the North
516 Kimberley, Western Australia. Nuytsia 23: 5–21.
517 Barrett, R.L. 2015. Fifty new species of vascular plants from Western Australia—celebrating fifty
518 years of the Western Australian Botanic Garden at Kings Park. Nuytsia 26: 3–20.
519 Beard, J.S. 1976. The monsoon forests of the Admiralty Gulf, Western Australia. Vegetatio 31: 177-
520 192.
521 Beard, J.S., Beeston, G.R., Harvey, J.M., Hopkins, A.J.M. & Shepherd, D.P. 2013. The vegetation of
522 Western Australia at the 1:3,000,000 scale. Explanatory memoir. Second edition.
523 Conservation Science Western Australia 9: 1–152.
22
S.E. Connor et al. Forgotten land-use impacts in NW Australia
524 Bierman, P.R., Albretch, A., Bothner, M.H., Brown, E.T., Bullen, T.D., Gray, L. & Turpin, L. 1998.
525 Erosion, weathering and sedimentation. In: Kendall, C. & McDonnell, J.J. (eds.), Isotope
526 Tracers in Catchment Hydrology, pp. 647–678. Elsevier, Amsterdam.
527 Blaauw, M., 2010. Methods and code for 'classical' age-modelling of radiocarbon
528 sequences. Quaternary Geochronology 5: 512-518.
529 Bolton, G.C. 1953. A survey of the Kimberley pastoral industry from 1885 to the present. University of
530 Western Australia, Perth.
531 Bond, W.J., Woodward, F.I. & Midgley, G.F. 2005. The global distribution of ecosystems in a world
532 without fire. New Phytologist 165: 525-538.
533 Bowman, D.M.J.S. 1988. Stability amid turmoil?: towards an ecology of north Australian eucalypt
534 forests. Proceedings of the Ecological Society of Australia 15: 149-158.
535 Bowman, D.M.J.S. 2003. Australian landscape burning: a continental and evolutionary perspective.
536 In: Abbott, I. & Burrows, N. (eds.) Fire in ecosystems of south-west Western Australia:
537 impacts and management, pp. 107–118. Backhuys Publishers, Leiden, The Netherlands.
538 Bowman, D.M.J.S., Brown, G.K., Braby, M.F., Brown, J.R., Cook, L.G., Crisp, M.D., Ford, F., Haberle,
539 S.G., Hughes, J., Isagi, Y., Joseph, L., McBride, J., Nelson, G. & Ladiges, P.Y. 2010.
540 Biogeography of the Australian monsoon tropics. Journal of Biogeography 37: 201–216.
541 Bowman, D.M.J.S., MacDermott, H.J., Nichols, S.C. & Murphy, B.P. 2014. A grass–fire cycle eliminates
542 an obligate-seeding tree in a tropical savanna. Ecology and Evolution 4: 4185-4194
543 Bowman, D. M. J. S., and W. J. Panton. 1993. Decline of Callitris intratropica Baker, R.T. and Smith,
544 H.G. in the Northern Territory: implications for pre-European and post-European
545 colonization fire regimes. Journal of Biogeography 20: 373-381.
23
S.E. Connor et al. Forgotten land-use impacts in NW Australia
546 Bowman, D.M.J.S., Wilson, B.A. & Hooper, R.J. 1988. Response of Eucalyptus forest and woodland to
547 four fire regimes at Munmarlary Northern Territory Australia. Journal of Ecology 76: 215-
548 232.
549 Bunting, M.J., Farrell, M., Brostrom, A., Hjelle, K.L., Mazier, F., Middleton, R., Nielsen, A.B., Rushton,
550 E., Shaw, H. & Twiddle, C.L. 2013. Palynological perspectives on vegetation survey: a critical
551 step for model-based reconstruction of Quaternary land cover. Quaternary Science Reviews
552 82: 41-55.
553 Carrión, J.S. & Fernández, S. 2009. The survival of the ‘natural potential vegetation’ concept (or the
554 power of tradition). Journal of Biogeography 36: 2202-2203.
555 Case, M.F. & Staver, A.C. 2017. Fire prevents woody encroachment only at higher-than-historical
556 frequencies in a South African savanna. Journal of Applied Ecology 54: 955-962.
557 Chiarucci, A., Araújo, M.B., Decocq, G., Beierkuhnlein, C. & Fernández-Palacios, J.M. 2010. The
558 concept of potential natural vegetation: an epitaph? Journal of Vegetation Science 21: 1172–
559 1178.
560 Craig, A.B. 1997. A review of information on the effects of fire in relation to the management of
561 rangelands in the Kimberley high-rainfall zone. Tropical Grasslands 31: 161–187.
562 Davies, S., Lamb, H. & Roberts, S.J. 2015. Micro-XRF core scanning in palaeolimnology: recent
563 developments. In: Croudace I., Rothwell R. (eds) Micro-XRF Studies of Sediment Cores.
564 Developments in Paleoenvironmental Research, vol 17. Springer, Dordrecht. pp. 189-226.
565 Douglas, M.M., Setterfield, S.A., McGuinness, K. & Lake, P.S. 2015. The impact of fire on riparian
566 vegetation in Australia’s tropical savanna. Freshwater Science 34: 1351–1365.
567 Dunlop, C.R. & Webb, L.R. 1991. Flora and vegetation. In: Haynes, C.D., Ridpath, M.G. & Williams,
568 M.A.J. (eds) Monsoonal Australia: Landscape, Ecology and Man in the Northern Lowlands,
569 pp.41-60. A.A. Balkema, Rotterdam.
24
S.E. Connor et al. Forgotten land-use impacts in NW Australia
570 Early, M.S.B., Cameron, K.C. & Fraser, P.M. 1998. The fate of potassium, calcium, and magnesium in
571 simulated urine patches on irrigated dairy pasture soil. New Zealand Journal of Agricultural
572 Research 41: 117-124.
573 Ely, L.L., Webb, R.H. & Enzel, Y. 1992. Accuracy of post-bomb 137Cs and 14C in dating fluvial deposits.
574 Quaternary Research 38: 196-204.
575 Fensham, R.J & Skull, S.D. 1999. Before cattle: a comparative floristic study of Eucalyptus savanna
576 grazed by macropods and cattle in North Queensland, Australia. Biotropica 31: 37-47.
577 Field, E., McGowan, H.A., Moss, P.T. & Marx, S.K. 2017. A late Quaternary record of monsoon
578 variability in the northwest Kimberley, Australia. Quaternary International 449: 119-135.
579 Gillieson, D., Smith, D.I., Greenaway, M. & Ellaway, M. 1991. Flood history of the limestone ranges in
580 the Kimberley region, Western Australia. Applied Geography 11: 105-123.
581 Haberle, S.G., Tibby, J., Dimitriadis, S. & Heijnis, H. 2006. The impact of European occupation on
582 terrestrial and aquatic ecosystem dynamics in an Australian tropical rain forest. Journal of
583 Ecology 94: 987-1004.
584 Haynes, C.D., Ridpath, M.G. & Williams, M.A.J. (eds) Monsoonal Australia: Landscape, Ecology and
585 Man in the Northern Lowlands. A.A. Balkema, Rotterdam
586 Head, L. & Fullager, R. 1992. Palaeoecology and archaeology in the East Kimberley. Quaternary
587 Australasia 10: 27-31.
588 Henderson, M.K. & Keith, D.A. 2002. Correlation of burning and grazing indicators with composition
589 of woody understorey flora of dells in a temperate eucalypt forest. Austral Ecology 27: 121-
590 131.
591 Hiscock, P., O’Connor, S., Balme, J. & Maloney, T. 2016. World’s earliest ground-edge axe production
592 coincides with human colonisation of Australia. Australian Archaeology 82: 2-11.
25
S.E. Connor et al. Forgotten land-use impacts in NW Australia
593 Hnatiuk, R.J. & Kenneally, K.F. 1981. A survey of the vegetation and flora of Mitchell Plateau,
594 Kimberley, Western Australia. In: Biological survey of Mitchell Plateau and Admiralty Gulf,
595 Kimberley, Western Australia, pp. 13–94. Western Australian Museum, Perth.
596 Ireland, A.W. & Booth, R.K. 2012. Upland deforestation triggered an ecosystem state-shift in a kettle
597 peatland. Journal of Ecology 100, 586–596.
598 Johnson, C.N., Balmford, A., Brook, B.W., Buettel, J.C., Galeti, M., Guangchun, L. & Wilmshurst, J.M.
599 2017. Biodiversity losses and conservation responses in the Anthropocene. Science 356: 270-
600 275.
601 Jones, T.P. & Rowe, N.P. (eds) 1999. Fossil plants and spores: modern techniques. Geological Society,
602 London.
603 Kent, M. 2012. Vegetation Description and Analysis: a practical approach. 2nd ed. Wiley-Blackwell,
604 Chichester.
605 Kirkpatrick J. 1999. A Continent Transformed: Human Impact on the Natural Vegetation of Australia.
606 2nd ed. Oxford University Press, South Melbourne.
607 Larsen, A., May, J.-H., Moss, P. & Hacker, J. 2016. Could alluvial knickpoint retreat rather than fire
608 drive the loss of alluvial wet monsoon forest, tropical northern Australia? Earth Surface
609 Processes and Landforms 41: 1583–1594.
610 Legge, S., Murphy, S., Kingswood, R., Maher, B. & Swan, D. 2011. EcoFire: restoring the biodiversity
611 values of the Kimberley region by managing fire. Ecological Management and Restoration
612 12: 84–92.
613 Loidi, J. & Fernández-González, F. 2012. Potential natural vegetation: reburying or reboring? Journal
614 of Vegetation Science 23: 596-604.
26
S.E. Connor et al. Forgotten land-use impacts in NW Australia
615 Lonsdale, W.M. 1994. Inviting trouble: introduced pasture species in northern Australia. Australian
616 Journal of Ecology 19: 345-354.
617 López-Marino, L., Martínez Cortizas, A. & López Sáez, J.A. 2010. arly agriculture and
618 palaeoenvironmental history in the North of the Iberian Peninsula: a multi-proxy analysis of
619 the Monte Areo mire (Asturias, Spain). Journal of Archaeological Science 37: 1978-1988.
620 Macphail, M. & Stevenson, J. 2004. Fungal Spores in Archaeological Context. PalaeoWorks Technical
621 Report 3. Department of Archaeology and Natural History, Australian National University,
622 Canberra.
623 Mariani, M., Connor, S.E., Fletcher, M.-S., Theuerkauf, M., Kuneš, P., Jacobsen, G., Saunders, K.M. &
624 Zawadzki, A. 2017. How old is the Tasmanian cultural landscape? A test of landscape
625 openness using quantitative land-cover reconstructions. Journal of Biogeography doi:
626 10.1111/jbi.13040
627 Maslin, B.R., Barrett, M.D. & Barrett, R.L. 2013. A baker’s dozen of new wattles highlights significant
628 Acacia (Fabaceae: Mimosoideae) diversity and endemism in the north-west Kimberley
629 region of Western Australia. Nuytsia 23: 543–587.
630 McGonigal. D. 1990. The Australian Geographic Book of the Kimberley. Australian Geographic, Terrey
631 Hills NSW.
632 Moritz, C., Ens, E.J., Potter, S. & Catullo, R.A. 2013. The Australian monsoonal tropics: An opportunity
633 to protect unique biodiversity and secure benefits for Aboriginal communities. Pacific
634 Conservation Biology 19: 343-355.
635 Morris, J.L., Higuera, P.E., Haberle, S.G. and Whitlock, C. 2017. Modern pollen from small hollows
636 reflects Athrotaxis cupressoides density across a wildfire gradient in subalpine forests of the
637 Central Plateau, Tasmania, Australia. The Holocene doi: 10.1177/0959683617702228
27
S.E. Connor et al. Forgotten land-use impacts in NW Australia
638 Mucina, L. & Daniel, G. (eds) 2013. Vegetation mapping in Northern Kimberley, Western Australia.
639 Curtin University, Bentley.
640 Murphy, B.P., Lehmann, C.E.R., Russell-Smith, J. & Lawes, M.J. 2014. Fire regimes and woody
641 biomass dynamics in Australian savannas. Journal of Biogeography 41: 133–144.
642 Ondei, S., Prior, L.D., Williamson, G.J., Vigilante, T. & Bowman, D.M.J.S. 2017. Water, land, fire, and
643 forest: multi-scale determinants of rainforests in the Australian monsoon tropics. Ecology
644 and Evolution 7: 1592–1604.
645 O’Neill, A.L., Head, L.M. & Marthick, J.K. 1993. Integrating remote sensing and spatial analysis
646 techniques to compare Aboriginal and pastoral fire patterns in the East Kimberley, Australia.
647 Applied Geography 13: 67-85.
648 Parham, J., Deng, S., Raun, W. & Johnson, G. 2002. Long-term cattle manure application in soil I.
649 Effect on soil phosphorus levels, microbial biomass C, and dehydrogenase and phosphatase
650 activities. Biology and Fertility of Soils 35: 328-337,
651 Patterson, W.A., Edwards, K.J. & Maguire, D.J. 1987. Microscopic charcoal as a fossil indicator of fire.
652 Quaternary Science Reviews 6: 3-23.
653 Payne, A.L., Watson, I.W. & Novelly, P.E. 2004. Spectacular Recovery in the Ord River Catchment.
654 Department of Agriculture and Food, Western Australia, Perth. Report 17/2004
655 Pepper, M. & Keogh, J.S. 2014. Biogeography of the Kimberley, Western Australia: a review of
656 landscape evolution and biotic response in an ancient refugium. Journal of Biogeography 41:
657 1443–1455.
658 Petheram, R.J., Kok, B. & Bartlett-Torr, E. 1986. Plants of the Kimberley region of Western Australia.
659 Revised ed. Western Australian Department of Agriculture, South Perth.
28
S.E. Connor et al. Forgotten land-use impacts in NW Australia
660 Pillans, B. 2007. Pre-Quaternary landscape inheritance in Australia. Journal of Quaternary Science 22:
661 439-447.
662 Proske, U., Heslop, D. & Haberle, S. 2014. A Holocene record of coastal landscape dynamics in the
663 eastern Kimberley region, Australia. Journal of Quaternary Science 29: 163–174.
664 Prosser, I.P., Rutherfurd, I.D., Olley, J.M., Young, W.J., Wallbrink, P.J. & Moran, C.J. 2001. Large-scale
665 patterns of erosion and sediment transport in river networks, with examples from Australia.
666 Marine and Freshwater Research 52: 81-99.
667 Radford, I.J., Gibson, L.A., Corey, B., Carnes, K. & Fairman, R. 2015. Influence of fire mosaics, habitat
668 characteristics and cattle disturbance on mammals in fire-prone savanna landscapes of the
669 northern Kimberley. PLoS ONE 10(6): e0130721.
670 Rangan, H., Bell, K.L., Baum, D.A., Fowler, R., McConvell, P., Saunders, T., Spronck, S., Kull, C.A. &
671 Murphy, D.J. 2015. New genetic and linguistic analyses show ancient human influence on
672 baobab evolution and distribution in Australia. PLoS ONE 10(4): e0127582
673 Ritchie, D. 2009. Things fall apart: the end of an era of systematic Indigenous fire management. In:
674 Russell-Smith, J., Whitehead, P.J. & Cooke, P.M. (eds.) Culture, ecology and economy of fire
675 management in north Australian savannas: rekindling the Wurrk tradition, pp. 23–40. CSIRO
676 Publishing, Collingwood.
677 Rowe, C. 2006. The Australasian Pollen and Spore Atlas User Document. PalaeoWorks Technical
678 Report 8. Department of Archaeology and Natural History, Australian National University,
679 Canberra.
680 Rull, V. 2015. Long-term vegetation stability and the concept of potential natural vegetation in the
681 Neotropics. Journal of Vegetation Science 26: 603-607.
682 Russell-Smith, J., Lucas, D., Gapindi, M., Gunbunuka, B., Kapirigi, N., Namingum, G., Lucas, K.,
683 Giuliani, P. & Chaloupka, G. 1997. Aboriginal resource utilization and fire management
29
S.E. Connor et al. Forgotten land-use impacts in NW Australia
684 practice in Western Arnhem Land, Monsoonal Northern Australia: notes for prehistory,
685 lessons for the future. Human Ecology 25: 159-195.
686 Russell-Smith, J., Yates, C., Edwards, A., Allan, G.E., Cook, G.D., Cooke, P., Craig, R., Heath, B. &
687 Smith, R. 2003. Contemporary fire regimes of northern Australia, 1997–2001: change since
688 Aboriginal occupancy, challenges for sustainable management. International Journal of
689 Wildland Fire 12: 283–297.
690 Skroblin, A. & Legge, S. 2012. Influence of fine-scale habitat requirements and riparian degradation
691 on the distribution of the purple-crowned fairy-wren (Malurus coronatus coronatus) in
692 northern Australia. Austral Ecology 37: 874-884.
693 Somodi, I., Molnar, Z. & Ewald, J. 2012. Towards a more transparent use of the potential natural
694 vegetation concept – an answer to Chiarucci et al. Journal of Vegetation Science 23: 590–
695 595.
696 Silcock, J.L., Witt, G.B. & Fensham, R.J. 2016. A 150-year fire history of mulga (Acacia aneura F.
697 Muell. ex Benth.) dominated vegetation in semiarid Queensland, Australia. The Rangeland
698 Journal 38: 391–415.
699 Stevenson, J. and Haberle, S.G. (2005) Macro Charcoal Analysis: A modified technique used by the
700 Department of Archaeology and Natural History. PalaeoWorks Technical Report 5.
701 Department of Archaeology and Natural History, Australian National University, Canberra.
702 Strona, G., Mauri, A., Veech, J.A., Seufert, G., San Miguel-Ayanz, J. & Fattorini, S. 2016. Far from
703 naturalness: how much does spatial ecological structure of European tree assemblages
704 depart from Potential Natural Vegetation? PLoS ONE 11(12): e0165178.
705 Telford, K., Maher, W., Krikowa, F., & Foster, S. 2008. Measurement of total antimony and antimony
706 species in mine contaminated soils by ICPMS and HPLC-ICPMS. Journal of Environmental
707 Monitoring 10: 136-140.
30
S.E. Connor et al. Forgotten land-use impacts in NW Australia
708 Tng, D.Y.P., Murphy, B.P., Weber, E., Sanders, G., Williamson, G.J., Kemp, J. & Bowman, D.J.M.S.
709 2011. Humid tropical rain forest has expanded into eucalypt forest and savanna over the last
710 50 years. Ecology and Evolution 2: 34–45.
711 Trauernicht, C., Murphy, B.P., Tangalin, N. & Bowman, D.M.J.S. 2013. Cultural legacies, fire ecology,
712 and environmental change in the Stone Country of Arnhem Land and Kakadu National Park,
713 Australia. Ecology and Evolution 3: 286–297.
714 Tyler, I. 2016. Geology and Landforms of the Kimberley. Bush Books, Department of Parks and
715 Wildlife, Kensington WA.
716 van Geel, B. & Aptroot, A. 2006. Fossil ascomycetes in Quaternary deposits. Nova Hedwigia 82: 313-
717 329.
718 Vigilante, T. 2001. Analysis of explorers’ records of Aboriginal landscape burning in the Kimberley
719 Region of Western Australia. Australian Geographical Studies 39: 135-155.
720 Vigilante, T. & Bowman, D.M.J.S. 2004. Effects of fire history on the structure and floristic
721 composition of woody vegetation around Kalumburu, north Kimberley, Australia: a
722 landscape-scale natural experiment. Australian Journal of Botany 52: 381–404.
723 Warren, S.D., Thurow, T.L., Blackburn, W.H. & Garza, N.E. 1986. The influence of livestock trampling
724 under intensive rotation grazing on soil hydrologic characteristics. Journal of Range
725 Management 39: 491–495.
726 Western Australian Herbarium 2017. FloraBase – The Western Australian Flora. Department of Parks
727 and Wildlife. http://florabase.dpaw.wa.gov.au/ [accessed 31 May 2017]
728 Whitlock, C. & Larsen, C. 2001. Charcoal as a fire proxy. In: Smol, J.P., Birks, H.J.B. & Last, W.M. (eds)
729 Tracking Environmental Change using Lake Sediments, Volume 3: Terrestrial, Algal and
730 Siliceous Indicators, pp. 75-97. Kluwer, Dordrecht.
31
S.E. Connor et al. Forgotten land-use impacts in NW Australia
731 Wilson, B.R. 1981. General introduction. In: Biological survey of Mitchell Plateau and Admiralty Gulf,
732 Kimberley, Western Australia, pp. 1-11. Western Australian Museum, Perth
733 Wohl, E.E., Fuertsch, S.J. & Baker, V.R. 1994. Sedimentary records of late Holocene floods along the
734 Fitzroy and Margaret Rivers, Western Australia. Australian Journal of Earth Sciences 41: 273-
735 280.
736 Woinarski, J.C.Z., Mackey, B., Nix, H.A. & Traill, B.J. 2007. The nature of northern Australia - natural
737 values, ecological processes and future prospects. ANU Press, Canberra.
738 Wright, A.J., Edwards, R.J., van de Plasschea, O., Blaauw, M., Parnell, A.C., van der Borge, K., de
739 Jonge, A.F.M., Roec, H.M., Selbyf, K. & Blackg, K. 2017. Reconstructing the accumulation
740 history of a saltmarsh sediment core: which age-depth model is best? Quaternary
741 Geochronology 39: 35–67.
742 Yibarbuk, D., Whitehead, P.J., Russell-Smith, J., Jackson, D., Godjuwa, C., Fisher, A., Cooke, P.,
743 Choquenot, D. & Bowman, D.M.J.S. 2001. Fire ecology and Aboriginal land management in
744 central Arnhem Land, northern Australia: A tradition of ecosystem management. Journal of
745 Biogeography 28: 325-343.
746 Ziembicki, M.R., Woinarski, J.C.Z., Webb, J.K., Vanderduys, E., Tuft, K., Smith, J., Ritchie, E.G.,
747 Reardon, T.B., Radford, I.J. (…) & Burbidge, A.A. 2015. Stemming the tide: progress towards
748 resolving the causes of decline and implementing management responses for the
749 disappearing mammal fauna of northern Australia. Therya 6: 169-225.
750
32
S.E. Connor et al. Forgotten land-use impacts in NW Australia
751
752 Fig. 1 Location of the main study site (MP11A -
753 triangle) and surface sampling sites (circles) in
754 relation to regional vegetation structure, NW
755 Kimberley Region, Western Australia.
756 Vegetation units after Mucina and Daniel (2013)
757
758
759
760
761
762
763 Fig. 2 Age-depth model for the MP11A
764 sediment core, showing calibrated ages and
765 associated errors from 210Pb and 14C isotopic
766 dating, with smooth spline interpolation
33
S.E. Connor et al. Forgotten land-use impacts in NW Australia
767
768 Fig. 3 Summary of major changes in environmental proxies at site MP11A, Mitchell Plateau, Western
769 Australia (see Appendix S2 for complete palynological and geochemical proxy diagrams)
34
S.E. Connor et al. Forgotten land-use impacts in NW Australia
770
771 Fig. 4 Scatterplots of plant abundance and savanna density parameters versus pollen abundances in
772 surface samples
773
774 Supporting Information
775 Appendix S1. Map of sampling locations, NW Kimberley Region
35
S.E. Connor et al. Forgotten land-use impacts in NW Australia
776 777
778
36
S.E. Connor et al. Forgotten land-use impacts in NW Australia
779 Appendix S2. Complete pollen and geochemical diagrams, MP11A record
e a la e e e d il a a c e b e e ia a h rn c c n ll s e m a a n y a f e n b a h u ) s a e a d p q te s a d e l a d F e r o S e e a s e o e e / a u g / ila l e a r h a u a S c t e e e M 2 y te e m s s u a r r a e g p e a / a la a a s s c / e e Z a e a u p u ic iz la a o a a a e e i c a m r c i ra e m u e e s s e a a a a / n n e a s ( c p h l a r p r h a a li a a a ta n o ig a e ia n o a r c c s c ia a ra a h in s e e ia i e ia c . s c s c ie r o s o c i in a e e s a e t a c m t c a b to h e e ti a u a r r e t l i c c r m m r a p m lo te o s o a h o p a s b is r n a c iu le c fe n ia a s s o e r g p c p e o n th la G G c n e ra g a a a r y la h s o g e c e o rm p ri p in s n s b s m r a i n a n ia il a a n m e lo n c o n o a s r iz a l ll ra ra m e ra r r a ic e h u p s z o l ro g d o s a o s o li e u b y lit u m o i o c v v m h y r id a o a h ri z c i A d an e a ge e a s o e e h s t c c ri ty i o y n u r o d c a r u Depth cal. BP r r r l s r d b a e l u c e t t d b p r i n 4 n r g r t m o l p p p r e i m h n t o e o d r r l u s re h e o a a e o u im c r a ri u re p n e o a u a h re e P 3 a o la m a s m o r a y y y e d ra tr y te la p o o p s p o o e e e s T S H C MyrtaceaeC C T D R T A BanksiaG M T B B S A P C F E B R A M C C P L F S L A A CaryophyllaceaeC D H C C T P In T U MyriophyllumN P TrileteP typeO M6 B S P S P S C G N A 0 -70 5 -60 10 -50 -40 15 -30 20 -20 -10 25 0 10 30 20 30 40 35 50 60 70 40 80 90 100 45 110 120 780 130 20 40 60 80 100 20 40 20 20 20 20 40 20 20 20 40 20 20 20 20 20 40 60 80 10002000300040005000 781 782
s s b e u Depth BP cal. r re h a a g T S Herbs Li Be Na Mg Al Si P K C Ti V Cr Fe Mn Co Ni Cu Zn Ge As Se Rb Sr Zr Cd Sn Sb B Au H Pb Bi Th U 0 -70 5 -60 10 -50 -40 15 -30 20 -20 -10 25 0 10 30 20 30 40 35 50 60 40 70 80 90 100 45 110 120 783 130 20 40 60 80 100 20 100 200 300 1000200030004000 200 400 600 800 500 1000 1500 1000 2000 3000 200 400 600 500 100015002000 200 400 600 50 100 150 200 100 200 300 500100015002000 100 200 300 400 20 20 40 60 80 20 40 60 20 40 60 0 20 20 20 40 0 100 200 300 400 0 0 20 0 784 785 786
37
S.E. Connor et al. Forgotten land-use impacts in NW Australia
787 Appendix S3. Principal Components Analysis result
788
789
38
S.E. Connor et al. Forgotten land-use impacts in NW Australia
790 Appendix S4. Pollen and vegetation data from surface samples
Pollen Morgan Basalt MP34 MP10 MBS MP11A MP12A MP10A MP37 MP12 River Water Swamp Corymbia 17 18 50 77 5 29 10 8 0 22 Myrtaceae 41 53 102 111 6 32 31 26 0 50 Callitris 4 6 1 7 0 8 0 0 1 0 Casuarina 2 4 1 1 0 0 1 0 0 0 Terminalia 6 1 6 6 0 2 2 0 0 0 Dodonaea 0 0 4 1 0 2 2 4 0 9 Rubiaceae 0 0 2 0 0 0 1 0 0 0 Timonius 1 0 0 1 0 0 0 4 0 3 Acacia 0 0 1 0 0 0 0 1 1 3 Banksia 0 1 2 0 0 2 0 0 0 0 Grevillea 1 2 0 2 0 0 1 2 0 2 Breynia 1 0 2 0 0 0 0 1 0 0 Spermacoce 3 1 2 0 0 1 1 2 0 1 Antidesma ghaesembilla 0 0 1 1 0 2 1 0 0 1 Petalostigma 1 0 1 0 1 0 0 1 0 0 Celtis 1 0 0 0 0 0 0 0 0 0 Mallotus 0 1 0 0 0 0 0 0 0 0 Moraceae 0 1 0 0 0 0 0 0 0 0 Goodeniaceae 0 0 1 0 1 0 0 0 0 0 Fabaceae 0 0 3 0 0 1 7 2 0 1 Lamiaceae 0 0 0 0 2 0 0 0 0 3 Barringtonia 0 0 0 0 1 0 0 0 0 0 Malvaceae 1 1 1 1 0 0 1 0 0 0 Brachychiton 0 0 0 0 0 0 1 1 0 0 Euphorbiaceae 0 0 0 0 0 0 1 1 0 1 Solanaceae 0 0 0 0 0 0 3 0 0 0 Rapanea type 0 0 0 0 0 0 2 0 0 0 Gentianaceae/Canscora type 0 0 0 0 0 0 13 0 0 0 Pandanus 3 2 12 13 1 7 7 23 0 16 Jasminum 0 0 0 0 0 0 6 0 0 0 Loranthaceae 6 27 2 1 0 1 0 0 0 0 CP4 (reticulated) 0 0 0 0 0 1 0 0 0 0 Small Grass 32 53 19 21 47 16 22 27 2 17 Large Grass 39 49 28 21 8 33 4 15 14 34 Asteraceae 0 8 1 0 0 0 0 0 0 0
39
S.E. Connor et al. Forgotten land-use impacts in NW Australia
Asteraceae (liguliflorae) 0 0 1 0 0 0 0 0 0 0 Amaranthaceae 1 2 1 1 0 1 3 1 0 2 Haloragis 3 2 0 3 0 0 0 2 0 4 Cyperaceae 9 9 11 8 62 14 1 5 0 4 Cyperaceae 2 0 1 0 0 0 6 0 2 0 3 Indeterminate 0 7 0 2 3 2 3 5 0 2 Liliaceae 0 35 1 2 0 0 1 1 0 0 Persicaria/Zygophyllaceae 0 0 0 1 0 0 0 1 0 0 Crinum 1 0 1 0 0 0 0 0 0 0 Utricularia 0 0 1 0 0 0 0 0 0 0 Myriophyllum 2 2 1 7 2 3 1 3 1 5 Drosera 0 0 0 0 0 0 0 0 0 1 Nymphaceae 0 0 0 1 2 3 2 0 0 1 Nymphoides 0 0 2 0 0 0 1 0 0 3 Pteris sp. 0 0 0 0 1 0 0 2 0 1 Ophioglossum 0 0 1 0 0 0 0 2 0 1 Ceratopteris 0 0 0 0 0 0 0 0 0 1 Trilete type 6 0 0 0 0 0 1 0 0 6 0 Botyrococcus 5 9 37 32 4 73 0 3 0 3 Pseudoschizaea 1 0 2 1 15 0 2 1 0 6 Sporormiella 0 2 0 0 1 0 1 1 0 1 Podaspora 0 0 2 0 8 0 1 1 0 2 Sordaria 2 3 4 1 6 5 3 2 0 9 Cercophora 0 0 0 0 6 0 0 0 1 3 Delitschia 0 0 1 0 0 0 1 2 0 1 Neurospora 2 0 1 1 67 0 3 4 0 2 Gelasinospora 0 0 0 4 1 1 2 0 0 3 Vegetation parameters for surface samples Morgan Basalt MP34 MP10 MBS MP11A MP12A MP10A MP37 MP12 River Water Swamp Canopy cover % 18.75 24.75 40.3125 30 21.66667 51.25 28.87143 22 18.33333 28.87143 Basal Area 3.875 4.8 6.0625 7.166667 5.333333 6.1 4.918333 5.375 3.333333 4.918333 Height (average) 11.225 12.295 15.725 13.23333 13.53333 14.22 10.99133 12.6 11.9 10.99133 Tree volume estimate 20.475 29 43.5125 42.58333 25.7 42.22 24.51167 32.95 20.33333 24.51167 Corymbia cover (DAFOR) 3 3.35 4 5 3 5 3 2.666667 2.25 3 Eucalypt cover (DAFOR) 2.25 3.2 3.714286 1 2.5 1 1 0 1.25 1 Terminalia cover (DAFOR) 0.25 2.25 2.714286 0 0.5 0 0 0 1 0 Poaceae cover (DAFOR) 4.5 5 4.857143 5 4.5 5 5 2.666667 5 5 Pandanus cover (DAFOR) 0.25 0.25 1.285714 3 2 3 2 3 1 2 791
40
S.E. Connor et al. Forgotten land-use impacts in NW Australia
792 Appendix S5. Geochemical data from surface samples
mg/kg Morgan R Basalt MP34 MP10 MBS MP11A MP12A MP10A MP37 MP12 Water Swamp Li 9 20 15 17 18 17 15 3 20 12 Be 1 1 1 1 2 2 1 0 1 1 Na 416 823 220 238 972 181 136 22 992 118 Mg 1297 9168 3920 4018 2294 3762 4390 108 13920 3480 Al 28405 44515 53612 51856 27583 48798 43287 10242 78854 34060 Si 582 523 165 203 300 312 178 239 192 281 P 133 347 100 518 896 298 172 25 67 66 K 671 1725 809 655 1832 726 484 145 1172 360 Ca 600 24664 2604 580 3412 1749 978 36 10655 715 Ti 300 748 464 490 52 495 510 103 624 439 V 102 293 118 124 17 117 121 37 226 95 Cr 50 53 66 192 17 198 196 24 75 134 Fe 29804 77635 28896 56956 8180 22927 33336 1224 58074 23757 Mn 476 6048 268 134 293 464 286 3 791 107 Co 15 48 18 15 13 20 21 1 29 14 Ni 15 40 34 45 8 50 48 4 45 39 Cu 16 63 44 32 23 40 32 5 69 25 Zn 13 56 35 34 17 38 33 3 51 26 Ge 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 As 2 7 2 4 1 2 3 1 3 2 Se 0.001 2 0.001 1 0.001 1 0.001 1 1 0.001 Rb 7 20 15 9 13 7 6 2 30 4 Sr 10 96 20 12 34 21 14 3 70 9 Zr 17 43 29 27 5 28 26 3 48 19 Cd 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 Sn 1 1 1 1 1 1 1 0.001 1 0.001 Sb 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 Ba 58 530 91 100 90 173 87 13 154 61 Au 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 Hg 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 Pb 4 14 8 12 6 13 11 1 12 8 Bi 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 Th 5 4 6 7 6 5 5 2 9 4 U 1 1 1 3 5 2 2 1 2 1
41
S.E. Connor et al. Forgotten land-use impacts in NW Australia
793 Appendix S6. Aerial photographs of study area in 1949 and 2017
794
42
Minerva Access is the Institutional Repository of The University of Melbourne
Author/s: Connor, SE; Schneider, L; Trezise, J; Rule, S; Barrett, RL; Zawadzki, A; Haberle, SG
Title: Forgotten impacts of European land-use on riparian and savanna vegetation in northwest Australia
Date: 2018-05-01
Citation: Connor, S. E., Schneider, L., Trezise, J., Rule, S., Barrett, R. L., Zawadzki, A. & Haberle, S. G. (2018). Forgotten impacts of European land-use on riparian and savanna vegetation in northwest Australia. JOURNAL OF VEGETATION SCIENCE, 29 (3), pp.427-437. https://doi.org/10.1111/jvs.12591.
Persistent Link: http://hdl.handle.net/11343/213924
File Description: Accepted version