|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 ECOLOGY AND CONSERVATION OF THE PURPLE CROWNED FAIRY WREN ( Malurus coronatus coronatus ) IN THE NORTHERN TERRITORY, AUSTRALIA By ANNAMARIA VAN DOORN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORI DA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007
2 2007 Annamaria Van Doorn
3 This work is dedicated to my son Lex
4 ACKNOWLEDGMENTS My work on the Purple crowned fairy wren in the Victoria River District would not have been possible without the help of many individuals and organizations My research would not have been possible without the commitment of Sarah Kerin who pursued the initial funding to undertake this work. My work was primarily funded by the Victoria River District Conservation Association and Tropical Savannas Cooperative Research Centre Additional support was received from Gregory National Park, the Key Centre for Tropical Wildlife Management and the Northern Territory Department of Natural Resources, Environment and the Arts. I am very grateful to my chair Patricia Werner, CoChair Mike Moulton and my committee members; Mike Avery, Alison Fox, and David Steadman. Patricia Werner provided constant support throughout my research and writing. I thank Mike Moulton for his support, especially during the earlier years of planning my research. My committee members greatly supported my research and showed enormous patience during my long absence from the University of Florida. In addition to my official committee, I would not have been able to complete my research or writing without the guidance, support, and encouragement of John Woinarski. I thank the administrative staff of the Department of Wildlif e Ecology and Conservation for all their help, especially Delores Tillman. I thank Craig Hempel for producing the maps. I also thank Barry Brook and Peter Whitehead for their support of my research. My field work was supported by many individuals includin g Darryl Hill, Sarah Kerin, and Murray Fuller who all provided invaluable insight into land management in the V ictoria R iver D istrict I received a lot of field support from the rangers of Gregory National Park who were always enthusiastic and a tremendou s help. Henri Robert spent a great deal of time with me conducting field work and his skill and sense of humor made him irreplaceable. I thank Luke Playford, Hafwen Pearce, Matt Reetz, and Vera Bosward for their assistance in the field. It
5 would not hav e been possible to conduct my research without the support of the stakeholders in the VRD including Auvergne Station, Coolibah Crocodile Farm, Fitzroy Station, Gregory National Park, and Victoria River Downs. The Victoria River Roadhouse, Helimuster and C oolibah Crocodile Farm (Blue and Janelle Pugh) were kind enough to let me spend many nights there during my field seasons. My parents have alway s been supportive during all my years of study and subsequent research and without their encourageme nt I would never have been able to reach my goals. Finally, I am very grateful to my partner Luke Playford. Without his help and strong support it would have been impossible for me to have completed my dissertation and for that I will always be grateful.
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 8 LIST OF FIGURES ................................ ................................ ................................ ......................... 9 ABSTRACT ................................ ................................ ................................ ................................ ... 12 CHAPTER 1 GENERAL INTRODUCTION ................................ ................................ .............................. 15 Background ................................ ................................ ................................ ............................. 15 Study Aims and Thesis L ayout ................................ ................................ ............................... 23 Obstacles and C onstraints ................................ ................................ ................................ ....... 24 2 THE VICTORIA RIVE R D ISTRICT AND THE PURP LE CROWNED FAIRY WREN; NATURAL ENVIRO NMENT AND HUMAN DIME NSIONS WITHIN THE STUDY AREA ................................ ................................ ................................ ....................... 30 The Victoria River District ................................ ................................ ................................ ..... 30 Natural Environment ................................ ................................ ................................ ....... 30 Human Dimensions ................................ ................................ ................................ ......... 32 The Purple Crowned Fairy wren in the Victoria River District ................................ ............. 34 Primary H abitat of the Purple crowned Fairy wren in the Victoria River District ......... 35 Study S ites ................................ ................................ ................................ .............................. 38 Selected Study Sites and C onditions ................................ ................................ ............... 38 3 ECOLOGY OF THE PURPL E CROWNED FAIRY WREN ................................ ............... 52 Introduction ................................ ................................ ................................ ............................. 52 Field Methods ................................ ................................ ................................ ......................... 55 Terminology ................................ ................................ ................................ .................... 55 Captures and Banding ................................ ................................ ................................ ...... 56 Morphology and Plumage ................................ ................................ ............................... 57 Group Size and Composition ................................ ................................ ........................... 58 Territory Shape and Size ................................ ................................ ................................ 59 Foraging Behavior ................................ ................................ ................................ ........... 60 Breeding Biology ................................ ................................ ................................ ............. 61 Annual Sur vival ................................ ................................ ................................ ............... 63 Results ................................ ................................ ................................ ................................ ..... 63 Captures and Banding ................................ ................................ ................................ ...... 63 Morphology and Pluma ge ................................ ................................ ............................... 64 Group Size and Composition ................................ ................................ ........................... 66
7 Territory Shape and Size ................................ ................................ ................................ 67 Foraging Behavior ................................ ................................ ................................ ........... 67 Breeding Biology ................................ ................................ ................................ ............. 69 Annual Survival ................................ ................................ ................................ ............... 73 Disc ussion ................................ ................................ ................................ ............................... 73 Ecology of the Purple crowned Fairy wren in The Victoria River District .................... 73 Comparisons between Victoria and Drysdale R iver Populations ................................ .... 76 4 RESPONSE OF TH E PURPLE CROWNED FAIRY WREN TO DISTURBANCE F ACTORS ................................ ................................ ................................ .............................. 93 Introduction ................................ ................................ ................................ ............................. 93 Methods ................................ ................................ ................................ ................................ .. 95 Vegetation and Site C ondition ................................ ................................ ......................... 95 Fire ................................ ................................ ................................ ................................ ... 97 Variation A mong S ites ................................ ................................ ................................ .... 97 Results ................................ ................................ ................................ ................................ ..... 98 Vegetation and Site C ondition ................................ ................................ ......................... 98 Fire ................................ ................................ ................................ ................................ 103 Variation Among S ites ................................ ................................ ................................ .. 103 Discussion ................................ ................................ ................................ ............................. 105 5 LANDSCAPE SCALE ANAL YSES OF HABITAT AND THE IMPLICATIONS FOR CONSERVATION OF THE PURPLE CROWNED FAIRY WREN ................................ 115 Introduction ................................ ................................ ................................ ........................... 115 Canegrass ( Chionachne cyathopoda ) D istribution ................................ ............................... 117 Methods ................................ ................................ ................................ ......................... 117 Results ................................ ................................ ................................ ........................... 119 Conservation and Management Recommendations ................................ .............................. 122 Grazing ................................ ................................ ................................ .......................... 123 Weeds ................................ ................................ ................................ ............................ 124 Fire ................................ ................................ ................................ ................................ 124 Fragmentation ................................ ................................ ................................ ................ 125 Stakeholder I nvolvement ................................ ................................ ............................... 125 Future Research D irections ................................ ................................ ........................... 127 Conclusion ................................ ................................ ................................ ............................ 128 .......... 134 A Bird S pecies ................................ ................................ ................................ .......................... 134 B Plant Species ................................ ................................ ................................ ......................... 137 REFERENCES ................................ ................................ ................................ ............................ 138 BIOGRAPHICAL SKETC H ................................ ................................ ................................ ....... 149
8 LIST OF TABLES Table page 2 1 Relationship between parent material, soils and topog raphy in the VRD. ....................... 43 2 2 Dominant vegetati on in the VRD ................................ ................................ ...................... 44 2 3 Characteristics of Chionachne cyathopoda and Mnesithea rottboelioides ...................... 45 2 4 Tenure and yearly conditions of field sites. ................................ ................................ ....... 46 3 1 Number of PCFW capt ures by year (2001 2003) ................................ ............................. 80 3 2 Morph ological measurements of captured individuals captured f rom 2001 2003 ........... 80 3 3 Summary table of comparisons in differences between male, female and juvenile morphological measurements. ................................ ................................ ........................... 81 3 4 Group size and composition ................................ ................................ .............................. 81 3 5 N est location and measurements ................................ ................................ ........................ 82 3 6 Fate of nests with eggs ................................ ................................ ................................ ...... 82 4 1 Pair wise differences between sites in the set of all env ironmental variables measured ................................ ................................ ................................ .......................... 110 4 2 Purple crowned f airy wren captures per site. ................................ ................................ .. 110 5 1 Canegrass ( Chionachne cyathopoda ) distribution, patch size and patch isolation by tenure ................................ ................................ ................................ ................................ 133
9 L IST OF FIGURES Figure page 1 1 Distribution f the two subspecies of Malurus coronatus ................................ ................... 26 1 2 Major drainage lines with in the distribution of the two species of the p urple crowned fairy wren (PCFW) Malurus coronatus ................................ ................................ ............ 27 1 3 Pandanus ( Pandanus aquaticus ). ................................ ................................ ....................... 28 1 4 Canegrass ( Chionachne cyathopoda ) ................................ ................................ ................ 29 2 1 Location of the Victoria River District and as sociated land tenures ................................ 47 2 2 Mean annu al rainfall in the study region ................................ ................................ ........... 48 2 3 Mean monthly rainfall in Timber Creek. ................................ ................................ ........... 49 2 4 Mean daily maximum temperature Timber Creek. ................................ ............................ 49 2 5 Broad vegetation classification of the study region. ................................ .......................... 50 2 6 Location of study sites. ................................ ................................ ................................ ...... 51 3 1 Adult male PCFW ................................ ................................ ................................ .............. 83 3 2 Adult female PCFW ................................ ................................ ................................ ........... 83 3 3 Eclipse male PCFW. ................................ ................................ ................................ .......... 84 3 4 Juv enile PCFW ................................ ................................ ................................ .................. 84 3 5 Location and shape of territories at Coolibah Crocodile Farm and Fitzroy Station .......... 85 3 6 Location and shape of territories within Gregory National Park (GNP1 and GNP2). ...... 86 3 7 Time spent foraging (%) in various substrates ................................ ................................ .. 87 3 8 Male and female foraging preferences ................................ ................................ ............... 88 3 9 Foraging behavior ................................ ................................ ................................ .............. 89 3 10 Prey i tems captured ................................ ................................ ................................ ............ 90 3 11 Number of nests found per month ................................ ................................ .................... 91 3 12 Nest entrance height ................................ ................................ ................................ ........... 92
10 4 1 Positioning example of 50 meter transects within territories. ................................ .......... 110 4 2 Frequency (%) of grazed Chionachne at individual study sites. ................................ ..... 111 4 3 Wallaby ( Macropus agilis ) and c ow ( Bos Indicus ) evidence at individual study sites. .. 112 4 4 Chionachne (c anegrass) stands in May 2002 at Coolibah before cattle were introduced. ................................ ................................ ................................ ....................... 113 4 5 Chionachne (c anegrass) stands, October 2002, four months after cattle arrived at Coolibah ................................ ................................ ................................ ........................... 114 4 6 Chionachne (c anegrass) stands after heavy grazing at Coolibah ................................ ..... 115 4 7 Groundcover frequencies (%) at study sites. ................................ ................................ ... 116 4 8 Freq uency of Chionachne cover (%) at individual study sites ................................ ........ 117 4 9 Height frequency distribution (%) of Chionachne at individual study sites .................... 118 4 10 Frequency (%) of Bareground (BG) at individual study sites. ................................ ........ 119 4 11 Frequency (%) of Bareground w/ debris (BGD) at individual study sites ....................... 120 4 12a Frequency (%) of Xanthium at individual study sites ................................ ...................... 121 4 12b River bank at Dashwood Crossing showing dense cover of Xanthium ........................... 122 4 13 Frequency (%) of total weeds ( Xanthium Ricinus and Passiflora ) at individual study sites ................................ ................................ ................................ ................................ .. 123 4 14 Canopy cover, expressed as a percentage at in dividual study sites ................................ 124 4 15 Canopy cover (species) at combined study sites ................................ .............................. 125 4 16 Dominant canopy cover composition (%) at individual study sites ................................ 126 4 17 Ordination of transects, based on their overall similarity in habitat variables measured ................................ ................................ ................................ .......................... 127 4 1 8a Fire at GNP1, 22 October, 2001. ................................ ................................ ..................... 128 4 18b Three months post fire at GNP1, January 31, 2002. ................................ ........................ 129 4 19 Average group sizes (adu lt birds only) at study sites ................................ ...................... 130 4 20 Adult survivorship at individual study sites ................................ ................................ ..... 131
11 4 21 Foraging preference illustrated by time spent foraging in relation to substrate abundance ................................ ................................ ................................ ........................ 132 5 1 Map showing width of c anegrass ( Chionachne cyathopoda ) along the Victoria River.. ................................ ................................ ................................ ............................... 130 5 2 Map showing the density of c anegrass ( Chionachne cyathopoda ) along the Victoria River. ................................ ................................ ................................ ................................ 131 5 3 Canegrass ( Chionachne cyathopoda ) patch size and distance to next patch ................... 132
12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ECOLOGY AND CONSERVA TION OF THE PURPLE CROWNED FAIRY WREN ( Malurus coronatus coronatus ) IN THE NORTHERN TE RRITORY, AUSTRALIA By Annamaria van Doorn August 2007 Chair: Partricia Werner Cochair: Michael Moulton Major : Wildlife Ecology and Conservation The purple crowned f airy wren ( Mal urus coronatus coronatus ) is a small endemic habitat specialist restricted to riparian vegetation in northern Australia. Its range extends from the Kimberley in Western Australia east to the Victoria River District in the Northern Territory. One of two sub species, M. c. coronatus is a threatened species with a fragmented distribution attributed to habitat loss and degradation. Along the Victoria River the purple crowned fairy wren inhabits stands of canegrass or river g rass ( Chionachne cyathopoda ). The re sults from my study show that, in th e Victoria River District, the purple crowned f airy wren is highly dependent on Chionachne for cover, foraging and breeding and was not found in any other dominant vegetation. This habitat association is in contrast to the si ngle previous study, where the purple crowned f airy wren was primarily depende nt on the aquatic p andanus ( Pandanus aquaticus ). Purple crowed f airy wrens live in small family groups and groups often remained in the same territory for consecutive years Territories are best described as polygons with an average size of 0.41ha. The average group size was 2.6, and although groups comprising a single breeding male and female were the most common (50%), groups with a single additional helper
13 were also com mon (41%). Purple crowned f airy wrens often foraged in loose family groups, usually within 2 met ers of the ground, and favored river grass and freshwater m angrove ( Barringtonia acutangula ) as foraging substrates. My study found purple crowned f airy wren s had a long breeding season of 22 weeks, the start of which coincided with the end of the wet season with a peak in May. During my study 65 nests were found. The majority of nests (85%) were located in river grass with an av erage nest entrance height o f 40.1cm. The average clutch size was 2.9 with 88% of clutches consisting of three eggs. Incubation was estimated at approximately 14 days and the nestling period 10 days. A high nest predation rate of at least 51% was found during my study with most p redation events occurring at the egg stage. Overall, the mean number of fledglings produced by a breeding female was 1.07 per year. One of the primary distu rbances investigated during my study was cattle grazing. Adult survivorship was heavily influenc ed by intense grazing resulting in a 74% loss of adult birds within one year in addition to a significant reduction in group size. Such a high mortality rate is very alarming for a species where the annual adult survival rates can be as high as 90%, with an average of 61%. Adult mortality is likely attribu ted to a loss of cover as mean r iver g rass height was greatly reduced and the percentage of bare ground increased at grazed sites. .A helicopter survey (encompassing 170 km of river frontage) along the Vi ctoria River revealed t hat the primary habitat of the purple crowned f airy wren is extremely fragmented and patch size extremely variable ranging from 0.23 hectares to 244 hectares with an average of 24.5 hectares. Distance between patches ranged from 0.3 km to 41 km with an average distance of 1.37 kilometers.
14 The population of the p urple cro wned f airy wren in the Victoria River District is characterized by a fragmented distribution and high sensitivity to intense grazing pressure. High nest predation rat es found during identify the need further investigation to determine primary nest predators. My study was conducted during high rainfall years thus cons study indicate that conservation efforts need to f ocus on reducing grazing pressure and increasing habitat connectivity to ensure the long term survival of this species in the Northern Territory.
15 CHAPTER 1 GENERAL INTRODUCTION This chapter provides a brief introduction to the biology of the species forming the focus of my study This chapter also introduces, justifies and contextualizes the primary research questions addressed in the remainder of the thesis. Background The Maluridae is a family of birds endemic to Australia and New Guinea; the fam ily includes five genera, three of which are found in Australia (Rowley and Russell, 1997). The fairy wrens ( Malurus ) have been an overall well studied genus in Australia, and have provided a rich and fertile research area for Australian ornithology lead ing to much of the information on cooperative breeding, longev ity, and brood parasitism ( Schodde, 1982; Tidemann, 1986; Rowley et al. 1991; Rowley and Russell, 1995; Brooker and Brooker, 1996; Rowley and Russell, 2002). For some of the southern Malurus s pecies, banded populations have been maintained for more than 13 years providing invaluable insight into the adaptation of this family to the highly variable environmental conditions of the Australian continent, and exemplifying life history traits of ende mic Australian passerines ( Rowley et al 1991; Brooker and Brooker, 1994; Russell and Rowley, 1998). The Australian continent has been subjected to severe landscape alteration since the arrival of first European settlers with the more populated, southern regions particularly affected (e.g. Ford et al. 2001; Hobbs, 2005). In this context, fairy wrens have been the subject of research examining the effects of habitat pertu rbations and fragmentation ( Brooker and Brooker, 1994; Russell and Rowley, 1998). As small territorial understory species with long lifespans, fairy wrens lend themselves well to any studies examining the effects of habitat alterations.
16 Despite the large number of studies on fairy wrens, the research focus has been very uneven and there ha s been little previous attention given to the only threatened Malurus species. There are two recognized subspecies of the p urple crowned fairy wren showing vari ation in plumage and size; the w estern ( M. c. coronatus ) and the Carpentarian ( M. c. macgillivra yi ) The two subspecies are separated by approximately 200 kilometers of unsuitable ha bitat (Figs. 1 1 and 1 2). My study does not directly consider the latter subspecies. The w estern race of the p urple crowned fairy wren ( M. c. coronatus ), hereafter PCF W, is currently listed as vulnerable by the Australian government (Environment Protection and Biodiversity Conservation Act, 1999). It has a patchy distribution in the Kimberley of Western Australia and the adjacent Victoria River District in the Northern Territory. The PCFW is a riparian habitat specialist that has experienced a substantial decline and is now exceedingly fragmented in distribution (Smith and Johnstone, 1977; Rowley, 1993; Garnett and Crowley, 2000). Due to its specific habitat requireme nts and threatened status, the PCFW is a prime candidate for studying the impacts of some of the primary threats to biodiversity in northern Australia, in particular in riparian zones. The status, ecology and conservation management of the PCFW are the su bject of my study The PCFW has been the focus of only one previous detailed study (Rowley, 1993; Rowley and Russell, 1993). That study was based on the Drysdale River in Western Australia (Fig. 1 2) in riparia n habitat dominated by aquatic p andanus ( Pan danus aquaticus ), a woody palm like shrub that may form dense thickets, typically with foliage 2 5 m high, that grows in and immediately adjacent to watercourses (Fig.1 3). In that study, wrens were seldom encountered more than 10 meters away from the riv er. Thus far, knowledge of the PCFW in the Northern Territory is restricted to only a few preliminary surveys conducted along the Victoria River
17 (Boekel, 1979; Parks and Wildlife Commission of the Northern Territory). As a threatened species with a fragm ented distribution this represents a serious gap in knowledge and one which my study aims to address. in the Kimberley, on the Victoria River and its tributaries, par ticularly in the lower reaches, the PCFW shows a strong aff inity for habitat dominated by River G rass Chionachne cyathopoda (more commonly referred to as c anegrass) [a tall (to 2 5m. robust grass forming almost monospecific and very dense stands: Fig. 1 3] and to a lesser extent northern c anegrass Mnesithea rottboellioides This apparent marked difference in habitat relationships affords a particular opportunity for comparative ecological study. Further, these apparently different requirements may necess itate site specific management guidelines, rather than perhaps inappropriate extrapolations for a differe nt environmental setting Purple crowned f airy wrens are territorial year round residents, living in either pairs or small family groups often maintain ing the same territory year after year (Rowley and Russell, 1993). Although socially monogamous, it is likely they are sexually promiscuous as are many other fairy wrens although genetic data confirming this is still lacking. Like other Malurus species, they are cooperative breeders. Helpers are often male progeny that remain in the natal territory after reaching maturity, while females usually disperse before the start of the following breeding season (Rowley and Russell, 1993). As a small endemic t erritorial species restricted to riparian habitats in northern Australia, the PCFW has without doubt been witness to many landscape changes since the first European settlers arrived in the 1880s (Rowley, 1993). As has been the case with many bird species in northern Australia, the potential response of the PCFW to disturbances is poorly understood, but
18 it is likely that the changes wrought by European colonists have been generally unfavorable to PCFWs (Thomas 1964; Rowley, 1993). This may be particularly so because the PCFW is a riparian specialist and European use of land has predominantly focused on rivers and riverine environments (Sattler, 1993). Relatively little information exists about the historical distribution of the PCFW in the Northern Terri tory. This can mostly be attributed to the remoteness of its range, the relative recency of European settlement, and the sparse human population of the region. To this day much of its range is largely inaccessible due to the topography and lack of roads. Even along the Victoria River, only a few areas are reliably accessible by road, and even these may be unreachable in the monsoonal wet season. Access by boat is hindered by the numerous rock bars that are dispersed throughout the river system, and river access is also complicated by the abundance of large saltwater crocodiles ( Crocodylus porosus ). The scarcity of accurate data on the historical distribution of the PCFW substantially constrains the assessment of population trends since settlement, and of current trends in the population. Although there are some anecdotal and observational data on the landscape changes that have occurred in riparian habitats since European settlement, until very recently there was little quantitative data available on t hese changes ( Lewis, 2002; Start and Handasyde, 2002; Sharp and Whittaker, 2003; Sharp and Bowman, 2004). In northern Australia, agriculture has been generally less intensive and the landscape remains more sparsely populated than in most other regions of Australia (Ridpath, 198 5). However, pastoralism ( livestock grazing) and changes to fire regimes have both been implicated in changes to faunal composition in northern Australia (Burbidge and McKenzie, 1989; Barnard, 1925; Franklin, 1998). Current resear ch is revealing that extensive and often subtle changes in
19 the landscape have altered the distribution and abundance of many species and that these changes are extremely compl ex and often interrelated ( Blakers et al 1984; Franklin, 1999; Fraser, 2000; Wil liams et al ., 2002). Further, from studies of the role of multiple disturbances, and in particular exogenous ones ( Dobkin et al ., 1998; Woinarski and Ash, 2002; Krueper et al ., 2003), it is clear that multiple and often interrelated disturbances are at th e root of these changes in observed patterns, making it difficult to identify one single causal factor. Riparian ecosystems encompass many diverse values including economic, social and biological (Gregory et al ., 1991; Malanson, 1993). Rivers and associate d riparian ecosystems play an integral role in the maintenance of healthy regional ecosystems, and have been referred et al ., 1991; Knopf and Samson, 1994). Riparian landscapes, as an ecotone between th e terrestrial and aquatic area, represent a zone of high biodiversity where habitat alterations can have extensive impacts ( Knopf et al ., 1988; Naiman et al. 1993; Saab, 1999; Krueper et al ., 2003; Scott et al ., 2003). Riparian zones provide habitat for both residential and migratory species and may provide a refuge or act as a corridor for dispersal for non obligate rip arian terrestrial species ( Machtans et al ., 1996; Bentley and Catterall, 1997; Fisher and Goldney, 1997; Skagen et al ., 1998; Woinarski et al ., 2000). Riparian corridors can also be the linkage between multiple habitats in a mosaic of landscapes, enabling species to move through the landscape accessing suitable habitat (Machtans et al ., 1996; Skagen et al ., 1998). This is especially import ant in northern Australia with its extended dry season, where the riparian vegetation can be of great importance to a large number of species when suitable habitats may be isolated from one another Furthermore, in northern Australia, riparian corridors m ay allow species to extend their range into lower rainfall areas (Woinarski et al ., 2000). For these reasons, it is not surprising that studies have shown that avian
20 diversity and abundances in riparian zones tend to be highly dynamic, particularly influe nced by seasonal cycles (Knopf and Samson, 1994; Woinarski et al. 2000). Riparian habitats are e xposed to both endogenous (natural) and exogenous ( unnatural) disturbances at more intense rates than surrounding h abitats ( Planty Tabacchi et al 1996). E ndogenous disturbances, such as flooding and other changes in water flow, are routine events in most river systems in northern Australia. However, exogenous disturbances, such as weed invasions and grazing, have been experienced by these systems only fair ly recently from an evolutionary perspective. In some instances disturbances can be considered both exogenous and endogenous depending on the circumstances. For example, fire caused by a lighting strike in the early wet season can be considered an endoge nous disturbance, although it can be classified as an exogenous disturbance when it is set in the middl for cattle. One of the primary disturbances being investigate d in n orthern Australia is fire and changes res ulting fro m altered burning regimes ( Braithwaite, 1987; Bowman, 1998; Yibarbuk et al ., 2001). Response to fire is dependent on fire regime (frequency, intensity, and timing) and ecological characteristics, thus research on multiple species and assemblage s has demonstrated decidedly differently responses to fire (Anderson, 1991; Griffiths and Christian, 1996; Williams et al ., 2002). Altered burning regimes appear to be at the core of major changes in the avian assemblage in Australia (Garnett, 1992; Woina rski, 1990; Woinarski and Recher, 1997; Woinarski and Ash, 2002), especia lly among granivorous birds in n orthern Australia (Franklin, 1999; Frase r, 2000). For the endangered gouldian f inch ( Erythrura gouldiae ), an increase in late season (i.e. hot) fires leading to a decline in suitable nesting trees was identified as one of the factors implicated in the widespread decline of this species in northern Australia (Tidemann,
21 1996). More recently, research has demonstrated that fire impacts gouldian f inch pop ulations by affecting seed availability in the short term and altered floristic structure in the long term (Price et al in press). Introduced species continue to be one of the primary threats to biodiversity throughout the world and Australia has not esc aped this problem (Soul, 1990; Vitousek et al. 1996; Pimental et al ., 2000). In Australia, an average of 10 plant species become naturalized each year (Groves, 1997), many of which have been intentional i ntroductions ( Lonsdale, 1994). These introduced plants have invaded, and continue to invade, native landscapes where they can have devastating results and in extreme cases alter ecosystem function (Braithwaite et al. 1989; Gordon, 1998). For example, in northern Australia, the introduced grass Andropo gon gayanus has become a threat to biodiversity in the region by modifying vegetation composition due to the effect it has had on the grass fire cycle (Rossiter et al. 2003). Riparian zones in particular may be more susceptible to plant invasions than ad joining landscapes due to combination of factors that include ease of propagule spread, habitat continuity, disturbed soil, and high nutrient levels (Griffin et al ., 1989; Malanson, 1993; Planty Tabacchi et al ., 1996). In northern Australia, pastoralism may have had a profound influence on ecosystem composition, function and structure (Archer et al ., 1988; Whitehead, 2000; Sharp and Whittaker, 2003). Riparian systems may be particularly affected: pressure is greatest during the extended dry season when c attle and feral animals congregate along the riparian corridors (Myers et al ., 2001; Lewis, 2002). Grazing has been implicated in influencing avian composition in riparian zones, most frequently resulting in an overall redu ction in species richness ( Taylo r, 1986; Saab et al ., 1995; Krueper et al ., 2003; Scott et al ., 2003). In addition, Ammon and Stacey (1997) recorded higher avian nest predation rates at grazed sites compared to ungrazed sites. Grazing
22 impacts may be particularly severe for avian specie s that occupy the lower levels of the vegetation rather than canopy dwelling species as structure of the understory can be greatly modified by grazing (Schultz and Leininger, 1990). For instance, foraging height was found to be a good indicator for determ ining avian species response to grazing, with the majority of ground dwelling species showing a decline with increased grazing pressure (Martin and Possingham, 2005). In many cases, it becomes hard to attribute major landscape changes to one single type of disturbance, and potential confounding effects make identifying causes even more difficult. Weed invasion, grazing and fire all seem to play a role in shaping the shifting vegetation composition within the riparian landscape. For example, the increase in woody vegetation in the riparian zone within the tidal portion of the Victoria River has been attributed to a lack of fire which has arisen as a consequence of extreme overgrazing (Sharp and Whittaker, 2003). In this instance both fire and grazing are at the crux of the observed change in vegetation composition that now seems to be irreversible. My dissertation has three objectives. Firstly, as our ecological knowledge of the PCFW thus far has been primarily based on a single study, I collected ecolo gical knowledge of the species in a contrasting setting that could then be used to fulfill the other objectives of my study The second objective was to examine the potential response of PCFW to grazing. Two other disturbances, weeds and fire, are probabl y also important but are dealt with more peripherally in this thesis. The third objective was conservation based and species specific; to develop a suitable conservation management strategy that aims to ensure the long term survival of the PCFW. Althoug h my third objective was species specific, by using a habitat specialist as an indicator
23 species, management recommendations will potentially benefit other species within the riparian habitat. A synthesis of the first two objectives was necessary to achie ve this objective. Study A ims and L ayout Chapter 2 provides a background of the study region including historical, environmental, and social factors that have shaped the region. For the PCFW, an historical background and habitat associations in the study are discussed. A description of the study sites is also provided in this chapter. The specific research questions addressed and approaches used are as follows: Is the ecology of PCFW in systems dominated by Chionachne (c anegrass) different to that descri bed in the only previous study (where the system was dominated by Pandanus )? This question is addressed in Chapter 3. Data on the general ecology of the PCFW were collected and compared to previous research based in a distinctly different habitat. The se data also provide the foundation for the subsequent questions posed in this thesis. What is the status of M. c. coronatus within the study area? This question is addressed in Chapter 5 with additional information from Chapter 3 (PCFW ecology) and Chapt er 4 (disturbances). Based on the data obtained from a helicopter survey, the distribution of Chionachne was mapped Territory and group sizes of PCFW at specific study sites were used to assess PCFW habitat requirements and densities. Other suitable hab itat areas, outside of the specific study sites, were surveyed to determine PCFW presence or absence. I used these data to estimate population size and to map the distribution of the PCFW in the study region. Is the current population of the p urple crowne d fairy wren stable?
24 This question is addressed in 5, but relies on results found in Chapter 3 and 4. The density, reproductive success and adult survivorship of the PCFW, supplemented by Chionachne distribution, are the key indications used to determine t he stability of the PCFW population. Do disturbances, in particular grazing, have an impact on the population of the p urple crowned fairy wren ? This question is addressed in Chapter 4. Grazing effects were analyzed using comparative means as grazing pr essure was distinct between sites. Habitat variables were used for these comparative analyses. Reproductive success, adult survivorship and group size are the variables analyzed to detect any differences between sites. Because initial observations indica ted that PCFW frequently used the introduced weed Xanthium as a foraging substrate, foraging observations were conducted to determine if this weed is a positive habitat component for the PCFW. What are the best management strategies for this species to e nsure long term survival? This question is addressed in Chapter 5. Management guidelines were developed based on (1) habitat requirements of the PCFW based on both the ecological and distribution data; (2) areas most suitable for rehabilitation and preserv ation with regard to connecting fragmented habitat based on the Chionachne distribution survey; and (3) recommendations for multiple stakeholder involvement based on personal observations, interactions that have occurred at field sites, and existing polici es and practices. Obstacles and C onstraints This is a largely correlative study, with little opportunity to manipulate experimental and disturbance variables. It was impossible to manipulate most of the disturbances that occurred at
25 study sites. For ex ample, my initial study design included a balanced representation of grazing intensity. However, without advance warning, one landholder introduced a high stocking rate of cattle to my main ungrazed sites during the course of my study By the end of the three year study, there were only two sites left that were not subjected to grazing and even these had seen an increase in some disturbance agents (e.g. an increase in feral buffalo). Resourcing constraints common to all PhD projects meant that my study was conducted over a three year period. This is a brief window through which to view the biology of a species likely to live more than six years; and in a region where irregular rainfall events may mean that there are no typical years or typical ecologic al responses. Despite these constraints, the data I collected of the data makes them more pertinent for devising a suitable management plan by giving an indication of important factors which may or may not be c ontrollable by mana gement. The implications of my research are that stakeholder attitudes toward protecting riparian zones will have a great influence over any future conservation efforts.
26 Figure 1 1. Distribution f the two subspecies of Malurus coron atus (adapted from Rowley and Russell, 1997). Note that the actual area of occupancy is far less than that indicated in this map, as the distribution is much more circumscribed to a network of narrow riparian strips (unmappable at this scale) rather than t he broad polygons indicated here.
27 Figure 1 2. Major drainage lines within the distribution of the two species of the p urple crowned fairy wren Malurus coronatus The w estern subspecies ( M. c. coronatus ) has been found from the Fitzroy River in Western A ustralia e ast to the Victoria River in the Northern Territory. The Carpentarian subspecies ( M. c. macgillivrayi ) has been found from the Leichhart River in Queensland north to the Roper River n the Northern Territory.
28 Figure 1 3. Pandanus ( Pandanus a quaticus ).
29 Figure 1 4. Canegrass ( Chionachne cyathopoda ).
30 CHAPTER 2 THE VICTORIA RIVER D ISTRICT AND THE PURP LE CROWNED FAIRY WREN; NATURAL ENVIRONMENT AND HUMAN DIMENSIONS WITHIN THE STUDY ARE A This Chapter introduces the study area providing a desc ription of location, climate, topography, geology and vegetation. From a human dimensions standpoint, it provides a description of the history, land use, and tenure of the region. In relation to the PCFW, the brief historical background of this species i n the Victoria River District (hereafter VRD) is a reflection of the scarcity of records. The habitat description of the PCFW highlights the lack of any intensive surveys of this species in this region. This Chapter concludes with a description of the loc ation and condition of the individual field sites used during my study The environmental, social and historical aspects of the VRD and the PCFW described in this chapter form an important base for this thesis. The Victoria River District Natural Environm ent The Victoria River Distric t, in the semi arid tropics of n orthern Australia, encompasses approximately 125,000 km 2 of land located 500 km south west of Darwin with multiple la nd tenures in place including national park, pastoral, a boriginal, defense, a nd private lands (Hill, 2000; Karfs, 2000) (Fig. 2 1). The Victoria River itself is approximately 720 km in length that originates on Riveren Station and eventually drains into the Joseph Bonaparte Gulf. The mouth of the river is more than 10 kilometers wide with an extensive tidal influence reaching more than 500 kilometers upstream, approximately 20 kilometers past Timber Creek (Kirby and Faulks, 2004). The Victoria River is a fairly shallow river with multiple sand and rock bars, which during the dry season separate the river into multiple billabongs, especially in the upper reaches. The river generally flows for only four to six months when flooding can be extensive. Annual flow
31 volume has been recorded as high as 4,184 million m 3 or a mean discharg e of 909m 3 per second at the Victoria River bridge, adjacent to one of the study sites, GNP1 (Kirby and Faulks, 2004). The average yearly rainfall for this region is approximately 800 mm (decreasing from the coastal north to the inland south: 938.5 mm in Timber Creek and 648.4 mm at Victoria River Downs) with a distinct wet and dry season (Bureau of Meteorology, Australia, 2005) (Figs. 2 2, 2 3). The wet season typically lasts from November through to March with the highest rainfall recorded in January an d February when flooding most frequently occurs. Daytime temperatures average around 30 0 C in the dry season and 37 0 C in the build up (the characteristically hot and humid time preceding the wet season) and wet season, although temperatures over 40 0 C are not uncommon in the build up (Fig. 2 4, Bureau of Meteorology, Australia, 2005). Humidity levels vary according to the distinct seasons, ranging from a mean relative low of 40% in the dry season to a mean relative high of 88% in the wet. There are five m ain catchments within the VRD, of which the Victoria River is the largest. The VRD encompasses characteristic mesas and sandstone plateaux, tropical savannas, and rocky outcrops with spinifex ( Triodia sp.). About half of the region consists of very shall ow soils (e.g. rugged stony outcrops, plateaux and hills), whilst the rest are erosional and alluvial plains that are favored for pastoralism (Kraatz, 2000). The shallow (or skeletal) soils belong to the Pinkerton land system whereas the alluvial floodpl ains belong to the Ivanhoe land system consisting of cracking clays (Stewart et al ., 1970). Major drainage lines, such as the Victoria River, are predominantly duplex clay dominated soils. Table 2 1 illustrates the relationship between the parent material and the associated soils found in the region. Multiple vegetation types characterize the VRD, and their distinction is associated with soils and annual rainfall (Fig. 2 5). Widespread vegetation types include Eucalyptus dominated
32 woodland s, mixed species low woodland, e ucalyptus dominated low open woodland and grasslands, each with their own associated understory grasses. Common tree and grass species found in association with each broad vegetation type are listed in Table 2 2. Less extensive than these vegetation types, there is a narrow network of riparian vegetation alongside the main river courses. This vegetation is described in more detail in the following sections. Human Dimensions Aboriginal people have lived in the VRD for at least 40,000 years and there are multiple clans in the district, with approximately 24 language groups, that have a strong physical, spiritual, and social connection to the land (Riddett, 1990; Mulvaney and Kamminga, 1999). Approximately 20% of the land in the VRD is a borig inal owned and there are about 27 communities in the region, although many of these consist of small family groups. There are also six a boriginal owned pastoral properties that cover approximately 10% of the district (Davis 2000). The first European to officially explore the Victoria River was John Stokes who arrived aboard the HMS Beagle in 1839 led by Captain John Wickam. However, they did not stay in the area very long nor travel far upstream and their descriptions of the flora and fauna are fairly superficial (Lewis, 2002). From 1855 to 1856 the North Australian Expedition led by Augustus Gregory explored the VRD in much greater detail. Ferdinand von Mueller, a botanist, was a member of this expedition and his reports have valuable descriptions of the vegetation in the district (Mueller, 1858). Pastoralism is one of the most important industries in the Northern Territory and there are approximately 1.4 million head of catt le in the t erritory with the majority (45%) destined for the live export trad e to South East Asia (Peart, 2001). In the VRD, pastoralism is the major industry
33 and 59% of the land is used for grazing. There are over 20 pastoral properties ranging in size from 1,100 km 2 to 12,000 km 2 ; and the total herd in the VRD is estimated to r ange from 515,000 to 750,000 head of cattle (Hill, 2000). Undoubtedly, pastoralism has had a profound influence on the VRD ecological landscape. Management depended on an open or free range system and overstocking was rampant with no established paddocks being present to control stock (Lewis, 2002). As a consequence of this primitive management practice, large numbers of cattle congregated along riparian corridors and billabongs causing great damage to these systems. According to Sharp and Bowman (2004), many of the changes that have occurred along the riparian alluvial plains along the Victoria River are irreversible. The region has been heavily influenced by pastoralism since the arrival of the first European settlers in 1883. The initial relationship betw een the early settlers and the a boriginal people was a troubled one that included many violent altercations (e.g. Davis, 2000; Riddett, 1990). However, by the 1950s, the cattle stations depended greatly on the local a boriginal people as workers althou gh they were only paid in rations and were in generally treated very poorly. T he situation of virtually free a boriginal labor came to an en d after the famous walk off by a boriginal people at Wave Hill station in 1966. This was followed by similar strike s throughout the district in the 1970s that revolutionized a boriginal rights (Davis, 2000). On Victoria River Downs Station alone, cattle numbers rose from 15,000 in 1889 and to between 119,000 to 170,000 in 1921. In addition to cattle, feral donkey and h orse numbers also soared putting further pressure on the ecosystem. For example, on Victoria River Downs in 1960 there were an estimated 20,000 donkeys and in the 1940s there were an estimated 20,0 00 scrub bulls (Lewis, 2002). With no supplementary wate r available away from the river, the
34 pressure on the riparian zones during the extended dry season must have been tremendous. Conditions improved dramatically in the late 1970s and early 1980s with the commencement of the Brucellosis and Tuberculosis Erad ication Campaign (BTEC) that resulted in improved fencing, control of feral cattle, and increasing provision of water away from riparian zones removing some of the pressure from these systems (Karfs and Trueman, 2005). The Purple Crowned Fairy wren in the Victoria River District Most of the information that exists for the region is in some way related to pastoralism and there is a paucity of information on biodiversity. Hence, in relation to the PCFW there are few data on distribution or abundance until f airly recently. However, accounts given by some of the earlier explorers give an indication of the type of vegetation found along the Victoria River and its tributaries. These accounts also give an insight into the condition of the riparian areas previou s to settlement and pastoralism, but only on a superficial level. Photographs comparing various sites in the VRD published by Lewis (2002) also provide an insight into changes that have occurred in the landscape although these mainly demonstrate an increa se in tree cover and any changes in riparian grass cover is not discernable. In addition, the majority of early pictures were taken when pastoralism was already well established and many of the sites had already been grazed so they give no indication of p re European settlement condition (Lewis, 2002). In relation to the PCFW, one of the most interesting observations that earlier explorers were obviously common and widespread as they are described and referred to on many different occasions by the early explorers in the region including John Stokes in 1839 and Augustus Gregory in 1855 (but see Lewis 2002, for multiple excerpts). Lewis (2002) has inferred that th ese reeds are most likely Chionachne cyathopoda and the description of them being almost
35 impenetrable, is in my experience, an accurate assessment. If the reed is indeed Chionachne then it points to a potentially widespread distribution of the PCFW in th e VRD before the arrival of European settlers in the region (see below for a description of the primary habitat of the PCFW). Primary hab itat of the Purple crowned Fairy wren in the Victoria River District In the past, there has been a great deal of confus ion in regard to the primary habitat of the PCFW, and this confusion extends to the VRD as well. In areas where the PCFW inhabits Pandanus aquaticus this has not been a problem but in the VRD where it inhabits riparian grasses multiple species have been e rroneously listed as the primary habitat of the PCFW. The majority of the confusion appears to lie in the varied usage of common plant names by different authors over time. In p articular, the common name c anegrass has been used to describe a variety of r iparian grass species in association with the PCFW. In the VRD the primary habitat of the PCFW is the riparian species river grass or c anegrass Chionachne cyathopoda sometimes in conjunction with northern c anegrass Mnesithea rottbeolioides (Table 2 3). Chionach n e cyathopoda is a rhizomatous perennial species with a widespread tropical and subtropical distribution and often forms dense thickets along riparian corridors with average height of 2.7 meters (at ungrazed sites), although it can reach a height o f up to 5 meters (Cowie et. al. 2000). In the VRD, Chionachne cyathopoda is an important source of food for many bird species and large f locks of finches (e.g. star f inch Neochmia ruficauda crimson f inch Neochmia pha eton yellow rumped m annakin Lonchu ra flaviprymna and chestnut breasted m annakin Lonchura castaneothorax ) congregate in the grass when it is seeding (pers. obs.)(see Appendix 1 for a complete list of bird species). Along the Victoria River, dense stands dominated by Chionachne can extend mo re than a kilometer away from the river along drainage lines, but it is generally found within 200 meters of
36 the river bank. During flooding events, Chionachne stands may be completely inundated and flattened. However, it is generally resilient, and reco vers rapidly following such flood events (pers. obs.). Areas that are inundated over long periods (e.g. drainage depressions) usually do not contain Chionachne suggesting it does not survive long periods of inundation (pers. obs.). I did not encounter a Mnesithea alone although these types of sites were surveyed. Several authors have listed Mnesithea as the primary habitat along the Victoria River ; however this seems highly unlikely as we surv eyed the same areas only to find Chionachne as the dominant grass species. I assume that this discrepancy is related to species identification and not a shift in habitat utilization. Habitat that was dominated by Pandanus aquaticus was also surveyed (appr oximately 30 Wickam River), however, in this habitat despite this vegetation type being abundant. In the past the PCFW has been found in Pan danus habitat along the Wickam River (Boekel, 1979), I surveyed this area but was unsuccessful in locating any wrens in that area. It is possible that the y no longer occur there or occur at very low densities. This is not particularly surprising as Boeke l (1979) described the low density of birds and expressed concern for the future survival of that population. There are reports of scattered populations in Pandanus along the upper reaches of the Victoria River but as was reported by Boekel (1979) in low numbers. The common name c anegrass has been widely used to describe a range of riparian grass species in as sociation with the PCFW throughout its range ( McGill, 1970 ; Boekel 1979; Rowley 1993 ; Goodfellow, 2001 ). As is often the case with common names, t hey can represent a large number of species which may vary in structure, ecology and distribution. Species that
37 have been referred to a s c anegrass in association with the PCFW include Ophiuros exaltatus (similar to Mnesithea rottboellioides ), native b ambo o Bambusa arnhemica and exotic b amboo Bambus sp., northern c anegrass Mnesithea ( Coelorhachis ) rottboelioides and various Sorghum sp. Within PCFW habitat, tree species consisted primarily of flooded box or c oolibah Eucalyptus microtheca river red g um Euc alyptus camaldulensis with some leichardt p ine Nauclea orientalis at the canopy level and freshwater m angrove Ba r ringtonia acutangula f icus Ficus coronulata and native white currant or d ogwood Flueggia virosa at the mid d les tory l evel Additional species found at lower de nsities at some sites included b auhinia Bauhinia cunninghamii c athormion Cathormion umbellatum gutta percha t ree Excoecaria parvifolia, weeping p aperbark Melalueca leucadendra and the aquatic p andanus Pandanus aquaticus (See Appendix 2 for a list of riparian plant species). Other species very common among the stands of Chionachne and Mne sithea included a variety of introduced weed species. These included noogoora b urr Xanthium strumarium castor o il Ricinus communis and wild p assionfr uit Passiflora foetida By far the most abundant weed and pr obably of greatest concern was noogoora b urr. Introduced in the 1970s around Pigeon Hole (Lewis, 2002), Xanthium has rapidly become widespread throughout the Victoria River catchment area. This species and its potential impacts on the PCFW are discussed in greater detail in Chapters 4 a nd 5. Two other weed species, h yptis Hyptis s u a veolens and rubber b ush Calotropis procera are also found commonly in this habitat but are generally restricted to vegetation further away from the river and outside of Chionachne stands.
38 Study S ites Selected Study Sites and C onditions During the first (preliminary) field season in 2001, I investigated various locations of the Victoria River a nd tributaries to identi fy potential study sites that contained PCFW and were accessible most of the year. A considerable amount of time was spent in discussion with stakeholders to familiarize them with the goals of this research project and to gain property access. I surveyed locations that contained both Chionachne and Pandanus as all the previous research indicated Pandanus was the primary habitat. A total of 12 sites were evaluated as three other sites, resulting in five field sites that were usable (Fig. 2 6). My aim was to establish both multiple un grazed and grazed (preferably with differing weed densities) sites for comparisons. Although this was the experimental design preferenc e, and initially sites were set up under those specifications, the following field season revealed that most of the sites no longer adhered to the presumed and/or original conditions, or sites could no longer be used. For example, after conversations with landholders several sites were selected on the understanding that they were fenced and hence ungrazed. However, subsequent visits revealed that most of the fences were not maintained and cattle were present. Due to the floods that occurred every year du ring my study fences were continuously washed away and damaged, deeming them ineffective for excluding cattle from the riparian zone. Maintenance of fencing was an issue throughout the Victoria River catchment and many sites that I surveyed during the pr eliminary field season had damaged fencing. During the first field season I chose five sites One of these, a pastoral property with heavy grazing pressure that still retained PCFW groups, I identified as suitable and fieldwork had commenced at this site when the landholder decided against allowing continued access to the
39 property. In lieu of this site, in the second year of my study I identified another site that had also been heavily grazed, Dashwood Crossing on Victoria River Downs was used for the rem ainder of my study (Fig. 2 6). Unfortunately, this site was less suitable due to the distance from other sites, lack of access at the beginning of each season, and small number of wren groups in residence. Changes in grazing pressure, erosion, and weed i nfestation during the different field seasons are summarized in Table 2 4. The first site (Coolibah) was located at the Coolibah Crocodile Farm (15.56256, 130.94884) on the eastern bank of the Victoria River downstream from Gregory National Park. This sma ll parcel of land is privately owned and the riparian zone is only used for river access and as a low impact campsite during the dry season, although only a small clearing is used for camping. This site initially (2001) contained a very dense understory o f Chionachne and habitat for PCFW with little erosion, low density of weeds and no known grazing pressure. In July 2002, approximately 20 head of cattle from the ne ighboring station gained access to this site and it was grazed intensively, the cattle remained on the site until the beginning of the west season (approximately October). In 2003, cattle once again gained access to this site and grazed the site but with less intensity. On both occasions, cattle grazed intensively on Chionachne and spent most of their time within the riparian zone where they had easy access to water. Once the Chionachne had been fed upon, less desirable plants (ones that are known to be toxic in large doses) were also eaten and included Xanthium and Ricinus My second site (Dashwood) was located at Dashwood Crossing on Victoria River Downs Station (16.33598, 131.11253), and was situated more than 200 km by road from the other sites and it was not possible to collect as much data from this site due to distance and limited access
40 during the wet season. Chionachne abundance at this site was low, especially in comparison to the other sites used in my study Chionachne was found in small patc hes and sometimes in single clumps spread inconsistently through the riparian zone. The Dashwood site had been consistently grazed for many decades, was heavily eroded and had high densities of weeds ( in particular Xanthium ) In recent times this site ha s been fenced and grazing pressure lowered but cattle still access the site at times, most likely due to damaged fences. According the current VRD manager and a long term pastoral facilitator (D. Hill pers. comm.) erosion has subsided greatly since many a reas have been fenced and new paddocks have been developed, resulting in reduced grazing pressure. The vegetation remained fairly co nstant at this site and no obvious changes took place year to year. The third site (Fitzroy) was located on Fitzroy Statio n (15.56195, 130.92899) on the western bank of the river not far downstream from Coolibah Crocodile Farm and is an Aboriginal owned station and a small family group resides at the site. In 2001 there was a very dense understory of Chionachne that was almo st impenetrable in many places. Although it had previously been used for grazing only a few head of cattle were present in 2001. PCFW were o bserved in large numbers and I deemed this site as high quality PCFW habitat. There was minimal activity in the r iparian area with the owners of the property only going down to the river for fishing and to refuel the water pump located next to the river. There were a few weeds present on the side of the dirt track leading down to the river and in some of the flood g ullies. In 2002 this site was leased to a pastoralist and cattle grazed the site at light to moderate levels until the beginning of the wet season. In 2003, some cattle were still present at the site but in very low numbers since paddocks and bores had b een established at sites away from the river reducing pressure in the riparian zone. However, in 2003 there was an influx of buffalo ( Bubalis
41 bubalis ) at this site and this resulted in some grazing and numerous tracks through the understory. The large nu mber of buffalo, due to their potential aggressive nature, made access difficult at times. In 2003 I also observed a large increase in weed density ( Xanthium primarily) at this site. Two sites (GNP1 and GNP2) were located 1.2 km apart within Gregory Nati onal Park (hereafter GNP), established in 1986 and covering approximately 13,000 km 2 The first site (GNP1) was located at the Victoria River bridge (15.61972, 131.12902) and is adjacent to a roadhouse with a large campground. A narrow dirt track leads d own to the river but is not used often and there was very little damage to the Chionachne from visitors. This is a favorite area for birdwatchers to look for the PCFW, but most choose to look for the birds from the highway and there was little disturbance to the birds or their habitat. The second site within the national park (GNP2) was located at the Victoria River Access road (15.63189, 131.13225) where a 4 wheel drive boat ramp was located, also on the western bank of the river. Recently (2005) this ha s been converted to a 2 wheel drive access area although great care was taken not to disturb the Chionachne habitat during construction and some stands that would have been destroyed were moved to a previously open area. Both sites were located in close v icinity to each other and had good access throughout most of the year. These two sites were chosen as they differed in Chionachne stand width and because their proximity to one another made it logistically feasible. At both these sites, there was some buf falo and pig ( Sus scrofa ) activity with buffalo numbers increasing toward the end of the study. Nevertheless, grazing impact at these sites was not evident although some distur bances from these species ( cl earings and tracks) were present Weeds (in parti cular Xanthium and Ricinus ) were present at both sites in moderate abundance.
42 Wild p assionfruit ( Passiflora foetida ) was also fo und in moderate abundance but only during the wetter time of year (November through April). Erosion was low to moderate at bot h the sites within the park but was particularly evident along the bank after the two big floods experienced in 2002 and 2003.
43 Table 2 1 Relationship between parent material, soils and topography in the VRD. ( Kraatz 2000). Parent Material Soils and To pography Basalt *red earths on sloping terrain *cracking clays (vertisols) on lower to flat slopes *rock outcrop and/or very stony surfaces common Limestone, dolomitic and calcareous sediments *earths and yellow earths on well drained slopes *cracking c lays (vertisols) on poorly drained, lower slopes Sandstone and calcareous sandstone *sandy red earths *some yellow earths and lateritic podzolic soils Non calcareous shales (northern VRD) *yellow earths *soils also prevalent on floodplain deposits Ca lcareous sedimentary rock or basalt *clays dominate alluvial plains
44 Table 2 2. Dominant vegetation in the VRD (data from Wilson et al. 1990) Dominant Vegetation Common Tree Species Common Grass Species Eucalyptus woodland E. tectifica E. terminalis E. tretrodonta E. miniata E. polycarpa Sorghum Chrysopogon fallax Sehima nervosum Plectracne pungens Low mixed species woodland E. pruinosa Terminalia arostrata Lysiphyllum cunninghamii Eulalia aurea Chrysopogon fallax Aristida sp. Dicanthium sp. Low ope n woodland E. brevifolia E. dichromophloia E. miniata Lysiphyllum cunninghamii Triodia pungens Plectrachne pungens Grassland Astrebla pectinata Chrysopogon fallax Dichanthium fecundum Astrebla sp.
45 Table 2 3. Characteristics of Chionachne cyathopoda an d Mnesithea rottboelioides (van Doorn and Low Choy, in prep.). Species Habitat Soils Gross Plant morphology Leaf attachment Reproductive (Inflorescence) morphology Reproductive (Inflorescence) location Chionachne cyathopoda Banks and tributaries of large Rivers Variety of soils from sand to loam to clay soils Rhizomatous perennial with stems branching and intertwining to 4 m high Throughout length of stem Relatively small Robust, enclosed by a prominent sheath Lower half of functionally female flowers upper half of functionally male flowers Terminal or in leaf axils Mnesithea rottboellioides Grassland / woodland associated with seasonal water (including creeks / wetlands perched on sandstone plateaus) Generally sandy loam soils Tussock perennial with leaves arising from base to 1 m high From base of plant Large forming bulk of plant Fine, in groups near the end of a long stem Sheath subtending flower groups not prominent Stem exerted from the plant base commonly to 2 m
46 Table 2 4. Tenure and year ly c onditions of field sites. Field Site Land Use Land Tenure 2001 Site Condition 2002 Site Condition 2003 Site Condition Coolibah Crocodile Farm (COOL) Small enterprise Freehold private Grazing none Weeds low Erosion med Grazing high Weeds med Erosion med Grazing low Weeds med Erosion high Dashwood Crossing (DASH) Pastoral Pastoral Lease Grazing med Weeds high Erosion high Grazing med Weeds high Erosion high Grazing low Weeds high Erosion high Fitzroy Station (FITZ) Pastoral Aboriginal (pastoral lea se 02 03) Grazing none Weeds low Erosion low Grazing low/med Weeds med Erosion low Grazing low Weeds med Erosion med Gregory 1 (GNP1) Conservation Northern Territory Government Grazing none Weeds med Erosion low/med Grazing none Weeds med/high Erosion lo w/med Grazing low Weeds med/high Erosion low/med Gregory 2 (GNP2) Conservation Northern Territory Government Grazing none Weeds med Erosion low/med Grazing none Weeds med/high Erosion low/med Grazing none Weeds med/high Erosion low/med
47 Figure 2 1 Location of the Victoria River District and associated land tenures. Note that Coolibah is under a perpetual pastoral lease.
48 Figure 2 2. Mean annual rainfall in the study region.
49 Figure 2 3. Mean monthly rainfall in T imber Creek (1981 2004) and Victoria River Downs (1885 2004). Data from the Australian Bureau of Meteorology, 2005. Figure 2 4. Mean daily maximum temperature Timber Creek (1981 2004) and Victoria River Downs (1885 2004). Data from the Australian B ureau of Meteorology, 2005.
50 Figure 2 5. Broad vegetation classification of the study region.
51 Figure 2 6. Location of study sites. Fitz = Fitzroy Station, Cool = Coolibah Crocodile Farm, GNP1 = Gregory National Park site 1, GNP2 = Gregory National Park site 2, Dash = Dashwood Crossing (Victoria River Downs Station). Note that indicated boundaries represent land tenure.
52 CHAPTER 3 ECOLOGY OF THE PURPL E CROWN ED FAIRY WREN This chapter provides a description of the ec ology of the PCFW, in the grass dominated riparian habitat in the VRD. This chapter also provides a comparison to the single in depth previous study conducted in a Pandanus dominated habitat. The ecology of the PCFW described in this current chapter provi des an important foundation for the two subsequent chapters. Introduction Many Australian passerines are characterized by their longevity, small clutch size, high incidence of multiple broods and low annual productivity (Fry, 1980; Woinarski, 1985; Yom To v, 1987; Russell, 1989). The incidence of cooperative breeding is also high in Australian passerines (Rowley, 1969; Russell, 1989), a life history trait that has theoretically and comparatively been linked to high adult survival (Hardling and Kokko, 2003). The Maluridae, an old endemic family that has been distinct for 35 40 m.y. (Sibley and Ahlquist, 1985), exhibits many of these characteristics, and has been the subject of many landmark studies that have helped define, distinguish and interpret the biolo gy of Au stralian birds in general ( Rowley et al ., 1991; Mulder, 1997; Cockburn et al ., 2003). Within the malurids, most such studies have been conducted on species occurring in the temperate south of the Australian continent. This bias gives some uncerta inty to the extent to which biological traits for the genus as a whole can be generalized. In particular, there has been only one relatively brief previous study of the largest species in the genus, the PCFW, and that study (Rowley and Russell, 1993) was restricted to one site. Given that this species has a high degree of habitat specificity and occurs in a monsoonal climate contrasting starkly to that in temperate Australia, aspects of the biology of this species may exhibit some characteristics that dif fer from that currently regarded as typical for the genus.
53 Some p revious long term studies ( Rowley et al. 1991; Brooker and Brooker, 2001) of temperate zone malurids (e.g. M. splendens M. pulcherrrimus ) have shown how some life history characteristics ( i.e. longevity and multiple broods) show some flexibility in response to variation in local conditions. It has been widely reported elsewhere that rainfall, for example, has been shown to influence productivity for many avian species as breeding is often influenced by food resources tied to the amount of precipitation (Lack, 1968; Perrins, 1970). Several Malurus species ( M. coronatus M. cyaneus M. pulcherrimus ) are known to reduce clutch size or abstain from breeding altogether following periods of low rainfall (Nias and Ford, 1992; Rowley and Russell, 1993; Brooker and Brooker, 2001). Many Malurus species have annual adult survival rates around 70%, hence, the strategy of surviving the bad years, such as low rainfall years, without breeding is a viable option (Rowley and Russell, 1997). Conversely, during favorable years, females will often produce multiple broods and breeding season may be extended, enabling them to exploit the favorable conditions and potentially compensating for bad years (Nias and F ord, 1992; Brooker and Brooker, 2001; Rowley and Russell, 2002). In the wet dry tropics of northern Australia, inter annual variation in rainfall can be very high (Ridpath, 1985; Taylor and Dunlop, 1985) resulting in a change of the breeding behavior of P CFWs (Rowley and Russell, 1993). Indeed, Rowley and Russell (1993) found no evidence of breeding of PCFW following two wet seasons marked by low rainfall. In addition to climatic variation, changes in habitat attributed to disturbances can also trigger ch anges in breeding behavior. For example, M. elegans displayed delayed breeding after a fire and productivity was lower the following year (Russell and Rowley, 1998)
54 This chapter describes aspects of the life history and behavior of PCFWs at a set of site s in the VRD. It specifically considers the following key issues. What are the basic characteristics of breeding biology for this species in the VRD? To what extent do these characters vary between years? What i s the diet and foraging behavio r of this sp ecies in my study area? To what extent do the breeding biology and other characteristics reported here for this species in the cane grass of the VRD contrast with those reported in the only previous study of this species, in a Pandanus dominated habitat in the e ast Kimberley? And, to the extent that there may be any marked differences, what environmental factors would contribute to such distinctions? Given that my study substantially amplifies the knowledge base of the PCFW the discussion in this chapter a lso considers the extent to which the breeding biology of this species falls within the characteristics typical for the genus as a whole. There are some important caveats that constrain the extent to which these questions could be addressed in my study an d the interpretation of the r esults obtained. My study occupied three field years, so there were limited opportunities for considering inter annual variation in environmental conditions and bird responses. Further, in contrast to the previous study of Ro wley and Russell (1993), rainfall conditions over the three years of my study were relatively similar. My skill at nest location and other factors increased over the course of the study, so results from the first year are more fragmentary than for subsequ ent years. Comparison among years was also hampered by the loss of one key study site for one year, because of unpredicted social issues with one landholder. I also acknowledge that my information on diet is limited: i n the extremely dense understor y of this area, there are few opportunities for prolonged observation of this furtive species. Comparison with the previous study of PCFWs by Rowley and Russell (1993) is an important component of this chapter. The ecological characteristics determined from on ly one
55 study site form a fragile base for characterizing biological features and conservation management there are significant differences in management interpretat ion. For example, Chionachne is more readily disturbed by cattle and is a less fixed structure than Pandanus ; Chionachne may extend further from riverbank and hence dispersion of PCFWs is less linear. There are some interpretational issues associated wit h the direct comparisons of these 2 studies: Rowley and constrained. In addition, due to structural differences between the riparian habitats some comparison s are also hindered. For example, they reported PCFW density as a linear measure in comparison to the area measurement I used. This chapter does not consider variation in biological characteristics among my study sites, associated with varying levels of d isturbance. That topic is the subject of Chapter 4. Field Methods Terminology ng progeny of the year and hence less than one year in age (1 characteristic non breeding plumage of adult males. Adult birds were more than one year in age (1+). Based on plumage, adult birds could be categorized into the following age groups; female (2 ), female (2+), male (2 ), male (2+), male (3+) (Rowley, 1988). To identify the breeding pair if a group was made up of more than a single pair, the female with brood patch was identified as the breeding female and the eldest male as her partner. The identity of a breeding female was also corroborated when she was frequently fed by males in her group. On a few occasions males within a group were almost identical in appearance and in
56 these instances I considered the se nior male to be the one that was most often found performing a duet (singing together) with the female. However, it is important to note that concept to unambiguously apply to cooperative breeders, and that even in cases where a domi nant and apparently paired male and female share a territory, other studies have shown that paternity is not necessarily straightforward (the females often sneaking out to mate outside the territory with a more distant male, Double and Cockburn, 2000). I u to denote various aspects of a family group over a one year period (i.e. productivity, composition, size) (Rowley and Russell, 1997). Only adult birds (1+) ces where a breeding pair did not remain together in the absence of the death of either bird (Rowley and Russell, 1993). Productivity was measured as the number of independent juveniles produced per group (i.e. breeding female) during the breeding season ( March September). Nests were termed as one fledgling to ind ependence (Rowley and Russell, 1997). I measured a nnual survival of adults as the proportion of birds present at the beginning of each breeding season that were still present at the beginning of the following breeding season Captures and Banding I captu red individuals with mist nets from 2001 through to 2003 at the five previously identified field sites (Coolibah, Dashwood, Fitzroy, GNP1 and GNP2 the two sites within Gregory National Park). I placed mist nets within known territories, and used a playba ck system to attract birds to the nets. At times only one net was used, but in open areas or where birds had proven hard to
57 catch, multiple nets were used to increase the probability of capturing all group members. Nets were placed away from territory bo undaries to avoid capturing multiple groups simultaneously or disturbing neighbouring groups. When possible, I placed nets along drainage lines and other open areas to avoid damaging wren habitat. Once captured, I removed birds and placed them individual cotton holding bags before processing. Nets were never left unattended and birds were customarily removed from the net within one to five minutes of capture. Although the aim was to capture the entire group, individual birds could be extremely wary and fledglings would often avoid approaching the nets altogether. I focused efforts on capturing birds that were un banded. In instances where all members of a group were already banded, groups were not purposely recaptured as the birds were easily identifie d by their color bands. I could not band birds at Fitzroy Station in 2003 due to restricted access and social issues. Birds were banded according to the Australian Bird Banders Association guidelines ( 1989 ). Each bird was banded with numbered aluminum b and (13 mm, size 02) and unique combination of three colored plastic bands for subsequent visual identification. Morphology and P lu mage Based on plumage, each captured bird was sexed and age (category) and molt condition for each bird was recorded. Durin g subsequent visits to field sites any observed banded birds were recorded and any chang es in plumage noted. F or each captured bird I recorded the following data : weight, wing length (flattened folded right wing), tail length and tarsus length. There wer e some occasions where morphological measurements were not obtained, a photograph was not taken, or film was damaged due to excessive heat resulting in missing data. Missing morphological data could be attributed to one the following; birds (primarily bre eding
58 females) became stressed or were suspected of carrying an egg and were only banded, birds escaped before all data were collected, or a measurement was overlooked at the time of capture. Photographic quality improved once a digital camera was used fro m 2002 onward. Differences between sexes (and between adults/juveniles) in dimensions and weights were assessed with t tests. As females may lay up to four clutches in one season, physical condition may deteriorate through the season. I compared female w eights between the beginning (May and June) and the latter part (July and August) of the breeding season to ascertain any average difference in body weight. For seasonal comparisons of weights, (three capture seasons), I kept samples independent by using only one measurement per female in the event of a recapture. I compared differences in weights using t tests. Group Size and C omposition I determined group size and composition by two methods. The first method was to identify group size and composition w hilst mist netting. Although the entire group was not always captured, the close proximity of other birds belonging to the same group made it fairly easy to identify them during mist netting attempts. Secondly, throughout the field season, groups would b e identifi ed and followed regularly ( during nest searches and foraging observations). Any birds encountered were recorded. Since most groups were followed on at least two occasions, group composition was well known. I compared group size across the thre e years o f my study using a Kruskal Wallis test. I used a Chi 2 ) test to determine if sex ratios varied from parity. I also compared the frequency distribution of group sizes, (total no. birds per group), in my study area with those recorded by Rowley and Russell ( 1993) for Drysdale River area, using a Kolmogorov Smirnov test
59 Territory Shape and Size One of the primary techniques for determining territory boundaries was by recording any territorial singing, either individually or as a duet. Birds were particularly active in territorial singing at the beginning of the breeding season and much less so at other times of the year when territories became less defined. At the beginning of the breeding season adjoining family groups would often respond to each other with s ong to mark boundaries. Further, groups rarely wandered far into adjoining territories during the breeding season. By repeatedly recording positions of singing birds, territory boundaries were often clear. Additionally, birds were followed regularly to monitor breeding and record foraging observations, making it was possible for me to determine territory boundaries of groups. During mist netting, territorial boundaries became clear when two adjacent groups were attracted to the net but rarely crossed int o each others territories. I constructed territory maps during the breeding season 2002 when the most time and resources were available and territories were well known. During the 2002 season, family groups were followed extensively and any territorial be havior recorded and marked using a GPS (Garmin GPS II Plus <15m). Data points were plotted as minimum convex polygons on Arcview 3.2 to determine territory shape and size and were subsequently mapped using ESRI ARC GIS 9.1. Territory sizes among sites we re compared using a Kruskal Wallis test. It was not possible to compare territory size for birds in my study area with those reported by Rowley for birds in the Drysdale River area, because Rowley reported only (streamside) length for the territories that he observed and not the width.
60 Foraging Behavior I recorded foraging behavior opportunistically by following individual birds as long as possible. The time when an individual was first seen was recorded as well as each time a foraging observation was mad e until the last foraging behavior was observed. The following data were collected for each observation: sex and identity of individual; foraging behavior following Rowley and Russell (1997), as either: hawk = flying bird takes flying prey, hover = bi rd remains stationary in air before or during taking prey, glean = bird is perched whilst taking prey, probe = bird inserts beak into substrate to obtain prey, pounce = bird flies down from perch to take prey on ground, or snatch = bird takes prey from substrate during flight, foraging substrate, as either Barringtonia (freshwater m angrove), C hionachne (c anegrass), Xanthium (noogoora b urr), Bareground, Bareground w/ debris, tree species, or other (as identified); and prey item.
61 Although I foll owed birds as long as possible, the density of the Chionachne habitat made it very difficult to follow any individual for an extended period of time. It was even harder to record prey items since their small size limited visibility and they were quickly i ngested. I used Chi Square ( 2 ) tests to compare the foraging observations for males and females to determine differences in foraging substrate. I also used Chi 2 ) tests to determine substrate foraging preferences based on vegetation abundances. It was not possible to make rigorous quantitative comparisons of the foraging behaviors recorded in my study with those reported by Rowley and Russell (1993) in the Drysdale River area, because they used very different methodologies (analysis of fecal pell ets), however broad qualitative comparisons are reported. Breeding B iology In each of the three years of my study rainfall during the wet season was above average and the primary habitat of the PCFW was flooded around February each year. From the beginn ing of each field season (March/April/May) I followed family g roups at each site to identify any nesting behavior. The recession of flood waters and access to field sites determined the start of observations at each sites. Some sites were accessible later than others depending on road conditions (i.e. Coolibah and Dashwood Crossing). The most common behaviors that indicated breeding were a female carrying nesting material or individual birds carrying prey items to nest sites. Females begging for food or being fed by other members of the group were other indications that they were breeding or may soon breed. On several instances breeding pairs were observed soon after copulation and once during copulation. These birds were then followed to determine the specific nest location. After id entification of a nest site, I determined the st atus of the nest when birds were not in attendance in order to minimize disturbance.
62 Each nest was monitored every four to six days (when possible) and n est status was determ ined ( stage of construction and nest content). Due to the distance between sites, this was sometimes impossible and nests were checked less frequently. Because checking nests at greater frequencies could increase the already high predation rates, by leav ing scent trails or being observed by potential nest predators, I arbitrarily decided to check nests every six days. During nest checks, I took different routes to the nests and spent as little time was as possible at nests. I also routinely checked for t he presence of any potential nest predators (e.g. Torresian c row Corvus orru pheasant c oucals Centropus phasianinus Varanus sp. etc.) before approaching a nest. Nest checks were conducted when group members were not in attendance when possible. If a fem ale was incubating at t he time of inspection, I would leave the site to avoid flushing the female and potentially increase the risk of nest abandonment. When nestlings were aged around six days, they were banded with a numbered aluminum band and three colo r bands as per adult birds. Once a nest was no longer in use ( after nest failure or successful fledging), I collected nest measurements. Nest measurements included; GPS location, substrate, entrance orientation, % cover (vegetation) within 1 m diameter o f the nest, cover description (plant species), nest base height above ground and nest entrance height above ground. There were some instances where nest measurements were not possible because nests had been completely destroyed and/or were inaccessible due to the presen ce of feral animals ( buffalo or scrub bulls ) that may have endangered observers. On only two occasions volunteers, could not locate previously found nests. Clutch size and productivity were compared between years, using a Kruskal Wallis test
63 Annual Survival I only included banded birds for annual survival rates because the identity, and therefore survival, of un banded birds was unknown. Note that this assessment may be confounded by movement of individual banded birds beyond the territory they held the previous year. Where such birds were re sighted elsewhere, these were included in the survival estimates. Survival rates for juveniles could not be adequately estimated as few nestlings were banded due to high predation rates, in addition t o low numbers of juveniles captured in nets. Additionally, the primary habitat of wrens was continuous at all sites, thus increasing the likelihood many juveniles may have dispersed into adjacent areas outside of my study sites before they could be locate d and banded. Median survival rates between years, 2001 to 2002 and 2002 to 2003, were compared with a Mann Whitney U t est. Results Captures and Banding A total of 177 PCFWs was captured and banded during the three field seasons comprising 54 females, 66 males and 25 juveniles less than one year of age (Table 3 1). In addition, nine nestlings (three clutches) were banded and 32 adult wrens were recaptured. During the first (2001) field season 22 birds were banded outside of the five field sites (see Chap ter 2), mainly due to restricted access to some sites the following field season (2002). Wrens were banded only at the five identified field sites during 2002 and 2003. In addition to the birds banded from 2001 through 2003, a small number of birds (six) were captured at two sites in 2000. There was no recorded mortality of birds due to captures. Only three (female) birds became stressed, the weights of two of these females suggested that they were carrying an egg; these females were released as soon as possible and located on subsequent surveys. Two minor
64 injuries were recorded during my study On one occasion an adult male suffered a small scratch on the tarsus when a Xanthium seed became lodged in the mist net. On the second occasion an adult male received minor injurie s attributed to a collared s parrowhawk ( Accipiter cirrhocephalus ) attempting to take the bird out of the net. Both birds survived were located on subsequent surveys. Morphology and P lumage I found morphological differences between m ales, females and juveniles (summarized in Tables 3 2 and 3 3). The average male was significantly larger than the average female in mass (males 11.1 0.6g, females 10.6 0.7g, t = 3.47, P < 0.01), wing length (males 54.0 1.7mm, females 51.7 1.5mm, t = 7.76, P < 0.001), tail length (males 71.7 5.4mm, females 68.7 6.5mm, t = 2.79, P < 0.01) and tarsus length (males 21 1.1mm, females 20.2 1.3mm, t = 3.26, P < 0.01). Juveniles were significantly smaller than adult males for all measurements wi th the exception of tail length. Juveniles were not significantly smaller than females except in weight. Interestingly, juveniles had longer tails on average (75.2 4.7mm) than either males (71.7 5.4mm: t = 2.98, P < 0.01) or females (68.7 6.5mm: t = 5.06, P < 0.001), despite a high degree of variation. The tails of juveniles were very short when they fledged and became progressively longer until the first molt. Many birds were molting during capture and the high degree of tail length variation is most likely attributable to molt cycle. Females were heavier at the beginning of the breeding season (weight in May June was 10.9 0.8 g (n = 20) than later in the season (weight in July August was 10.5 0.7 g, n = 40: t = 2.07. P < 0.05) Breeding mal es were easily identified by facial plumage (Fig. 3 1). Males have a characteristic purple crown with a black oblong spot in the centre, wide black facial mask that
65 extends around the nape and black bill. Adult females have grey crowns (ranging from a dar k bluish grey to grey with varying amounts of brown), chestnut ear coverts, complete white eye ring, distinct white eye brow, and brown bill (Fig. 3 2). Tails in both males and females are blue sometimes tinted greenish. The upper bodies of males and fem ales are cinnamon brown and the lower (chin, belly and breast) are cream with varying shades of buff. Eclipse males could be identified by their grey (sometimes partly brown) crowns, and grey to blackish ear coverts (Fig. 3 3). Some males kept their bree ding plumage throughout the year and a few males were still in eclipse plumage well into the breeding season although were known to be breeding based on the fact that they were the senior male or only male in the group. Juvenile birds (less than one year of age) were identifiable by their lack of male or female plumage (Fig. 3 4). Juveniles have brown crowns with none or very little grey, chestnut to brown ear coverts, brown (often light brown) bills, incomplete white eye circles, and lacked the obvious w hite eye stripe found in adult females (although some white may be present). In addition, juveniles often had longer tails (Table 3 3) and could sometimes be identified by their behavior. As juveniles matured, younger males were easier to identify than fe males as the former would acquire a few grey to blackish feathers in the ear coverts, purple or black feathers on the crown and a blackish bill. However, care had to be taken to not confuse juvenile males starting to acquire adult plumage with eclipse mal es or males molting. In general, I knew which groups had juveniles present and the dominant male within that group so identifying young males was easier. Confusion between juveniles and females can also occur and plumage varies greatly among individuals. Although adult female PCFWs usually have a grey crown, I found this to be highly variable, with numerous females displaying varying degrees of grey and brown plumage
66 on their crown. The degree of grey plumage appears to be linked to moult and not age (e. g., one female was at least six years old but yet her crown had a substantial amount of brown). Another adult (2+) female was first captured in 2001 and had a predominantly grey crown, but when she was recaptured in 2002 a lot more brown was present and f inally in 2003 her crown was a distinct jet grey. In addition to head coloration, breeding females could often be identified by bent (curved) rectrices from incubating, thus further distinguishing females from juveniles. Group S iz e and C omposition PCFW li ve in small family groups ranging in size from two to five birds, consisting of a breeding female and senior male and commonly included additional helpers (Table 3 4). The most frequently recorded group composition (in 50% of cases) was a single breeding male and female. The total number of males in these groups (88 group years) was 137, and of females 93: 2 = 4.611, DF = 1, P < 0.05). However, groups with one additional helper were also common (41%). Mean group size was 2.6 0.7 (median 2.5) based on 88 group years. Towards the end of the breeding season the number of individuals in a group could rise up to eight or more including progeny of the year (note that these latter were not included in this tabulation). Helpers in groups were predominantly males, and ranged in number from one to three; 44.3% of group years had one or more male helpers. Female helpers were rare and only 5 (5.7%) group years had a female helper, and none of these were present for the duration of the breeding season. Male helpers (usually progeny of the previou s year) frequently stayed in their natal territory for consecutive years. During my study 16 males remained helpers for at least two consecutive years. Pair bonds were often maintained over consecutive years with replacement of individuals of a pair occ urring after either death or divorce. I recorded 10 breeding pairs that remained
67 together for three consecutive breeding seasons and three breeding pairs remained together for two breeding seasons. Divorce was recorded three times and all six birds involv ed stayed within three territories of where they were first recorded. It is likely that more breeding pairs remained together for consecutive breeding seasons but as some birds were not banded it was impossible to determine the fidelity of those birds. Te rritory Shape and Size I mapped 30 territories in 2002, nine at Coolibah, five at Fitzroy, five at GNP1 and 12 at GNP2. Territories at both Coolibah and Fitzroy were mapped before cattle arrived. Territories were not mapped at Dashwood due to time constr aints. Territories are best described as polygons with an average size of 0.41 ha (4,135m 2 ). The shape of territories was highly dependent on width and distribution of the Chionachne and some sites had multiple territories perpendicular to the river (Figs 3 5, 3 6) Territories were smallest at Coolibah (0.26 ha) and Fitzroy (0.43 ha ) and larger within the two national park sites, GNP1 (0.55 ha) and GNP2 (0.45 ha), the difference in mean territory size was statistically significant (H = 8.62, P < 0.05). Foraging Behavior I recorded 963 minutes (16.05 hrs) of foraging observations at various locations during the 2002 and 2003 field seasons, 350 minutes were recorded for females, 550 for males, and 63 for juveniles. The greater number of observations made for males can be attributed to a larger number of males being present at the study sites and greater ease of following males with their conspicuous plumage. The lower number of observations for juveniles can be attributed to the fact that tracking juvenil es was more difficult as they were generally more timid and many had not yet been banded. Wrens tended to forage in loose groups and I often observed juveniles following adults as the group moved through vegetation during the breeding season. On occasion s where the
68 dominant male and female where busy re nesting after having produced a successful clutch, juveniles were often found with helpers and on numerous occasions I observed helpers feeding juveniles. When birds were foraging as a group, contact call s could be heard as the birds moved through dense understory. Wrens tended to forage individually when feeding nestlings or juveniles. Birds s pent the most time foraging in c anegrass Chionachne (60%), usually within 2 meters of the ground (Fig. 3 7). In terestingly, the introduced herbaceous weed, Xanthium (13%) was the second most common foraging substrate followed by freshwater m angrove Ba r ringtonia (12%), a small tree 5 8 m in height, and bareground with debris (8%). Less time was also spent foraging in tree species other than Barringtonia (3%), bareground (2%) and other substrat es such as the introduced weed castor o il Ricinu s and the small native shrub Flueggia (2%). In total, these comprised only a small portion of time compared to other substrates. Comparisons between vegetation cover and foraging substrate indicated that Chionachne and Barringtonia were the preferred substrates ( 2 = 123.981, DF = 4, P < 0.001). Overall, birds did occasionally forage in tall trees, but by far the majority of time was spent within 4 meters of the ground. Observations during floods also suggest that taller species (e.g. Ba r ringtonia Eucalyptus spp Ficus ) are important foraging substrates during floods when Chionachne is inundated. 2 = 15.282, DF = 5, P < 0.01). Males showed a greater affinity for Ba r ringtonia in comparison to females, whereas females showed a stronger affinity for Chionachne in comparison to males (Fig. 3 8). Although the sample size for juveniles was too small to analyze statistically, there were some notable differences in foraging preferences based on th e limited sample size (n = 63 min).
69 Juveniles spent approximately the same percent of time foraging in Chionachne as adults (57%) but also favored bareground with debris (19%) and a lower preference for Ba r ringtonia than either males or females (6%). Of t he 110 foraging behaviors recorded (Fig. 3 9), gleaning was by far the most common (60%) followed by snatching (18%). The following behaviors were recorded less frequently: pounce (8%), probe (7%), hover (4%) and hawk (3%). On 27 occasions the prey item c aptured by a bird was identified (Fig. 3 10). Lepidopterans (butterflies and moths) were the most common prey item captured (48%), followed by Orthopterans (grasshoppers, crickets and katydids 19%), Hymenopterans (ants, bees, and wasps 19%), Hemiptera ns (bugs, aphids and hoppers 7%), and Dipterans (flies and mosquitoes 7%). Breeding B iology I recorded a long breeding season of 22 weeks, the earliest month an active nest was recorded was in March (2002) and the latest a nest was found was in Septem ber (2003) with the majority of nests found in May (Fig. 3 11). No evidence of breeding was found from October to February. During the three field seasons, 65 active nests were found; eight in 2001, 22 in 2002 and 35 in 2003. Of the 65 nests found, 24 w ere found at the GNP1, 24 at GNP2, 10 at Coolibah and 6 at Dashwood. Nest searches were not conducted at Fitzroy (see methods). Purple crowned Fairy wren nests are large domed structures typical in the Malurus genus. In my study, only females built nest s, but females were often accompanied by the dominant male or the entire group during the building process. A female often showed an affinity for a specific so for some females nesting in the same territory over successive years. Nests were also frequently
70 placed close to clearings (e.g. drainage areas, roads, and tracks), perhaps to provide a better escape route for females from predators. Nests situated on the edges of drainage lines (i.e. generally small erosion valleys caused by flooding) were also generally higher, as nests would be placed in Chionachne on top of the bank, and would be more difficult to access for some predators (i.e. non avian predators) Purple crowned Fairy wrens showed a strong affinity for Chionachne as a nesting substrate; 85% (55) of nests were located in Chionachne The remaining nests were placed in Mne sithea 14% (9), another clumping riparian grass similar to Chionachne and on e single nest (1%) was found in a Ba r ringtonia sapling. Nests were constructed of varying amounts of Chionachne blades (the principal nesting material), Mne sithea blades, strips of bark, and leaves with other small rootlets, twigs, and spider webs added (usually in the nest lining). Nests were generally located low to the ground near the base of either Chionachne or Mne sithea grass clump. The average height of the base of the nest from the ground was 30.7 cm and the average height of the nest entrance w as 40.1 cm above the ground (Fig. 3 12). Nest entrance orientation varied but predominantly faced east (73%). The amount of cover (vegetation) surrounding nests ranged from 30% to 90% with an average coverage of 64.7%. Re nesting was common after either failed or successful nests and females would lay as many as four clutches in any given breeding season. From 2001 to 2003, I found 13 (15%) groups that had at least two nests, 17 (19%) groups that had at least three nest and five (6%) groups that had at l east four nests in a single season. Five groups (6%) of 88 group years) successfully fledged young from two clutches in a single year. All of these belonged to older females that were at least three years old.
71 Clutch size was determined for 49 of the 65 active nests. Eggs were faded white in color and had brown/reddish splotches predominantly on the rounded end. The mean clutch size was 2.9 ( 0.4), with 88% of clutches consisting of three eggs, 10% of two eggs and 2% of one egg. Only females incubated and eggs were laid on successive days. When a female was incubating, the senior male would sometimes follow the female to the nest and wait in nearby vegetation. However, most of the time the female would approach the nests alone and rejoin the group at the end of each incubating session. On occasions when females were flushed, they consistently moved off the nest at the last moment. I estimate incubation to be approximately 14 days. The nestling period last ed approximately 10 days in my study. Fledgl ings remained cryptic during the first week and stayed well hidden in dense undergrowth where they could often be heard begging for food as other group members approached. When potential predators were nearby, fledglings would be led away by group members Further, all group members, including helpers exhibited decoy 1997). The rodent run was also exhibited when I approached a nest or when juveniles were captur ed in mist nets. During my study high nest predation rates were found. Of the 51 nests where at least one egg was present; 10% of the clutches did not hatch or were abandoned and 39% of clutches were failed at the egg stage (predation) before hatching, on ly 51% of the clutches hatched. Of the 26 hatched clutches, 19% failed at the nestling stage (predation) but 46% fledged young. The outcome of nine nests was unknown (i.e. it was not possible to confirm whether nestlings fledged or were victims of pre d a t ion, however, no fledglings were found). The combined predation rate
72 for the 51 clutches was 51% with the possibility that the rate could be as high as 67%. I found no evidence of nest para sitism during the course of my study. I was not able to identify predators for any of the nests with the exception of two where cats ( Felis catus ) were identified by paw prints and both nests were completely destroyed. Predation by avian species is often indicated by a hole at the back of the nest; one nest was found in this condition suggesting an avian predator. Complete destruction of nests was very rare and nests were usually found intact or with only minor disturbance to the nest entrance. Nests were usually found completely empty with no sign of egg shell meani ng the eggs were taken whole. Potential predators, all of which were frequently ob served at study sites, include blue winged k ookaburra Dacelo leachii pheasant c oucals Centropus phasianinus Torresian c rows Corvus orru g oannas Varanus spp., several snak e species, and rats (including Rattus rattus ). In addition to the outcome data for the 51 nests with eggs described above, it was possible to determine which groups had fledged at least some offspring during the breeding season and calculate group annual p roductivity. Overall, the mean number of fledglings produced per group year (i.e. per breeding female of 88 groups) was 1.07 based on the number of juveniles I encountered and witnessed fledging. Productivity was very similar among years and the mean num ber of fledglings produced per female was 1.05 in 2001 (n = 22), 1.09 in 2002 (n = 34) and 1.06 2003 (n = 32) (H = 0.21, P = 0.9). It is quite possible that my productivity level is an underestimate as some juveniles may have dispersed or been predated be fore I could find them. For example, one group at Coolibah in 2003 fledged three young, however, each second day this group was visited the number of fledglings was reduced by one and no fledglings were left after six visits. Presumably, as these
73 fledgli ngs were still dependent on adults and too young to disperse, they were taken by predators. Annual Survival Adult annual survivorship was relatively high with 61% (68 of 112) adult banded birds surviving from 2001 to 2003. Survival rates were higher from 2001 to 2002 (70%) than from 2002 to 2003 (56%) (U = 358.5, P < 0.05). The lower survivorship rate from 2002 to 2003 can be attributed to the high number of birds lost from Coolibah, the intensely grazed site, during this time (71%) as discussed in detai l in Chapter 4. Discussion Ecology of the Purple crowned Fairy wren in The Victoria River District In my study of the PCFW in the VRD, average adult survival was 61%, mean clutch size was 2.9, incidence of successful multiple broods was 6%, annual produc tivity was 1.07 fledglings per female, and the breeding season lasted 22 weeks. These values are consistent with the basic life history traits of the Malurus genus and characteristic of Australian passerines. Purple crowned f airy wrens occurred in small f amily groups in well defined territories that were often maintained throughout the year and often for successive years. As with previous studies (Emlen, 1991; Rowley and Russell, 1997; Brooker and Brooker, 2001) PCFW family groups in the VRD consist of a breeding pair and often included one or more helpers, most likely retained progeny. Most helpers were males. Female offspring probably dispersed earlier than males, a notion supported by research on other Malurus species (Rowley and Russell, 1997). I di senior male survival and only three years of observations.
74 The PCFW show ed a strong affinity for Chionachne and this species plays a crucial role in all aspects of PCFW ecology (i.e. foraging, breeding and cover). Wrens were rarely found outside Chionachne habitat and when they were, only rarely would they stray more than 30 meters beyond the nearest Chionachne patch. Xanthium is a highly invasive weed easily spread by burrs (Smith, 1995) and was ubiquitous at all sites, often forming dense thickets along the riparian corridor and especially along drainage lines and erosion c hannels (i.e. areas with abundant bare ground). While it could be assumed that Xanthium is a favorable component within the PCFW habitat mosaic because it would provide foraging and cover opportunities, this species is suitable only for a short period as it quickly dries, leaving little cover or food. Moreover, although birds will forage in Xanthium it was not a preferred substrate. Because this weed may competitor of the preferred grass, Chionachne Xanthium should be regarded as a negative component w ithin PCFW habitat. Although the PCFW has been known to breed throughout the year when conditions are suitable (Rowley and Russe ll, 1997), the results from my study indicate a clear breeding season that commences at the beginning of the dry season and end s when conditions are at their driest in September. The initiation of breeding appears to coincide with the recession of flood waters and subsequent recovery of Chionachne once waters recede. During the three years of my study, the Victoria River flooded and the majority of Chionachne habitat was inundated to some extent. It is possible that breeding may occur at different times of the years when the river does not flood and Chionachne is suitable for breeding. Nest predation rates were found to be at l east 51% but could be as high as 67%. Nest predation rates for other Malurus species have been reported to be as high as 55% (Mulder, 1992) however predation rates may vary greatly among and within species, depending on
75 climatic and habitat c onditions ( Br ooker and Brooker, 2001). Nevertheless, the predation rate found during my study is one of the highest recorded for Malurus species (Rowley and Russell, 1997). Aside from two predation events attributed to cats and one to an avian predator, nest predator s were not identified but could potentially include a range of species commonly seen at all sites. However, native rats (e.g. Rattus villosissimus ) and exotic rats ( Rattus rattus ) were found at study sites and could potentially be a primary predator of wr en nests. Clutch predation by rats characteristically leaves nests intact, such as were found during my study and elsewhere rats have been shown to be significant predators of avian clutches (Major, 1991; Laurance and Grant, 1994). Despite high nest pred ation rates, females managed to fledge an average of 1.07 young per year (i.e. per group year). Undoubtedly, this would not be possible without the large number of replacement clutches, frequently four that females laid in a season. Although it is true t hat during my study females re nested up to three times after a failed attempt, it is also very possible that during less favorable breeding seasons this may not be possible. Thus the likelihood of females producing young with such high predation rates w ould be very small. The high levels of predation found during my study are of particular concern and need further investigation. In particular, comparisons between sites and identification of nest predators are crucial. The overall combined survival rat e for all sites was 61%, although survival rates were higher for 2001 through 2002 than 2002 through 2003. Overall adult survivorship is similar to other Malurus species (Rowley and Russell, 1997), although the survival rates at two of the sites within Gr egory National Park are remarkably high, potentially linked to very favorable site conditions in combination with high rainfall.
76 The high annual adult survival rates, extended breeding season, and multiple broods found during my study suggest the PCFW is able to exploit favorable conditions when they arise and could produce a high recruitment rate. However, variable adult survival rates found between sites indicate that even in favorable years, adult survival can be compromised. Further, my study took pl not reflect long term trends of the population in the VRD. Details of differences (i.e. annual adult survival, group size, and productivity) among sites are dealt with in detail in Chapter 4. Comparisons between Victoria and Drysdale River Populations There are clear structural differences between the riparian grass, Chionachne inhabited by the PCFW in my study and the woody palm like shrub, Pandanus inhabited by the P CFW studied by Rowley and Russell (1993). Consequently, the data reveal considerable differences between the VRD population and the previously studied Drysdale River population. Some of these differences can be attributed to habitat utilization and are not surprising, if we consider the structural dissimilarity between the two habitat types (e.g. nest height, territory shape and size). Although not directly addressed in my study preliminary differences in median body size between the two PCFW population s indicate further research is needed to determine the importance of these differences in regard to subspecies classification. At Drysdale River, in Pandanus habitat, group size of PCFWs was smaller (2.28; Rowley and Russell, 1993) in comparison to grou ps along the Victoria River (2.61) (D = 0.3020, P < 0.001). At Drysdale River, 80% of groups consisted of a sole breeding pair (Rowley and Russell, 1993) whereas along the Victoria River 50% of groups consisted of a breeding pair and an additional 41% had one helper.
77 The shape of territories of the PCFW between Pandanus and Chionachne habitat also differed according to the distribution of each species. Because Pandanus exists in narrow strips along the banks of the river, but Chionachne forms dense stands that can vary greatly in shape and width, PCFWs at Drysdale River were rarely found more than 10 meters from the river, whereas, if Chionachne was present, they could be found more than a kilometer from the river in the VRD. Chionachne in some areas (e.g GNP2, Fig. 3 6) was wide enough that it could support two wren territories extending back from the river. Hence, given a length of river, Chionachne habitat can potentially support more wrens than Pandanus habitat. However, this advantage may only be s ignificant in undisturbed areas as Chionachne is also more prone to disturbances (e.g. grazing) than Pandanus (Officer, 1964; Boekel, 1979). The prey items consumed by PCFWs reported by Rowley and Russell (1993) differ from the most common prey items cons umed during my study In my study Lepidopterans (butterflies and moths) and Orthopterans (grasshoppers, crickets etc.) were the most common observed prey items taken by wrens, whereas Rowley and Russell (1993) reported a large number of Coleopterans (bee tles), Hymenopterans (ants) and Hemipterans (bugs) as prey items. However, these differences are probably largely because of different methodologies. I used observational data that would be biased towards records of large conspicuous prey items in compar ison to the fecal analyses used by Rowley and Russell that could in turn be biased toward prey with hard exoskeletons (1993). Breeding data from the Drysdale River study indicated a bimodal breeding pattern, similar to the pattern found in some other tropi cal passerines and potentially related to food availability (Rowley and Russell, 1993; Noske and Franklin, 1999). Although there is a possibility that the difference in breeding pattern between the VRD and Drysdale River is due to site variables
78 (habitat or climate), Rowley and Russell (1993) suspected that the bimodal breeding pattern might simply reflect survey effort during the breeding season. Other Malurus species studied in depth (e.g. M. splendens and M. pulcherrimus ) show no signs of a bimodal bre eding pattern although neither one of these species are tropical in distribution (Rowley et al. 1991; Brooker and Brooker, 2001). However, data from my study also indicate that a bimodal breeding pattern is also unlikely for the PCFW, at least in Chionac hne dominated habitat. Consistent with primary habitat use, PCFWs at Drysdale River primarily used Pandanus as a nesting substrate (96%) and Chionachne or Mne sithea ( 99 %) in the VRD. Consequently, nests in the VRD were lower (307 mm) than those at Drysda le River (880 mm) consistent with nest substrate structure and height. Although nest predation rates at Drysdale River were unknown, higher nest location and placement in Pandanus (the serrated foliage of which may offer some additional protection) could h ave resulted in lower predation rates in Pandanus habitat. Despite the high predation rates encountered in Chionachne habitat in the VRD, average productivity was higher in this habitat (1.07) in comparison to the Pandanus habitat (0.78) at Drysdale River (Rowley and Russell, 1993). There was high degree of variation in productivity between years at Drysdale River ranging from 0.26 (young per group), following a year with low rainfall to as high as 1.22 (young per group) following a year with high rainfal l. There was little variation between years in the VRD (1.05, 1.09 and 1.06 young per group respectively from 2001 through 2003), however all three breeding seasons were preceded by high rainfall. If rainfall influences productivity in the VRD as much as it did at Drysdale River, productivity could potentially be extremely low in the VRD in low rainfall years if predation rates as high as they were in my study
79 Annual adult survival was high in both Pandanus (70.7%) and Chionachne (60.7%) habitat, varyin g between years in both habitats, and in the VRD between sites (Chapter 4). Both habitat types can provide good cover against predators and predator evasion has been recorded in both habitat types. For example, in my study PCFW evaded predators on numero us occasions by diving down into the dense Chionachne vegetation. During flooding, however, Pandanu s habitat would be less likely to be completely inundated and thus, be more resistant to destruction compared to Chionachne Thus, during and following floo ding events, other emergent vegetation may be important for the survival of wrens in Chionachne habitat. The differences found in my study in comparison to previous research highlight the need for a broad range of site specific studies in order to underst and the biology and ecology of a species, especially when the habitat a species occupies and external influences within that habitat could differ significantly.
80 Table 3 1. Number of PCFW captures by year (2001 2003), excluding preliminary captures in 2 000. Individuals 2001 2002 2003 Female 19 25 10 Male 24 28 14 Juvenile 8 4 13 1 st capture/banded 51 57 37 Recaptures 1 20 11 Total Birds Captured 52 77 48 Table 3 2. Morphological measurements of captured individuals captured fro m 2001 2003. Mean SD (range). Individuals Weight (g) Wing Length (mm) Tail Length (mm) Tarsus Length (mm) Female (n = 54) 10.6 0.7 (9 12.5) 51.7 1.5 (47.5 55) 68.7 6.5 (49.09 87.9) 20.2 1.4 (17.5 23.5) Male (n = 66) 11.1 0.6 (9.75 12.5) 54 1.7 (50 57) 71.7 5.4 (58.3 83.3) 20.98 1.1 (18.27 23.26) Juvenile (n = 26) 10.3 0.6 (9.5 11.5) 51.6 1.3 (49 54) 75.2 4.7 (62.69 85.9) 19.9 0.7 (18.51 21.27)
81 Table 3 3. Summary table of comparisons in differences between ma le, female and juvenile morphological measurements. F = female (n = 54), M = male (n = 66), J = juvenile (n = 25), W = weight, WL = wing length, TL = tail length, and TA = tarsus length (t tests, significant differences identified at P< 0.05 by *, P < 0.00 1 by **, not significant by NS). Weight (g) Wing length (mm) Tail length (mm) Tarsus length (mm) F M J F M J F M J F M J F ** ** NS ** ** ** NS M ** ** ** ** ** ** ** ** J ** NS ** ** ** NS ** Table 3 4. Group size and composition (n = 88 group years). Group Size and composition 2001 (n = 22) 2002 (n = 34) 2003 (n = 32) Total no. (n = 88) Total group sizes 2 (1M, 1F) 9 16 19 44 44 3 (2M, 1F) 12 10 12 34 36 3 (1M, 2F) 0 2 0 2 4 (3M, 1F) 1 2 0 3 6 4 (2M, 2F) 0 3 0 3 5 (4M, 1F) 0 1 1 2 2
82 Table 3 5. Nest location and measurements 2001 2003. Nest substrate (n = 65) Chionachne 85%, Mnesithea 14%, Barringtonia 1% Nest entrance (n = 33) Mean 40.5 cm (SD 12.2) Nest base (n = 35) Mean 30.7 cm (SD 14.2) Nest entrance orientation (n = 33) East 73%, West 12%, North 9%, South 6% % Vegetation cover (n = 36) Mean 65 % (SD = 15.1) Table 3 6. Fate of nests with eggs (n = 51). Fate of clutches Number of nests (n = 51) Fledged at least 1 young 12 Abandoned (eggs un hatched) 5 Failed at egg stage (predation) 20 Failed at nestling stage (predation) 5 Unknown (but fledglings not located) 9
83 Figure 3 1. Adult male PCFW, full breeding plumage. Figure 3 2. Adult female PCFW.
84 Figure 3 3. Eclipse male PCFW. Figure 3 4. Juvenile PCFW.
85 Figure 3 5. Location and shape of territories at Coolibah Crocodile Farm and Fitzroy Station, 2002. Coolibah Fitzroy Territories
86 Figure 3 6. Location and shape of territories within Gregory National Park (GNP1 and GNP2). GNP1 GNP2
87 Figure 3 7. Time spent foraging (%) in various substrates (n = 963 min). Approximate height of primary foraging substrates; Chionachne (c anegrass) = 1 3m, Xanthium (noogoora b urr) = < 1m, Ba r ringtoni a (freshwater m angrove) = 4 8 m.
88 Figure 3 8. Male and female foraging preferences. Percent of time spent in each substrate. CHIO = Chionachne XAN = Xanthium BAR = Ba r ringtonia BGD = Bareground w/debris, BARE = Baregrou nd (n = 900 min.).
89 Figure 3 9. Foraging behavior. Number of times each behavior was observed, expressed in %. (n = 110).
90 Figure 3 10. Prey items captured. Percent of total prey types, by insect order (n = 27).
91 Figur e 3 11. Number of nests found per month, 2001 2003 (n = 65).
92 Figure 3 12. Nest entrance height (n = 33).
93 CHAPTER 4 RESPONSE OF THE PURP LE CROWNED FAIRY WREN TO DISTUR BANCE FACTORS The aim of this chapter is to investigate the primary threats to the PCFW population in the VRD. Specifically, I focus on the effects of cattle grazing as this disturbance has been implicated as a cause of recorded structural changes in riparian and other ve getation, and also been correlated with changes in avian diversity and abundance. The VRD presents an excellent opportunity for comparative measures as grazing was present in riparian areas at varying degrees. Weeds and fire are additional potential thre ats to the PCFW; however these are only discussed superficially in comparison to grazing for compelling reasons. Weeds occurred at all sites, making comparative analyses more difficult. Fire was witnessed only on a single occasion at a very small scale a nd is discussed in this context. Introduction In the southern regions of Australia, there has been a significant decline in avian diversity, and in the extent and quality of native vegetation. In some regions, habitat modification can be as high as 90%, a nd this loss has been implicated as one of the driving forces of bird decline (Recher and Lim, 1990; Ford et al. 2001 ). Although the changes in faunal composition may not be as obvious as they are in the southern region, recent studies in northern Austra lia are demonstrating large scale yet subtle changes to avian populations as well ( Franklin, 1999; Garnett and Crowley, 200 0; Woinarski and Catterall, 2004 ; Hobbs, 2005). Although species responses vary, pastoralism and changing fire regimes may be the u nderlying causes of these changes (Woinarski, 1990; Crowley and Garnett, 19 99; Woinarski and Ash, 2002). The impacts of some disturbances, most notably fire, has been studied for several Malurus species (e.g. Malurus elegans and Malurus splendens ) (Brooke r and Brooker, 1994; Russell and Rowley, 1998). Habitat fragmentation has also been studied in detail for species
94 such as Malurus pulcherrimus (Brooker and Brooker, 2001) and the effects of fragmentation and fire have been modelled for the splendid f airy wren, Malurus splendens (Brooker and Brooker, 1994). The Purple crowned fairy wren however, has not been studied in this context even though it is confined to riparian habitat prone to disturbances such as weed invasions and grazing. Anecdotal evidence has suggested that grazing, erosion, and weeds are the primary causes of PCFW population declines and fragmentation (Smith and Johnstone, 1977; Garnett, 1992; Rowley, 1993) but until the current study, there has been no concrete evidence of such effects. E lsewhere, the effects of grazing on bird species in general have been the subject of numerous studies that have overwhelmingly concluded that grazing has a negative influence on avian diversity and nest suc cess ( Croonquist and Brooks, 1993; Ammon and Stace y, 1997; Krueper et al ., 2003; Scott et al ., 2003). Grazing in particular changes understor y composition and struct ure ( Fleischner, 199 4 ; Jansen and Robertson, 2001; Martin and Possingham, 2005; Werner, 2005), these changes may be one of the most importan t effects of grazing on the PCFW as i t is dependent on the understor y for foraging, breeding and cover. For example, Martin and Possingham (2005) found that foraging substrate was a good indicator of how an avian species would cope under grazing pressure. Thus, species foraging at lower and medium levels of the vegetation strata would be influenced by grazing at higher levels than species tha t forage in the upper canopy. In the case of the PCFW, the results from Chapter 3 reveal that the majority of forag ing takes place at lower levels of the vegetation with little time spent in the middle to upper veg etation levels. Riparian zones tend to be exposed to higher rates of disturbances than other habitats and riparian obligates such as the PCFW have evolved w ith endogenous disturbances such as annual flooding and occ asional fires. However, exogenou s disturbances within the riparian zone, such as grazing by domestic stock, are relatively new and as a species that spends the
95 majority of its time within four met ers of the ground, the PCFW may be particularly susceptible to changes (structurally and floristic ally) in the understor y. In the VRD, 59% of the land is used for pastoralism, and supports 545,000 cattle (Karfs and Trueman, 2005). Aside from the economic benefits of pastoralism, the industry carries strong historical ties to an era whose sentiment is still felt strongly throughout Australia. However, the industry has contributed to the modification of savannas in northern Australia and this modification has influenced faunal composition (Hobbs, 2005). I used a comparative approach to determine the threats to the PCFW and its primary habitat focusing on grazing effects, using site specific habitat variables and site specific ecological and life history par ameters ( group size, adult survivorship, and productivity). I focused on comparisons between grazed and ungrazed sites in addition to pre and post grazing comparisons, to determine relative differences. The roles of fire and weeds are included superficia lly based on the limited data available for these variables, and the lack of scope in my study for manipulating these variables. Methods Vegetation and Site C ondition Grazing pressure by cattle varied between sites. Coolibah was heavily grazed (from 2002 ); Dashwood had a history of heavy grazing during my study but was now lightly grazed; Fitzroy was lightly/moderately grazed; and the two sites within Gregory National Park (GNP1 and GNP2) were ungrazed. These grazing categories were subjective, based pri marily on the number of cattle observed at sites; nevertheless, they provided a baseline for comparative analyses (also see Table 2 3, Chapter 2). At each of the five fixed study sites, I mapped vegetation using a point intercept method along 50 meter tran sects. T ransects (two in each PCFW territory) were placed i n a diagonal line, parallel to each other to ensure the maximum coverage possible and were
96 placed at least 20 meters apart (Fig. 4.1 ). At each one meter intersect (point data) I recorded the foll owing; substrate (plant species, bareground, bareground w/ de bris, or dead trunk/branch), substrate height canopy cover (species) and height. In addition, the substrate located within 20 cm radius of the intercept (plot data) was also recorded. To assess wallaby and cattle densities I recorded the presence or absence of wallaby feces and cattle f eces and hoof print s within a 20 cm radius of each o ne meter intercept Unfortunately, it was not possible to conduct two transects for all PCFW territories due t o the high prevalence of buffalo and scrub bulls which made it impossible to conduct surveys safely. On another occasion it was impossible to finish all transects at Coolibah due to the interferenc e of an adjoining stakeholder. A ll data were collected du ring the 2002 dry season (June September) when all five sites had been established and PCFW territories had been identified. I chose this time for logistical reasons as three of the sites were not accessible durin g the wet season. Further, this period coi ncided with the PCFWs breeding season a time when vegetation cover and structure might be more impor tant for nesting and foraging. Differences among sites for each environmental variable were analyzed using ANOVA. Tests were followed by a posteriori Tuke variables were significant in the ANOVA at the 0.05 level. Percentage values were arcsine transformed prior to statistical analyses (Fowler et al. 1998). O verall habitat diff erences among sites were analysed u sing ordination (multi dimensional scaling) to portray variation across all measured habitat variables described above (excep t for the two variable cattle f eces and wallaby feces) This was undertaken in the program PRIMER ( Clarke and Gorley, 2001 ), and w as based on the Kylczynski similarity index. A two dime nsional solution is presented.
97 The extent to which the matrix of pair wise similarity values was associated with site was examined using ANOSIM ( Clarke and Green, 1988 ). Fire Over the course of my st udy there was only one incidence of fire, and the extent was minimal with only three territories affec ted to any great extent. Although I witnessed the fire, it occurred before vegetation data was obtained Photographic evidence does show the regeneratio n of Chionachne at this site and it was also possible to document PCFW mortality and re colonization post fire. Variation A mong S ites The methods I used to collect ecological data on the PCFW are detailed in Chapter 3. Differences in PCFW biology and eco logy among sites were analyzed using all sites un less sample sizes were too small, these are indicated where present. At GNP1 and GNP2 (both ungrazed), all sites and transects were used. At Coolibah pre grazing and post grazing measures were possible for a number of PCFW variables (i.e. group size, adult survival and productivity). Sample sizes at Fitzroy for PCFW biological and ecological data were too small for comparative analyses Dashwood had been continuously grazed during all my field seasons maki ng any comparis ons between years superfluous. F oraging preferences of PCFW at one site (GNP2) served as the base line data to examine this variable I did not want to collect data at grazed sites to avoid any potential confounding effects of grazing Da ta to determine foraging preferences were based on number of minutes per foraging substrate (see Chapter 3 for a full description of methods for foraging data) and substrate abundance s were based on plot data based on 17 transects conducted at this site (G NP2). A d escription of vegetation transects was described previously in this chapter
98 Comparisons among sites of PCFW ecological variables (i.e. annual adult survival, group size, productivity) were tested using either a Mann Whitney U test (for compariso ns between two groups/sites) or a Kruskal Wallis test (for comparisons involving more than two sites): I used thes e non parametric tests as the assumption of normality could not be made for these variabl es. I used a Chi square test to compare the frequenc y distribution of foraging sites (substrates) against the actual proportion of those sites in the environment (using Minitab 13.1, Minitab Inc., Coventry, UK). Results Vegetation a nd Site C ondition The incidence of grazed Chionachne was highly variable be tween sites; two sites (GNP1 and GNP2) displayed no evidence of grazed Chionachne two sites had similar amounts of grazed Chionachne (Dashwood; 31 % and Fitzroy; 39 %) and at Coolibah the majority (93%) of Chionachne was grazed (Fig. 4 2 F 4,70 = 135.2 P < 0.001). Cow and wallaby density varied among sites (Fig. 4 3). Cow evidence (f eces and hoof print) was the most prevalent at Coolibah and was higher than any of the other sites (F 4,70 = 31.7 P < 0.001). Cow evidence was similar for Dashwood and Fitzroy and non existent at GNP1 and GNP2. Wallaby evidence (f eces) also varied between sites and was highest at Dashwood ( 40 %) followed by Coolibah (32%) and Fitzroy ( 20 %). Wallaby evid ence was much lower at GNP1 (2 %) and GNP2 (0.2%) than any of the other sites (F 4,70 = 92.8 P < 0.001). Figures 4 4 through 4 6 illustrate the change in Chionachne cover that occur r ed when cattle were moved onto the site at Coolibah in May, 2002 Figure 4 4 illustrates the site just before cattle arrived and Figs 4 5 and 4 6 aft er cattle had been at the site for approximately 4 months. Although the base of the Chionachne is still present, the change in cover is clearly
99 vis ible. Note that the timing of cattle arriving, at the end of May coincided with the peak of the breeding s eason. I sampled seventy five veg etation transects in 2002 representing 39 te rritories at the five fixed field sites. Twenty two transects (14 PCFW territories) were conducted at Coolibah, five (2.5 PCFW territories) at Dashwood, 14 (seven PCFW territorie s) at Fitzroy, and 17 (eight and a half PCFW territories) at both GNP1 and GNP2. There were only a few groups present at the Dashwood Crossing site accounting for the low numbe r of transects at that site. Note that in the corresponding figures, Coolibah, Dashwood and Fitzroy represent the grazed sites and GNP1 and GNP2 represent the two ungrazed sites located within the national park. Ground cover was not very diverse, dominated by one plant species, Chionachne cyathopoda (Fig. 4 7 ). Overall, Chionachne ac counted for 19% of point data and 39% of plot data For point data Chionachne frequency ranged from 10 % at Dashwood to 26 % at GNP2 (Fig. 4 8 ). Chionachne frequency (plot data ) was significantly lower at Coolibah than GNP2 but not GNP1 (F 4,70 = 5.8, P < 0.001). Chionachne frequency of plot data was simila r to that of point data, ranging from 21 % at Dashwood to 49 % at GNP2 Dashwood had the overall lowest coverage in comparison to all other sites (F 4,70 = 6.4, P < 0.001). Chionachne height, at height i ntervals of <1 m, 1 2 m, and >2 m, differed among sites (Fig. 4 9 ). Chionachne greater than two meters in height was most prevalent at the two ungrazed sites, GNP1 (40%) and GNP2 ( 46 %), no Chionachne greater than two meters was recorded at either Coolibah or Dashwood. Although Chionachne more than two meters was less common at Fitzroy (32 %), this difference was not significant. Coolibah had a disproportionately larger number of Chionachne stands l ess than one meter in height (70 %) in comparison to all si tes except Dashwood (F 4,70 = 19.7 P < 0.001). In contrast, the two sites that were not exposed to any grazing by cattle (GNP1 and GNP2) had very few stands
100 that were l ess than 1 meter in height (14% and 8 % respectively). Medium height (1 2 meters) Chion achne stands were less prevalent at Coolibah than the remaining sites (F 4,70 = 2.8, P = 0.031). Baregrou nd accounted for an overall 22% of point data (Fig. 4 7 ). Bareground was most common in erosion gullies and along drainage lines where it was often ex tensive. I found a substantial difference in bareground incidence between sites most notably at Coolibah where bareground was prevalent at both at points ( 42 %) and plots (33%). In contrast, GNP2 had a lower frequency of bareground; ba reground accounted for only 8% of point data and 5% of plot data (Fig. 4 10 ). Relative to the other sites, Coolibah had a higher frequency bareground than Fitzroy, GNP1 and GNP2 but not less than Dashwood ; further, GNP2 had a higher frequency of bareground than Dashwood and Fitzroy but not GNP1 ( point data F 4,70 = 21.3, P < 0.001).Within plots Coolibah had an even higher frequency of bareground in comparison to all other sites including Dashwood, and Fitzroy had a higher incidence of bareground than either of the ungrazed sites GNP1 or GNP2 (plot data F 4,70 = 38.7, P < 0.001). Bare ground with debris (e.g. leaves and dried Chionachne blades) was the dominant non vegetative cover and accounted for 53 % of point data and 26% of plot data (Fig. 4 7 ). Debris was often very thi ck, especially at sites where Chionachne was very dense when a thick layer of dried blades was often present (e.g. GNP2). The amount of bareground with debris varied between sites with the lowe st frequency found at Coolibah (41 %) and the highest at GNP2 ( 59 %) ( Fig. 4 11, point data F 4,70 = 5. 8, P < 0.001). However, the difference between sites using the plot data was less distinct although Fitzroy had more bareground with debris than Coolibah and GNP1 ( plot data F 4,70 = 3.6 P = 0.01) The invasive weed Xanthium was t he second most common understor y plant species encountered at study sites ( overall frequency ; 2% of point data, 9% of plot data Fig. 4 17 ).
101 Xanthium frequency at points was greatest at Dashwood (6 %) followed by GNP2 (5 %) in comparison to th e sites, Fitzroy and GNP1 which had the same amount of coverage (2 %) and no Xanthium was present at Coolibah (Fig. 4 12a, point data F 4,70 = 12.6 P < 0.001). Despite relative low frequency of Xanthium at points Xanthium was pres ent at higher frequency w ithin plots, although the pattern of frequencies between sites remained constant. More Xanthium was found at Dashwood (2 6%) than either Coolibah (2%), Fitzroy (8 %) and GNP1 (6 %) but not significantly more than at GNP2 (18 %) which had more Xanthium than Co olibah and GNP1 (plot data F 4,70 = 9.7 P < 0.001). Figure 4 12 b illustrates the extent of Xanthium alo ng the river bank at Dashwood. In addition to Xanthium two other weed species, Passiflora foetida and Ricinus c ommunis were also found at some of th e study sites The frequency of the three combined weeds ( Xanthium Passiflora Ricinus ) followed a similar pattern to the distribution of Xanthium alone (Fig. 4 1 3 ). Dashwood and GNP2 had the highest incidence of weeds both points (6% at each site) and within plots (26% and 22 %). Although Coolibah had the lowest incidence of weeds at both points (0.3%) and within plots (3 %) in comparison to all sites, the di fference was significantly less than at Dashwood and at GNP1, but not less than at Fitzroy (2%, 8 %) and GNP1 (2 %, 7 %) w hich had similar weed frequencies (point data F 4,70 = 10.0 P < 0.001; plot data F 4,70 = 11.6, P < 0.001). Total canopy cover varied among sites (F 4,70 = 13.8 P < 0.001), with the h ighest cover found at GNP1 (72%), GNP2 (67 %) and Dashwood (65%) (Fig. 4 14 ). Coolibah and Fitzroy had statistically signifi cantly lower canopy covers, 24% and 38 % respectively, in comparison to the national park sites GNP1 and GNP2, however, Fitzroy did not differ significantly statistically from Dashwo od. F our species comprised the canopy cover. These included two Eucalypt species, E. microtheca (coolibah or f looded Box) and E. camaldulensis (river red g um), Ba r ringtonia
102 acutangula (freshwater m angrove) and Ficus coronulata (river f ig). E. m icrotheca (58 %) was the most common canopy species followed by E. camaldulensis (27 %), Ba r ringtonia (6 %) and Ficus (4 %) (Fig. 4 15 ). Cano py species differed among sites, primarily in the relative abundances of E. microtheca and E. camaldulensis ( E. microtheca F 4,70 = 4.5 P = 0.003; E. camaldulensis F 4,70 = 4.7 P = 0.002 Fig. 4 16 ). Eucalyptus m icrotheca was the dominant canopy cover at all sites with the exception of GNP2 where E. camaldulensis was more abundant. E. microtheca was especially dominant at Dashwoo d, where E. camaldulensis was completely absent. Ba r ringtonia was more common at GNP1 than elsewhere but this difference was significantly larger only than th at at Coolibah (F 4,70 = 3.9 P = 0.006). Ficus was present in low densities at all but one site (Dashwood) and was most common at GNP2 but abundance differences among sites did not vary greatly (F 4,70 = 0.52, P = 0.719). Overall, there was substantial variation in PCFW habitat across the set of five sites as determined by an ordination of all transe cts by their ha bitat variables ( Figure 4 17) This ordination has an acceptable stress value of 0.20. The variation is most marked between the two national park sites (GNP1, GNP2) on one hand (with no overlap of groups of clusters) and the heavily graze d site (Coolibah) on the other t he less int ensively grazed sites (Dashwood and Fitzroy ) fall between these clusters. This distinction among sites was also corroborated by ANOSIM, where the Global R statistic was 0.60 (p<0.001), indicating that there was a substantial overall variation in environmental variables among sites. Pair wise comparisons among sites are given in Table 4 1 ; these show significant environmental variation between all pair of sites (other than between Dashwood (where sample size was small) and Fitzroy): differences were most marked between Coolibah and each of the two national park sites (GNP1 and GNP2).
103 Fire An area of GNP2 was burned in October 2001 in an effort to create a fire break from an approaching fire As it crossed the stu dy site, t he fire was initially very hot (consuming some of the canopy) but soon (within 50 meters) slowed down and burned only patches of ground level vegetation. The understor y of three PCFW territories was burnt severely (Fig. 4 18a). During the fire one adult female disappeared ; she was never seen again and was presumed to have perished. T he male of the original pair was subsequently located two territories from his original territory Regeneration of Chionachne occurred rapidly and three months la ter ( December 5, 2002 ), a reasonable understor y was present (Fig. 4 18b ) and birds were found foraging in the area B y February 2002, this territory was once again occupied by a breeding pair breeding season all three territories tha t had been heavily burned in 2001 were once again occupied by breeding pairs. Variation A mong S ites The number of PCFW captured and banded at the five field sites is summarized in Table 4 2. Average g roup size among sites for combined years (2001 2003) var ied significantly between sites (Fig. 4 19 ) (H = 2.0 P = 0.058) Fitzroy was not included in the analyses as the sample size was very small ( n = 4 ) Numerically, the lowest average group size was found at Dashwood (2.1) and GNP2 had the highest (2.8). G roup size was the same at Coolibah and GNP1 (2.7) and slightly lower at Fitzroy (2.4) C omparisons of group sizes at Coolibah pre and post grazing indicate d a significant reduction in group size when cattle entered the site ; average group size pre grazin g was 3 .0 in contrast to post grazing where mean group size was 2.2 ( U = 141.5, P = 0.014). I found substantial differences in adult survivorship among sites ( H = 13.7, P = 0.003, Fig. 4 20 ). Adult survivorship for two combined years (2001 2002 and 2002 2 003) was
104 highest within the national park, 96% at GNP1 and 83% at GNP2. Coolibah and Dashwood had lower survival rates, 51% and 30% respectively. Survival rate was lowest at Fitzroy (29 %) however this was based on only one year (2001 to 2002) and as this site was accessed infrequently it was not included in the statistical analysis. S urvival rate at Coolibah was much higher from 2001 through 2002 (90%) before cattle arrived in comparison to 2002 though 2003 after cattle arrived (26%), a 64% loss of adult banded birds. The median difference between these two years was statisti cally significant ( U = 53.5, P < 0.01). Purple crowned f airy wrens showed a preference for particular foraging substrates as time spent foraging in particular substrates was not pro portional to the amount of substrate present (Fig. 4 21 ) 2 = 15.282, P < 0.01 ). Most notably, Chionachne and Ba r ringtonia were preferred foraging substrates and bareground and bareground with debris were avoided, the difference between Xanthium abundance and foraging time was not significantly different. I found some difference between group productivity among sites, the number of fledglings produced per breeding female was highest at Dashwood (1.36), Coolibah (1.29) and GNP2 (1.22) with l ower rates were found at GNP1 (0.83) and Fitzroy (0) (Figure 4 22 ) however, this difference was not statistically signifi cant (H = 1.94, P = 0.586). Although Dashwood had the highest group productivity, this does not reflect the density of PCFWs at that site. Dashwood had the lowest mean group size and high adult mortality rates, thus the actual density of PCFWs at this site was very low in comparison to other sites. The unexpected low productivity rate at GNP1, within ungrazed Gregory National Park, may be attributed to high predation rates at this site. Of the 20 clutches found at GNP1 10 (50%) were eaten and three (15%) were abandoned; only four (20%) were successful and the fate of three (15%) was unknown but likely eaten.
105 Discussion I found some significant di fferences between sites in relation to grazing pressure and my results support the premise that grazing can have a profound effect on riparian vegetation structure and the PCFW population along the Victoria River. The differences found in the understory b etween grazed and ungrazed sites are not unexpected, especially at Coolibah where grazing pressure was intense, but the differences among sites was further bolstered by changes seen over time at a single site where birds were studied pre and post grazing. The reduction in group size and adult survivorship of the PCFW at the intensely grazed site occurred swiftly and were alarming. That grazing results in significant changes in vegetation structure and that these changes in vegetation vary according to gra zing intensity is, in itself, not particularly surprising. It is consistent with previous studies that have shown changes in vegetation structure as a result of grazing, and that high stocking rates cause the most dama ge to riparian vegetation ( McIntyre an d Lavorel, 1994; Yates et al. 2000; Jansen and Robertson, 2001 ; Krueper et al. 2003 ). During the dry season, cattle congregate along the river where they have easy access to water. Chionachne is a palatable grass and when other species of grasses becom e dry, Chionachne often remains green for a longer period, making it an attractive food source for cattle along the riparian corridor. Coolibah, despite being heavily grazed, still retained stands of Chionachne suggesting that one season of grazing is no t enough to destroy the root base. In contrast, at Dashwood, a site grazed over a long period of time, there were fewer stands of Chionachne than any other sites (both grazed and ungrazed) implying that long term grazing has an influence not only on Chion achne height but also on overall abundance. Chionachne height varied significantly among sites being much shorter at grazed sites, thereby reducing the amount of cover available for PCFWs.
106 At grazed sites, wallaby density was higher than at both ungrazed sites. Wallabies were also present at one of the national park sites (GNP1) where they congregate at the watered campsite adjacent to the river where green grass is always available. The thick reed like properties of healthy Chionachne may not be attract ive to wallabies, and, indeed I found no evidence of Chionachne being eaten by wallabies at GNP1. However, the regrowth after cattle grazing, floods and fire is soft and tender, making it attractive to wallabies and during this time there may be substanti al influence of wallaby grazing on Chion a chne cover (Meers and Adams, 2003). Agile w allabies ( Macropus agilis ) are known to eat a diverse diet allowing them to exploit a vast array of resources in a variable environment (Stirrat, 2002), particularly in th e dry season, so Chionachne regrowth may be a valuable resource to them. However, any potential grazing pressure by wallabies on Chionachne regrowth would most likely have little effect on the grass abundance compared to intense grazing by cattle; furthe r their effects on soil would be less severe (Bennett, 1999). This notion is supported by research on pasture grazing by black striped w allabies ( Macropus dorsalis ) where, despite a widespread belief in the significant effect of wallaby grazing on pasture the greatest effect on vegetation structure was cattle grazing (Baxter et al. 2001). Weeds, in particular Xanthium are pervasive through the riparian corridor of the Victoria River and Xanthium occurred at most of the study sites. At Coolibah, Xanthiu m was less abundant, because high stock densities caused an overall reduction in all herbaceous ground cover due to trampling, resulting in the large proportion of bareground encountered at this site (results compar able to studies elsewhere: Fleischner, 19 94; Jansen and Robertson, 2001). Xanthium was most prevalent at Dashwood, however, one of the sites within the national park (GNP2) also had a very high incidence of Xanthium This is perhaps a reflection of the large number of open areas (drainage lines and flood valleys) present at the national park site (GNP2) site. Although PCFWs foraged in Xanthium this was not a
107 preferred foraging substrate compared to Chionachne and Ba r ringtonia Perhaps some cover is b etter than none ( bareground). Importantly, my study indicated that PCFWs will exploit this resource if it is available, but it is equally important to consider that Xanthium is only suitable for foraging only f or a short period of time ( from the time it is tall enough to provide cover until it drie s), and that this time does not extend through the breeding season. Furthermore, this species is not suitable for nesting, as the single stems could not support a nest, and may be negative component if it competes with Chionachne Difference in canopy cov er may be attributable to flood dynamics as extensive flooding in this region is common and, based on water flow dynamics, floods are an important moderato rs of riparian vegetation ( Start and Handasyde, 2002). During flooding events, canopy may become mo re important as Chionachne may be inundated. Opportunistic observations of PCFWs during flooding events found birds foraging in most canopy species found within the riparian zone. During flood events, which occurred during all three field season, PCFW wer e observed in Bar r ingtonia Ficus and E. microtheca substrates not noted to be used frequently during the non flood times that comprised the bulk of my fieldwork. Although fire was only witnessed once during the course of my study I did record some assoc iated adult mortality in the PCFW territory that burned the hottest. Chionachne regenerated quickly after fire, with adequate rainfall, and was inhabited the following season. It is important to note that regeneration of Chionachne after fire was depende nt on the timing of rainfall subsequent to the fire. One of the primary concerns about fire in Chio n a chne stems from the fact that it grows in long connected corridor usually without appropriate fire breaks, allowing fire to spread throughout a large area Because Chionachne often exists in monospecific stands, a widespread fire would leave little cover for PCFWs in a large area until regeneration. Furthermore, any presence of grazing animals at a site following fire could further hamper regeneration of Chionachne
108 For a species whose individual annual survival rates can be over 90% (e.g. within Gregory National Park), the loss of 74% of adult birds in one breeding season (e.g. at Coolibah when cattle were introduced) is great c ause for concern. No similar decrease in adult survival was recorded at either national park sites, and it is unlikely that some large scale environmental factor was responsible for the change to PCFW recorded at Coolibah. On the other hand, despite the decrease in adult survivorship and group size at the intensely grazed site (Coo l ibah), group productivity did not decrease. The remaining birds still produced an average of 1.09 juveniles per group (i.e. per breeding female), very close to the average ann ual productivity across all sites. The same was true at Dashwood, where despite relatively high adult mortality rates and smaller group sizes, group productivity rates were similar to those at ungrazed sites. The long term perspective of the PCFWs at hea vily grazed sites may not be as secure as at ungrazed sites, had I been able to determine juvenile survival. It is likely that juvenile survival at Coolibah after grazing began may have been very low, judging by the fact that very few birds were retained as helpers within their natal territory. It is important to note that despite high productivity at Dashwood and Coolibah the total number of birds produced at these sites was low, hence, the se sites retain overall low PCFW densities in comparison to the ot her ungrazed sites. At one of the national park sites (GNP1), group productivity was low, even though adult survival was high. This site had good habitat qualities; it was not subjected to grazing, Chionachne was abundant, and weed presence was lower th an the other national park site (GNP2). This suggests that habitat quality may not be, at least directly, totally responsible for a low productivity rate. However, nest predation rates seemed to be very high at this site (Chapter 3), and perhaps this fac tor resulted in lower productivity rates than the second site within the national park (GNP2).
109 In summary, changes attributable to intense grazing (i.e. at Coolibah) resulted in a rapid reduction in Chionachne height and hence cover for the PCFW, which sp ends most of its time in lower vegetation (Chapter 3). Most importantly, an extremely high level of adult mortality marked the transition of one site from ungrazed to grazed within a single breeding season, strongly suggesting that grazing is a primary thr eat to PCFW populations in the VRD. Long term grazing (i.e. at Dashwood) was associated with a reduction in Chio n a chne cover, an increase in weed abundance and high adult mortality of PCFWs. Purple crowned f airy wren conservation management is the topic o f the next chapter.
110 Table 4 1. Pair wise differences between sites in the set of all environmental variables measured. Values in body of table give R, with associated probability levels marked as ns = not significant, p<0.05, ** p<0.01 and *** p<0.0 01. Site GNP1 GNP2 Coolibah Fitzroy GNP2 0.16 ** Coolibah 0.92 *** 0.92 *** Fitzroy 0.27 *** 0.40 *** 0.65 *** Dashwood 0.36 0.32 0.84 *** 0.001 ns Table 4 2. Purple crowned f airy wren captures per site. Year/Site Coolibah (grazed ) Dashwood (grazed) Fitzroy (grazed) GNP1 (ungrazed) GNP2 (ungrazed) 2001 12 2 8 4 7 2002 29 11 9 12 20 2003 24 6 0 9 9 Total no. captured 65 19 17 25 36 Figure 4 1. Positioning example of 50 meter transects within ter ritories. Territory A Territory B
111 Figure 4 2. Frequency (%) of grazed Chionachne at individual study sites (n = 75). COOL = Coolibah, DASH = Dashwood, FITZ = Fitzroy, GNP1 and GNP2 = Gregory National Park sites. Grazed Ungrazed
112 Figure 4 3 Wallaby ( Macropus agilis ) and c ow ( Bos Indicus ) evidence at individual study sites. Wallaby evidence meas ured as % feces and Cow as % f eces and hoof prints found (plot data) (n = 75). COOL = Coolibah, DASH = Dashwood, FITZ = Fitzroy, GNP1 and GNP2 = Gregory National Park sites. Grazed Ungrazed
113 Figure 4 4 Chionachne (c anegrass) stands in May 2002 at Coolibah before cattle were introduced.
114 Figure 4 5 Chionachne (c anegrass) stands, October 2002, four months after cattle arrived at Coolibah, Octob er 200 2.
115 Figure 4 6 Chionachne (c anegrass) stands after heavy grazing at Coolibah, October 2002.
116 Figure 4 7 Groundcover frequencies (% ) at study sites. BGD = Bareground w/ debris, CHIO = Chionachne cyathopoda BG = bareground, and XAN = Xanthium strumarium (n = 75).
117 Figure 4 8 Frequency of Chionachne cover (%) at individual study sites (n = 75). COOL = Coolibah, DASH = Dashwood, FITZ= Fitzroy, GNP1 and GNP2 = Gregory National Park sites. Grazed Ungrazed
118 Figure 4 9 Height frequency distribution (%) of Chionachne at individual study sites (n = 75). COOL = Coolibah, DASH = Dashwood, FITZ = Fitzroy, GNP1 and GNP2 = Gregory National Park sites. Grazed Ungrazed
119 Figure 4 10. Frequency (%) of Bareground (BG) at individual study sites (n = 75). COOL = Coolibah, DASH = Dashwood, FITZ = Fitzroy, GNP1 and GNP2 = Gregory National Park sites. Grazed Ungrazed
120 Figure 4 11. Frequency (%) of Bareground w/ debris (BGD) at individual study sites (n = 75). COOL = Coolibah, DASH = Dashwood, FITZ = Fitzroy, GNP1 and GNP2 = Gregory National Park sites. Grazed Ungrazed
121 Figure 4 12 a Frequency (%) of Xanthium at individual study sites (n = 75). COOL = Coolibah, DASH = Dashwood, FITZ = Fitzroy, GNP1 and GNP2 = Gregory National Park sites. Grazed Ungrazed
122 Figure 4 12 b. River bank at Dashwood Crossing showing dense cover of Xanthium August 2002 ( Xanthium has already dried, though mid wa y through the Dry Season ).
123 Figure 4 13. Frequency (%) of total weeds ( Xanthium Ricinus and Passiflora ) at individual study sites (n = 75). COOL = Coolibah, DASH = Dashwood, FITZ = Fitzroy, GNP1 and GNP2 = Gregory Nati onal Park sites. Grazed Ungrazed
124 Figure 4 14 Canopy cove r, expressed as a percentage at individual study sites (n = 75). COOL = Coolibah, DASH = Dashwood, FITZ = Fitzroy, GNP1 and GNP2 = Gregory National Park sites. Grazed Ungrazed
125 Figure 4 15. Canopy cover (species) at combined study sites expressed as a percentage E.MIC = Eucalyptus microtheca E.CAM = Eucalyptus camaldulensis BAR = Barringtonia acutangula FIC = Ficus coronulata CAT = Cathormium umbellatum (n = 75).
126 Figure 4 16 Dominant canopy cover composition (%) at individual study sites. E.MIC = Eucalyptus microtheca E.CAM = Eucalyptus camaldulensis BAR = Ba r ringtonia acutangula FIC = Ficus coronulata (n = 75). COOL = Coolibah, DASH = Dashwood, FITZ = Fitzroy, GNP1 and GNP2 = Gregory National Park sites. Grazed Ungrazed
127 Figur e 4 17 Ordination of transects, based on their overall similarity in habitat variables measured. Symbols mark the five separate study sites (C = Cooliba h, D = Dashwood, F = Fitzroy, G1 = GNP1, G2 = GNP2).
128 Figure 4 18 a. Fire at GNP1, 22 October, 2001. Picture taken approximately two hours post fire.
129 Figure 4 18 b. Three months p ost fire at GNP1, January 31, 2002. Note that the broad leaved vege tation is weeds ( Xanthium and Ricinus ) in addition to the dominant Chionachne seen clearly in the forefront
130 Figure 4 19 Average group sizes (adult birds only) at study sites 2001 2003 (n = 88 group years). COOL = Cooli bah, DASH = Dashwood, FITZ = Fitzroy, GNP1 and GNP2 = Gregory National Park sites. Grazed Ungrazed
131 Figure 4 20 Adult survivorship at individual study sites from 2001 2002 and 2002 2003 (n = 40 group years). COOL = Coolibah, DASH = Dash wood, GNP1 = Gregory National Park Site 1, GNP2 = Gregory National Park Site 2. Note that survivorship at Dashwood was 0 from 2001 through 2002. Fitzroy was not included as visits were insufficient to accurately determine survivorship. Grazed Ungrazed
132 Figure 4 21 Foraging preference illustrated by time spent foraging in relation to substrate abundance (based on transect data 20 cm intercepts and minutes foraging). CHIO = Chionachne XAN = Xanthium BGD = Bareground w/ debris BARE = bare ground, BAR = Barringtonia
133 Figure 4 22 Productivity (number of fledglings per breeding female) at individual study sites, 2001 2003 (n = 80 group years). COOL = Coolibah, DASH = Dashwood, FITZ = Fitzroy, GNP1 and GNP2 = Gregory National Park sites. Note that this graph represents productivity per bird but does not reflect the total number of fledglings produced or the total number of breeding females present at each site. Grazed Ungrazed
115 CHAPTER 5 LANDSCAPE SCALE ANAL YSES OF HABITAT AND THE IMPLICATIONS FOR CONSERVATION OF THE PURP LE CROWNED FAIRY WREN This final chapter presents information on the distribution of the primary habitat of the PCFW in the larger landscape, and discusses implications for future conservation management. This chapter concludes with management recommendations for the PCFW population in the VRD in addition to future research directions. Introduction Indicator and/or focal species have been used in conservation biology in various contexts although their ut ility has been the subject of substantial debate (Ricketts et al ., 1999; Lindenmayer et al. 2002; Mac Nally and Fleishman, 2004; Bani et al. 2006). However, Lawler et al (2003) found that at risk species functioned we ll as an indicator group in by thems elves. Their results concurred with Groombridge (1992), suggesting that managing and conserving at risk species could help preserve biodiversity as a whole. The focal species approach uses species most sensitive to a particular threat, for example, habita t specialists, and can be an important conservation tool (Lambeck 1997). The utility of this approach was demonstrated by a study of woodland birds in southeastern Australia where conservation of the habitat of two focal species was shown to also provide for the needs of at least 95% of other woodland birds (Watson et al. 2001). Because the PCFW is a riparian obligate, highly dependent on Chionachne in the VRD, it is reasonable to examine the wider distribution of this grass as an important indicator of the conservation potential of the PCFW itself as an indicator species for riparian zones. Lambeck (1997) included three categories of focal species: area or habitat limited species, movement limited species, and
116 management limited species. Based on its e cology, distribution, and response to grazing, the PCFW fits all of these categories making it useful indicator species. However, as the PCFW along the Victoria River is confined to a single understory grass species it may be more useful to include a host of riparian obligate species as a focal group in order to ascertain riparian health. structure (Wiens 1989). Habitat fragmentation results in isolated patches of habitat in a variegated landscape (Andr n, 1994; Hobbs, 2005). The effects of fragmentation are often studied in relation to patch size and patch isolation, with models often based on a binary s implification of the environment in (unsuitable lands) (e.g. Hill and Caswell, 1999; Debinsky and Holt, 2000). However, fragmentation studies that have more biological realism (i.e. species specific attributes and habitat q uality) will most likely lead to much clearer conclusions (e.g. Debinsky and Holt, 2000; Cooper and Walters, 2002; Wiegand et al. 2005). For example, in the woodlands of south eastern Australia, the response of species to fragmentation varied in differen t landscapes, suggesting that the matrix plays can also play an important role (Watson et al. 2005). The effect of habitat fragmentation has been studied for several Malurus species (e.g. Malurus pulcherrimis and Malurus splendens ) in Australia (Brooker and Brooker, 1994; Brooker and Brooker, 2001). For these malurids, Brooker and Brooker (2001) found that smaller patches had higher adult mortality rates than larger patches, but that reproductive success was higher in smaller patches (perhaps due to diff erences in brood parasitism rates). The authors concluded that both small and large patches were important for the conservation of Malurus pulcherrimus In a
117 model developed to predict the response of Malurus splendens to fire and fragmentation, Brooker and Brooker (1994) found that the likelihood of populations going extinct in patches was highly dependent on fire. That is, if there is a high risk of wildfire, very large patches were needed to sustain the population but if fire were excluded, patches si ze less than a quarter the size of a large patch could be viable. The strict adherence of the PCFW to Chionachne habitat along the Victoria River coupled with the potential impact of grazing on its primary habitat makes the PCFW especially prone to fragme ntation effects. Long distance dispersal capabilities of the PCFW are most likely limited however, the wrens possibly have adaptations to natural disturbances such as flooding which, necessitate frequent local movements. The ability to disperse short dis tances from local disturbances may benefit the PCFW when it is necessary to respond to other novel habitat disturbances (e.g. grazing). The variable nature of the PCFW habitat suggests that the structure of non Chionachne habitat may be crucial in any dis persal movements and that the conservation of non Chionachne habitat may also play an important role in PC FW conservation (e.g. understor y cover), as dispersal must occur over these non Chionachne areas. Canegrass ( Chionachne cyathopoda ) D istribution Meth ods In order to determine the abundance of Chionachne in the broader landscape, and the degree of fragmentation of PCFW habitat, an estimate of Chionachne abundance and distribution along the Victoria River was required. Existing aerial photography was exa mined but as it was predominantly black and white, and it was impossible to distinguish Chionachne from adjacent vegetation. Additionally, most of the photography was also outdated. Satellite imagery was also examined but the resolution was too coarse a scale to accurately identify Chionachne Higher
118 resolution scenes, which may have diagnostically shown Chionachne patches, were unaffordable. An opportunity to examine Chionachne via helicopter was made available, and this proved to be a very efficient a nd effective method of mapping Chionachne I conducted mapping in August 2003, a time of year when Chionachne is clearly visible and distinct as it stays greener for a longer period than other ground vegetation components. Due to funding constraints, onl y a limited area could be surveyed. I decided to focus on the main river course and cover as much area as possible that potentially could contain PCFW habitat. Earlier surveys by boat indicated that Chionachne does not extend much further downstream from the town of Timber Creek (15.631181 S, 130.47612 E), where the Victoria river is subject to tidal influence (and hence the nature of the riparian vegetation changes), and so this was chosen as the starting point. The end point of the survey, approximatel y 170 km upstream and the limit of the distance that could be covered, was Dashwood Crossing (16.33771 S, 131.10987 E), above which the Victoria River becomes a much smaller watercourse. Both sides of the river were surveyed simultaneously using two obser vers (A. van Doorn and H. Henri), one on each side of the helicopter. I conducted a p reliminary survey to determine feasibility and parameters (width and density of Chionachne ) to be used for mapping Chionachne distribution. Sections along the river wher e Chionachne occurred were measured for length along the river using hand held GPS (Garmin GPS II Plus, <15m). The sections were not entirely continuous, of course, so patches of Chionachne were also mapped. For each patch, the start ing and end points we re marked and the following two variables were recorded; density (sparse, medium, dense) and width (narrow, medium, wide). Although it was impossible to estimate exact width using this method, I was able to
119 assign approximate width categories based on comp aring known (measured) widths at field sites with estimates from the preliminary survey, as follows: narrow = <50m, Medium = 50 100m, and wide = 100 200m. Observers were calibrated during an initial trial. It was not possible to adequately survey trib utaries that may have held additional Chionachne although where visible Chionachne extended into the tributary were classified as being wide (i.e. 100 200m). The area of each Chionachne patch was estimated by multiplying patch length with estimated patch width. For Chionachne area calculations, I used the upper limit for each width parameter, that is, narrow = 50 m, medium = 100m and wide = 200, and hence estimated total area represents the upper limit of the estimate of Chionachne area. Distributional m aps of Chionachne were c reated using ESRI ARCGIS 9.1. I e stimated PCFW total population size for the Victoria River from Timber Creek through to Dashwood Cr ossing by dividing the total Chionachne area by the average area occupied by individual birds at st udy site (based on density data, Chapter 3). Assumptions underlying these estimates included: 1. Chionachne is the only primary habitat of the PCFW in this system (an assumption corroborated by substantial searches of this system in non Chionachne habitat) 2. Small and fragmented patches were inhabited by PCFWs, as well as the larger patches (as were studied in detail on the ground), 3. PCFW density does not vary with Chionachne density (sparse, medium and dense). Results Along the approximately 170 km s tretch of river surveyed, the total area of Chionachne based on the widths assigned to each patch size was 1,222 ha. However,
120 the Chionachne distribution was severely fragmented into a total of 52 patches, with varying length, widths a nd densities (Figs. 5 1, 5 2). The average patch size was 24.5 ha ranging from 0.23 ha to 244 ha (Table 5 1). The average distance from a patch to its nearest neighbor was 1.37 km, ranging from 0.33 to 41 km. Most patches, regardless of size, were near others, but there wer e also a substantial number of smaller isolated patches (Fig. 5 3). Chionachne distribution was not equally distributed along the river and appeared to vary, especially between areas under different tenure (i.e. according to land use) (Figs. 5 1, 5 2, Tab le 5 1). Of the total amount of Chionachne found along the river, the largest percentage of Chionachne was found within Gregory National Park (40%) and within this area Chionachne was also the least fragmented (Figs. 5 1, 5 2). Further, in Gregory Nation al Park, Chionachne had the largest average patch size by far, more than twice as large as that of the two pa storal properties (Table 5 1). The largest number of small isolated patches was found on Bradshaw (this area had been pastoral property but was ta ken over by the defense department in 1996), nevertheless, this area still has a substantial amount (13%) of the total Chionachne area (Table 5 1). Fitzroy Station, also used for pastoralism, had a large area of Chionachne (32% of the total found) but was highly fragmented in areas (primarily downstream). Coolibah Station had several larger patches and also contained a considerable abundance of Chionachne (14% of the total area), particularly in the area adjoining the national park. The smallest amount o f total Chionachne surveyed was found at Coolibah Croco dile Farm (1%) mainly because it consists of a very small area (Table 5 1) One of the most interesting finding s was the lack of suitable habitat for approximately 40 kilometers, from the last patch o f Chionachne in Gregory National
121 Park, south to Dashwood Crossing on Victoria River Downs. This a rea is accessible only by air and hence prior to this survey, the extent of Chionachne in this stretch was unknown. A distance of 40 km is a significan t barr ier for a small understor y bird with weak flight and strong habitat preferences, thus effectively isolating the se two areas from one another (e.g. Brooker and Brooker, 1997). Only one small patch of Chionachne was found on Victoria River Downs and this wa s located exactly at the end point of my survey (Dashwood Crossing). Additional areas of Chionachne do exist past the survey limit at Dashwood Crossing point but are apparently very patchy in distribution and do not occur in any great abu ndance (D. Hill, pers, comm.). Average territory size of the PCFW was 0.41 ha (based on four sites, Coolibah, Fitzroy, GNP1 and GNP2; Chapter 3) and average group size was 2 .6 (based 88 group years; Chapter 3). Based on 1,222 ha total area of Chionachne 2,955 family gro ups ( or 7,713 individuals ) could be present along this 170 km stretch of the Victoria River. Because the total area of Chionachne was used for this estimate of unlikely that the severely fragmented or very small patches would contain as high of densities of PCFWs as measured in Chapter 3, and the total populat ion is actually much smaller. It is unlikely that a small patch isolated from other patches can sustain PCFWs in the long term. In addition, as mentioned above, I used the upper limit of the width categories in calculating the total area of Chionachne further ensuring a maximum size for the estimate of total PCFW population size. Also, the four study sites used here a s a basis for average PCFW density estimates were all in moderate to high quality habitat and moderate to large patch size, so mean density across the broader habitat is probably appreciably less than this calculated figure.
122 To give an indication of how o ur estimate may be optimistic, if we consider that within a single season after cattle were introduced at Coolibah, the average group size dropped to 2.2 (Chapter 3), and if we applied this lower group size to the area of Chionachne used for pastoralism, t he optimistic estimate of the total population size (above) would decrease by 712 birds (9%). Further, if we apply the fact that the number of birds at Coolibah decreased by 64% in one season when cattle were introduced to the site (Chapter 3), and apply this percentage to the total area of Chionachne used for pastoralism, then the total population estimate for the PCFW 5,430 bir ds. Despite the uncertainty of the population size estimate based on the Chionachne survey, the distribution is very useful for comparative measures to identify key areas for conservation and differences between land uses. The distribution map of Chionachne from my study Conservation and Management Recommendations Purple crowned Fairy wrens depend on Chionachne for foraging, breeding and cover and hence any conservation action should focus on Chionachne preservation. Below I have outlined management recommendations based on findings from my study with each potentially threatening process being addressed individually. Grazing is dealt with explicitly as my study focused on this particular element. Fragmentation of habitat is also dealt with as a threatening process, understanding that fragmentation is most likely the product of several factors. Based on my own experiences with this project, I have outlined some issues and strategies to address stakeholder involvement, a key issue in a multiple land use area such as the VRD.
123 Finally, I have also inc luded a section on future research directions as a guidance that may serve as a guide to other researchers or government agencies. Grazing Based on results from my study grazing is one of the most important threats to the PCFW in the VRD. Vegetation at grazed sites offered far less cover for PCFWs than ungrazed sites, primarily due to the difference in Chionachne height (Chapter 4). Of greatest significance was the alarming increase in adult mortality resulting in fewer breeding pairs, within one seaso n at the newly intensely grazed site, Coolibah. Such a large increase in adult mortality over such a short period of time demonstrates the acute impact that grazing can have if it is conducted at high intensities. Unfortunately, survival rates at Fitzroy, which was grazed lightly, are not well known but Chionachne height was not comprom ised as much as at Coolibah, and it is reasonable to expect that survival rates might be more favorable at lower grazing pressures. The low incidence of Chionachne at the D ashwood site, in areas where Chionachne might be expected to occur (Chapter 4), suggests that long term grazing may eventually lead to a loss of Chionachne stands themselves. Fencing the entire riparian corridor along the Victoria River and tributaries wo uld most likely provide the best conservation measure for the PCFW, however, it is extremely unlikely this will be accomplished, for a number of reasons. The remote location of the region and frequent flooding make fencing and maintaining fences over larg e areas very difficult, costly, and time consuming. For most stations, the cost of fencing make it economically unattractive, and perhaps in some cases, prohibitive. Nevertheless, fencing of the riparian corridor should be encouraged and with government subsidies these costs are much reduced. In all likelihood a combination of land management techniques will need to be incorporated to reach a compromise
124 and undoubtedly these will need to be integrated into a larger catchment conservation management plan Grazing by feral animals (i.e. buffalo, pigs, scrub bulls) should also not be overlooked as these animals can also severely affect riparian vegetation. Control of feral animals will not only benefit riparian conservation but will also benefit pastoralism by reducing competition with cattle Weeds Weeds are widespread throughout the VRD. Weeds pose a threat not only to biodiversity (e.g. by competing with desirable native species), but also impact pastoralism by potentially sickening stock, reducing desi rable pasture species and hampering access to water. In relation to the PCFW, weeds can be for cover and foraging, and are preferable to bareground. However, weeds are not a preferred substrate, at least in the case of Xanthium where Chionachne is availa ble, nor do birds nest in weeds (Chapter 4). The control of most weeds in the district is not plausible and attention should focus on the eradication of new invasive s pecies before they can spread. Fire Despite the limited data available for the effects of fire on Chionachne and PCFW, we can infer from the width and connectivity of Chionachne in some areas (e.g. Gregory National Park), that a late season (hot) burn could cover a very large area and have severe consequences for the PCFW (Chapter 4). Alth ough Chionachne regenerates rapidly after fire, a wide scale fire could result in high mortality rates of the PCFW. Firebreaks would provide some protective measures in areas of high connectivity. However, management actions will need to consider the pot ential of
125 larger fire breaks to further fragment habitat and that fire breaks may become infested with weeds. Fragmentation The survey conducted during the course of my study has demonstrated the fragmented nature of Chionachne distribution. Because the largest a eri al extent of Chionachne is within Gregory National Park from a conservation standpoint this area is critical for the long term survival of the PCFW. Both Coolibah Station and Fitzroy Station directly abut the national park, and have considera ble amount of Chionachne including larger patches that are separated by only small distances. Small patches of Chionachne located between larger patches would ideally be conse rved to increase connectivity. Observations made during my study indicated that there is a dispersal threshold beyond which re colonization is unlikely. Two large patches of dense but isolated, Chionachne were found in the Victoria River catchment that had no PCFWs present. Conversations with land managers revealed that these patches of Chionachne had only recently regenerated after having been grazed intensely for a long time. Potentially, these areas of suitable habitat could be used as re introduction sites once the habitat suitability is determined. Future research to determine dispersal threshold, in combination with an investigation into dispersal habitat suitability, will allow for more accurate management actions to determine patches of high conservation value. I suggest that the dispersal threshold will not be very high, ju dging from PCFW behavior and morphology. Stakeholder I nvolvement In remote location such as the VRD, garnering community support may be a pivotal element for any long term success. Purple crowned f airy wrens are
126 distributed in multiple land use areas, mak ing stakeholder involvement increasingly important as connectivity undoubtedly plays a major role in long term PCFW conservation. In the case of the PCFW, there are multiple avenues for stakeholder involvement to bolster support for any required conservat ion action. Government subsidies for conservation management priorities (fencing, weed and erosion control) are extremely important for the future management of riparian zones in the VRD and full use should be made of these funds. The Victoria River Cons ervation Association (VRDCA) can make an important contribution to any future conservation efforts by continued efforts in funding applications and liaison with stakeholders. Notably, the PCFW is a good candidate for such communication and conservation ac tion, as it is an aesthetic and high profile species, well known to landholders and other stakeholders. Ecological knowledge alone will not be sufficient to develop or implement an adequate management plan. Without participation of, and cooperation among the stakeholders, the habitat will remain fragmented and the region will continue to have small isolated populations of PCFW. In addition, because barriers are often lacking between adjacent properties there are always flow on effects from one property to the next. For example, if one stakeholder is actively controlling weeds, but the property upstream is not, then the effectiveness of weed control will always be compromised. Although this point may seem logical and straightforward, it emphasizes the n eed for a long term commitment and continuous stakeholder involvement. Conservation Management Recommendations Below I have outlined a series of management recommendations that should be included in the development for a conservation management plan for t he VRD. A. Population Estimation and Monitoring Activities
127 Identify key areas for conservation (i.e. areas that will increase connectivity by linking patches) and instigate conservations methods at these sites (e.g. fencing, erosion control). Sampling Chi onachne patches of varying size and isolation, to determine PCFW absence or presence in addition to determining PCFW densities in accordance with habitat quality, allowing for a more accurate population estimate. A widespread survey of the Victoria River catchment to determine PCFW distribution and habitat associations. Establishment and maintenance of population monitoring, where possible linked to assessment of efficacy of management inputs. B. Land Use and Habitat Assessment Activities Preservation of healthy Chionachne stands (fencing, fire breaks) Control of feral animals across all sites Amelioration of grazing pressure in riparian areas using a variety of methods including reduced stocking rates, increased artificial water sources, fencing core ar eas of Chionachne and restricted access. Erosion control measures to avoid losing Chionachne root bases. Long term monitoring of vegetation monitoring to assess habitat condition. C. Stakeholder, Public, and Government Involvement Enhanced development of stakeholder involvement (regular contact, dissemination of information, involvement in decision making). An assessment of fencing costs in cooperation with stakeholders. Enhanced communication and publicity to increase public awareness and interest t o engender community support. At a broader scale, liaison with government and research agencies in Western Australia for collaborative research and management strategies. Future research directions Investigation into nest predation rates and causes. Resear ch on dispersal dynamics (e.g. maximal dispersal distances, dispersal mortality, dispersal habitat use). Further investigation into PCFW ecology (e.g. paternity, incest avoidance).
128 Metapopulation theory applications to examine population dynamics of the PCFW. Population viability analyses (PVA) to investigate disturbance effects and develop predictive models. Development of behavior based (based on fitness maximization) population models which may produce more accurate predictions. Investigation into the role of Chionachne as an erosion control agent. Further investigations into the role of weeds, fire, grazing and other disturbances (e.g. interactions, confounding effects). Further research into pastoral management in riparian zones (e.g. stocking rates, resting periods, exclusion methods) that is more compatible with preserving Chionachne Conclusion Although many aspects of the ecology of the PCFW and identification of threatening processes were ex amined during the course of my study, there are still ma ny areas that need further investigation. As is the case with most studies, answering one question leads to many more. However, the re sults found during my study have provided a substantial basis for conservation management initiatives. Common sense cons ervation initiatives, such as riparian corridor protection, should be initiated as soon as possible, using coordinated efforts combining solid science, community and stakeholder education and participation, and if necessary using any legislation available by government agencies to protect this unique specie s and its distinctive habitat. It is possible to continue sitting on the fence, touting the need for more data, however, that stance will do nothing to conserve a species with a fragmented distribution w hose habitat needs immediate protection. We need to implement conservation strategies to conserve existing riparian habitat, and in so doing, will preserve not only the PCFW, but potentially also other dependent riparian species.
1 29 The other alternative, re habilitation at some later date, is an expensive proposition in an area as vast as the Victoria River District.
130 Figure 5 1. Map showing width of c anegrass ( Chionachne cyathopoda ) along the Victoria River. Note that this map does not include one stud y site, Dashwood Crossing south of Gregory National Park, as the next patch of Chionachne was located a t the end of the survey
131 Figure 5 2. Map showing the density of c anegrass ( Chionachne cyathopoda ) along the Victoria River. Note that this map does not include one study site, Dashwood Crossing south of Gregory National Park, as the next patch of Chionachne was located at the end of the survey
132 Figure 5 3. Canegrass ( Chionachne cyathopoda ) patch size and distance to next patch (n = 52).
133 Table 5 1. Canegrass ( Chionachne cyathopoda ) distribution, patch size and patch isolation by tenure (from helicopter based survey from Timber Creek to Dashwood Crossing, not i ncluding tributaries, based on c anegrass width approximations). Note: Bradshaw was a pa storal property prior to 1996. Site name Tenure Land use Total Canegrass area (ha) (%) No. of patches Ave. patch size (ha) Ave. distance to next patch (km) PCFW conservation value Potential implementation of conservation Gregory National Park Conservatio n Conservation 486.09 (40%) 7 69.44 1.74 High High Fitzroy Station Aboriginal Pastoral (varied across area) 389.36 (32%) 16 24.33 0.99 High Low/medium Coolibah Station Pastoral lease Pastoral 175.65 (14%) 6 29.27 1.93 High Low Bradshaw Defense Defense t raining 153.56 (13%) 21 7.31 1.22 Medium High Coolibah Crocodile Farm Private Private (crocodile farming) 17.32 (1%) 2 17.32 1.93 Medium High
134 APPENDIX A BIRD SPECIES Scientific name Common name Coturnix ypsilophora (Bosc) Brown q uail* Dendrocygna e ytoni (Eyton) Plumed w histling duck Tadorna r adjah (Lesson) Radjah s helduck Anas superciliosa (Gmelin) Pacific b lack duck Anhinga melanogaster (Pennant) Darter Phalacrocorax melanoleucos (Vieillot) Little pied c ormorant Pelecanus conspicillatus (Temmi nck) Australian p elican Egretta novaehollandiae (Latham) White faced h eron Ardea p acifica (Latham) White necked h eron Egretta garzetta (Linnaeus) Little e gret Ardea Alba (Linnaeus) Great e gret Ardea intermedia (Wagler) Intermediate e gret Nycticorax c aledonicus (Gmelin) Nankeen night h eron Platalea regia (Gould) Royal s poonbill Platalea flavipes (Gould) Yellow billed s poonbill Ephippiorhynchus asiaticus (Latham) Black necked s tork Aviceda subcristata (Gould) Pacific b aza Elanus axillaris (Latham) Black shouldered k ite Milvus migrans (Boddaert) Black k ite Haliastur sphenurus (Vieillot) Whistling k ite Haliastur indus (Boddaert) Brahminy k ite Haliaeetus leucogaster (Gmelin) White bellied sea e agle Accipiter fasciatus (Vigors & Horsfield) Brown g oshawk Accipiter novaehollandiae (Gmelin) Collared s parrowhawk Falco berigora (Vigors & Horsfield) Brown f alcon Falco peregrinus (Tunstall) Peregrine f alcon Falco cenchroides (Vigors & Horsfield) Nankeen k estrel Grus rubicunda (Perry) Brolga Actitis hypoleucus (Linnaeus) Common s andpiper Burhinus grallarius (Latham) Bush s tone curlew Elseyornis melanops (Vieillot) Black fronted d otterel Vanellus miles (Boddaert) Masked l apwing Geopelia striata (Linnaeus) Peaceful d ove* Geopelia humeralis (Temminc k) Bar shouldered d ove* Calyptorhynchus banksii (Latham) Red tailed Black c ockatoo Cacatua roseicapilla (Vieillot) Galah Cacatua sanguinea (Gould) Little c orella* Cacatua galerita (Latham) Sulphur crested c ockatoo Trichoglossus haematodus (Linnaeus) R ainbow l orikeet Psitteuteles versicolor (Lear) Varied l orikeet Aprosmictus erythropterus (Gmelin) Red winged p arrot Chrysococcyx osculans (Gould) Black eared c uckoo Chrysococcyx basalis (Horsfield) c uckoo Eudynamys scolopacea (Linna eus) Common k oel Scythrops novaehollandiae (Latham) Channel billed c uckoo
135 Centropus phasianinus (Latham) Pheasant c oucal* Ninox connivens (Latham) Barking o wl Ninox novaeseelandiae (Gmelin) Southern b oobook Tyto alba (Scopoli) Barn o wl Podargus strig oides (Latham) Tawny f rogmouth Eurostopodus argus (Hartert) Spotted n ightjar Aegotheles cristatus (Shaw) Australian o wlet nightjar Alcedo azurea (Latham) Azure k ingfisher Dacelo leachii (Vigors & Horsfield) Blue winged k ookaburra Todiramphus sanctus ( Vigors & Horsfield) Sacred k ingfisher Merops ornatus (Latham) Rainbow b ee eater Eurystomus orientalis (Linnaeus) Dollarbird Climacteris melanura (Gould) Black tailed t reecreeper Malurus coronatus (Gould) Purple crowned f airy wren* Malurus melanocephal us (Latham) Red backed f airy wren* Pardalotus striatus (Gmelin) Striated p ardalote* Smicrornis brevirostris (Gould) Weebill Philemon buceroides (Swainson) Helmeted f riarbird Philemon argenticeps (Gould) Silver crowned f riarbird Philemon citreogularis (Gould) Little f riarbird Lichenostomus unicolor (Gould) White gaped h oneyeater* Lichenostomus flavescens (Gould) Yellow tinted h oneyeater* Melithreptus albogularis (Gould) White throated h oneyeater Lichmera indistincta (Vigors & Horsfield) Brown h oneye ater* Ramsayornis fasciatus (Gould) Bar breasted h oneyeater Conopophila rufogularis (Gould) Rufous throated h oneyeater Certhionyx pectoralis (Gould) Banded h oneyeater Microeca fascinans (Latham) Jacky w inter Pomastostomus temporalis (Vigors & Horsfiel d) Grey crowned b abbler Daphoenositta chrysoptera (Latham) Varied s itella Colluricincla harmonica (Latham) Grey s hrike thrush Myiagra rubecula (Latham) Leaden f lycatcher Myiagra alecto (Temminck) Shinning f lycatcher Myiagra inquieta (Latham) Restless f lycatcher* Grallina cyanoleuca (Latham) Magpie lark Rhipidura fuliginosa (Sparrman) Grey f antail Rhipidura rufiventris (Vieillot) Northern f antail* Phipidura leucophrys (Latham) Willie w agtail* Coracina novaehollandiae (Gmelin) Black faced c uckoo shr ike Coracina papuensis (Gmelin) White bellied c uckoo shrike Lalage sueurii (Vieillot) White winged t riller Oriolus flavocinctus (Vigors) Olive backed o riole Artamus leucorynchus (Linnaeus) White breasted w oodswallow Artamus cinereus (Vieillot) Black faced w oodswallow Artamus minor (Vieillot) Little w oodswallow Cracticus nigrogularis (Gould) Pied b utcherbird Corvus orru (Bonaparte) Torresian c row* Chlamydera nuchalis (Jardine & Selby) Great b owerbird Taeniopygia bichenovii (Vigors & Horsfield) Dou ble barred f inch* Poephila acuticauda (Gould) Long tailed f inch* Poephila personata (Gould) Masked f inch*
136 Neochmia phaeton ((Hombron & Jacquinot) Crimson f inch* Neochmia ruficauda (Gould) Star f inch* Lonchura flaviprymna (Gould) Yellow rumped m annikin Lonchura castaneothorax (Gould) Chestnut breasted m annikin* Heteromunia pectoralis (Gould) Pictorella m annikin* Dicaeum hirundinaceum (Shaw) Mistletoebird Cisticola exilis (Vigors & Horsfield) Golden headed c isticola* Bird species observed at ripar ian study sites. Species marked with were found within Chionachne cyathopoda stands, all other species were located in the riparian habitat within 50m of Chionachne cyathopoda or in the canopy above it. Classification follows Christidis and Boles (1994)
137 APPENDIX B PLANT SPECIES Botanical name Common name Acacia holosericea (A. Cunn.) Candelabra w attle Barringtonia acutangula (L.) Gaertn. Freshwater m angrove Bauhinia cunninghamii (Benth.) Benth. Bauhinia Cathormiom umbellatum (Vahl) Kostermans Cat hormiom Chionachne cyathopoda (F. Muell.) River grass (c anegrass) Crotalaria novae hollandiae (DC) New H olland r attlepod Eucalyptus camaldulensis (Dehn.) River red g um Eucalyptus microtheca (F. Muell.) Flooded box (c oolibah) Excoecaria parviflora (Mue ll. Arg.) Gutta percha t ree Ficus coronulata (F. Muell.) River f ig Flueggia virosa (Roxb.ex Willd) Voigt Dogwood Hyptis suaveolens (L.) Hyptis* Melalueca leucadendra (L.) Cadjeput Mnesithea rottboelioides (R. Br.) Northern caneg rass Nauclea oriental is (L.) Leichardt p ine Pandanus aquaticus (F. Muell.) River p andanus Parkinsonia aculeata (L.) Parkinsonia* Passiflora foetida (L.) Wild p assionfruit* Ricinus communis (L.) Castor oil p lant* Xanthium strumarium (L.) Noogoora b urr* List of plant speci es found within Purple crowned fairy wren habitat. Note that the list is based on vegetation transects within Purple crowned fairy wren territories (described in Chapter 3). Vegetation transects were conducted during the dry season resulting in absence s ome species (e.g. annual wet season grasses). Introduced species are indicated with an *. Classification of botanical names follows Petheram and Kok (2003).
138 REFERENCES Ammon, E. M. and P. B. Stacey. 1997. Avian nest success in relation to past grazing r egimes in a montane riparian system. The Condor 99:7 13. Anderson, A. N. 1991. Responses of ground foraging ant communities to three experimental fire regimes in a savanna forest of tropical Australia. Biotropica 23:575 585. Andrn, H. 1994. Effects of hab itat fragmentation on birds and mammals with different proportions of suitable habitat: a review. Oikos 71:355 366. Archer, S., C. Seifres, C. Bassham and R. Maggio. 1988. Autogenic succes sion in a subtropical savanna: R ates, dynamics and processes in the conversion of grassland to thorn woodland. Ecological Monographs 58:111 127. Bani, L, D. Massimino, L. Bottoni, and R. Massa. 2006. A multiscale method for selecting indicator species and priority conservation area: A case study for broadleaved forests in Lombardy, Italy. Conservation Biology 20(2):512 526. Barnard, C. A. 1925. A review of the bird life on Coomooboolaroo Station, Duaringa District, Queensland, during the past 50 years. Emu 24:252 265. Baxter, G. S., E. J. Moll and A. T. Lisle. 2001. Pa sture grazing by black striped wallabies ( Macropus dorsalis ) in central Queensland. Wildlife Research 28:269 276. Bentley, J. M. and C. P. Catterall. 1997. The use of bushland, corridors, and linear remnants by birds in southeastern Queensland, Australia. Conservation Biology 11(5):1173 1189. Bennett, M. B. 1999. Foot areas, ground reaction forces and pressures beneath the feet of kangaroos, wallabies and rat kangaroos (Marsupialia: Macropodidae). Journal of Zoology, London 247:365 369. Blakers, M., S. J. J F. Davies and P. N. Reilly. 1984. The atlas of Australian birds. Melbourne University Press, Carlton. Boekel, C. 1979. Notes on the status and behaviour of the p urple crowned fairy wren Malurus coronatus in the Victoria River Downs area, Northern Terr itory. Australian Bird Watcher 8(3):91 97. Bowman, D. M. J. S. 1998. Tansley Review No. 101. The impact of a boriginal landscape burning on the Australian biota. New Phytologist 140:385 410. Braithwaite, R. W. 1987. Effects of fire regimes on lizards in th e wet dry tropics of Australia. Journal of Tropical Ecology 3:265 275. Braithwaite, R. W., W. M. Lonsdale and J. A. Estbergs. 1989. Alien vegetation and native biota in tropical Australia: The spread of Mimosa pigra Biological Conservation 48:189 210.
139 Bro mham, L, M. Cardillo, A. F. Bennett and M. A. Elgar. 1999. Effects of stock grazing on the ground invertebrate fauna of woodland remnants. Australian Journal of Ecology 24:199 207. Brooker, M. G. and L. C. Brooker. 1996. Acceptance by the splendid fairy wr en of parasitism by cuckoo: further evidence of evolutionary equilibrium in brood parasitism. Behavioral Ecology 7:395 407. Brooker, M. G. and L. C. Brooker. 2001. Breeding biology, reproductive success and survival of blue breasted fair y wrens in fragmented habitat in the Western Australian wheatbelt. Wildlife Research 28:205 214. Brooker, L. C. and M. G. Brooker. 1994. A model for the effects of fire and fragmentation on t he population viability of the splendid f airy wren. Pacific Conse rvation Biology 1:344 358. Brooker, L. C. and M. G. Brooker. 1997. Dispersal and population dynamic of the blue breasted fairy wren, Malurus pulcherrimus in fragmented habitat in the Western Australian wheatbelt. Wildlife Research 29:225 233. Burbidge, A. A. and N. L. McKenzie. 1989. Patterns in the modern decline of Western Conservation 50:143 198. Christidis. L. and W. E. Boles. 1994. The taxonomy and species of birds of Australia and its territories. Royal Australasian Ornithologists Union Monograph 2. RAOU, Melbourne. Clarke, K. R. and R. N. Gorley. 2001. PRIMER v5: User Manual/Tutorial PRIMER E, Plymouth. Clarke, K. R. and R. H. Green. 1988. Statistical design an study. Marine Ecology Progress Series 46 : 213 26. Cockburn, A., H. L. Osmond, R. A. Mulder, D. J. Green, and M. C. Double. 2003. Divorce, dispersal, and incest avoidance in the cooperatively breeding superb f airy wren M alurus cyaneus Journal of Animal Ecology 72(2):189 202. Cooper, C. B. and J. R. Walters. 2002. Experimental evidence of disrupted dispersal causing decline of an Australian passerine in fragmented habitat. Conservation Biology 16(2):471 478. Cowie, I. D., P. S. Short and M. Osterkamp M adsen. 2000. Floodplain flora: A flora of the coastal floodplains of the Northern Territory, Australia. Australian Biological Resources Study, Canberra/PWCNT, Darwin. Croonquist, M. J. and R. P. Brooks. 1993. Effects of habit at disturbance on bird communities in riparian corridors. Journal of Soil and Water Conservation 48(1):65 70.
140 Crowley, G. and S. Garnett. 1999. Seeds of the annual grasses Schizachyrium sp p. a s a food for tropical granivorous birds. Australian Journal of Ecology 24:208 220. Davis, R. 2000. Aboriginal lands and communities. Pages 15 16. In M. Kraatz (Ed.), Managing for healthy country in the VRD. Tropical Savannas CRC, Darwin. Debinsky, D. M. and R. D. Holt. 2000. A survey and overview of habitat fragment ation experiments. Conservation Biology 14(2):342 355. Dobkin, D. S., A. C. Rich, and W. H. Pyle. 1998. Habitat and avifaunal recovery from livestock grazing in a riparian meadow system of the northwestern Great Basin. Conservation Biology 12(1):209 221. D ouble, M. and A. Cockburn. 2000. Pre dawn infidelity: Females control extra pair mating in superb fairy wren. Proceedings of the Royal Society of London 267:465 470. Emlen, S. T. 1991. Evolution of cooperative breeding in birds and mammals. Pages 301 3 35. In J. R. Krebs and N. B. Davie s (Eds.), Behavioural ecology: A n evolutionary approach, Blackwell, Oxford. Fahrig, L. 2003. Effects of habitat fragmentation on the extinction threshold: A synthesis. Ecological Applications 12:345 353. Fisher, A. M. and D. C. Goldney. 1997. Use by birds of riparian vegetation in an extensively fragmented landscape. Pacific Conservation Biology 3:275 288. Fleischner, T. L. 1994. Ecological costs of livestock grazing in western North America. Conservation Biology 8(3):629 644. Ford, H. A., G. W. Barrett, D. A. Saunders, and H. F. Recher. 2001. Why have birds in the woodlands of Sou thern Australia d eclined? Biological Conservation 97:71 88. Fower, J., L. Cohe n, and P. Jarvis. 1998. Prac tical statistics for field biologists John Wiley & Sons, Chichester. Franklin, D. C. 1998. Patterns in the decline of granivorous birds in the northern Australian savannas. Masters Thesis, Northern Territory University, Darwin. Franklin, D. C. 1999. Evidence of disarray amongst granivorou s bird assemblages in the savannas of northern Australia, a region of sparse human settlement. Biological Conservation 90:53 68. Fraser, F. J. 2000. The impacts of fire and grazing on the partridge p igeon: the ecological requirements of a declining tropica l granivore. PhD dissertation, Australian National University, Canberra. Fry, C. H. 1980. Survival and longevity among tropical landbirds. Proceedings of the Pan African Ornithological Congress 176(4):333 343.
141 Garnett, S. 1992. The action plan for Austral ian b irds Australian National Parks and Wildlife Service, Canberra. Garnett, S. T. and G. M. Crowley. 2000. The action plan for Australian birds 2000. Environment Australia, Canberra. ubfowl Press, Parap. Gordon, D. R. 1998. Effects of invasive, non indigenous plant species on ecosystem proces ses: L essons from Florida. Ecological Applications 8(4):975 989. Gregory, S.V., F. J. Swanson, W. A. McKee and K.W. Cummins. 1991. An ecosystem pe rspective of riparian zones. Bioscience 41:540 541. Griffin, G. F., D. M. Stafford Smith, S. R. Morton, G. E. Allan, K. A. Masters and N. Preece. 1989. Status and implications of the invas ion of t amarisk ( Tamarix aphylla ) on the Finke River, Northern Terri tory, Australia. Journal of Environmental Management 29:297 315. Griffiths, A. D. and K. A. Christian. 1996. The effects of fire on the frillneck lizard ( Chlamydosaurus kingii ) in northern Australia. Australian Journal of Ecology 21:386 398. Groves, R. H. 1997. Recent incursions of weeds to Australia 1971 1995. Technical Series No. 3. Cooperative Research Centre for Weed Management Systems, Adelaide. Groombridge Cambridge. Har dling, R. and H. Kokko. 2003. Life history traits as causes or con sequences of social behaviour: W hy do cooperative breeders lay small clutches? Evolutionary Ecology Research 5:6 700. Hill, D. 2000. Pastoral lands and enterprises. Page 14. In M. Kraatz (Ed .), Managing for healthy country in the VRD. Tropical Savannas CRC, Darwin. Hill, M. E. and H. Caswell. 1999. Habitat fragmentation and extinction thresholds on fractal landscapes. Ecology Letters 2:121 127. Hobbs, R. J. 2005. Landscapes, ecology and wildl ife management is a highly modified environment an Australian perspective. Wildlife Research 32:389 398. Jansen, A. and A. I. Robertson. 2001. Relationship between livestock management and the ecological condition of riparian habitats along an Australian floodplain river. Journal of Applied Ecology 38(1): 63 75. Karfs, R. A. 2000. Geology, geomorphology and soils. Pages 9 10. In Kraatz, M (ed.), Managing for healthy country in the VRD. Tropical Savannas CRC, Darwin.
142 Karfs, R. A. and M. Truemann. 2005. T racking changes in the VRD pastoral District, Northern Territory, Australia. Report to the Australian Collaborative Rangeland Information System (ACRIS) Management Committee. NT Department of Natural Resources, Environment and the Arts. Kirby, S. L. and J. J. Faulks. 2004. An assessment of the physical and ecological condition of the Victoria River and its major tributaries. Technical Report No. 44/2004. Top End Waterways Project, Katherine. Knopf, F. L., J. A. Sedgewick, and R.W. Cannon. 1988. Guild stru cture of a riparian avifauna relative to seasonal cattle grazing. Journal of Wildlife Management 52:280 290. Knopf, F. L. and F. B. Samson. 1994. Scale perspectives on avian diversity in western riparian ecosystems. Conservation Biology 8:669 676. Kraatz, M. 2000. Managing for healthy country in the VRD. Tropical Savannas CRC, Darwin. Krueper, D., J. Bart, and T. D. Rich. 2003. Response of vegetation and breeding birds to the removal of cattle on the San Pedro River, Arizona (U.S.A). Conservation Biology 1 7(2):607 615. Lack, D. 1965. Evolutionary e cology. Journal of Ecology 53(2):237 245. Lack, D. 1968. Ecologica l adaptations for breeding in b irds. Methuen, London. Lambeck, R. J. 1997. Foc al species: A multispecies umbrella for nature conservation. Conser vation Biology 11:849 856. Laurence, W. F. and J. D. Grant. 1994. Photographic identification of ground nest predators in Australian tropical rainforest. Wildlife Research 21:241 248. Lawler, J. J., D. White, J. C. Sifneos, and L. L Master. 2003. Rare spec ies and the use of indicator groups for conservation planning. Conservation Biology 17(3):875 882. lands a case study from the Victoria River District, Norther n Territory. Tropical Savannas CRC, Darwin. Lindenmayer, D. B., A. D. Manning, P.L. Smith, H.P. Possingham, J. Fischer, I. Oliver and M. A. McCarthy. 2002. The focal species approach and l andscape restoration: A critique. Conservation Biology 16(2):338 345 Lonsdale, W. M. 1993. Rates of spread of an invading species Mimosa pigra in northern Australia. Journal of Ecology 81:513 521. Lonsdale, W. M. 1994. Inviting trouble: I ntroduced pasture species in northern Australia. Australian Journal of Ecology 19: 345 354.
143 McGill, A. R. 1970. Australian w arblers. The Bird Observers Club, Melbourne. McIntyre, S. and S. Lavorel. 1994. Predicting richness of native, rare, and exotic plants in response to habitat and disturbance variables across a variegated landscape. Conservation Biology 8(2):521 531. Mac Nally, R. and E. Fleishman. 2004. A successful predictive model of species richness based on indicator species. Conservation Biology 18(3):646 654. Machtans, C. S., M A. Villard, and S. J. Hannon. 1996. Use of riparia n buffer strips as movement corridors by forest birds. Conservation Biology 10(5):1366 1379. Major, R. E. 1991. Identification of nest predators by photography, dummy eggs, and adhesive tape. Auk 108:190 195. Malanson, G. P. 1993. Riparian landscapes. Camb ridge University Press, New York. Martin, T. G. and H. P. Possingham. 2005. Predicting the impact of livestock grazing on birds using foraging height data. Journal of Applied Ecology 42(2):400 408. Meers, T. and R. Adams. 2003. The impact of grazing by eas tern grey k angaroos ( Macropus giganteus ) on vegetation recovery after fire at Reef Hills Regional Park, Victoria. Ecological Management and Restoration 4(2):126 132. Mueller, F. 1858. Botanical report on the North Australian Expedition under the command of A. C. Gregory, Esq. Journal of the Proceedings of the Linnean Society, Botany 2:137 168. Mulder, R. A. 1992. Evolutionary e cology of the mating system of superb f airy wrens. Unpublishe d Ph D thesis. Australian National University, Canberra. Mulder, R. A. 1997. Extra group courtship displays and other reproductive tactics of superb f airy wrens. Australian Journal of Zoology 45:131 143. Mulvaney, J. and J. Kamminga. 1999. Prehistory of Australia. Allen and Unwin, Sydney. Naiman, R. J., H. Decamps, and M. Pollock. 1993. The role of riparian corridors in maintaining regional diversity. Ecological Application 3:209 212. Nias, R. C. and H. A. Ford. 1992. The influence of group size and habitat on reproductive success in the superb f airy wren Malurus cyaneus E mu 92:238 243. Noske, R. A. and D. C. Franklin. 1999. Breeding seasons of land birds in the Australian monsoo nal tropics: D iverse responses to a highly seasonal environment. Australian Biologist 12(1):72 90. Officer, H R. 1964. Adaptability of the purple crowned w ren to loss of habitat. Emu 63:340 342.
144 Peart, J. 2001. Pastoral production in wetlands and riparian areas. Page 4. In wise use of wetlands in northern Australia : G razing management in wetlands and riparian habitats. Proceedings of a Workshop, Da rwin, N.T. 255. Petheram, R. J. and B. Kok. 2003. Plants of the Kimberley region of Western Australia. University of Western Australia Press, Crawley, Western Australia. Pimental, D., L. Lach, R. Ziniga, and D. Morrison. 2000. Environmental and ecological costs of nonindigenous species in the United States. Bioscience 50:53 65. Planty Tabacchi, A., E. Tabacchi, R. J. Naiman, C. Deferrari and H. Decamps. 1996. Invasibility of species ri ch communities in riparian zones. Conservation Biology 10(2):598 607. Price, O., Palmer, C., Hooper, D., Dostine, P., Lewis, M., and Woinarski, J. (in press). Evaluation of a population monitor ing program for the endangered gouldian f inch in the Yinberri e Hills, northern Australia. Emu. Recher, H. F. and L. Lim. 1990. A review of current ideas of the extinction, conservation and Society of Australia 16:187 301. Ricketts, America? Patterns of species richness and the utility of indicator taxa for conservation. Bioscience :369 381. Riddett, L.A. 1990. Kine, kin, and country: The Victoria River Dis trict of the Northern Territory 1911 1966. Australian National University, Northern Territory. Ridpath, M. G. 1985. Ecology of the wet dry tropics: how different? Proceedings of the Ecological Society of Australia 13:3 20. Roberge, J M and P. Angelstram. 2004. Usefulness of the umbrella species as a conservation tool. Conservation Biology 18(1):76 85. Rossiter, N. A., S. A. Setterfield, M. M. Douglas, and L. B. Hutley. 2003. Testing the grass fire cycle: alien grass invasion in the tropical savannas of n orthern Australia. Diversity and Distribution 9(3):169 176. y by a passerine, the superb blue w ren Malurus cyaneus (L.). Behaviour 19:170 176. Rowley, I. 1969. Communal species of Australian birds. Bonner Zoologische Beitrage 19:361 168. Rowley, I. 1988. The purple crowned f airy wren Malurus coronatus RAOU Conservation Statement. RAOU Report No. 34.
145 Rowley, I. 1993. The p urple crowned f airy wren Malurus coronatus I. History, distribution and present s tatus. Emu 93:220 34. Rowley, I., M. Brooker and E. Russell. 199 1. The breeding biology of the splendid f airy wren Malurus splendens : the significance of multiple broods. Emu 91:197 221. Rowley, I. and E. Russell. 1993. The purple crowned f airy wren Maluru s coronatus II. Breeding biology, social organization, demography, and management. Emu 93:235 50. Rowley, I. and E. Russell. 199 5. The breeding biology of the white winged f airy wren Malurus leucopterus leuconotus in a Western Australian coastal heathla nd. Emu 95:175 184. Rowley, I. and E. Russell. 1997. Bird families of the world; Fairy wrens and g rasswrens. Oxford University Press, Oxford. Rowley, I. and E. Russell. 2002. A population study of the blue breasted f airy wren, Malurus pulcherrimus at Drya ndra, Western Australia. Emu 102:127 135. Russell, E. M. 1989. Co operative breeding a Gondwanan perspective. Emu 89:61 62. Russell, E. and I. Rowley. 1998. The effects of fire on a pop ulation of red winged f airy wrens Malurus elegans in k arri forest in southwestern Australia. Pacific Conservation Biology 4:197 208. Saab, V. 1999. Importance of spatial scale to habitat use by breed ing birds in riparian forests: A hierarchical analysis. Ecological Applications 9:135 151. Saab, V. A., C. E. Brock, T. D. Ric h and D. S. Dobkin. 1995. Livestock grazing effects on migratory birds in western North America. Pages 311 353. In T. E. Martin and M. Finch (eds.), Ecology and management of migratory b irds: A synthesis and review of critical issues, Oxford University Pr ess, New York. Sattler, P. S. 1993. Riparian zone management in Queensland and the Northern Territory. Pages 157 171. In Bunn, S. E., B. J. Pusey and P. Price (eds.), Ecology and management of riparian zones in Australia. Land and Water Resources Research and Development Corporation, Canberra. Schodde, R. 1982. The f airy wrens: a monograph of the Maluridae. Lansdowne, Melbourne. Scott, M. L., S. K. Skagen and M. F. Merigliano. 2003. Relating geomorphic change and grazing to avian communities in riparian fo rests. Conservation Biology 17(1):284 296. Schultz, T. T. and W. C. Leininger. 1990. Differences in riparian vegetation structure between grazed areas and exclosures. Journal of Range Management 43:295 299. Setterfield, S. A., M. M. Douglas, L. B. Hutley, and M. Welch. 2005. Effects of canopy cover and ground disturbance on establishment of an invasive grass in an Australian s avanna. Biotropica 37(1):25 31.
146 Sharp, B. R. and D. M. J. S. Bowman. 2004. Net woody vegetation increase confined to seasonally inun dated lowlands in an Australian tropical savanna, Victoria River District, Northern Territory. Austral Ecology 29:667 683. Sharp, B. R. and R. J. Whittaker. 2003. The irreversible cattle driven transformation of a seasonally flooded Australian savanna. Journal of Biogeography 30:783 802. Sibley, C. G. and J. E. Ahlquist. 1985. The phylogeny and classification of the Australo Papuan passerine birds. Emu 85:1 14. Skagen, S. K., C. P. Melcher, W. H. Howe, and F. L. Knopf. 1998. Comparative use of riparian corridors and oasis by migrating birds in southeastern Arizona. Conservation Biology 12(4):896 909. Smith, L. A. and R. E. Johnstone. 1977. Status of the purple crowned w ren ( Malurus coronatus ) and buff sided r obin ( Poecilodryas superciliosa ) in Western A ustralia. Western Australian Naturalist 13:185 188. Smith, N. M. 1995. Weeds of natural ecosystems: a field guide to environmental weeds of the Northern Territory, Australia. Environment Centre, NT. Darwin, NT. Stacey, P.B. and J.D. Ligon. 1991. The benef its of philopatry hypothesis for the evolution of cooperative breeding: variation in territory quality and group size. American Naturalist 137:831 846. Stewart, G. A., R. A. Perry, S. J. Paterson, D. M. Traves, R. O. Slatyer, P. R. Dunn, P. J. Jones and J R. Sleeman. 1970. Lands of the Ord Victoria area, W.A. and N.T. Land Research Series No. 28, CSIRO, Melbourne. Soul, M. E. 1990. The onslaught of alien species, and other challenges in the coming decades. Conservation Biology 4:233 239. Start, A. N. an d T Handasyde. 2002. Using photographs to document environmental change: T he effects of dams on the riparian vegetation of the lower Ord River. Australian Journal of Botany 50:465 480. Stirrat, S. C. 2002. Foraging ecology of the agile wallaby ( Macropus a gilis ) in the wet dry tropics. Wildlife Research 29:347 361. Taylor, D. M. 1986. Effects of cattle grazing on passerine birds nesting in riparian habitat. Journal of Rangeland Management 39:254 258. Taylor, J. A. and C. R. Dunlop. 1 985. Plant communities of the wet dry t ropics of Australia: the Alligator Rivers region. Proceedings of the Ecological Society of Australia 13:83 127. Thomas, F. 1964. The pastoral industries of the w est Kimberley area. Pages 192 209. In General report on lands of the West Kimberley area, W.A. CSIRO Land Research Series No. 9.
147 Tidemann, S. C. 1986. Breeding in three species of fairy wren ( Malurus ): do helpers really help? Emu 86:131 138. Tidemann, S. C. 1996 Causes of the decline of the gouldian f inch Erythrura gouldiae Bird Conservation International 6:49 61. Van Doorn, A. and Low Choy, J. In prep. Canegrass: a clarification of the primary habitat of the purple crowned f airy wren ( Malurus c. coronatus ) in the Victoria River District, Northern Territory. Vitousek, P. as global environmental change. American Scientist 84:468 478. Watson, J., D. Freudenberger and D. Paull. 2001. An assessment of the focal species approach for conserving birds in variegated landscapes in southeastern Australia. Conservation Biology 15(5):1364 1373. Watson, J. E. M., R. J. Whittaker and D. Feudenberger. 2005. Bird community responses to habi tat fragmentation: H ow consistent are they across landscapes? Journa l of Biogeography 32:1353 1370. Werner P A 2005 Impact of feral water buffalo and fire on growth and survival of ma ture savanna trees: a n experimental field study in Kakadu National Park, northern Australia Austral Ecology 30: 625 647. Whitehead P. 20 00. The clever country: W here cows manage wildlife. Pages 155 168. In R. Dixon (ed.), Business as usual? local c onflicts a nd global challenges in n orthern Australia A Collection of Papers at the Challenges of the North Conference, Darwin, N.T. Wiegand, T ., E. Revilla and K. A. Moloney. 2005. Effects of habitat loss and fragmentation on population dynamics. Conservation Biology 19(1):108 121. Wiens, J. A. 1989. The ecology of bird communities, volume 2, processes and patterns. Cambridge University Press, Cambridge. Williams, R. J., A. D. Griffiths, and G. E. Allen. 2002. Fire regimes and biodiversity in the savannas of northern Australia. Pages 281 304. In Bradstock, R. A., J. E. Williams and A. M. Gill s (eds.), Flammable Australia: T he fire regimes and bi odiversity of a continent, Cambridge University Press, Cambridge. Wilson, B. A., P. S. Brocklehurst, M. J. Clark and K. J. M. Dickson. 1990. Vegetation survey of the Northern Territory, Australia. Technical Report No. 49, Conservation Commission of the NT. Wilson, L. O. 1979. Public forum. Pages 77 87. In Cope, O. B. (Ed.), Grazing and riparian/stream ecosystems, Trout Unlimited, Denver, Colorado.
148 Woinarski, J. C. Z. 1985. Breeding biology and life history of small insectivorous birds in Australia forests : R esponse to a stable environment? Proceedings of the Ecological Society of Australia 14:159 168. Woinarski, J. C. Z. 1990. Effects of fire on the bird communities of tropical woodlands and open forests in northern Australia. Australian Journal of Ecolog y 15:1 22. Woinarski, J. C. Z. and C. P. Catterall. 2004. Historical changes in the bird fauna at Coomooboolaroo, northeastern Australia, from the early years of pastoral settlement (1873) to 1999. Biological Conservation 116:379 401. Woinarski, J. C. Z. a nd H. F. Rech er. 1997. Impact and response: A review of the effects of fire on the Australian avifauna. Pacific Conservation Biology 3:183 205. Woinarski, J. C. Z, C. Brock, M. Armstrong, C. Hempel, D. Cheal, and K. Brennan. 2000. Bird distribution in rip arian vegetation in the extensive natural landscape of Australia savanna: A broad scale survey and analysis of a distributional data base. Journal of Biogeography 27:843 868. Woinarski, J.C.Z and A.J. Ash. 2002. Responses of vertebrates to pasto ralism, military land use and landscape position in an Australian tropical savanna. Austral Ecology 27(3):311 323. Yates, C. J., D. A. Norton, and R. J. Hobbs. 2000. Grazing effects on plant cover, soil and microclimate in fragmented woodlands in south wes tern Australia : I mplications for restoration. Austral Ecology 25(1):36 47. Yibarbuk, D. M., P. J. Whitehead, J. Russell Smith, D. Jackson, A. Fisher, P. Cooke, D. Choquenot, and D. J. M. S. Bowman. 2001. Fire ecology and Aboriginal land management in central Arnhem Land, Northern Australia : A tradition of ecosystem management. Journal of Biogeography 28:325 343. Yom Tov, Y. 1987. The reproductive rates of Australian passerines. Australian Wildlife Research 14:319 330.
149 BIOGRAPHICAL SKETCH Annamaria van Doorn was born in Rhenen, The Netherlands on June 15, 1971. She attended Menlo College in California where s he received her Ba chelor of Science in Biological Sciences in 1992. In 1994 she commenced graduate school at the University of Florida and earned her Masters of Science degree in Wildlife Ecology and Conservation in 1997. Her Masters research focused on introduced parrots in Florida and further developed her interest in birds. Annamaria continued her graduate studies at the University of Florida where her focus shifted from parrots to fairy wrens. Annamaria conducted her PhD research on p urple crowned f airy wrens in the Northern Territory of Australia. Annamaria continues to live in the Northern Territory where she will be conducting further research on purple crowned f airy wrens and working as an independent environmental consultant.