Bird population recovery 22 years after intensive logging near Eden, New South Wales

Transcription

1 CSIRO PUBLISHING Emu, 2003, 103, Bird population recovery 22 years after intensive logging near Eden, New South Wales Rodney P. Kavanagh and Matthew A. Stanton Forest Research Division, State Forests of New South Wales, PO Box 100, Beecroft, NSW 2119, Australia. Abstract. Intensive logging for sawlogs and woodchips in south-eastern Australia is known to cause an immediate impact on bird populations. We measured bird abundance and species turnover during the medium term (13 and 22 years) after intensive alternate-coupe logging in a forest area that had been studied previously for short-term (0 4 years) impacts. The study provides a series of snapshots of bird population recovery following one of the most intensive logging operations in the region when, unlike the current practice, no old trees were retained in logged areas and no unlogged forest was retained in strips along all drainage lines. The aim was to assess the longer-term effects of logging on the bird assemblage and to determine whether recovery had occurred. We found that recovery had occurred for a large component of the avifauna within 22 years of intensive logging (as practised in 1976 when few or no old trees were retained in logged areas). Most bird species that forage among canopy foliage, in the air, among the understorey and on the ground had recovered. Five common forest birds were significantly more common on logged coupes than unlogged coupes. The Bell Miner, Manorina melanophrys, a honeyeater often associated with disturbed forest and implicated in eucalypt canopy dieback, had established colonies on some logged coupes 22 years after intensive logging. However, some hollow-nesting bird species (e.g. treecreepers, cockatoos) had still not fully recovered. A number of other common forest birds occurred so rarely in the study area (on both logged and unlogged coupes) that few conclusions could be made about their responses to logging. Evidence suggested that intensive logging may have had an effect at a local landscape (compartment) level as well as at the level of individual logged coupes. This was indicated by changes to the avifauna of the unlogged coupes over time (increasing numbers of species) as the forest regenerated on the adjacent logged coupes. Introduction Intensive logging has been shown to reduce populations of forest birds in south-eastern Australia for up to 15 years after logging (Loyn 1980, 1985; Kavanagh et al. 1985; Smith 1985; Recher et al. 1991; Taylor and Haseler 1995; Taylor et al. 1997). Few studies in south-eastern Australia have examined the medium- and long-term effects of intensive logging except retrospectively, in which a range of age classes following logging (or wildfire as a surrogate for logging) were compared (Kutt 1996; Loyn 1998). Others have used ecological studies to infer the effects of logging on birds (Recher et al. 1985; Ford et al. 1986). No medium- or longterm studies, in which pre-logging data were obtained for logged plots, have been published. Taylor et al. (1997) monitored changes in bird assemblages for seven years on two logged plots, only one of which had pre-logging data. Kavanagh et al. (1985) compared bird assemblages between a series of logged and adjacent unlogged coupes each year for five years beginning immediately after intensive logging. Logged coupes were found to have greatly altered populations of some bird species. The trend remained for the duration of the study ( ), but was most pronounced during the first two years of the study. Bird numbers on the unlogged coupes were also low but increased with time, always remaining greater than on the logged coupes. By 1980, when the regenerating forest had reached four years of age, 78% of the species recorded at that time on the unlogged coupes were also present on the logged coupes (Kavanagh et al. 1985). Some species were recorded only on the logged coupes. The greatest adverse effects of logging were on birds that required tree hollows for nesting, on species that normally foraged in the forest canopy and on birds that forage among moist ground litter. These effects were greater upon non-passerines than on passerines. Two groups of birds were advantaged by logging. During (1 2 years after logging), these were species that are normally associated with open or disturbed situations. By 1979 and 1980 (3 4 years after logging), species that are normally associated with a dense shrub layer were abundant on the logged coupes. The aim of the current study was to assess the longer-term effects of intensive logging on bird assemblages and to determine whether recovery had occurred by 13 and 22 years after logging. We compared the bird assemblages occurring on the same series of adjacent logged and unlogged coupes that were studied by Kavanagh et al. (1985). Recovery by the Royal Australasian Ornithologists Union /MU /03/030221

2 222 Emu R. P. Kavanagh and M. A. Stanton local avifauna over the period 4, 13 and 22 years after logging was examined at both the coupe (~15 ha) and the compartment (~ ha) scale. Methods Study area The study area was located in open forest on ridges dominated by Silvertop Ash (Eucalyptus sieberi), Blue-leaved Stringybark (E. agglomerata) and White Stringybark (E. globoidea). The alternate logged and unlogged coupes were located in an area bounded by Banksia Road and the Imlay Forest Road in East Boyd State Forest about 30 km south of Eden, New South Wales (see Fig. 1; also Kavanagh et al. 1985). Each coupe was about ha in size, and adjacent logged unlogged coupes were treated as matched pairs for analysis. Coupes were paired with others, usually those adjacent, that had the most similar vegetation type and aspect (Fig. 1, Table 2). The study area was the first compartment in the region to be logged in the alternate-coupe manner. Previously, much larger areas were harvested in a single operation. Intensive logging for sawlogs and woodchips in half of the study area was carried out in 1976 according to the standard protocols of the day, which did not include the retention of any old or mature trees to improve habitat conditions for wildlife, or the retention of any unlogged forest along drainage lines. These latter provisions did not become standard practice in the region until 1981, and have gradually evolved into the conservation protocols now routinely applied (Anon. 1999). The study area was burnt by wildfires in 1952 and No postlogging burns were undertaken, but the area was subject to a fuelreduction burn in 1994 (42% coverage with 9% crown scorch). Regeneration assessments made two years after logging showed that high levels of tree stocking occurred naturally and represented the original tree species present on each coupe (South-east Region Forestry Office, Eden, records). Log dumps were ripped and planted with Yellow Stringybark (E. muelleriana), which also occurred naturally in the study area. In 1980, the understorey on logged coupes consisted of a dense layer of shrubs dominated by Goodenia ovata, Daviesia latifolia, Acacia longifolia, A. mucronata and A. terminalis under an overstorey of regenerating eucalypts about 6 m high. Over time, the cover of understorey species thinned appreciably so that, by 1998, the understorey was easy to walk through, except for some patches of Wiry Ricegrass (Tetrarrhena juncea). By 1998, the regenerating eucalypts had almost reached the height of trees on the unlogged coupes. Sampling methods The 14 logged and unlogged coupes (7 matched pairs) established by Kavanagh et al. (1985) and last sampled in 1980 were re-sampled in 1989 (13 years after logging) and 1998 (22 years after logging). Sampling plots within each coupe were based on a point located near the centre of the coupe on an upper mid-slope area suitable for harvesting (Fig. 1). Pairs of plots were located in adjacent logged and unlogged coupes. The 1980 data are reproduced in this paper to facilitate comparisons with data collected in 1989 and Two additional coupes providing an 8th pair of logged unlogged coupes were sampled in 1998 to increase the sample size for future studies, but data from these coupes are not considered further here. Birds were counted on 1-ha plots in the logged and unlogged coupes in early November of each year. Four observers were rotated among the plots, with each spending 1 h on each plot recording all birds seen and heard within a 60-m radius of a central point. Each 1-ha plot was marked with flagging tape at the four compass points to reduce observer error when estimating the plot boundaries. Each plot was censused four times, once per observer, over a minimum of two mornings, usually by means of two early (before 0900 hours) and two late (after 0900 hours, but before 1200 hours) counts. Analysis Differences between matched pairs of logged and unlogged coupes were tested using Wilcoxon s signed-rank test (S-plus: MathSoft 1999). Comparisons were made within years for total species richness, total number of individual birds, mean species richness, and for the abundance of individuals of each species. The data for each count on each coupe were entered into the analysis in the order in which they were collected. This was done to minimise the influence of prior knowledge about the coupes that may have accumulated among the observers. Thus, the first count on one coupe was matched with the first count on another, and so on. The power of the (non-parametric) test to find no significant differences between treatments was estimated, for each year, by plotting the total counts for each species in the unlogged coupes against the total counts in the logged coupes. The significance level (P < 0.05) for each species was displayed and used to indicate the range of sample sizes that were adequate for the test to find significant differences, should E B2 O N G P Q R F M L B D S Fig. 1. The pattern of alternate small (~15 ha) coupe logging in the Banksia Road study area near Eden. Letters indicate the location of bird sampling sites metres B Sample site Roads Drainage Logged Unlogged N S W

3 Bird population recovery after logging Emu 223 they exist. Using this form of sensitivity analysis, it was possible to differentiate between bird species that showed no significant difference between treatments from those for which insufficient data were available for statistical testing. Species turnover (changes in species composition and abundance) within the study area was assessed by multi-dimensional scaling of the bird assemblages occurring on the logged and unlogged coupes. Semistrong hybrid multi-dimensional scaling (SSHMDS) based on the Bray Curtis association measure (Belbin 1995) was used to provide a relational picture (an ordination scatter plot) of the species assemblages occurring during each coupe-year census. The degree of similarity in assemblage composition between each coupe-year is indicated by their inter-site distances. Coupe-years close together shared the same species with similar abundance distributions. Before analysis, species abundance data were log 10 -transformed. Analysis of similarities (ANOSIM: Clarke 1993; in PATN: Belbin 1995) was used to test for significant differences in bird assemblages between treatments in each year and between years for each treatment. This test uses a randomised permutation procedure (99999 iterations) to compare the level of similarity within designated groups to those produced by chance. Results A total of 76 bird species was recorded during the three sampling years from 1980 to 1998 (Table 1). Five species (Striated Thornbill, Yellow-faced Honeyeater, Brown Thornbill, White-naped Honeyeater and Grey Fantail) made up 46.0% of all birds recorded during the study. Thirty-seven species made up 95.3% of all records, with a further 39 species making up the remainder of the data set. During 1980, 1989 and 1998, 22 species were recorded in each sampling year on both logged and unlogged coupes, 29 species were recorded in each sampling year on logged coupes, and 27 species were recorded in each sampling year on unlogged coupes (Table 1). As indicated above, this contrasts with the relatively large number of species that were recorded only rarely in the study area. For example, 17 species were recorded only in a single year and on a single treatment (Table 1). Three of these were raptors, which would have large ranges and live at naturally low population densities. One was a common nocturnal bird that is infrequently detected during diurnal surveys. Differences in bird abundance A summary of the abundance of birds in each year and on each logged and unlogged coupe is provided in Table 2. These data show a pattern of increasing numbers of birds, and bird species, on the logged coupes during the study. However, on the unlogged coupes, bird abundance remained fairly stable but the number of species increased considerably. In 1989, 13 years after intensive logging, there were significantly more birds, and more species, on the unlogged coupes than on the logged coupes (Table 3). By 1998, 22 years after logging, these differences had disappeared (Table 3), indicating that bird population recovery had occurred at the community level, but not for all species (see below). Species responses For each species, the mean number of birds recorded per hectare per hour on the logged and unlogged coupes is shown in Table 1. In total, 20 species (each indicated by an asterisk) showed a significant response to logging (at P < 0.05) during at least one year of the study. In 1980, four years after logging, eight species (Fan-tailed Cuckoo, Black-faced Cuckoo-shrike, Scarlet Robin, Eastern Yellow Robin, Rufous Whistler, Grey Shrike-thrush, Striated Thornbill and Whitethroated Treecreeper) were significantly less abundant on logged coupes and three species (Pilotbird, Yellow-faced Honeyeater and Red-browed Finch) were significantly more abundant on the logged coupes (Table 4). (Note: scientific names for each species are provided in Table 1). In 1989, 13 years after logging, eight species (Eastern Yellow Robin, Golden Whistler, Grey Fantail, Striated Thornbill, Whitethroated Treecreeper, Red-browed Treecreeper, Varied Sitella and White-naped Honeyeater) were significantly less abundant on logged coupes and no species were significantly more abundant on the logged coupes (Table 4). In 1998, 22 years after logging, five species (Gang-gang Cockatoo, Fan-tailed Cuckoo, White-throated Treecreeper, Red-browed Treecreeper and Red Wattlebird) were significantly less abundant on logged coupes and five species (Rufous Whistler, Golden Whistler, Eastern Whipbird, Striated Thornbill and Bell Miner) were significantly more abundant on the logged coupes (Table 4). The three species that in 1980 showed a positive response to logging (Pilotbird, Yellow-faced Honeyeater and Redbrowed Finch) had each declined in abundance on the logged coupes by 1989 and 1998 (Table 1). For these species, there appeared to be no differences in their abundance between the logged and the unlogged coupes, but data for statistical testing were adequate only for the Yellow-faced Honeyeater (Table 4). Similarly, of the eight species in 1980 that showed a negative response to logging (see above), three (Fan-tailed Cuckoo, Rufous Whistler and Grey Shrike-thrush) showed no differences between logged and unlogged coupes in 1989 (the data for Scarlet Robin and Black-faced Cuckoo-shrike were insufficient for statistical testing), and by 1998 two of these species (Rufous Whistler and Striated Thornbill) were actually more abundant on the logged coupes (Table 4). One species (Fan-tailed Cuckoo) had reverted to being less common on the logged coupes in Thus, four species (Fan-tailed Cuckoo, Eastern Yellow Robin, Striated Thornbill and White-throated Treecreeper) were slow to recover from logging. However, by 1998, one species (Eastern Yellow Robin) had recovered, another (Striated Thornbill) became more abundant on the logged coupes, but two species (Fan-tailed Cuckoo and White-throated Treecreeper) still remained more abundant on the unlogged coupes, although numbers of the White-throated Treecreeper were slowly increasing on the logged coupes (Table 4). Several

4 224 Emu R. P. Kavanagh and M. A. Stanton Table 1. Number of birds (mean ha 1 h 1 ) for logged and unlogged coupes 4, 13 and 22 years after logging Nomenclature and sequence follows Christidis and Boles (1994). indicates a detection rate lower than * indicates a significant result at P < 0.05 (see Table 4). Total species: 1980, 50; 1989, 54; 1998, 61; , 76 Species Logged Unlogged Brown Goshawk Accipiter fasciatus Collared Sparrowhawk Accipiter cirrhocephalus Australian Hobby Falco longipennis Common Bronzewing Phaps chalcoptera Brush Bronzewing Phaps elegans 0.1 Wonga Pigeon Leucosarcia melanoleuca 0.1 Yellow-tailed Black-Cockatoo Calyptorhynchus funereus Gang-gang Cockatoo Callocephalon fimbriatum* Rainbow Lorikeet Trichoglossus haematodus Musk Lorikeet Glossopsitta concinna Little Lorikeet Glossopsitta pusilla Australian King Parrot Alisterus scapularis Crimson Rosella Platycercus elegans Brush Cuckoo Cacomantis variolosus Fan-tailed Cuckoo Cacomantis flabelliformis* Horsfield s Bronze-Cuckoo Chrysococcyx basalis 0.1 Shining Bronze-Cuckoo Chrysococcyx lucidus Australian Owlet-nightjar Aegotheles cristatus Laughing Kookaburra Dacelo novaeguineae 0.1 Sacred Kingfisher Todiramphus sanctus Superb Lyrebird Menura novaehollandiae White-throated Treecreeper Cormobates leucophaeus* Red-browed Treecreeper Climacteris erythrops* Superb Fairy-wren Malurus cyaneus Spotted Pardalote Pardalotus punctatus Striated Pardalote Pardalotus striatus 0.1 Pilotbird Pycnoptilus floccosus* White-browed Scrubwren Sericornis frontalis Brown Gerygone Gerygone mouki 0.1 Brown Thornbill Acanthiza pusilla Buff-rumped Thornbill Acanthiza reguloides 0.1 Striated Thornbill Acanthiza lineata* Red Wattlebird Anthochaera carunculata* Little Wattlebird Anthochaera chrysoptera Bell Miner Manorina melanophrys* Lewin s Honeyeater Meliphaga lewinii Yellow-faced Honeyeater Lichenostomus chrysops* White-eared Honeyeater Lichenostomus leucotis 0.3 Yellow-tufted Honeyeater Lichenostomus melanops 0.2 Brown-headed Honeyeater Melithreptus brevirostris White-naped Honeyeater Melithreptus lunatus* Crescent Honeyeater Phylidonyris pyrrhoptera New Holland Honeyeater Phylidonyris novaehollandiae 0.1 Eastern Spinebill Acanthorhynchus tenuirostris Scarlet Honeyeater Myzomela sanguinolenta Jacky Winter Microeca fascinans Scarlet Robin Petroica multicolor* Flame Robin Petroica phoenicea Rose Robin Petroica rosea Eastern Yellow Robin Eopsaltria australis* Eastern Whipbird Psophodes olivaceus* Spotted Quail-thrush Cinclosoma punctatum 0.4 Varied Sittella Daphoenositta chrysoptera* Crested Shrike-tit Falcunculus frontatus 0.1 Continued next page

5 Bird population recovery after logging Emu 225 Table 1. Continued Species Logged Unlogged Golden Whistler Pachycephala pectoralis* Rufous Whistler Pachycephala rufiventris* Grey Shrike-thrush Colluricincla harmonica* Black-faced Monarch Monarcha melanopsis Leaden Flycatcher Myiagra rubecula Rufous Fantail Rhipidura rufifrons Grey Fantail Rhipidura fuliginosa* Black-faced Cuckoo-shrike Coracina novaehollandiae* White-bellied Cuckoo-shrike Coracina papuensis 0.1 Cicadabird Coracina tenuirostris 0.1 Olive-backed Oriole Oriolus sagittatus White-browed Woodswallow Artamus superciliosus 1.4 Dusky Woodswallow Artamus cyanopterus 0.1 Grey Butcherbird Cracticus torquatus 0.3 Pied Currawong Strepera graculina Australian Raven Corvus coronoides Satin Bowerbird Ptilonorhynchus violaceus Red-browed Finch Neochmia temporalis* Mistletoebird Dicaeum hirundinaceum Welcome Swallow Hirundo neoxena 0.1 Silvereye Zosterops lateralis Common Blackbird Turdus merula Total species species, notably the Red-browed Treecreeper and the Ganggang Cockatoo, showed negative effects as late as 22 years after logging, because they were only just beginning to recolonise the unlogged coupes within the study area (Table 1). The Scarlet Robin was notable for its apparent decline throughout the study on both the logged and the unlogged coupes (Tables 1, 4). In total, 26 species showed no significant response to logging during at least one year of the study (Table 5). Four species (Brown Thornbill, Eastern Spinebill, Spotted Parda- Table 2. Summary of bird census results for logged and unlogged coupes in 1980, 1989 and 1998 The bird census in 1980 was 4 years after logging, that in 1989 was 13 years after logging and that in 1998 was 22 years after logging Treatment/coupe Mean individuals ha 1 h 1 Mean species ha 1 h 1 Total species ha 1 (4h) Logged B Unlogged D Logged B Unlogged E Logged F Unlogged G Logged O Unlogged N Logged Q Unlogged P Logged M Unlogged R Logged S Unlogged L Logged plots mean Unlogged plots mean

6 226 Emu R. P. Kavanagh and M. A. Stanton Table 3. Differences in bird populations between logged and unlogged coupes near Eden Wilcoxon s signed-rank test with correction used to test for differences. No. of pairs refers to the number of matched-pair comparisons. **, significant at P < 0.01 No. of pairs Z P Z P Z P Individuals ha 1 h <0.001** Species ha 1 h ** ** Total species ha 1 (4h) lote and White-browed Scrubwren) were unaffected by logging (P > 0.05) during all three sampling years from 1980 to 1998 (i.e. during the period 4 22 years after logging). By 1998, 20 species were equally common on logged and unlogged coupes and, as mentioned above, a further five species (Golden Whistler, Rufous Whistler, Striated Thornbill, Bell Miner and Eastern Whipbird) were more abundant on the logged coupes (Tables 4, 5). Insufficient data for statistical testing were available for another 39 species. Bird population changes Evidence for bird population recovery within the study area can be demonstrated by the increasing similarity, or overlap, among the bird assemblages occurring on the logged and unlogged coupes (Fig. 2). The degree of overlap in multivariate space increased from 1980 to 1989, but not significantly from 1989 to 1998, for polygons representing the avifauna of logged and unlogged coupes. This is demonstrated by ANOSIM within-year comparisons contrasting logged and unlogged coupes which gave a significant (P < 0.05) result in 1980, but no significant differences in 1989 and 1998 (Table 6). Of great significance, however, was the changing avifauna of all coupes, both logged and unlogged, over time. This is demonstrated by the gradual movement of all coupes in multivariate space towards the unlogged end of the spectrum (i.e. the lower left-hand corner of Fig. 2). Also, the ANOSIM within-treatment comparisons showed a significant (P < 0.05) difference in all contrasts for unlogged coupes between years and for logged coupes between years (Table 6). This indicates that the avifauna of the unlogged coupes was also changing with time, suggesting that bird population recovery was also occurring at the local landscape (or compartment) scale. That is, forest regrowth on the logged coupes has recovered to the stage Table 4. Bird species displaying a significant difference between treatments in at least one year during Wilcoxon s signed-rank test with correction at P < 0.05 significance level. Response: ns, not significant (P > 0.05); na, insufficient data for analysis; none, the species was not detected in that year;, the species was detected less on logged coupes; +, the species was detected more on logged coupes Species (4 years post-logging) (13 years post-logging) (22 years post-logging) Response P Response P Response P Gang-gang Cockatoo none Fan-tailed Cuckoo ns White-throated Treecreeper <0.001 < Red-browed Treecreeper none Pilotbird na na Striated Thornbill < <0.001 Red Wattlebird na na Bell Miner na na Yellow-faced Honeyeater ns ns White-naped Honeyeater ns ns Scarlet Robin na na Eastern Yellow Robin ns Eastern Whipbird na na Varied Sittella na na Golden Whistler ns Rufous Whistler <0.001 ns Grey Shrike-thrush ns ns Grey Fantail ns ns Black-faced Cuckoo-shrike na ns Red-browed Finch na na

7 Bird population recovery after logging Emu 227 where it is now contributing to the capacity of the unlogged coupes to support additional bird species. Increasing numbers of uncommon species were recorded as the study progressed, particularly on the unlogged coupes. The total number of species recorded on logged and unlogged coupes, respectively, was 41 and 40 for 1980, 45 and 45 for 1989, and 46 and 59 for 1998 (Table 1). Thus, in 1998, 28% more species were recorded on the unlogged coupes than on logged coupes. Throughout the study, the total number of species recorded in each year increased from 50 in 1980, to 54 in 1989, and to 61 in Discussion No other studies in Australia have reported the pattern of succession of bird species and bird numbers following intensive logging over the medium (or long) term on the same study plots. The most comparable study (Williams et al. 2001) was undertaken in the Karri (Eucalyptus diversicolor) forests of Western Australia; this study reported the pattern of succession of bird species and bird numbers following intensive logging over 14 years on the same study plots. That study had only one replicate for each treatment (clearfelled and unlogged), but pre-logging data were collected and the immediate (1 5 years post-logging) and short-term (12 14 years post-logging) effects were reported. Several other studies have taken a comparative, or chronosequence, approach that has included forests of similar, or older, age classes following intensive logging or, more usually, following wildfire (Loyn 1980, 1985, 1998; Smith 1985; Kutt 1996). For none of these studies, including the present study, were pre-logging data available for comparison with later results, but this study did collect data from all coupes immediately after logging (see Kavanagh et al. 1985). The main finding by all studies is that bird population recovery is rapid during the first 3 7 years after logging, with many bird species, but not all, recorded in logged areas by about 15 years after logging (Loyn 1980, 1985; Kavanagh et al. 1985; Smith 1985; Taylor and Haseler 1995; Taylor et al. 1997; Williams et al. 2001). Chronosequence studies of bird populations within forests regenerating after wildfire or clearfelling suggest that full recovery of most bird species will occur within years, but that some species are not expected to fully recover for more than 100 years, unless some old trees are present in the stand to Table 5. Bird species displaying no significant difference between treatments in at least one year during Wilcoxon s signed-rank test with correction at P < 0.05 significance level. Response: na, insufficient data for analysis; significant responses (P < 0.05) indicated in bold type;, the species was detected less on logged coupes (up to P < 0.1); +, the species was detected more on logged coupes (up to P < 0.1) Species (4 years post-logging) (13 years post-logging) (22 years post-logging) Response P Response P Response P Grey Fantail Brown Thornbill White-naped Honeyeater Golden Whistler Eastern Spinebill Spotted Pardalote White-browed Scrubwren Crimson Rosella na Rufous Fantail na na Superb Fairy-wren na White-eared Honeyeater na na Yellow-faced Honeyeater Grey Shrike-thrush Rufous Whistler < Fan-tailed Cuckoo Eastern Yellow Robin Black-faced Cuckoo-shrike na Rainbow Lorikeet na na Crescent Honeyeater na na Pied Currawong na na Superb Lyrebird na na King Parrot na na Olive-backed Oriole na na Shining Bronze-cuckoo na na Musk Lorikeet na na Satin Bowerbird na na 0.630

8 228 Emu R. P. Kavanagh and M. A. Stanton Stress = Fig. 2. Similarity of bird assemblages on logged and unlogged coupes in 1980, 1989 and 1998 using SSHMDS (Belbin 1995). Polygon groupings are made by treatment within each year. Logged coupes grouped by broken lines, unlogged coupes grouped by solid lines L L L U U U provide a source of hollows for nesting (Loyn 1980, 1985; Williams et al. 2001). Population recovery Four years after logging in this study there was little similarity between the bird populations of the logged and unlogged coupes. Twelve species showed clear differences in abundance; nine of these were affected adversely. The species that were disadvantaged occupied a wide range of niches, including the forest canopy (eucalypt seed eaters, foliage-gleaning insectivores), tree trunks (bark-foraging insectivores) and the ground litter layer (insectivores) (Recher et al. 1985; Ford et al. 1986). Two of these species require hollows in old trees for nesting. Species advantaged during the first four years after logging included grass seed feeders, and insectivores inhabiting dense shrubby regrowth. The overall differences in the bird assemblages of logged and unlogged coupes were clearly displayed as non-overlapping points in the ordination representing data for all coupes and all years. Thirteen years after logging, birds were still more abundant on the unlogged coupes, as found by Smith (1985) and Williams et al. (2001) in year-old regrowth, but there was considerable overlap in the bird assemblages of the logged and unlogged coupes, as shown by the ordination display. However, at least eight species had still not recovered to pre-logging abundances. Again, these species occupied a wide range of niches from the forest canopy through the subcanopy to the ground. Species previously advantaged by logging had equalised across logged and unlogged coupes as the regrowth assumed the structure of a young forest and became less open. The spread of coupes in the ordination analysis indicates the greater variability of bird populations on the logged than on the unlogged coupes. Twenty-two years after logging, convergence had largely occurred between the bird assemblages of the logged and unlogged coupes. This was due to the recovery by most remaining common, but previously disadvantaged, species. All foliage-gleaning insectivores, aerial foraging insectivores, ground-foraging insectivores (except the Fan-tailed Cuckoo), and nectar feeders (except the Red Wattlebird) appeared to have recovered their numbers on the logged coupes. However, the slow recovery of the Fan-tailed Cuckoo, and the apparent general decline of the Scarlet Robin in the study area (including on the unlogged blocks), suggests that the open ground and understorey conditions Table 6. ANOSIM pairwise comparisons of bird species assemblages between treatments in each year and between years for each treatment ns indicates no significant difference (P > 0.05) between plot groups Year/Treatment Contrast P Within-year comparisons 1980 Unlogged Logged Unlogged Logged ns 1998 Unlogged Logged ns Within-treatment comparisons Unlogged Unlogged Unlogged Logged Logged Logged

9 Bird population recovery after logging Emu 229 required by these forest and woodland birds had still not returned. Both species have been reported as absent or uncommon in young (<15 years) regrowth (Loyn 1980; Smith 1985). Nectar-feeding birds (Red Wattlebird, Eastern Spinebill, Crescent Honeyeater, Rainbow Lorikeet, Musk Lorikeet and Little Lorikeet) increased in numbers in 1998, probably due to the heavy flowering of E. globoidea that occurred during sampling in that year. Several species responded inconsistently to logging. For example, the Golden Whistler was unaffected by logging in 1980, was adversely affected in 1989, but then was more abundant in logged coupes in Similarly, the Grey Fantail and White-naped Honeyeater showed no differences between treatments 4 years and 22 years after logging, but both were significantly more abundant in unlogged coupes 13 years after logging. These inconsistencies may be due to structural changes in the forest as it regenerates, including likely changes in the abundance of invertebrate prey, and its influence on the habitat requirements of particular species. A small number of common species, in particular the White-throated Treecreeper, remained disadvantaged by intensive logging. This species, and most others that remained less abundant on logged coupes, require tree hollows for nesting. Birds that nest in tree hollows are well known to be the main group requiring special attention by managers (Loyn 1980, 1985; Recher et al. 1980; Smith 1985; Williams et al. 2001). Several species, notably the Bell Miner, increased greatly in numbers on some logged coupes. In the case of the Bell Miner, total counts increased on the logged coupes from 1 in 1980, to 4 in 1989, to 46 in 1998, especially on coupes Q and M, which were situated among moist forest types at the heads of two gullies. During the same period, total Bell Miner counts on the unlogged coupes (all on the adjacent coupe R) were 1, 0 and 3. This provides support for the hypothesis (Stone 1999) that logging disturbance can be a contributing factor in creating the habitat conditions required by the Bell Miner. It is noteworthy that no signs of eucalypt dieback were evident. The similarity of bird assemblages on logged and unlogged coupes 22 years after logging contrasts with findings for flying mammalian insectivores (bats) on the same study sites (Law and Chidel 2001): bat activity in regrowth coupes was less than half that recorded in unlogged coupes. Local landscape effects The changing nature of the bird assemblages in the study area, especially on the unlogged coupes, suggests that either there has been a significant turnover of species irrespective of the logging treatments, or that intensive logging 22 years previously may have caused a loss of species and a decline in populations of some bird species from the unlogged coupes, as well as on the logged coupes. Both scenarios are impossible to verify because pre-treatment sampling was not undertaken in this study. However, the first-time occurrences and the increasing abundances of several species (such as the Yellow-tailed Black Cockatoo, White-bellied Cuckooshrike, Crested Shrike-tit, and Red-browed Treecreeper) on the unlogged coupes in 1998 suggest that the second scenario is more plausible: that logging may have had a broader impact than just on the logged coupes. Other species (such as the Cicadabird, Superb Lyrebird, and Crescent Honeyeater) are now more abundant in the study area on both logged and unlogged coupes than previously. The general impression is that the unlogged coupes are now capable of supporting more birds because the surrounding forest regrowth on the logged coupes is contributing additional habitat for them. This is likely to be significant for species with large home ranges, particularly for those with home ranges larger than the size of the unlogged coupes. Although Williams et al. (2001) also noted a change in bird community structure on their unlogged block immediately after logging, the likelihood that intensive logging may have longer-term effects on birds at the local landscape scale has not been reported previously. Species that are characteristic of open areas (e.g. Australian Pipit) and those that rapidly colonise recently disturbed forests (e.g. Flame Robin, Jacky Winter, Dusky Woodswallow) have almost disappeared from the study area. One small, tree-hollow-nesting bird, the Striated Pardalote, may have declined due to likely increased competition from other species for the few remaining tree hollows that are available in this study area. In comparison, the closely related Spotted Pardalote, a ground-burrow nester, is common on both the logged and the unlogged coupes. Nectar-feeding birds were abundant on both the logged and the unlogged coupes during This increase was associated with a major flowering event for White Stringybark and probably not with forest recovery in the study area. Among the increasers listed above, only the Crescent Honeyeater is a nectarivore. The possibility that 1998 was a good year for birds, and thus not related to forest recovery per se, was examined by inspection of the six-month average (April September) Southern Oscillation Indices (SOI) for the three sampling years 1980, 1989 and 1998 ( In each year, the indices were close to average, suggesting that unusual weather conditions were unlikely to be a factor in the observed changes in bird assemblages. Management implications The effects of intensive logging for most common birds dissipates within 22 years in the dry open-forests of southeastern New South Wales, provided that older-aged forest is present nearby. However, complete recovery of all bird species had not occurred within this time frame. This may be due to the absence of old, hollow trees on all logged areas in this study, despite the presence of unlogged forest adjacent to

10 230 Emu R. P. Kavanagh and M. A. Stanton the logged coupes. It is now well recognised that species recovery will be enhanced if some old trees are left on logged coupes at the time of logging (e.g. Smith 1985). Thus, our results may represent an extreme case because the logging prescriptions applied in 1976 were different to those applied routinely today. Logging prescriptions in New South Wales now require a number of old trees to be retained at the time of logging in all areas logged and the retention of unlogged forest along all drainage lines (Anon. 1999). These and other changes to management procedures appear to have significantly improved recovery of bird populations following logging (RPK, unpublished data). However, without the presence of some old trees nearby, recovery for some species might not be expected for at least 165 years when suitable tree hollows begin to form (Mackowski 1984; Wormington and Lamb 1999; Lindenmayer et al. 2000). A similar conclusion was reached by Loyn (1980). The influence of other possible effects of logging on the birds, such as changes to fecundity, are unknown. This study was unable to conclude anything about the effects of logging for a large number of species that were uncommon in the study area. This is not unexpected (see Williams et al. 2001) since animal count data are typically highly skewed such that, at any one location, relatively few species make up the bulk of the counts and many other species are recorded much less frequently. Nonetheless, further work is required to assess the sensitivity to logging for these species. This study found evidence to suggest that, initially, a wider effect of logging occurred than that which occurred on logged areas. The addition of new species to the unlogged areas as the surrounding forest regenerated is in accordance with current thinking about the importance of the quality of habitat for species in the surrounding matrix (Laurance 1994; Loyn 1998). This highlights the need for managers to undertake conservation planning at the local landscape and broader regional scales. Unless unlogged or older-aged forest is distributed throughout the landscape, and in close proximity to logged areas (which is now the case in New South Wales), much longer rotation lengths will be required. Acknowledgments We thank Stephen Debus, Dick Turner, Graham Turner and Rick Webster, who formed part of our bird-census team during 1989 and/or Wyn Jones, Jim Shields and Martin Schulz helped to collect the data in Traecey Brassil advised on running the PATN analysis programme. Andrew Hayward is thanked for helpful discussion about methods for estimating the power of non-parametric statistical tests. Brad Law, Mike Bullen, Bob Bridges, Steve Dodds, Vic Jurskis and two anonymous reviewers are thanked for their comments on the manuscript. References Anon. (1999). Terms of Licence under the Threatened Species Conservation Act 1995 Eden. Licence conditions for forestry operations in the Eden Regional Forest Agreement (see Integrated Forestry Operations Approvals on pdf/eden_tsl.pdf.) Belbin, L. (1995). PATN. Pattern Analysis Package. (CSIRO Wildlife and Ecology: Canberra.) Christidis, L., and Boles, W. E. (1994). The Taxonomy and Species of Birds of Australia and its Territories. Monograph No. 2. (Royal Australasian Ornithologists Union: Melbourne.) Clarke, K. R. (1993). Non-parametric multivariate analyses of changes in community structure. Australian Journal of Ecology 18, Ford, H. A., Noske, S., and Bridges, L. (1986). Foraging of birds in eucalypt woodland in north-eastern New South Wales. Emu 86, Kavanagh, R. P., Shields, J. M., Recher, H. F., and Rohan-Jones, W. G. (1985). Bird populations of a logged and unlogged forest mosaic at Eden, N.S.W. In Birds of Eucalypt Forests and Woodlands: Ecology, Conservation, Management. (Eds A. Keast, H. F. Recher, H. Ford and D. Saunders.) pp (Royal Australasian Ornithologists Union & Surrey Beatty: Sydney.) Kutt, A. S. (1996). Bird population density in thinned, unthinned and old lowland regrowth forest, East Gippsland, Victoria. Emu 96, Laurance, W. F. (1994). Rainforest fragmentation and the structure of small mammal communities in tropical Queensland. Biological Conservation 69, Law, B. S., and Chidel, M. (2001). Bat activity 22 years after first-round intensive logging of alternate coupes near Eden, New South Wales. Australian Forestry 64, Lindenmayer, D. B., Cunningham, R. B., Pope, M. L., Gibbons, P., and Donnelly, C. F. (2000). Cavity sizes and types in Australian eucalypts from wet and dry forest types a simple rule of thumb for estimating size and number of cavities. Forest Ecology and Management 137, Loyn, R. H. (1980). Bird populations in a mixed eucalypt forest used for production of wood in Gippsland, Victoria. Emu 80, Loyn, R. H. (1985). Bird population in successional forests of Mountain Ash Eucalyptus regnans in central Victoria. Emu 85, Loyn, R. H. (1998). Birds in patches of old-growth ash forest, in a matrix of younger forest. Pacific Conservation Biology 4, Mackowski, C. M. (1984). The ontogeny of hollows in Blackbutt (Eucalyptus pilularis) and its relevance to the management of forests for possums, gliders and timber. In Possums and Gliders. (Eds A. P. Smith and I. D. Hume.) pp (Surrey Beatty & the Australian Mammal Society: Sydney.) MathSoft (1999). S-PLUS 2000 Professional Edition for Windows, Release 2. (MathSoft Inc.: Seattle, WA.) Recher, H. F., Rohan-Jones, W., and Smith, P. (1980). Effects of the Eden Woodchip Industry on terrestrial vertebrates with recommendations for management. Research Note No. 42, Forestry Commission of New South Wales, Sydney. Recher, H. F., Holmes, R. T., Schulz, M., Shields, J., and Kavanagh, R. (1985). Foraging patterns of breeding birds in eucalypt forest and woodland of southeastern Australia. Australian Journal of Ecology 10, Recher, H. F., Kavanagh, R. P., Shields, J. M., and Lind, P. (1991). Ecological association of bird species and habitats during the breeding season in southeastern New South Wales. Australian Journal of Ecology 16,

11 Bird population recovery after logging Emu 231 Smith, P. (1985). Effects of intensive logging on birds in eucalypt forest near Bega, New South Wales. Emu 85, Stone, C. (1999). Assessment and monitoring of decline and dieback of forest eucalypts in relation to ecologically sustainable forest management: a review with a case study. Australian Forestry 62, Taylor, R. J., and Haseler, M. E. (1995). Effects of partial logging systems on bird assemblages in Tasmania. Forest Ecology and Management 72, Taylor, R., Duckworth, P., Johns, T., and Warren, B. (1997). Succession in bird assemblages over a seven year period in regrowth dry sclerophyll forest in south-east Tasmania. Emu 97, Williams, M. R., Abbott, I., Liddelow, G. L., Vellios, C., Wheeler, I. B., and Mellican, A. E. (2001). Recovery of bird populations after clearfelling of tall open eucalypt forest in Western Australia. Journal of Applied Ecology 38, Wormington, K., and Lamb, D. (1999). Tree hollow development in wet and dry sclerophyll eucalypt forest in south-east Queensland, Australia. Australian Forestry 62, Manuscript received 17 January 2002; accepted 12 May