1
Increasing biodiversity in urban green spaces through simple vegetation interventions
1
Caragh G. Threlfall
a*
caragh.threlfall@unimelb.edu.au Phone: +61 4437493519
2
Luis Mata
b
luis.mata@rmit.edu.au
3
Jessica Anne Mackie
c
jessicaanne.mackie@gmail.com
4
Amy K. Hahs
d
hahsa@unimelb.edu.au
5
Nigel E. Stork
c
nigel.stork@griffith.edu.au
6
Nicholas S.G. Williams
a,d
nsw@unimelb.edu.au
7
Stephen J. Livesley
a
sjlive@unimelb.edu.au
8
a
School of Ecosystem and Forest Sciences, The University of Melbourne, 500 Yarra Boulevard,
9
Richmond, Victoria 3121, Australia
10
b
Interdisciplinary Conservation Science Research Group, School of Global, Urban and Social
11
Studies, RMIT University, 124 La Trobe Street, Melbourne, Victoria 3000, Australia
12
c
Environmental Futures Research Institute, Griffith School of Environment, Griffith University, 170
13
Kessels Rd, Nathan, Queensland 4111, Australia
14
d
Australian Research Centre for Urban Ecology, Royal Botanic Gardens Victoria c/o The School of
15
Biosciences, The University of Melbourne, Parkville, Victoria 3010, Australia
16
17
*Corresponding author
18
Running title: Urban green space biodiversity
19
20
21
2
Summary
22
1. Cities are rapidly expanding worldwide and there is an increasing urgency to protect urban
23
biodiversity, principally through the provision of suitable habitat, most of which is in urban
24
green spaces. Despite this, clear guidelines of how to reverse biodiversity loss or increase it
25
within a given urban green space is lacking.
26
2. We examined the taxa- and species-specific responses of five taxonomically and functionally
27
diverse animal groups to three key attributes of urban green space vegetation that drive
28
habitat quality and can be manipulated over time: the density of large native trees, volume of
29
understorey vegetation and percentage of native vegetation.
30
3. Using multi-species occupancy-detection models we found marked differences in the effect of
31
these vegetation attributes on bats, birds, bees, beetles and bugs. At the taxa-level,
32
increasing the volume of understorey vegetation and percentage of native vegetation had
33
uniformly positive effects. We found 30–120% higher occupancy for bats, native birds,
34
beetles and bugs with an increase in understorey volume from 10 to 30%, and 10–140%
35
higher occupancy across all native taxa with an increase in the proportion of native vegetation
36
from 10 to 30%. However, increasing the density of large native trees had a mostly neutral
37
effect. At the species-specific level, the majority of native species responded strongly and
38
positively to increasing understorey volume and native vegetation, whereas exotic bird
39
species had a neutral response.
40
4. Synthesis and applications. We found the probability of occupancy of most species examined
41
was substantially reduced in urban green spaces with sparse understorey vegetation and few
42
native plants. Our findings provide evidence that increasing understorey cover and native
43
plantings in urban green spaces can improve biodiversity outcomes. Redressing the
44
dominance of simplified and exotic vegetation present in urban landscapes with an increase
45
in understorey vegetation volume and percentage of native vegetation will benefit a broad
46
array of biodiversity.
47
48
Key-words: biodiversity loss; green infrastructure; green space vegetation; large old trees; multi-
49
species site occupancy-detection models; native vegetation; understorey vegetation; urban
50
biodiversity; urban ecosystems; vegetation management
51
3
Introduction
52
Biodiversity in cities is under increasing threat (Aronson, La Sorte, Nilon et al. 2014) from habitat loss,
53
the introduction of competitive or predatory exotic species, climate change and ecosystem
54
degradation (Grimm, Faeth, Golubiewski et al. 2008). By 2030, urban land cover is expected to triple
55
what is was in 2000, leading to loss of locally significant or threatened habitats, especially in highly
56
biodiverse regions of the world (Seto, Güneralp & Hutyra 2012). Devising ways to protect and
57
increase biodiversity in cities, whilst allowing for continued urban growth is critical as nearly 60% of
58
required urban land cover is yet to be built (Fragkias, Güneralp, Seto et al. 2013).
59
60
The retention of native vegetation is an effective strategy to conserve urban biodiversity (Hahs,
61
McDonnell, McCarthy et al. 2009; Aronson, La Sorte, Nilon et al. 2014). A recent meta-analysis of
62
factors influencing urban biodiversity (Beninde, Veith & Hochkirch 2015) concluded that remnant
63
vegetation patches had to be >50 ha to retain threatened or urban-sensitive species. Additionally, city
64
governments worldwide are planting vegetation in streets, parks, gardens and on roofs to help offset
65
the negative impacts of increased urban density (Tzoulas, Korpela, Venn et al. 2007). Collectively,
66
protecting existing and planting new native vegetation has great potential to curb urban biodiversity
67
loss. However, despite the costs of managing urban green spaces and the widespread planting of
68
new vegetation, little is known about which vegetation attributes provide quality habitat for a wide
69
range of taxa (Lin & Fuller 2013).
70
71
Urban green spaces often contain simplified habitats that lack large hollow-bearing trees,
72
decomposing logs or native ground and mid storey vegetation (Le Roux, Ikin, Lindenmayer et al.
73
2014; Threlfall, Ossola, Hahs et al. 2016). Urban green space managers need information on how to
74
better manage vegetation to retain complex habitats but to date, research has focussed on habitat
75
quality for birds (Chace & Walsh 2006; Beninde, Veith & Hochkirch 2015). The loss of large, native
76
trees (Chace & Walsh 2006; Stagoll, Lindenmayer, Knight et al. 2012) and declines in habitat
77
complexity (Evans, Newson & Gaston 2009) have a negative impact on urban bird communities but
78
little is known about the impacts on other taxa (Beninde, Veith & Hochkirch 2015).
79
80
4
We examine the extent to which vegetation attributes in urban green spaces influence the taxa- and
81
species-specific responses of insectivorous bats (Suborder Microchiroptera), bees (Order
82
Hymenoptera; Superfamily Apoidea), nine beetle families (Order Coleoptera), heteropteran bugs
83
(Order Hemiptera; Suborder Heteroptera) and birds (Subclass Neornithes). We focus on three key
84
vegetation attributes known to influence fauna habitat quality; 1) large native tree density; 2)
85
understorey volume; and 3) percent of native vegetation (Chace & Walsh 2006; Stagoll, Lindenmayer,
86
Knight et al. 2012; Beninde, Veith & Hochkirch 2015). Our previous research suggests these variables
87
are important drivers of species richness and community composition between green space types for
88
bees (Threlfall, Walker, Williams et al. 2015), bats and birds (Threlfall, Williams, Hahs et al. 2016) and
89
bugs (Mata, Threlfall, Williams et al. in press). However, aggregate community measures such as
90
species richness may not adequately account for species-specific responses to vegetation quality.
91
Here, we use multi-species site occupancy-detection models to provide a more nuanced
92
understanding of urban biodiversity patterns by accommodating both taxa-level and species-specific
93
responses. We use this approach to identify beneficial management actions to support a wide range
94
of taxonomically and functionally diverse native biodiversity.
95
96
MATERIALS AND METHODS
97
Study design and location
98
The study was conducted in Melbourne, Australia’s second most populous city (4 million people). A
99
large proportion of metropolitan Melbourne is covered in low density single storey detached houses. It
100
also has a relatively high density of public parks and nearly 50 km
2
of golf courses. Differences in soil
101
type, rainfall, topography and native vegetation were minimised by restricting the study area to the
102
Gippsland Plain bioregion in south-eastern Melbourne. This bioregion also occupies a significant
103
portion of metropolitan Melbourne.
104
105
We sampled biodiversity within the three dominant urban green space habitats in south-east
106
Melbourne (Figure 1): golf courses, public parks and residential neighbourhoods. These habitats have
107
the greatest scope for management intervention to improve urban biodiversity. We did not sample
108
remnant habitats as these are not a dominant or evenly distributed feature in the study area. We
109
identified triplets of green spaces (golf course, public park and residential neighbourhood) that were
110
5
developed in the same decade and suburb by examining historical aerial imagery, and consulting
111
municipal land release and construction records. This allowed us to standardise green space age,
112
location and type (See Appendix A1 for a description of each site). We excluded sites that contained,
113
or were within one kilometre of natural waterways. We selected 39 green spaces for sampling (13 golf
114
courses, 13 public parks and 13 residential neighbourhoods) to maximise variation in vegetation
115
structure and composition. Our focus was to sample from a range of green spaces that differed
116
substantially in relation to our three key vegetation attributes, rather than to focus on a comparison of
117
green space type per se.
118
119
Survey methods
120
Urban green space vegetation
121
Within our 39 green spaces, we randomly established 247 plots to measure vegetation variables: 104
122
in golf courses, 104 in residential neighbourhoods and 39 in small urban parks. A minimum of two
123
plots were established in green space sites <5 ha in size. Two additional plots were established for
124
every 5 ha increase in green space size, up to a maximum of eight plots in large residential
125
neighbourhoods or golf courses (for details of plot placements see Appendix A2). In golf courses and
126
urban parks plots were 20 x 30 m in size (600m
2
). To represent residential neighbourhoods, plots
127
consisted of the front garden, the pavement and road verge (if present) out to a midway point of the
128
road directly in front of the property. The width and depth of each front garden primarily dictated the
129
size of each plot in the residential neighbourhoods (ranging from 211 to 870 m
2
) as we could only
130
sample properties for which we had permission. To characterise the effect of management practices
131
on vegetation within each green space plot, we measured: 1) the density of all trees; 2) volume of
132
ground- and mid-storey vegetation, hereafter referred to as understorey vegetation and 3) vegetation
133
composition (native and exotic). Vegetation at the plot level was used to calculate the average for that
134
green space site via averaging the values recorded for plots within each green space (n=39).
135
136
Density of trees
137
For each tree (defined as having a diameter >8 cm) within a plot we measured stem diameter at
138
breast height (DBH) at 1.3 m above ground level. We then categorised each species as being native
139
or exotic to Australia. Preliminary analysis suggested the density of all native trees per hectare had a
140