Introduction
As urbanised land area expands across the globe and human populations within them grow (Seto et al. 2012), natural habitats are becoming increasingly fragmented and degraded (Lambin et al. 2001; McKinney 2006; Burgin et al. 2016). While this has undoubtedly impacted the sustainability of both terrestrial (McKinney 2002) and water dependent (Hill et al. 2015, 2017) biodiversity, there is growing evidence that some wildlife including mammals, reptiles, birds and aquatic invertebrates can persist in both remnant patches of natural habitat (water, vegetation etc.) and highly modified habitats within the urban landscape (Koenig et al. 2001; Stokeld et al. 2014; Hill et al. 2017; Cooper et al. 2020; Van Helden et al. 2020a; 2020b). The capacity of these habitats to support animal wildlife, mostly due to species’ behavioural flexibility (Sol et al. 2013) and ability to exploit the novel resources on offer (Van Helden et al. 2021a; 2021b), has fuelled belief that urban landscapes, and particularly residential gardens, present a significant conservation opportunity for some wildlife (Goddard et al. 2010; Aronson et al. 2017; Soanes and Lentini 2019; Soanes et al. 2019).
One approach to maximise the conservation potential of residential gardens is through wildlife friendly gardening (WFG). Wildlife friendly gardening includes the provision of appropriate food, shelter and water resources in an attempt to support the needs of animal wildlife (Goddard et al. 2010). Associations between the presence of particular garden features (e.g., shelter, food and a lack of natural predators) and animals in gardens provide promising evidence that the manipulation of garden features could offer conservation benefit (e.g., Baker et al. 2003; Daniels and Kirkpatrick 2006; Fontaine et al. 2016; Threlfall et al. 2016; Van Helden et al. 2020b). As knowledge of the factors that influence wildlife presence in urban landscapes has increased, so too has the interest in and uptake of WFG (Gaston et al. 2007; Davies et al. 2009; Cox and Gaston 2018).
Advertisement
This growing interest in WFG is supported by strong promotional programs (see Goddard et al. 2010; Larson et al. 2022) that advocate for the protection, enhancement and addition of wildlife resources in gardens, including the supplementation of water sources. Supplementary provision of water within the residential landscape may particularly benefit conservation, as urban wetlands, lakes and rivers are often only seasonally available to animals (Gaston et al. 2005b), have been strongly altered from their natural state (e.g., Walsh et al. 2005), and the landscape-scale connectivity among them has been reduced by urban land use change (Burgin et al. 2016). In contrast, water sources in gardens are often permanently watered, can be numerous (Gaston et al. 2005b; Davies et al. 2009) and implemented by a substantial proportion of residents (Gaston 2007; Fardell et al. 2022). In the United Kingdom for example, there are an estimated 2.5–3.5 million private garden ponds (Davies et al. 2009) and multiple wildlife conservation programs encourage the public to make water available in their gardens as a means of supporting animal wildlife, especially through periods of hot, dry weather (e.g., RSPB 2021; The Wildlife Trusts 2021).
Animals use water for a multitude of reasons, including habitat, drinking, bathing and reproduction. As human populations within urban areas grow, natural aquatic habitats are increasingly becoming threatened (Walsh et al. 2005; Dudgeon et al. 2006; Vörösmarty et al. 2010; Reid et al. 2019), and their capacity to support biodiversity across multiple functional levels is reduced (Reid et al. 2019). In many regions, climate drying and warming compound these threats, further altering the capacity of natural habitats to support wildlife (Woodward et al. 2010). In areas experiencing both climate warming/drying and urbanisation, novel water sources including natural and constructed water bodies in public spaces and residential gardens have been identified as potentially important for supporting animal wildlife (Chester and Robson 2013; Cleary et al. 2016; Hill et al. 2017). While public blue spaces within urban areas are known to support a variety of animals (Lynn et al. 2006; Hill et al. 2017; Oertli and Parris 2019; Xie et al. 2022), knowledge of the potential value of garden water sources for particular animal taxa is comparatively scarce. The few studies that have explored the use of garden water sources by animal wildlife have concentrated on the use of baths by birds (although with some notable exceptions, see Beebee 1979; Gaston et al. 2005a). These studies demonstrate that a variety of both small and larger-bodied native birds utilise watered bird baths (e.g., Gehlbach 2012; Miller et al. 2015; Cleary et al. 2016), especially those that have adapted to human-dominated habitats (Cleary et al. 2016), and that supplementary water can enhance the value of other wildlife friendly gardening actions such as native plantings (Coetzee et al. 2018) or bird feeding stations (Miller et al. 2015).
While these studies make good progress in evaluating the importance of bird baths, they do not consider the use of other water sources, such as ground water-bowls and ponds, nor the broader range of taxa known to inhabit residential gardens, including mammals (Van Helden et al. 2020b), insects (Jones and Leather 2012), amphibians (Beebee 1979) and reptiles (Koenig et al. 2001). Indeed, evidence that the number and variety of species present in gardens increases with water availability (Fardell et al. 2022) and that garden water sources are used by amphibians (Beebee 1979) and small mammals (Miller et al. 2015), suggests that supplementary water in gardens may benefit a broader array of animal wildlife than is already known. To ensure we are best providing for animal wildlife in the face of climate change in an increasingly urbanised world, it is important to better understand which species use garden water sources, the frequency of use, how these differ among water source types and whether garden water sources support different wildlife compared to larger public blue spaces; one urban water source for which biodiversity value has already been established (Hill et al. 2015, 2017; Chester and Robinson 2013).
We addressed these knowledge gaps using (i) an online questionnaire to establish the prevalence and variety of water sources provided within gardens and (ii) animal wildlife surveys to determine species richness, wildlife visitation, and species assemblages in gardens and small urban lakes in Hertfordshire in South-East England. This study provides evidence of the potential conservation value of garden water sources, and highlights future research required to fully explore and understand their potential use, conservation benefit and application in broader wildlife conservation strategies.
Advertisement
Materials and methods
Study area
The study was conducted within a 20 km radius of St Albans city, Hertfordshire, South-East England, United Kingdom (Fig. 1). Hertfordshire covers 1643 km2, has a total population of approximately 1.2 million people and a population density of 728 people km− 2 (2020 estimates, ONS 2022). The study area included St Albans and nearby townships of Hemel Hempstead, Hatfield and Harpenden. St Albans supports a population of approximately 149,000 people within an area of 161 km2 (population density = 926 people km− 2) (2020 estimates, ONS 2022).
South-East England is one of the driest and warmest regions in the United Kingdom (Met Office 2016a, b). Both air temperature and precipitation vary seasonally; temperature is highest in July (mean maximum temperature = 22.1 °C) and lowest in January (mean maximum temperature = 7.1 °C), and rainfall is lowest in March (mean total rainfall = 42.7 mm) and highest in October (mean rainfall = 78.2 mm) (Rothamsted climate station; period of record 1991–2020; Met Office 2022a). The United Kingdom is predicted to experience significant climate drying and warming in summer, with rainfall declining by up to 47% and air temperature increasing by up to 5.4 °C by 2070 (Met Office 2022b).
Fig. 1
Location of residential gardens (red dot •) and urban lakes (blue dot •) within Hertfordshire County (---), United Kingdom, where wildlife surveys were undertaken during June and July 2021. Shading on the main map indicates a 20 km radius from St Albans within which all survey sites and questionnaire respondents were located. Inset shows location of Hertfordshire (red dashed line ---) north of London (•) within the United Kingdom. Photos illustrate examples of garden water sources including (a) a bird bath, (b) a pond, (c) a ground water-bowl and urban lakes (d-f) where wildlife surveys were undertaken
×
![]()
Questionnaire
An online questionnaire (JISC software; https://www.jisc.ac.uk/about) was developed as part of a broader study (EKG unpublished data) to determine the prevalence and variety of garden water sources, how residents supplemented or modified these sources (i.e., addition of running water, animals and plants), the bird, mammal, amphibian and insect wildlife observed at garden water sources and the willingness of residents to participate in the wildlife observational study. We asked residents to report the number and dimensions of any ‘pond’ (water body constructed at or close to ground level), ‘ground water-bowl’ (bowl of water at ground level), ‘bird bath’ (bowl of water raised above ground level), ‘water feature’ (sculptural or artistic structure with flowing or standing water) and ‘other water source’ (any standing water source not previously defined) that was present in their garden. The survey was circulated via St Albans wildlife groups (Wilderhood Watch) and St Albans Facebook community pages (All Things St Albans, St Albans Eco, St Albans Friends of the Earth), for a period of 8 weeks between 14 June and 10 August 2021. Our release strategy intentionally targeted the ‘conservation-minded’ demographic to increase survey responses from residents who provided supplementary water in their gardens. Given the targeted release strategy, the proportion of respondents that provided supplementary water may not be representative of the broader demographic within the study area.
Wildlife surveys
We undertook timed animal wildlife (hereafter ‘wildlife’) surveys at 13 garden sites (selected from a total of 34 respondents indicating their willingness to undertake wildlife surveys) and six urban lakes within Hertfordshire (N = 19, Fig. 1). Garden sites were selected to include gardens with a pond (N = 5), ground water-bowl (N = 4) or bird bath (N = 4). Urban lakes (size range 200 m2 to 10,000 m2, N = 6) were selected to represent the variety of available urban waterbodies (small, large, natural or constructed). Each lake contained a variety of habitats including open water, reed beds, mud banks and overhanging vegetation. All urban lakes were located within 200 m of a garden site to reduce the potential for species’ geographical ranges to influence differences in taxa visitation among sites.
Wildlife surveys in residential gardens were undertaken by citizen scientists (i.e., the residents) and at urban lakes by an ecological researcher (EKG). At each site, we aimed to complete three, one-hour observations each week for four weeks in June-July 2021 (total = 12 observation hours/site). Citizen scientists were advised to undertake observations at one of their pre-existing garden water sources (bird bath, pond or ground water-bowl) for all observation periods. We were unable to account for variability in the number or type of water sources, additional WFG actions, or garden characteristics that may have also influenced wildlife observations at these sites. Wildlife surveys were undertaken at varying times during daylight hours between 0500 and 2100 h. To standardise methods employed, all citizen scientists were sent an information pack, pre-recorded video and an invitation to a live question and answer session to ensure their understanding of how to complete the wildlife survey. These resources also contained images of wildlife species likely to be encountered by participants during wildlife surveys to increase their accuracy of species identification (Mason and Arathi 2019).
During each observation period, all visits to the focal water source were recorded for birds, mammals, insects, amphibians and reptiles. A visit was defined as an individual interacting directly with the water (i.e., drinking, bathing, resting etc.). Consecutive visits by the same individual within 5 min were recorded as a single visit. Citizen scientists were instructed to observe their water source from approximately 5 m away to minimise disturbance to wildlife whilst allowing accurate identification and recording of species as small as a hoverfly (Syrphidae). At urban lakes, we restricted wildlife observations to a 200 m2 area to standardise wildlife detectability at different sized lakes. The observed area at each urban lake was representative of the diversity and abundance of available habitats. The observer was positioned less than 5 m from the water’s edge with unobstructed vision of the study area using binoculars.
Data analysis
From the full questionnaire developed as part of a broader study, we isolated information on the number, variety and dimensions of water sources, and any modifications that had been made to them, in respondents’ gardens. Any responses that were incomplete or from residences outside the study area (Fig. 1) were removed from analyses. Differences in the volume of each identified water source type was tested using a one-way analysis of variance with post-hoc Tukey multiple comparisons of means using a Bonferroni adjustment on log transformed water volumes. All analyses of questionnaire responses were performed in RStudio, R Statistical Software (V4.1.2; R Core Team 2021) and the significance level was set at alpha = 0.05.
Citizen-derived data on wildlife visitations to garden water sources were screened for accuracy and consistency among individual citizen scientists. As all citizen-derived wildlife survey data were based on focal observation, species identification could not be verified. We accepted most species data, because citizens have been shown to accurately identify morphospecies when given training (Mason and Arathi 2019). To increase reliability of data we made the following adjustments to some species records used for analysis. Records of resident species (e.g., pond snails Lymnaea stagnalis) or any taxa smaller than a hoverfly were omitted from analyses. Due to differences in the use of scientific and common names among citizen scientists, visitation records were recoded to higher taxonomic levels for some birds (e.g., ‘pigeon’ and ‘sparrow’) and some insects (e.g., ‘bees and wasps’, ‘hoverflies’, ‘damselflies’, ‘dragonflies’, and ‘butterflies and moths’). These taxonomic groups were subsequently applied to the visitation data recorded at urban lakes also.
Taxa richness and visitations were standardised by time and expressed as taxa hr− 1 and visits hr− 1 respectively. We also divided the full visitation data matrix (‘all wildlife’) into three data matrices to include ‘small wildlife’, ‘small birds’ and ‘insects’. For each of these subset data matrices, visitation rates of large water dependent birds (e.g., species of geese and ducks) were omitted on the basis that these taxa were unlikely to be able to access garden water sources due to their larger body size. ‘Small wildlife’ therefore included visitation rates for all taxa (except large birds), and ‘small birds’ and ‘insects’ only included visitation rates for small birds and insects respectively. Where possible, we compared wildlife visitation between existing public blue spaces (lakes or wetlands) and gardens with water sources (i.e., between ‘locations’), and between individual ‘water source types’ (i.e., urban lakes, bird baths, ground water-bowls and ponds) due to fundamental differences in size, depth, and position (e.g., elevated versus on ground).
We used a two-sample t-test with equal variance to test for differences in taxa richness and visitation rates among ‘locations’ and a one-way analysis of variance with post-hoc Tukey multiple comparisons of means using a Bonferroni adjustment to test for difference among water source types. For visitation rate, we ran these analyses on ‘all wildlife’ and ‘small wildlife’ data matrices. Both these analyses were undertaken in RStudio, R Statistical Software (V4.1.2; R Core Team 2021) on log-transformed data that met the assumptions of both data normality (Shapiro-Wilk’s p > 0.05) and homogeneity of variance (p > 0.05).
Differences in assemblages of ‘all wildlife’, ‘small wildlife’, ‘small birds’ and ‘insects’ among ‘locations’ and ‘water source types’ were examined using non-metric multidimensional scaling (nMDS) based on Bray Curtis similarity matrices of square-root transformed standardised visitation rates for all taxa in each assemblage type (Primer V7; Clarke and Gorley, 2015). The Analysis of Similarity (ANOSIM) routine in Primer V7 was used to determine significant differences in species assemblages between ‘locations’ and ‘water source types’, and the Similarity Percentages routine (SIMPER) was used to identify the species contributing most (> 5% contribution) to any significant dissimilarities among these groups.
Results
Questionnaire responses
Of the 105 questionnaire responses, 75 indicated the presence of at least one pre-existing water source in their garden. Of a total of 122 supplementary water sources reported, bird baths (n = 38), groundwater bowls (n = 35) and ponds (n = 33) were more common than garden water features (n = 14) and ‘other’ water source types (n = 2) (χ23, N=122 = 11.8, p = 0.008). The total number of all water sources in the gardens varied from one to six (mean = 1.16 ± 0.1 SE) and the proportion of gardens with one (n = 40) and more than one (n = 35) water source was similar. The total volume of supplementary water provided in gardens ranged from 0.45 to 726,000 L (mean = 12071.94 L ± 10072.01 L, N = 72) and the volume of each water source type (not including ‘other’) varied (F3, 99 = 75, p < 0.001); all source types differed from one another (all p < 0.01) except for ground water-bowls and bird baths. The mean volume of bird baths was 12.26 L (± 2.83 L, n = 37), garden water features was 1226.76 L (± 1026.93 L, = 13), ground water-bowls was 7.53 L (± 2.83 L, n = 34) and ponds was 27497.98 L (± 17107.22 L, n = 31).
Modifications were made by some residents to all water source types, except for ground water-bowls. The addition of running water was common for garden water features (78.57% of features; N = 14), although it was also added to some bird baths (7.89%; N = 38) and ponds (21.2%; N = 33). Plants were commonly added to ponds (90.91%) and garden water features (28.57%) whereas animals were only added to ponds (27.27%). Fish (goldfish Carassius auratus, koi carp Cyprinus rubrofuscus and chub Leuciscus cephalus) and frogs (including adults, tadpoles or spawn) were most commonly added to ponds (18.18% and 15.15% of ponds respectively), although newts (Lissotriton vulgaris) and snails (Gastropoda) had also been added to one pond each.
Wildlife surveys
A total of 207 h of observation were undertaken, including 135 h in 12 gardens (mean = 11.25 h site− 1 ± 1.194 SE) and 72 h at six lakes (mean = 12 h site− 1). Bird baths and ground water-bowls were each observed for a total of 40 h and ponds for a total of 55 h (Table 1).
Across all sites we recorded a total of 43 taxa including 31 birds, 7 insects, 2 mammals, 2 amphibians and 1 reptile (Table 1). Of these, five are known exotic species, and two taxonomic groups potentially include exotic species (Table 1). Twenty-seven taxa were recorded in garden water sources; 16 birds, 7 insects, 2 mammal and 2 amphibian taxa. Thirty taxa were recorded at lakes; 23 birds, 4 insects and 1 mammal, amphibian and reptile. Approximately 50% of the species recorded in each location were unique to that location (gardens = 48.15%, N = 27; urban lakes = 53.33%, N = 30). Mean standardised taxa richness (species hr− 1) recorded at lakes was 0.89 species hr− 1 ± 0.08 SE (N = 6), at ponds was 0.52 species hr− 1 ± 0.07 SE (N = 5), at bird baths was 0.95 species hr− 1 ± 0.20 SE (N = 4) and at ground water-bowls was 0.64 species hr− 1 ± 0.12 SE (N = 4). Standardised taxa richness did not vary between locations (mean = 0.69 ± 0.09 SE, N = 13) (t17 = -1.770, p = 0.095) nor among water source types (all pairwise comparisons p > 0.05).
A total of 3505 animal visitations to water sources were recorded during the 207 h of observation. Of these, 2727 visitations (77.80%) were recorded at lakes and 778 visitations (22.20%) were recorded at garden water sources (Table 1). Mean visitation rate of ‘all wildlife’ was significantly higher at lakes (37.88 visits hr− 1± 12.66 SE) compared to gardens with water sources (5.32 visits hr− 1± 1.34 SE) (t17 = -4.791, p < 0.001) and to each garden water source type (F3, 15 = 8.445, p = 0.002; all pairwise comparisons p < 0.05) but did not vary among garden water source types (Fig. 2). Mean visitation rates for ‘small wildlife’ did not vary between locations (lakes: 6.43 visits hr− 1± 1.95 SE; gardens: 37.88 visits hr− 1± 12.66 SE) (t17 = 0.016, p = 0.988) or among water source types (F3, 15 = 0.475, p = 0.705) (Fig. 2).
Table 1
Total visitation rates (visits hr-1) by each species to each water source type recorded during wildlife surveys undertaken in June and July 2021 in Hertfordshire, United Kingdom. The mean hours of observation for each source type was: bird bath, 10.00 h ± 2.915 SE (n = 4); ground water-bowl, 10 h ± 1.000 SE (n = 4); pond, 11 h ± 0.77 SE (n = 5); lake, 12 h (n = 6)
Taxa | Bird bath | Ground water-bowl | Pond | Lake | Total |
|---|---|---|---|---|---|
Birds
| |||||
Black-headed gull (Chroicocephalus ridibundus) | 1.500 | 0.522 | |||
Blackbird (Turdus merula) | 0.350 | 0.425 | 0.109 | 0.083 | 0.208 |
Blackcap (Sylvia atricapilla) | 0.014 | 0.005 | |||
Blue tit (Cyanistes caeruleus) | 0.650 | 0.100 | 0.042 | 0.159 | |
Canada goose (Branta canadensis) * | 13.319 | 4.633 | |||
Chaffinch (Fringilla coelebs) | 0.625 | 0.121 | |||
Collared dove (Streptopelia decaocto) * | 0.025 | 0.018 | 0.010 | ||
Coot (Fulica atra) | 3.167 | 1.101 | |||
Dunnock (Prunella modularis) | 0.025 | 0.125 | 0.029 | ||
Goldfinch (Carduelis carduelis) | 0.350 | 0.200 | 0.055 | 0.014 | 0.126 |
Great crested grebe (Podiceps cristatus) | 0.014 | 0.005 | |||
Great tit (Parus major) | 0.350 | 0.056 | 0.087 | ||
Green woodpecker (Picus viridis) | 0.025 | 0.005 | |||
Greenfinch (Chloris chloris) | 0.075 | 0.014 | |||
Grey wagtail (Motacilla cinerea) | 0.111 | 0.039 | |||
Herring gull (Larus argentatus) | 0.111 | 0.039 | |||
House martin (Delichon urbicum) | 1.319 | 0.459 | |||
Jackdaw (Coloeus monedula) | 0.025 | 0.005 | |||
Jay (Garrulus glandarius) | 0.050 | 0.010 | |||
Little grebe (Tachybaptus ruficollis) | 1.083 | 0.377 | |||
Magpie (Pica pica) | 0.075 | 0.025 | 0.014 | 0.024 | |
Mallard (Anas platyrhynchos) | 7.181 | 2.498 | |||
Moorhen (Gallinula chloropus) | 2.639 | 0.918 | |||
Mute swan (Cygnus olor) | 0.319 | 0.111 | |||
Pigeon (native and feral) # | 1.175 | 0.300 | 0.473 | 0.278 | 0.507 |
Pochard (Arythya farina) | 1.194 | 0.415 | |||
Robin (Erithacus rubecula) | 0.625 | 0.225 | 0.036 | 0.069 | 0.198 |
Sparrow (Passer sp.) | 0.325 | 0.200 | 0.083 | 0.130 | |
Starling (Sturnus vulgaris) | 1.350 | 1.850 | 0.109 | 0.647 | |
Swift (Apus apus) | 0.028 | 0.010 | |||
Tufted duck (Aythya fuligula) | 0.917 | 0.319 | |||
Insects
| |||||
Bee/wasp (various spp.) # | 0.125 | 0.091 | 0.048 | ||
Bumblebee (Bombus) | 0.050 | 0.050 | 0.091 | 0.042 | 0.058 |
Butterfly/moth (various spp.) | 0.275 | 0.181 | 0.116 | ||
Damselfly (Zygoptera) | 1.600 | 3.181 | 1.531 | ||
Dragonfly (Anisoptera) | 0.036 | 0.417 | 0.155 | ||
Hoverfly (Syrphidae) | 2.018 | 0.536 | |||
Longhorn beetle (Rutpela maculata) | 0.018 | 0.005 | |||
Amphibian
| |||||
Common frog (Rana temporaria) | 0.073 | 0.019 | |||
Common newt (Lissotriton vulgaris) | 2.000 | 0.264 | 0.623 | ||
Mammal
| |||||
Brown rat (Rattus norvegicus) * | 0.075 | 0.222 | 0.092 | ||
Grey squirrel (Sciurus carolinensis) * | 0.025 | 0.050 | 0.014 | ||
Reptile
| |||||
Terrapin (Emydidae) * | 0.014 | 0.005 | |||
Grand Total
|
6.575
|
3.625
|
6.727
|
37.875
|
16.932
|
Fig. 2
Mean standardised visitation rate (visits hr-1) of all wildlife (grey bar) and ‘small wildlife’ (white bar) to water source types recorded during wildlife surveys undertaken in June and July 2021 in Hertfordshire, United Kingdom. Whiskers represent SE
×
![]()
Species assemblages at gardens and lakes
Multivariate analyses indicated no difference in insect assemblages between locations (Table 2; Fig. 3d), nor among the ‘small bird’ assemblages visiting each of the water source types (Table 2; Fig. 3c). In contrast, assemblages of ‘all wildlife’, ‘small wildlife’ and ‘small birds’ varied among locations, and assemblages of ‘all wildlife’, ‘small wildlife’ and ‘insects’ differed among water source types (Tables 2, Fig. 3a-d). For these latter wildlife groups, we detected no significant differences among source types in gardens (bird bath, ground water-bowl and pond); only differences between garden water sources and urban lakes.
The difference in the assemblage of ‘all wildlife’ among locations and among water source types (Table 2; Fig. 3a) was driven mostly by the absence of large aquatic birds visiting garden water sources, and the higher visitation rates of damselfly to urban lakes compared to garden water sources (Table 1). The difference in the assemblage of ‘small wildlife’ (Fig. 3b) among locations was influenced most by the visitation of house martin (Delichon urbicum) and black-headed gull (Chroicocephalus ridibundus) to urban lakes only, starlings (Sturnus vulgaris) to gardens only, higher visitation rates of insects (dragonfly and damselfly) to urban lakes and higher visitation rates of pigeons to gardens (Table 1). Differences in visitation of ‘small wildlife’ to urban lakes compared to bird baths and ground water-bowls (Fig. 3b, d) were influenced most by the variations in the visitation of starling, damselfly, black-headed gull, house martin, dragonfly, pigeon, robin and blackbird (Tables 1 and 2).
Differences in the ‘small bird’ assemblages between locations (Fig. 3c) were influenced most by the visitation of house martin and black-headed gull to urban lakes only, the visitation of starlings to gardens only and the higher visitation rates of all other bird species (contributing > 5% to the dissimilarity) to gardens (Table 1). For ‘insects’, differences in visitation to urban lakes and bird baths (Table 2; Fig. 3c) were influenced most by visitation of damselfly and dragonfly to lakes, the visitation of bees/wasps to bird baths only and the higher visitation rate of butterflies/moths to bird baths (Table 1). Differences in visitation of insects to urban lakes and ground water-bowls were influenced mostly by the visitation of damselfly and dragonfly to urban lakes only and the higher visitation rate of bumblebee to groundwater bowls (Table 1).
Fig. 3
Non-metric multidimensional scaling (nMDS) plot showing variation in assemblages of (a) ‘all wildlife’ (b) ‘small wildlife’ (c) ‘small birds’ and (d) ‘insects’ between types of water sources in gardens (ground water-bowl (○), bird bath (∆) and pond (□)) and urban lake (•)
×
![]()
Table 2
Results of multivariate analysis of taxa assemblages (visits hr− 1) recorded at garden water sources and urban lakes during wildlife surveys undertaken in June and July 2021 in Hertfordshire, United Kingdom. R statistics, p-values, % dissimilarity and taxa contributing to dissimilarity among groups (> 5% contribution up to 80% dissimilarity) is shown for (A) locations (gardens versus lakes) and (B) all water source types, and for four assemblage subgroups (i) all wildlife, (ii) small wildlife, (iii) insects and (iv) small birds. Only significant pairwise comparisons are shown for significant global tests. Abbreviation GWB=ground water-bowl
Wildlife | R statistic (p value) | Dissimilarity (%) | Contributing taxa (%) |
|---|---|---|---|
A. Garden versus Lake
| |||
i) All wildlife
|
0.554 (0.001)
| 89.57 | Canada goose 12.49; mallard 10.47; damselfly 8.28; moorhen 7.26; coot 6.54 |
ii) Small wildlife
|
0.363 (0.005)
| 84.57 | Damselfly 15.64; black-headed gull 9.68; house martin 8.04; dragonfly 6.28; pigeon 6.22; starling 5.78 |
iii) Insects
|
0.139 (0.087)
| N/A | N/A |
iv) Small birds
|
0.452 (0.006)
| 81.61 | House martin 14.79; black-headed gull 13.27 pigeon 11.55; starling 9.69; blackbird 8.52; robin 7.29; sparrow 6.08 |
B. Water Source Types
| |||
i) All wildlife
|
Global: 0.510 (0.001)
| ||
Bird bath/lake: 0.750 (0.001) | 90.48 | Canada goose 11.63; mallard 9.74; damselfly 8.33; moorhen 6.75; coot 6.15 | |
GWB/lake: 0.806 (0.005) | 91.76 | Canada goose 12.76; mallard 10.71; damselfly 9.35; moorhen 7.44; coot 6.69 | |
Pond/lake: 0.525 (0.009) | 87.09 | Canada goose 12.97; mallard 10.87; moorhen 7.55; damselfly 7.33; coot 6.79; hoverfly 5.49 | |
ii) Small wildlife
|
Global: 0.342 (0.009)
| ||
Bird bath/lake: 0.520 (0.014) | 87.01 | Damselfly 14.17; starling 9.09; black-headed gull 8.09; pigeon 6.91; house martin 6.88; dragonfly 5.64; robin 5.11 | |
GWB/lake: 0.548 (0.014) | 88.03 | Damselfly 17.33; black-headed gull 10.32; house martin 8.45; dragonfly 6.89; starling 6.50; blackbird 6.23 | |
iii) Insects
|
Global: 0.548 (0.003)
| ||
Bird bath/lake 0.651 (0.036) | 91.87 | Damselfly 37.01; bee/wasp 21.14; butterfly/moth 18.29; dragonfly 16.70 | |
GWB/lake 0.719 (0.036) | 94.7 | Damselfly 45.92; dragonfly 21.54; bumblebee 17.36. | |
iv) Small birds
|
Global: 0.032 (0.346)
| N/A | N/A |
Discussion
Aquatic ecosystems are experiencing unprecedented pressure from land and climate change and their value in providing ecosystem and biodiversity services is becoming increasingly compromised (Dudgeon et al. 2006; Vörösmarty et al. 2010; Reid et al. 2019). The provision of supplementary water as a wildlife friendly garden feature may play an important conservation role through increasing the availability of water as well as the connectivity among watered habitats in regions experiencing climate warming and drying, and urban expansion. Beyond the already known prevalence of residential bird baths (Gaston et al. 2007) and ponds (Davies et al. 2009), our study indicates the likely prevalence of other water source types including ground water-bowls and water features that may also provide resources to wildlife. Our wildlife surveys demonstrated that while large-bodied water birds more frequently visited urban lakes, garden water sources were visited by small animal wildlife as frequently as urban lakes, but often by a different assemblage of species. Our results suggest that garden water sources can increase water dependant biodiversity across the urbanised landscape and, that for small-bodied taxa at least, garden water can supplement existing urban blue spaces, and support the diversity of species that use them. Together, our results contribute new evidence that supports the conservation potential of garden water sources.
The provision of supplementary water in gardens has become a popular and strongly promoted form of WFG. It represents an important proportion of WFG actions undertaken by residents (Gaston et al. 2005b, 2007) and has resulted in the establishment of significant water resources across some urban landscapes (Davies et al. 2009). In the United Kingdom for example, up to 3.5 million ponds are estimated to contribute around 349 additional hectares of water surface area across the residential landscape (Davies et al. 2009). Our results suggest these figures likely underestimate the availability of supplementary garden water sources, as we found at least 50% of residents with a water source had more than one source, and residents provided water source types in addition to ponds, including bird baths, ground water-bowls, and water features that have not been incorporated into existing estimates. Given that the engagement in WFG, including water supplementation, varies demographically (Gaston et al. 2007), and the fact that our results are from a survey that targeted a relatively small cohort of conservation-minded participants, investigation of how much additional water may be provided through other water source types warrants further consideration.
The value of supplementary water provision lies in its ability to support wildlife, both locally at the point of source (e.g., Gaston et al. 2005a; Gehlbach 2012; Miller et al. 2015; Cleary et al. 2016), and more broadly through providing connectivity among the variety of water sources that occur across the urban landscape (sensu Goddard et al. 2010). Interestingly, despite strong advocacy for the provision of garden water sources (Goddard et al. 2010; Larson et al. 2022), and a growing body of evidence for the conservation value of garden habitats for terrestrial fauna (see for example Goddard et al. 2010; Maclagan et al. 2018; Van Helden et al. 2020b; 2021b; Gazzard et al. 2022), there remains a scarcity of quantitative evidence of the conservation value of garden water sources (Gaston et al. 2007; Davies et al. 2009). Until now, most studies have explored the use of bird baths by birds (Gehlbach 2012; Miller et al. 2015; Cleary et al. 2016), although some have also looked at the biodiversity value of ponds (Gaston et al. 2005a) or incidentally recorded the use of garden water sources by other taxa (Beebee 1979; Miller et al. 2015). Our data show that a variety of garden water sources are used by multiple taxa including birds, insects, amphibians, reptiles and mammals and that species richness and visitations are not influenced by source type, and also do not vary between gardens and lakes for small-bodied taxa at least. Our results suggest that garden water sources can elevate water dependant biodiversity across the urbanised landscape and provide additional resources to the suite of wildlife already occurring in established blue spaces.
Interestingly, respondents to our questionnaire reported making modifications to the water sources within their gardens including the addition of plants, running water and a variety of animal life. Habitat diversity and complexity strongly influences aquatic community structure across all levels of organisation (Soukup et al. 2022), and indirectly communities of terrestrial fauna that rely on aquatic food sources (e.g., Dahlin et al. 2021). Although it remains unexplored in the broader literature, whether modification of garden water sources provide additional biodiversity value is a worthy line of future research. Given the importance of behavioural flexibility to wildlife’s capacity to exploit novel resources within residential gardens (Sol et al. 2013; Van Helden et al. 2021b), understanding how different wildlife taxa utilise garden water sources may also provide insight into their value, as well as their design.
Despite significant growth in urban wildlife research over the last two decades, there remains a focus on few ‘fields of study’ and taxa, predominantly birds and mammals (Magle et al. 2012; Collins et al. 2021). In our study, we focussed on a seldom explored field of aquatic urban wildlife research and attempted to broaden the taxonomic focus to include ‘all wildlife’. Due to difficulties in accessing privately owned land, our approach was to engage citizen scientists for monitoring garden water sources, as their involvement has proven beneficial across a diversity of ecological research programs (Gallo-Cajiao et al. 2018; Paloniemi et al. 2018; Steven et al. 2019, 2021; Lloyd et al. 2020). We see several components of our approach that could be improved in future studies, specifically our reliance on observational wildlife surveys, undertaken by citizen scientists in one season, only during daylight hours and in a relatively small sample of garden and urban wetland sites. Despite residents being supplied with training materials and support, we found variation in survey ability among residents particularly for small, cryptic species. In relation to our study, this affected recording of some animal species (some birds and some insects) that we combined into higher taxonomic classifications due to inconsistencies in nomenclature and species discrimination. While focal observations can provide quantitative information on some taxa (e.g., birds), this approach is unlikely to provide comprehensive data on semi aquatic and aquatic fauna. Wildlife such as these often require sampling techniques other than observation, including animal handling rendering their utility in citizen-based sampling programs problematic due to animal welfare issues. While our focus on visitations of wildlife to water sources during summer, when wildlife demand for water is greatest, is ecologically relevant, exploring seasonal variability in the use of garden water sources (e.g., Van Helden et al. 2021a) would provide a more comprehensive understanding of their biodiversity value. All wildlife surveys were conducted during daylight hours and therefore potentially missed other nocturnal and crepuscular species. Future research could utilise additional sampling methods, such as motion triggered cameras, that have proven beneficial to other garden-based wildlife surveys (Van Helden et al. 2020b) and are suitable for deployment by citizen scientists.
Conclusions
This study represents one of very few quantitative studies demonstrating anthropogenic water sources in domestic gardens are utilised by a variety of animal taxa. We demonstrate that these water sources can both increase biodiversity within urban landscapes and provide additional water sources to support the suite of taxa already utilising urban blue spaces. Our results provide an important baseline on which future investigations can expand examination of the biodiversity value of garden water sources for all animal wildlife, how the use of these water sources varies temporally, the array of behaviours and requirements they support, and how they interact with other garden and landscape scale components of urban conservation strategies.
Acknowledgements
We thank residents of Hertfordshire for their enthusiastic participation in our research as citizen scientists. We thank Wilderhood Watch, All Things St Albans, St Albans Eco and St Albans Friends of the Earth community groups for promoting our project, particularly the questionnaire to the residents of Hertfordshire. We thank Dr Erik Kiviat and an additional anonymous reviewer for their helpful suggestions on an earlier version of this manuscript.
Declarations
Ethics approval
This research was approved by the Human Ethics Committees of Bristol University and The University of Western Australia (2021/ET000595), and the Animal Ethics Committees of Bristol University (UIN/21/020) and The University of Western Australia (F18979).
Consent to participate
Participation of citizen scientists in this research was voluntary. Participation was supported by a Participant Information Form for the questionnaire and training information and opportunities for the wildlife survey. Completion of each component was considered evidence of consent to participate.
Competing interests
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.