Integrating biodiversity considerations into urban golf courses: managers’ perceptions and woody plant diversity in the Eastern Cape, South Africa

ABSTRACT

Rapid rates of urbanisation affect biodiversity through habitat fragmentation and loss. Because urban golf courses are large green spaces, they potentially harbour much biodiversity if managed for such. The area of untransformed land of golf courses in the Eastern Cape (South Africa) was determined using Geographic Information System (GIS), the woody plant composition of a subsample was determined by field sampling and the greenkeepers were interviewed. There was a significant relationship between climate and woody cover, species richness and percentage native plants but not species diversity. There was no relationship between management scores and species richness, diversity and percentage native. Significant relationships were evident between woody plant species richness and course income, number of grounds staff, number of club members and greenkeepers’ years of experience. These results suggest a complex suite of factors that play a role in the woody plant composition of urban golf courses and their contribution to urban biodiversity.

KEYWORDS:

• Biodiversity
• greenkeeper
• green space
• management
• out of play
• urbanisation

1. Introduction

Urbanisation poses many challenges for the natural environment and biodiversity, mostly through land use change, habitat fragmentation and reconfiguration, and pollution of land, water and air (Jeon, Olofsson, & Woodcock, 2014; Scolozzi & Geneletti, 2012; Tzoulas et al., 2007). Habitat fragmentation in urban areas is often viewed as causing declines in species richness, abundance and overall biodiversity of urban areas; much of this is due to the loss of natural vegetation (McKinney, 2002; Van Der Veken, Verheyen, & Hermy, 2004). Yet, there is increasing evidence that many towns and cities harbour remarkable species richness, and that with appropriate consciousness, policies and actions, the diversity and abundance of species in urban areas can be improved (Puppim de Oliveira et al., 2011), including for endemic species (Hodgkison, Warnken, & Hero, 2007a). Thus, new perspectives are emerging regarding the possibility of ecosystem service generation from urban green infrastructure, and that land sharing as opposed to sparing (Grau, Kuemmerle, & Macchi, 2013) is possible for a range of urban land uses. An important strategy to achieve this goal is the provision of sufficient and networked urban green spaces (Alvey, 2006).

The amount, nature and configuration of urban green spaces vary between urban areas and countries (Dobbs, Nitschke, & Kendal, 2014). However, it is a common lament that there is usually insufficient green space for human needs and the average size is too small to foster sufficient biodiversity (e.g. Hepcan, 2013). Yet, most of this commentary relates to public urban green spaces, and the contribution of private urban green spaces, that is, gardens, has been overlooked. This is despite the bulk of urban green space being under private tenure (González-García & Gomez-Sal, 2008; Shackleton et al., 2017) and often harbouring high levels of biodiversity (Goddard, Dougill, & Benton, 2010; Loram, Tratalos, Warren, & Gaston, 2007), both native and non-native species. Nonetheless, most private gardens are relatively small and in many regions are enclosed, which limits the movement of species and hence the abundance and types of species supported (Goddard et al., 2010). Certain species, such as birds and small mammals, may also be at risk from the depredations of domestic pets (Baker, Molony, Stone, Cuthill, & Harris, 2008).

In contrast, most urban golf courses do not suffer from these drawbacks. First, they are large, typically over 50 ha (Terman, 1997), and so can potentially harbour species and support ecosystem processes at a larger scale. Second, they rarely suffer the same level of negative impacts from domestic pets. Third, because they exist to serve paying members, they generally have far higher financial and human resources to manage the green space than do most public green spaces. Fourth, most golf courses have a large proportion (30–70%) of their land under relatively minimal management intervention; these are the out of play areas (Tanner & Gange, 2005). Fifth, their attractiveness and use can be buffers to their transformation as urban areas densify and so they may be relatively long-lived (Vaz, Caetano, & Nijkamp, 2011) and thereby harbour particularly large trees. Last, although there are already between 25,000 and 31,500 (Hudson & Bird, 2009; Tanner & Gange, 2005) golf courses globally, the number is increasing rapidly (Colding & Folke, 2009; Hudson & Bird, 2009), which presents an opportunity to proactively influence their design, construction and management to better promote biodiversity. The management of protected areas is expensive, and many governments cannot afford to manage them effectively; therefore, within the urban boundaries, willing stewards have a potentially substantial role in the conservation of biodiversity (Colding, Lundberg, & Folke, 2006). Golf course managers could become such stewards; all costs would be at the expense of the golf courses, but they too would enjoy the benefits of a more environmentally friendly course while increasing biodiversity and ecosystem service provision for the greater urban area. Examples include the use of golf courses for recreation and exercise by walkers and bird watchers, active introductions or management of a variety of wildlife species (Terman, 1997), carbon sequestration and ameliorating the urban heat island effect, in other words, ecosystem services bundles.

Given this combination of features, it is understandable that urban golf courses are receiving increasing attention regarding their potential for land sharing to simultaneously provide high-value recreational spaces for golfers and contribute to biodiversity conservation and ecosystem services in urban settings. Often viewed as sites of habitat and biodiversity loss when created (Jeon et al., 2014), with strategic management it is possible for golf courses to attract more anthrophobic species (species that are poorly adapted to human disturbance) and hence add to urban biodiversity conservation (Colding & Folke, 2009). This is not to ignore that some golf courses may also engage in actions that have environmentally negative impacts, such as widespread use of chemicals, high water use and introduction of alien species (Wheeler & Nauright, 2006). Nonetheless, from a perspective of maintaining and promoting biodiversity in urban settings they have considerable potential. For example, Tanner and Gange (2005) surveyed nine golf courses in Surry (UK) and reported higher abundance and richness of birds, bees, beetles and trees than neighbouring farmlands; there was no difference in the diversity of herbaceous plant species. Such differences are likely to be even starker in urban areas. Hodgkison et al. (2007a) revealed the value of golf courses in providing habitat to threatened species in Queensland, Australia.

To date, all but one case study regarding the biodiversity of urban golf courses have been situated in developed world contexts, mostly North America, Europe and Australia. The meta-analysis review by Colding and Folke (2009) identified only one case study from a developing country. Whether or not such a macro-level differentiation will result in different dynamics around urban golf courses and their ability to foster biodiversity is currently unknown, but merits consideration. On the surface there are a number of differences between developed and developing world contexts that influence urban greening (Shackleton, 2012), some of which are likely to be applicable to golf courses, including higher biodiversity, demand for provisioning services (such as firewood, traditional medicines and fruits) from urban green areas, lower information and knowledge about the importance of urban biodiversity, limited financial resources to ameliorate the effects of harsh climates and the potential for damage by urban livestock (Shackleton, 2012). Thus, it is necessary that consideration of the contributions of golf courses to urban biodiversity and ecosystem services also be examined in developing world contexts.

At a more localised scale, two additional key factors are likely to have considerable influence on the biodiversity of golf courses. These are the policy framework of the golf club and the management actions on the course (Hammond & Hudson, 2007). The position that intersects these two is the role of the greenkeeper or grounds manager. Thus, the knowledge and attitude of the greenkeeper about biodiversity can potentially promote or constrain the biodiversity outcomes with respect to species richness, diversity, structure and supportive ecological processes. For example, Hodgkison, Hero, and Warnken (2007b) and Burgin and Wotherspoon (2009) showed how, in Australia, appropriate management could promote biodiversity without constraining the needs of the golf players, through activities such as amphibian-friendly fringes around water features, leaving dead wood in situ and increasing the mowing height in the rough area. Yet, few of the studies on golf course biodiversity have considered the greenkeepers’ interest and actions pertaining to biodiversity. Thus, what is the relative contribution of management actions and perceptions to the biodiversity status of urban golf courses? The exception is the work of Hammond and Hudson (2007), who showed that the majority of golf course managers surveyed in the UK agreed that golf courses were important areas for wildlife and 43% had sections about wildlife in the management plan for their course.

Within the context mentioned earlier, we sought to ascertain the potential contribution of urban golf courses in the Eastern Cape province, South Africa, to urban biodiversity conservation. The corresponding objectives were to (1) determine the size and cover of urban golf courses in the province, (2) assess their woody plant diversity, (3) determine greenkeepers’ perceptions of biodiversity conservation and management on golf courses and (4) investigate links between management perceptions, actions and woody plant structure and diversity.

2. Study area

This study took place in the Eastern Cape, South Africa, home to approximately 6.5 million people, and increasing at approximately 0.4% per year over the last decade. Just over one-third (36.6%) are urban dwellers (StatsSA (Statistics South Africa), 2011). The unemployment rate in 2011 was 37.5% (StatsSA (Statistics South Africa), 2011). The Eastern Cape includes seven out of the nine biomes of South Africa with at least 6164 different plant species, of which 316 are threatened (CSIR, 2004). The study sites fall within the forest, Albany thicket, Nama karoo and grassland biomes.

The first phase of the study consisted of a GIS analysis of 43 urban golf courses in the Eastern Cape province. The second phase involved using a subsample of 12 urban golf courses for vegetation surveys. The criteria for selection of the subsample were that they be situated within two broad belts from the coast inland (Figure 1), so to include a range of bioclimatic settings in the study. The climatic conditions vary with increasing distance from the ocean, with the coastal areas experiencing mild, temperate conditions, whereas the inland areas experience harsher conditions (hotter and drier, but with lower winter temperatures) and more seasonal growing conditions. The altitude, size, mean annual rainfall and mean high and low temperatures for each site are included in Table 1.

Table 1. Attributes of the 12 sites for woody plant determination (SA Explorer, 2016; World Weather Online, 2016).

Golf course location	Altitude (m)	Size (ha)	Mean rainfall/annum (mm)	Mean high and low temperatures (°C)
Port Alfred	21	62.0	485	23.8; 12.6
East London	47	82.6	593	24.8; 14.3
Alexandria	213	26.6	529	23.1; 12.3
Alicedale	283	55.2	334	24.0; 7.8
King William’s Town	382	63.5	502	23.0; 11.4
Adelaide	592	29.0	369	22.8; 9.9
Bedford	730	24.8	390	23.3; 9.3
Stutterheim	844	33.5	598	21.2; 10.8
Cradock	886	31.4	248	28.2; 10.1
Queenstown	1090	99.2	399	23.3; 9.3
Cathcart	1198	33.2	454	21.2; 10.8
Dordrecht	1609	33.4	466	23.4; 6.8

Figure 1. Location of study sites along two belts, each from inland to coast.

3. Methodology

This study made use of both quantitative and qualitative data collected from fieldwork as well as interviews with the greenkeepers. Prior to work in the field, a GIS analysis was done on 43 Eastern Cape urban golf courses identified using online inventories and snowball sampling. This sample is likely to include more than 90% of urban golf courses in the province. A subsample of 12 of these courses was chosen for further investigation along the two broad belts depicted in Figure 1. Fieldwork consisted of assessing the percentage cover of each woody species within predetermined random plots so that woody plant species diversity could be calculated. Interviews were conducted with the greenkeepers at the 12 courses to determine their perceptions of biodiversity conservation on urban golf courses.

3.1. Data collection

The ground cover of all 43 Eastern Cape urban golf courses was assessed using aerial photographs and ArcGIS. Polygons were created around the area of each of the golf course and 50 m × 50 m grids were created within each polygon. Forty per cent of the blocks were randomly selected and analysed for six ground cover classes: rough (area outside of greens and fairways), woody vegetation, greens (greens and fairways), water (water features on the course), infrastructure and cement/roads. The size of all 43 golf courses was determined as well as the mean percentage cover of the ground cover classes.

Field data collection was done at a subsample of 12 golf courses to assess the woody plant species cover of 10 plots per golf course. Most golf courses are shaped in a way that they have woody vegetation in the rough surrounding the fairways and greens, but they may also have thin strips of woody vegetation between fairways. Therefore, 7 out of the 10 plots were randomly placed in the rough and the remaining 3 were randomly placed between fairways. Random points were generated by Google Earth Pro. The plots in the rough were 10 m × 10 m and the plots between the fairways were 5 m × 100 m. The reason for the plots differing in size was that the wooded areas between the fairways tend to be long and relatively narrow (often a single row of trees). In each plot, each woody plant was identified (in the field if known or a voucher specimen was collected for identification) and the total aerial cover was visually estimated for each woody species as a percentage of the total plot area, as was the total cover across all woody plant species. Woody plant species richness (count of species), percentage cover of each species, mean percentage cover for rough and fairway plots, species diversity for rough and fairway plots and percentage of native woody plants were determined for each golf course. Species diversity was calculated using Simpson’s Diversity Index (1–D) where D = ∑ (n/N)² with n = cover of woody species x and N = total cover of all woody species (Simpson, 1949).

Direct interviews were conducted with the head greenkeepers of each of the 12 golf courses. Interviews consisted of both open- and closed-ended questions. The line of inquiry was aimed towards how greenkeepers perceive the role of urban golf courses in biodiversity conservation. Hence topics included greenkeepers’ understanding of biodiversity, opinions on biodiversity conservation on urban golf courses, perceived importance of high plant diversity, biodiversity conservation targets/objectives, access to biodiversity management information, management practices, training, education and experience (Appendix 1). Interviews were very conversational ranging between 1 and 3 h, with some greenkeepers taking the researchers on a tour around parts of the golf course to point out certain features, illustrate some of their answers or ask questions of the researchers. The list of possible biodiversity friendly practices was from Hammond and Hudson (2007). Rating scores were (such as the influence of club members on course management) self-ratings by the respondents. For those greenkeepers who stated that they had limited understanding of the term biodiversity, we explained it as the number of different species, their relative abundance in a place (such as a golf course) and the interactions between them that support the variety of life. All interviews were transcribed, and permission to use responses was given by interviewees.

3.2. Data analysis

All analyses were done using Statistica 13. Simple linear regressions were done to determine whether there were significant relationships between altitude and species richness, species diversity, percentage ground cover and percentage of native vegetation. Altitude was used as a quantitative proxy for climate because climates at low-altitude sites near the coast are warmer, moister, frost-free, with a longer growing season and overall more tropical than the inland, higher altitude sites, with a continuum between the opposite ends.

Interview data included both quantitative and qualitative data. Quantitative data were in the form of management scores, which were determined from a list of management practices for which interviewees had to rank the frequency at which they applied each practice (Table 2). Two scoring systems were used, an overall management score and a native species management score. The overall management score used the seven management practices in Table 2 to obtain a score out of 28 for each course based on the ranking given by greenkeepers. The native-friendly management score used only the last three management practices from Table 2 to get a score out of 12. Simple linear regressions were run to determine whether there were significant relationships between management scores, derived from a list of management practices said to promote biodiversity, and species richness, percentage native woody plants and species diversity. Multiple regression analyses were also done where the dependent variables were species richness, species diversity, percentage native woody plants and mean percentage cover, and the predictor variables were overall management score, native-friendly management score, annual fees paid by golf club members, number of golf club members, number of grounds staff and years of experience in green keeping.

Table 2. Management practices that are regarded to promote woody plant biodiversity (Hammond & Hudson, 2007). Greenkeepers were asked to rank the degree to which each is implemented on their golf course.

Practices	Rank each practice on scale of 1 to 4:
Leave large natural wooded areas undisturbed	1 = never
Maintain undergrowth and leaf litter in wooded areas	2 = seldom
Designate ‘no mow areas’ in out-of-play areas	3 = often
Leave old/dead trees in situ	4 = always
Removal/reduction of non-native tree species
Alien invasive species control
Use native plants when planting

4. Results

4.1. Vegetative cover of Eastern Cape urban golf courses

The mean and median size of 43 Eastern Cape urban golf courses were 34.0 ± 25.53 ha and 25.0 ha, respectively. The total area of all golf courses combined was 1461 ha. Six ground cover classes were observed: greens, rough, woody, infrastructure, water and cement/road with most (52%) being under rough and greens (34%) (Figure 2). The mean percentage cover of rough and woody classes combined contributed 66.6% of the vegetation cover. Therefore, assuming that rough and woody ground cover classes offer high potential for biodiversity on golf courses, there is a large proportion of Eastern Cape urban golf courses that could potentially be managed for biodiversity conservation.

Figure 2. Mean proportions (±SD) of six different ground cover classes on 43 Eastern Cape golf courses representative of 40% of the total size.

4.2. Woody plant diversity

Woody plant species richness ranged between 8 and 28 and the percentage of native woody plants ranged from 0% to 86% (Table 3). Between the 12 golf courses, 104 woody plant species were recorded; 61 were native to South Africa and 43 were non-native, of which some were declared invasive alien species (Table 4). It was noted that all courses had at least one declared invasive woody plant species (some had as many as six), which by South African legislation should be removed (Department of Environmental Affairs, 2014).

Table 3. Woody plant diversity status of the 12 golf courses (FW = plots between fairways; Plot = plots in the rough; D = Simpson’s Diversity Index (1-D).

Golf course	Altitude (m)	Species richness	Mean FW cover (%)	Mean plot cover (%)	FW D	Plot D	Native (%)
Port Alfred	21	25	133.0	123.4	0.877	0.822	80
East London	47	21	64.3	103.4	0.807	0.741	76
Alexandria	213	28	83.3	145.5	0.703	0.877	71
Alicedale	283	14	67.3	80.3	0.651	0.704	86
King William’s Town	382	22	49.5	63.4	0.656	0.871	64
Adelaide	592	7	74.0	69.2	0.553	0.629	43
Bedford	730	21	64.7	83.1	0.489	0.892	62
Stutterheim	844	13	54.7	49.4	0.568	0.504	23
Cradock	886	15	61.0	69.7	0.777	0.834	47
Queenstown	1090	11	36.0	46.1	0.553	0.824	18
Cathcart	1198	7	74.0	63.0	0.597	0.591	0
Dordrecht	1609	8	35.7	58.2	0.741	0.590	25

Table 4. Woody plant species found at each golf course (indigenous species are in bold).

Species	Port Alfred	East London	Alexan-dria	Alice-dale	King William’s Town	Adelaide	Bedford	Stutter-heim	Cradock	Queens-town	Cath-cart	Dord-recht
Acacia dealbata											x
Acacia longifolia			x
Acacia mearnsii							x	x			x
Acacia saligna								x			x
Allophylus natalensis	x	x
Azima tetracantha				x
Brachychiton populneus						x
Brachylaena discolor	x	x
Brachylaena elliptica							x
Brachylaena transvaalensis		x
Callistemon sp.					x
Cassia sp.												x
Cassuarina cunninghamania		x			x		x	x	x
Carissa bispinosa	x	x					x
Celtis africana			x				x		x
Cestrum laevigatum			x
Chrysanthemoides monilifera			x
Clerodendrum glabrum	x
Coddia rudis	x						x
Colpoon compressum	x
Cordia caffra		x
Corymbia ficifolia			x
Cotoneaster frigidus								x
Cotoneaster sp.	x
Cupressus sp.							x					x
Cussonia spicata					x
Diospyros dichrophylla	x
Diospyros lycioides			x
Dodonaea viscosa	x
Dovyalis caffra	x
Ehretia rigida				x	x
Ekebergia capensis		x	x
Elaeodendron croceum	x
Erythrina caffra			x
Erythrina lysistemon		x
Eucalyptus baueriana										x
Eucalyptus beyeri					x						x
Eucalyptus camaldulensis												x
Eucalyptus cinerea											x	x
Eucalyptus gomphocephala			x
Eucalyptus grandis		x		x	x	x		x		x	x
Eucalyptus maculata												x
Eucalyptus sideroxylon		x
Euclea natalensis	x	x
Euclea racemosa		x
Ficus natalensis		x
Ficus sp.	x		x
Ficus sur		x
Ficus thonningii	x	x
Fraxinus sp.									x
Gleditsia triacanthos										x
Grevillea robusta						x			x
Grewia occidentalis			x				x
Gymnosporia buxifolia		x
Gymnosporia heterophylla				x	x	x	x		x
Gymnosporia polyacantha									x
Harpephyllum caffrum		x
Jacaranda mimmosifolia							x
Lantana camara	x	x	x		x
Leonotis leonurus	x
Ligustrum lucidum								x
Lycium acutifolium				x					x
Lycium sp.				x	x		x
Manilkara sp.			x
Melia azedarach					x		x			x
Mystroxylon aethiopicum	x		x
Nicotiana sp.	x
Olea africana			x	x		x		x
Olea capensis							x		x
Pappea capensis				x
Pinus radiata		x	x			x	x	x		x	x	x
Pittosporum viridiflorum		x
Plumbago auriculata				x	x		x
Podocarpus falcatus					x
Populus canescens										x
Quercus acutissima								x
Quercus palustris								x
Quercus sp.							x		x	x
Quercus suber									x
Rhamnus prinoides			x		x
Rhigozum obovatum			x
Rhoicissus sp.			x		x
Rubus sp.			x
Salix babylonica ‘Crispa’									x
Salix sp.									x	x
Salvia africana-lutea	x
Schinus molle				x					x
Schotia latifolia								x
Scutia myrtina	x		x		x		x
Searsia glauca			x		x
Searsia crenata	x	x			x
Searsia incisa	x		x
Searsia laevigata			x	x	x
Searsia lancea					x				x
Searsia pyroides	x						x			x		x
Searsia refracta				x
Searsia sp.				x			x
Senagalia galpinii												x
Solanum sp.					x
Thuja occidentalis							x
Ulmus pumila								x		x
Vachellia karoo			x	x	x	x	x	x	x	x
Withania somnifera	x
Xanthoxylon capense			x

Significant negative relationships were found between altitude and woody plant species richness (R² = 0.565, p = 0.005), percentage of native woody plants (r² = 0.747, p = 0.0001) and mean percentage cover for both fairway plots (r² = 0.392, p = 0.029) and rough plots (r² = 0.520, p = 0.008). There were no significant relationships between altitude and fairway plot species diversity (r² = 0.140, p = 0.230) or rough plot species diversity (r² = 0.205, p = 0.140) (Figure 3). Altitude was used as a quantitative proxy for climate; therefore, climate had an effect on species richness, percentage of native plants and percentage cover of woody plants, but not on species diversity. Thus, drier and cooler inland sites had fewer species, most of which were non-native and lower woody plant cover relative to coastal sites.

Figure 3. Relationships between four variables: (a) species richness, (b) percentage of native trees, (c) mean percentage cover of fairway and rough, (d) species diversity; and altitude. (FW = plots between fairways; Plot = plots in the rough; D = Simpson’s Diversity Index).

4.3. Greenkeepers’ perceptions of biodiversity conservation on urban golf courses

Half of the greenkeepers self-rated themselves as having none (33%) or little (17%) understanding of the term ‘biodiversity’ and the other half had a moderate (17%) or good (33%) understanding. Irrespective, all of the greenkeepers agreed with the statement ‘Urban golf courses can play a role in biodiversity conservation’. A variety of reasons were provided for their agreement; some were more general, for example, golf courses provided natural beauty, increased habitat leads to increased biodiversity and that they were contained and well maintained areas that act as a ‘parkland in an urban area’. One comment was that golf courses could afford to maintain their area therefore making them a more ‘economical’ way for biodiversity conservation; another comment was that educated management of golf courses can build biodiversity.

The greenkeepers were asked to rank the importance of high plant diversity on golf courses; 59% believed it was important, 33% believed that it was of medium importance and 8% thought it was not important. General reasons why those that found high diversity to be important include the aesthetic value of nature, increased plant diversity could provide golfers with extra challenges and that monoculture is not sustainable for ecosystem dependents (e.g. birds and bees). Reasons for those that felt high plant diversity was of medium or low importance include higher maintenance, water usage and increased obstructions which might affect golfers’ ability to play.

The greenkeepers were also asked to rank the influence of the members of the golf club on the vegetative composition of their course. More than half (58%) said that the member influence was weak, 25% said it was of medium strength and 17% said that members had a strong influence. This shows that woody plant vegetative composition of the golf courses was largely a product of the management actions decided by the greenkeepers.

4.4. Biodiversity conservation targets/objectives

A total of 3 of the 12 golf courses had biodiversity conservation objectives/targets in place for their courses. These objectives/targets all involved the removal of non-native vegetation and increasing the number of native trees on the golf course. One golf course in particular had specific policies against planting non-native tree species, and their objectives included the removal of non-native tree species while inter-planting with native tree species. When asked if they would like to see more being done to promote biodiversity on their golf course, eight greenkeepers responded positively. Their interests were mainly centred on including more vegetation that is native and introducing wildlife to their courses, but these examples were followed by reasons why such interests are difficult to realise. Challenges included lack of finances, limited availability of water and the need for ‘more committed management’. Adelaide’s golf course was a good example of how challenges prohibit management for biodiversity; there is little access to water for irrigation as well as livestock grazing and browsing on the course inhibiting the survival and growth of newly planted trees. The other four greenkeepers who responded negatively did so because they were satisfied with the current biodiversity status of their golf courses.

4.5. Training and accessibility of information on biodiversity management

One-third of the greenkeepers had received formal training in green keeping in the form of Technicon education on turf grass management or short courses on topics such as pesticide/fungicide use. It is noteworthy that none of the formal training included biodiversity teachings. The other two-thirds had not received any formal training; their knowledge was in the form of knowledge and experience from their predecessors, self-taught knowledge and farming knowledge.

Opinions on the accessibility of information on good biodiversity management practices for golf courses were almost equally split, seven thought that such information was easily accessible and five thought that it was not. All greenkeepers admitted that such information was available to anyone who specifically sought it out. One greenkeeper commented that small town golf courses seldom consider biodiversity in greenkeeping. Another view was that of the available information very little is in relation to managing golf courses for biodiversity conservation but instead its focus is more on management of the greens with lessons mostly based on golf courses in the US or Europe. The most common ways in which the greenkeepers from all 12 courses obtained management information was through books, golfing magazines, the Internet and through networking and sharing information with other greenkeepers. A few of the greenkeepers made mention of different greenkeeper associations where they were able to network with other greenkeepers but one interviewee felt that biodiversity management was ‘not actively promoted in greenkeeping associations’.

4.6. Relationships between golf course management and current diversity status

Regression analyses between management scores and species richness, percentage native woody plants and diversity indices of both fairway and rough plots (Figure 4) revealed that there were no significant relationships. This indicates that the different management practices assessed had little effect on percentage of native species, species richness or species diversity. The multiple regression analyses did not reveal any significant relationships between species diversity or percentage cover and several possible predictor variables (Table 5). There were significant relationships between species richness and most of the predictor variables (r² = 0.995, F _{8, 3} = 38.83, p < 0.01) (Table 4). All but one of the predictor variables showed a significant relationship to species richness, that was the overall management score (p = 0.431).

Table 5. Predictor variables, p-values and significance for regression analysis of woody plant species richness.

Predictor variables	p-Value	Significant
Percentage native	0.003	Yes
Overall management score	0.387	No
Native-friendly management score	0.012	Yes
Annual membership fees	0.018	Yes
Number of members	0.008	Yes
Number of grounds staff	0.006	Yes
Years of green keeping experience	0.006	Yes

Figure 4. Relationships between three variables: (a) species richness, (b) percentage of native trees, (c) species diversity; and management scores. FW = plots between fairways; Plot = plots in the rough; D = Simpson’s Diversity Index.

5. Discussion

Using the example of urban golf courses, this work illustrates a number of challenges and opportunities for land use science (Müller & Munroe, 2014) to assess and promote multifunctional parcels of land within urban settings. Urban land use dynamics and how to maintain or even promote biodiversity and the supply of ecosystem services in towns and cities are pressing questions globally, but which have been particularly neglected in sub-Saharan Africa. As such, this study touches on core elements of the Global Land Programme (Verburg et al., 2015), such as land use management in urban settings, and the assumed or real land use trade-offs. The notion of incorporating other land use objectives and ecosystem service benefits from urban golf courses, currently managed solely for recreational services to urban golfers, offers new dimensions for the land-sharing debates through identifying and managing bundles of ecosystem services compatible with the primary land use, i.e. golf.

5.1. Potential for biodiversity conservation on Eastern Cape golf courses

The ability of urban golf courses to harbour biodiversity is dependent on several factors such as size, habitat heterogeneity and climate (Hammond & Hudson, 2007; Hudson & Bird, 2009). The average size of Eastern Cape golf courses is almost 34 ha, which is smaller than Terman’s (1997) report of 54 ha, but this was a global estimate comprising a larger and more developed world sample population. Thirty-four hectares is still a large area of contiguous green space within an urban setting considering that the average urban park is no bigger than 10 ha (Kaczynski, Potwarka, & Saelens, 2008), and for Eastern Cape towns, McConnachie, Shackleton, and McGregor (2008) reported the mean size of public green spaces to be 1.7 ± 0.2 ha. Furthermore, on average, 66% of Eastern Cape golf courses consist of woody vegetation and rough, which are areas that are considered out-of-play, and which consequently could be managed for biodiversity. This falls within the 40–70% range of out-of-play area estimated by Tanner and Gange (2005) in the UK. Out of play areas can consist of various habitat types, much of which are relatively undisturbed by course managers and with conscious application could be managed for biodiversity conservation and ecosystem service provision in a multifunctional land use system.

5.2. Relationships between climate and woody plant diversity

The study golf courses occurred along gradients from inland to the coast, which allowed for examination of the effect of macro-climate on species richness, percentage ground cover, percentage of native woody plants and species diversity. The general trend was that species richness decreased with increasing altitude and distance from the ocean. A similar trend was also seen for both the percentage ground cover and native woody plants. This is most likely due to inland climates being harsher in comparison to the more temperate coastal ones. Another factor to consider is the vegetation types or biomes in which the study sites occur. The sites occurred in the particular areas of the Eastern Cape where the biomes include: forest, Albany thicket, savannah, grassland and fynbos; sites closer to the coastal areas are in the wooded biomes (forest or thicket) and the inland sites fall within the less wooded savannah, grassland and fynbos biomes (CSIR, 2004; Rutherford, Mucina, & Powrie, 2006). Forest and Albany thicket are characterised by high woody cover, increasing the probability that golf courses situated in these biomes have greater woody plant species diversity, percentage ground cover and percentage of native woody plants. The grassland and fynbos biomes naturally have fewer woody plants; they are predominantly open grassland with shrubby vegetation and sparse trees (Lubke, Everard, & Jackson, 1986). Species diversity is a function of species richness and species evenness/abundance. The fact that climate and species diversity did not share a relationship means that although the number of species increases across the sites the abundance/evenness of species may not.

5.3. Greenkeepers’ perceptions of biodiversity conservation on urban golf courses

According to Hammond and Hudson (2007), greenkeepers’ management techniques and attitudes towards biodiversity conservation are important and can play a vital role in the species composition of a golf course. Before a greenkeeper can attempt to increase or conserve biodiversity, it is necessary for him/her to have some understanding of the concept and its importance. The study revealed a wide range of self-reported understanding of biodiversity among the respondents; half rated themselves as having no to little understanding and the other half had a moderate understanding of biodiversity. This is likely to be a reflection of the type of training/education received by the greenkeepers. Only one-third of the greenkeepers had received formal training, none of which included any teachings on biodiversity. This suggests that there is a poor link between golf course management and biodiversity management, indicating areas for improvement. Urban golf courses are likely to have a greater chance of increasing local biodiversity levels if greenkeepers are informed on the latest and most effective management practices and are exposed to practices that foster biodiversity (Colding & Folke, 2009).

Greenkeepers stated that information on biodiversity management for golf courses was not easily accessible. It was only accessible to those who were willing and able to seek it and, even if information was located, it was mostly about North American or European courses and strategies. To conserve indigenous biodiversity on South African urban golf courses, managers and greenkeepers need to have readily accessible information on good biodiversity management practices for the South African context. Some of the greenkeepers networked and shared information with fellow greenkeepers. Communication and information sharing between greenkeepers could have many positive outcomes for biodiversity conservation on golf courses and should be encouraged; it should however augment information that is readily available to all greenkeepers. Greenkeeper associations in South Africa could do more to promote biodiversity management. Social learning or community of practice models could be a useful approach in this regard.

Initially, only half of the greenkeepers claimed to understand the term ‘biodiversity’. However, after it was explained, all the greenkeepers agreed that urban golf courses could play a role in biodiversity conservation. The reasons given for their agreement were the same reasons as reported in previous studies, such as a ‘parkland in an urban area’ and that golf courses create habitats for a variety of wildlife. Tanner and Gange (2005) and Hodgkison et al. (2007a) support these same views. One interviewee’s comment on how courses are an ‘economical’ method of biodiversity conservation supported the view of Colding et al. (2006) that golf courses are businesses that have the means and potential to be stewards of biodiversity conservation. Although in some cases the term biodiversity was not fully understood, there are still links between why greenkeepers feel golf courses can play a role in conservation and information in the literature (Colding & Folke, 2009).

Most of the greenkeepers believed that a high level of woody plant diversity on golf courses was of importance, although some believed it was limited. Besides the aesthetic appeal that high plant diversity could bring to a golf course, some greenkeepers were aware of how high plant diversity may contribute to a better and healthier ecosystem. Colding and Folke (2009) suggest that golf courses in urban areas are perceived to have a higher ecological value than golf courses in rural areas because their ecological value increases with increased anthropogenic influence; many courses are effectively islands within an urban matrix. We argue that this island analogy demands that urban golf courses actively become more than just sites for playing golf, but embrace a more land use science approach in managing the courses to supply a variety of ecosystem services to both golfers and the broader urban society. The ecological value of individual golf courses can be further enhanced by increasing landscape heterogeneity and connectivity on courses, which allows for species movement between patches of vegetation, acting as corridors that link other urban patches within the urban ecosystem (Colding, 2007). Landscapes can be made more heterogeneous by altering the composition and arrangement so that patches are linked to create ecologically functional units (Colding, 2007). In the context of urban golf courses, this can be achieved in part by increasing the woody plant diversity and cover. However, golf courses are first and foremost golf courses and hence managers and greenkeepers may be pressurised by club members to configure the golf course in a certain way because they are paying members (Wheeler & Nauright, 2006). However, most of the greenkeepers said that members of the club had little influence on management. In this regard, interviewees saw increasing woody plant diversity as both positive and negative because this may create more of a challenge to golfers.

The majority of the cases revealed that golf club members’ influence on the woody plant composition and management thereof was of weak to medium strength. This indicates that the greenkeepers and golf club committees are the ones that have the most say in how an individual golf course looks and is managed. Therefore, the prospect of conserving biodiversity on golf courses relies, for the most part, on golf course management within the constraints of financial and climatic contexts.

5.4. Biodiversity management

Only 25% of the golf courses had any explicit biodiversity conservation objectives, but more than 60% of the greenkeepers said that they would like to see more being done to conserve biodiversity on their golf courses. Terman (1997) and Hunter, Kealy, and Forrest (2010) believe that natural golf courses are the best way to promote biodiversity conservation on urban golf courses, albeit under good management. Designing or even restoring a golf course to be more natural helps to support meta-populations of native fauna and flora as well as the ecosystems on which they depend (Terman, 1997). The biodiversity conservation objectives mentioned by the greenkeepers included the reduction of non-native vegetation and the planting of more native tree species. Terman (1997) further states that native vegetation provides the highest quality habitat for local fauna and that the best management plan is to manage golf courses so that the habitat requirements of native species are met. Therefore, a major management goal for all golf courses should be to increase the proportion of native woody plants and limit the proportion of non-native ones, and eradicate invasive alien species. Non-native woody plants have a negative effect on local biodiversity because they alter the composition, structure and complexity of an ecosystem (Tanner & Gange, 2005). Hunter et al. (2010) found that ecosystem functionality could be greatly improved on golf courses where native tree numbers are low by planting more native trees.

We found no relationship between the coarse, qualitative management scores and the species richness, percentage native woody plants and species diversity. There are three possible explanations for this lack of a relationship. First is that species richness, percentage native woody plants and species diversity are influenced more by factors beyond the control of greenkeepers, such as climate and soils. Second is that the list of management practices was not comprehensive enough and did not cover fully the management practices that could promote biodiversity. Third is that since most of the managers did not undertake specific actions to promote biodiversity, the variation across the management scores was insufficient, especially considering the small sample size. However, woody plant species richness was significantly correlated with actions such as native-friendly management, annual golf club membership fees, number of members, number of grounds staff and the length of experience of the greenkeepers. It is easy to assume that the number of grounds staff and number of club members is a reflection of the relative income of a course. The assumption is that the more members a golf club has, the greater the annual fee income to the club and therefore the greater financial resources for hiring grounds staff, tree planting and maintenance which in turn can contribute to better biodiversity outcomes. Hammond and Hudson (2007), Burgin and Wotherspoon (2009), Colding and Folke (2009) and Hudson and Bird (2009) all highlight that through effective management, golf courses can increase their biodiversity, and that managing golf courses for biodiversity conservation contributes to increasing biodiversity in the broader urban setting.

6. Conclusion

South Africa experiences a wide range of threats to its many habitats and species. Land use change and urbanisation are among these. As a mega-diversity country it is important that authorities and citizens respond to these threats in an informed, considered and integrated manner. With diminishing areas of public green space as towns and cities densify and expand, biodiversity-friendly strategies need to be promoted on the remaining green areas on both private and public lands. Urban golf courses can potentially play a major role because of their relatively large size for an urban green space, having a large proportion of the area (>60%) under limited use, and relative to many public urban green spaces, have reasonable financial resources to support management objectives.

This study has shown that the sampled urban golf courses in the Eastern Cape of South Africa are large and have a large proportion of their area as out of play zones. This means that they could be managed for urban biodiversity conservation alongside the needs of the golfing members. Second, the botanical survey results show that a wide richness of tree and shrub species are supported on golf courses, but this is compromised by the presence of alien invasive species (at every course) and for some courses, a predominance of non-native species. This potentially reduces the nature or quantity of additional ecosystem services provided. For example, urban bird species richness and abundance in streets is lower for non-native trees than indigenous ones (Shackleton, 2016). The presence of invasive alien species means that they can act as sources of seed for invasion of neighbouring properties.

Lastly, the survey of greenkeepers revealed limited knowledge of biodiversity and strategies to manage or increase it, but high willingness. Implementing biodiversity conservation on urban golf courses requires informed and committed greenkeepers. This requires available information and training courses that are locally relevant. The training courses available to greenkeepers in South Africa would benefit from specific modules on the importance of biodiversity and strategies to promote it in different biomes of the country, rather than having to rely on information from northern countries. In the shorter term, relatively ubiquitous strategies would include favouring indigenous plants over non-native ones, control of invasive alien plants, increasing structural complexity, limiting physical activities in the out of play areas, labelling trees, limiting use of chemicals and engendering participation of club members in citizen science approaches in monitoring of species of particular interest. National conservation NGOs are likely to be able to assist with designing context-specific materials and training greenkeepers and club members. Such actions will not only increase biodiversity on urban golf courses, but will bolster other ecosystem services (such a recreational and educational values, carbon sequestration and mitigation of urban temperature extremes), as well as providing a model for other privately owned lands such as corporate grounds and perhaps public institutions (e.g. libraries, schools and hospitals).

Implementation of such strategies can be prompted or achieved by conservation agencies and relevant city authorities raising awareness among urban golf course managers and working with them to implement more biodiversity and environmentally friendly management strategies. The first step would be building awareness through training courses and materials and, in time, requiring urban golf courses to include statements and objectives regarding biodiversity in the course management plans.

Acknowledgements

This work was sponsored by the South African Research Chairs Initiative of the Department of Science and Technology and the National Research Foundation of South Africa. Any opinion, finding, conclusion or recommendation expressed in this material is that of the authors and the NRF does not accept any liability in this regard. We are grateful to the greenkeepers for permission to work on golf courses and for their willingness to share information with us. We are appreciative for comments on an earlier draft of this paper by James Gambiza.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was sponsored by the South African Research Chairs Initiative of the Department of Science and Technology and the National Research Foundation of South Africa. Any opinion, finding, conclusion or recommendation expressed in this material is that of the authors and the NRF does not accept any liability in this regard.

Appendix 1.

Golf course manager interview -

A) Background

1. What is your age and gender?

2. How long have you worked in golf course management?

3. Have you received any training or education in this field? If so, what?

4. How old is the golf course?

5. How many grounds staff are employed on your golf course?

6. What do you understand by the term biodiversity?

7. Are you aware of any species of conservation concern on or surrounding the golf course?

B) Management

1. Are there any biodiversity conservation objectives/targets in place on your golf course; if so what are they?

2. The following is a list of management practices that are regarded to promote plant biodiversity; please tick and rank the practices that are implemented on your golf course. Ranking works as follows: 1-not applicable; 2-never; 3-seldom; 4-often; 5- always.

Table

Practices	Yes or no (tick)	Rank (1-5)
Leave large natural wooded areas undisturbed
Maintain undergrowth and leaf litter in wooded areas
Designate ‘no mow areas’ in out-of-play areas
Use native plants when planting
Removal/reduction of non-native tree species
Alien invasive species control
Leave old/dead trees in situ

C) Manager perceptions of biodiversity conservation on golf courses

1. Rank the level of agreement with the following statement: ‘Urban golf courses can play a role in biodiversity conservation’ (1-strongly disagree; 5-strongly agree). Please provide a reason for the given ranking.

2. Rank the importance of high plant diversity on golf courses (1-not very important; 5-very important). Please provide a reason for the given ranking.

3. How strong is the golf course members’ influence on the vegetative composition of the golf course? (1-not very strong; 5-very strong).

4. Would you like to see more being done to promote biodiversity on your golf course? If so, what?

5. Do you feel that information on good biodiversity management practices for golf courses is easily accessible? Please elaborate.

6. Where do you access information about biodiversity management on golf courses?