Local peoples’ knowledge and perceptions of Australian wattle (Acacia) species invasion, ecosystem services and disservices in grassland landscapes, South Africa

ABSTRACT Many alien tree species were introduced into grassland ecosystems in South Africa by the commercial forestry industry for paper and timber for furniture. Over decades some of these introduced species escaped into neighbouring farms and community land. Adult trees from these alien species now provide other ecosystem services, notably fuelwood. Depending on the spatio-temporal context, many of these species can also negatively affect ecosystem services. We collected interview data from commercial and communal farmers in the upper Umzimvubu catchment in South Africa to compare farmers’ knowledge and perceptions of invasive wattle species invasion and their associated ecosystem services and disservices. Fuelwood and fencing poles were the most common uses of wattle by commercial (83%; 67%) and communal (99%; 49%) farmers. On the other hand, the reduction of grass cover and loss of grazing land were the most commonly mentioned negative impacts of wattles by commercial (83%; 75%) and communal (92%; 80%) farmers. Although both groups recognise the importance of wattles in providing ecosystem services, most communal farmers perceived wattles to have more negative effects than benefits. The findings demonstrate that both farmer groups highly depend on ecosystem services and are affected by disservices of wattles. However, while large-scale commercial households favour the presence of wattles in the landscape, communal households prefer complete removal of the wattles from the landscape. This may be due to lack of locally available alternative options or inability to replace or purchase ecosystem services affected by wattles from other sources or markets like commercial farmers.


Introduction
In this rapidly changing and uncertain world, people's well-being, biodiversity, and ecosystem services are interconnected, and consequently understanding social factors and processes is crucial for achieving effective and sustainable environmental governance (Bennett et al. 2015(Bennett et al. , 2017;;Reyers et al. 2015;Vaz 2018;IPBES 2019).Ecosystem services (ES) are the benefits humans obtain from natural and transformed ecosystems, either directly or indirectly (Costanza et al. 1997(Costanza et al. , 2017;;Bennett et al. 2015;Díaz et al. 2018).Invasion by alien species is one of the main drivers of global environmental change affecting many ecosystems and the services they provide, potentially increasing with climate change and international trade (Zimmermann et al. 2014;Castro-Díez et al. 2019).Alien species are those introduced, intentionally or accidentally, by humans to new locations (Zimmermann et al. 2014).Many such species were brought into new areas for specific ES.However, little or no attention was given to their potential to spread and become a problem elsewhere (Pejchar and Mooney 2009;Rouget et al. 2016;Liu et al. 2017;Russell and Blackburn 2017).Consequently, their ability to spread from sites of introduction to surrounding areas and further afield often causes unprecedented and diverse impacts on social-ecological systems (Richardson et al. 2011;Castro-Díez et al. 2019;Kapitza et al. 2019).Depending on the spatiotemporal context, some invasive alien plants (IAPs) negatively affect existing ES and create ecosystem disservices (EDS) (Kull et al. 2011;Potgieter et al. 2019;Shackleton et al. 2019d.Ecosystem disservices are defined as 'the ecosystem generated functions, processes and attributes that result in perceived or actual negative impacts on human well-being' (Shackleton et al. 2016).The EDS from IAPs range from preventing access to significant ES such as freshwater, grazing for livestock, and medicinal plants by local users, to changes in ecosystems processes (such as altered fire regimes and nutrient cycles), causing negative financial implications to agriculture, forestry, and fishery sectors to adverse impacts on economic growth and food security (van Wilgen et al. 2001(van Wilgen et al. , 2008;;Wise et al. 2012;Castro-Díez et al. 2019).The ecological impacts of IAPs on some ecosystems are often very difficult to reverse, depending on the magnitude, scale, and extent of invasion (Le Maitre et al. 2011;Ndhlovu et al. 2011;Yapi et al. 2018;Castro-Díez et al. 2019).
In South Africa many alien plant species were introduced to supply various ES, with little knowledge about their potential to spread beyond their intended areas.As a result of their invasiveness, some IAPs have invaded large portions of South Africa's natural ecosystems, which threatens the sustainability of these ecosystems and the livelihoods that are dependent on ecosystem goods and services (De Neergaard et al. 2005;van Wilgen et al. 2008;Le Maitre et al. 2011;Ndhlovu et al. 2011;Shackleton et al. 2015b).To date, many of these IAPs are already well established in some areas, while others are in the early stages of invasion (van Wilgen and Richardson 1985;Nel et al. 2004;van Wilgen et al. 2008;Kotzé et al. 2010).
Until recently, most IAPs research and control programmes globally and in South Africa have focused on the ecological and economic impacts of the IAPs, paying less attention to the significance these species may have for human livelihoods and well-being (Shackleton et al. 2019e).Yet, invaded landscapes are embedded in social-ecological systems where land user actions affect ecological processes (Gaertner et al. 2014;Vaz 2018).Some landscapes can become more vulnerable to IAPs than others due to the characteristics of the social systems to which they are connected (Bennett et al. 2015;Yletyinen et al. 2021).Changes in ecological conditions resulting from IAPs may lead to negative or positive human perceptions and attitudes towards IAPs.As a result, arguments for managing IAPs increasingly consider the human perceptions of the invading species rather than just the biophysical impacts (Vaz et al. 2017;Novoa et al. 2018;Castro-Díez et al. 2019).Perceived impacts of IAPs are driven by a combination of factors, including attributes of the invading species and the invaded landscape as well as the social context (Potgieter et al. 2019;Shackleton et al. 2019d).Yet, until recently, research that is designed to highlight preferences and perceptions of IAPs across stakeholder groups has been limited.There is a need to examine multi-stakeholder perceptions of IAPs to guide the design of management interventions that are consistent with various economic situations (Shackleton et al. 2015a).
South Africa presents an ideal opportunity to conduct such case studies due to its historically diverse socio-economic landscapes, which generate different social-ecological systems alongside one another (Shackleton et al. 2001(Shackleton et al. , 2015a;;Cumming et al. 2014a;Hamann et al. 2015).This study explores local people's knowledge and perceptions of IAPs.We identified two land tenure systems with substantively different socio-economic challenges, including communal systems and commercial farming systems.The land tenure system has long been thought to play an important role in determining natural resource use (Shackleton et al. 2001).In South Africa, communal land tenure systems are characterised by lowincome households highly dependent on local ecosystems for direct access to natural resources either to supplement their income or for their subsistence.There are often very limited economic opportunities through which to secure natural resources from elsewhere.Communal rural communities in South Africa have been historically disadvantaged under the racially discriminatory colonial and the apartheid regimes due to a lack of government support and remain comparatively underdeveloped in terms of essential service delivery (Nnadozie 2011).Land management practices are largely influenced by the policies of the past and the present democratic government.Additionally, the issues surrounding land tenure reform policies have yielded a multitude of government institutions with overlapping mandates to govern access to common resources.In some cases, the conflicting agendas between government institutions and traditional authorities can have undesirable effects of weakening the ability of the traditional authorities to effectively govern communal lands and resources, undermining the resilience of the social-ecological systems to respond to changes.
On the other hand, commercial farming systems in South Africa are generally located on freehold private tenure lands that were advantaged during the colonial and apartheid periods.Thus, they are characterised by large land parcels, with low population densities and reasonably secure incomes.Because they are generally wealthy, they are able to secure their basic needs for essential ES through market supplies from distant ecosystems (e.g.piped water, purchased fodder and fertilisers).As a result, commercial farming systems have low direct dependence on local environment for ES.Commercial farms involve more active management compared to communal land.Land management practices are concerned with maximising economic returns resulting in a grazer dominated systems with moderate but constant stocking rates and a low fire frequency.
These two tenure and economic system types face very different sustainability challenges.Focusing on two different but adjacent social-ecological systems provides important insights, thus allowing policies to be better targeted to address the sustainability challenges faced in different land tenure systems.Firstly, invasion by IAPs is likely to have different effects amongst different social groups operating on the two land tenure systems (Shackleton et al. 2007;Kannan et al. 2014).Secondly, the level of use and reliance on IAPs may differ between different socioeconomic groups.Therefore, the presence and management or removal of IAPs from ecosystems are likely to engender different effects between the two systems.
Australian wattles include the world's most prolific invasive plants ( Le Maitre et al. 2011;Richardson et al. 2011).They have invaded large areas of grassland ecosystems in South Africa, with the potential to continue to spread into new areas (Gouws and Shackleton 2019b).Yet, although some of these species are desirable and provide a vital livelihood resource in local communities (Kull et al. 2011), negative perceptions surrounding wattles persist as they inevitably increase in abundance and extent of invasion (de Neergaard et al. 2005;Shackleton et al. 2007;van Wilgen et al. 2011;Witt et al. 2018;Ngorima and Shackleton 2019).In South Africa there is substantial evidence of the negative impacts of Australian wattles on biodiversity, ecosystem services, and livelihoods (Le Maitre et al. 2011;Gwate et al. 2016Gwate et al. , 2021;;Oelofse et al. 2016;Yapi et al. 2018).These include reduced livestock grazing capacity, stream flows, biodiversity, and residents' sense of place and culture.From a social perspective, issues relating to the safety and security of vulnerable groups (mainly women and children) have led to negative perceptions surrounding wattle stands around villages (De Neergaard et al. 2005;Kull et al. 2011;Ngorima and Shackleton 2019).As with many other IAPs that provide benefits and EDS, conflicting perceptions underlying the management of these species hamper the effectiveness of control actions.
Accordingly, ecosystem managers are increasingly seeking frameworks that bring balance; enhancing the provisioning services of IAPs while minimizing the negative impacts.Hence, local stakeholder views, values, and demands on the management for both ethical and conservation purposes should always be considered within this context before responsive management interventions are designed (Gaertner et al. 2016).
Within this context, this study sought to compare and contrast the knowledge and perceptions of species of wattle invasion between commercial and communal land tenure systems using a lens of ecosystem services and EDS.

Study area
The study was conducted in the Umzimvubu catchment, which lies in the northern region of the Eastern Cape and on the eastern border of the KwaZulu-Natal province of South Africa.From the foot of the Drakensberg Mountains, at the Lesotho border, the catchment makes its way in a southeast direction into the Indian Ocean.Sites were selected on the bases of being in the current range of wattle invasion in the upper section of the Umzimvubu catchment (upper Umzimvubu catchment), including Matatiele, Cedarville, Kokstad, and Mount Frere.The area includes sites with juxtaposed private and communal land tenure systems.Matatiele, Mount Frere, and Kokstad are the three main towns in the study area with populations of 12 466, 5 252, and 51 561, respectively.Cedarville is a small town (with fewer than 5 000 people) within the study area comprised mainly of private commercial farms.The different land use systems represent the area's prevailing environmental and socio-economic conditions.
The total population of the Umzimvubu catchment is approximately 2.2 million people, with 85% residing in rural communal tenure, settlements and dependent, to varying extents, on directly accessible natural resources for livelihoods.Many people in the area depend on government social grants as their primary cash income.The main livelihood activities in the area include extensive livestock grazing in rangelands and crop farming which is coupled, in some parts, with state food projects.Most rural households in communal areas report cash incomes of just over R600 (approx.US$38) per month, mostly in the form of social grants.
The climate in the area ranges from temperate, with high frost frequency in the northern, higher altitudes, to sub-tropical along the coastal belt.It is a summer rainfall area, with annual average rainfall ranging from approximately 650-1 000 mm in the higher-lying areas of the upper catchment and along the coastal areas, respectively (Mucina et al. 2006).The vegetation is mostly grassland, occupying approximately 69% of the catchment, along with savanna (21%), coastal (8%), forest, and thicket (2%) biomes covering the remainder.Grasslands cover the northern and central areas of the catchment, while savanna is represented in the central and western areas (Mucina et al. 2006).

Australian wattles as invasive alien species
Australian wattles (Acacia dealbata, A. decurrens, and A. mearnsii) were introduced into South Africa for a variety of purposes, including soil stabilisation, supply of fuel, timber, and in the case of Acacia mearnsii (black wattle), to produce tanbark for the leather industry (Dunlop and MacLennan, 2002;Forestry South Africa 2009).Hence, the distribution and range of these species is largely attributable to dissemination by humans (Wilson et al. 2007;Richardson et al. 2011;Hirsch et al. 2017).Australian wattles are generally fast-growing trees that can reach reproductive maturity within 2-3 years (Gibson et al. 2011).These traits give them a competitive edge over other plant species.On the other hand, the same traits, namely fast growth, make them desirable to some land users.Consequently, many local communities have adopted wattles for uses beyond the initially intended purposes (Kull et al. 2011).For instance, in the grassland biome, where trees are naturally limited, wattles become an important source of shade for both people and livestock, may become an essential source of fuelwood, construction material, fencing poles, and an income source when harvested for sale (Ngorima and Shackleton 2019).

Data collection
Two groups of farmers were identified, namely communal land users and commercial livestock farmers on private tenure farms.Communal households were randomly selected across three villages: Colana (near Mount Frere), Mvenyana, and KwaMzongwana (near Matatiele).The villages were selected in collaboration with the local NGO, Environment and Rural Solutions (ERS), based on density of wattle invasion, and current and historical wattle clearing activities.Commercial private farms were sampled in the vicinity of Matatiele, Kokstad, and Cedarville.A list of commercial farmers was obtained from the ERS.This was followed by a snowball approach (Biernacki and Waldorf 1981), where participants were asked to suggest names of other commercial livestock farmers with wattle species or a history of wattle invasion on their farms.Interviews were conducted in the home language of the respondents (Xhosa or English).A translator was used where the respondents were not fully conversant in Xhosa or English.In each household, the oldest available and willing member of the family was interviewed.A total of 100 interviews was sought; however, due to the small number of commercial farmers that met the selection criteria of having wattles on their lands, this number could not be achieved.Consequently, seventy-seven interviews were conducted across different sites, including 65 communal land users and 12 commercial farmers.Interviews were based on a semistructured questionnaire comprising a mix of openended, closed-ended, and ranking questions (Creswell 2009), designed to understand land users' perceptions of the effects of wattle invasion on livelihoods and the environment.The open-ended questions allowed respondents' perceptions of wattles to emerge without pre-empting, while the closed-ended questions allowed for quantitative analysis of the key constructs.Respondents were asked about (i) the type of livestock they keep (including reasons for keeping livestock); (ii) their knowledge of wattles, perceptions, and awareness of wattle invasion (including causes of spread, rate of spread, and abundance in different landscapes); (iii) whether they view wattles as providing any benefits (including perceived benefits); (iv) whether their level of use of wattles has changed over time; (v) whether wattles have negative impacts on people or the environment (including perceived negative impacts); and (vi) the sociodemographic information of the household (including ownership of household assets such as cars, tv, and fridge).Surveys in communal areas took place from 08h30 to 17h30 during weekdays.Interview appointments with commercial farmers were prearranged via phone call and so were not limited to weekdays.Interviews took approximately 40 minutes per respondents.
Ethics approval for the study was via the Departmental Screening Committee of the Department of Environmental Science, Rhodes University (Reference Number (ES17/26).Participation was voluntary -participants were informed about their rights to refuse to answer any questions and to withdraw from the interview at any time should they wish to do so.Informed consent was obtained from the participants and anonymity and confidentiality were explicitly granted.Questionnaires did not include any personal information that could be used to identify individuals, and thus, all data was anonymised prior to analysis.

Analysis
We used a combination of qualitative and quantitative statistical approaches.Questionnaire responses were grouped into various categories according to the themes in the questionnaire, describing stakeholders' knowledge and perceptions of wattle invasion, benefits, and costs associated with wattles invasion.We first tested for association between stakeholder groups and participants' responses to closed-ended questions.We used thematic coding and a grounded theory approach to identify emergent themes in land users' responses to open-ended questions (Corbin and Strauss 1990).Subsequently, statements were coded as 0 or 1, corresponding to the absence or presence, respectively, of a theme in an interview script.This enabled us to examine associations between themes, stakeholder groups, and participant's responses to closed-ended questions.Chisquare tests were used to determine differences between perceptions, benefits, and costs associated with wattle invasion and perceptions of rangeland ecosystem services between stakeholder groups.The analysis was run using demographic variables (Table 1) as independent variables and the responses as dependent variables.A wealth index was constructed as the sum of the number of wage earners, social grants, total assets such as cars, television, refrigerator, and livestock (all standardised to between 0 and 1 against the highest value) per household.Subsequently, the wealth index was ranked into quartiles (poorest, poor, wealthy, and wealthiest).All analyses were performed using STATISTICA ver.14 software (TIBCO Software Inc. 2019).

Demographics, education and economic data of the two stakeholder groups
Most of the respondents from both commercial and communal settings were male -with an average age of just over 50 years (Table 1).The oldest respondent was 87 years old at the time of the interview.Most communal farmers were born in the area, except for a very few individuals from other villages (primarily women relocating after marriage).Commercial households generally have higher levels of education than communal households, as reflected by the average number of people with tertiary education.Many communal households in the area depend on government grants for cash income, whereas commercial farmers rely on wages and income from the farm business.

Local people's knowledge and perceptions of wattle invasion
All respondents (100%) knew what wattle is and almost all communal respondents (99%) and all commercial farmers (100%) perceived it to be invasive (Table 2).Some respondents in communal areas mentioned that they gained their knowledge of wattle invasiveness through participating in the Working for Water clearing programmes.Respondents in communal areas referred to wattle by its vernacular name, i.e. idywabasi in isiXhosa.
The percentage cover of wattle in communal areas differed among landscape types, with streams and grazing areas being the most invaded as expressed by the majority of the respondents (63% and 86%, respectively) (Figure 1).When asked if they wanted wattle in the landscape, significantly (p < 0.001) over half (58%) of communal farmers expressed their dislike of wattle and did not want it anywhere in the landscape, whereas all commercial farmers (100%) considered wattle as mostly beneficial (Table 2).All communal farmers (p < 0.001) perceived wattle to be spreading in different parts of the landscape compared to 58% of commercial farmers (Table 2).More than 90% of commercial farmers perceived wattle as most common in the riparian areas and in grazing areas (67%).Whereas people in communal areas perceived wattle invasion as most abundant in grazing areas (86%), followed by riparian areas (63%).Other invaded landscapes mentioned by less than 50% of respondents included croplands and near homesteads (Figure 1).
Six different causes of wattle spread were mentioned between the two land user groups (Figure 2).However, the number of causes mentioned differed between the commercial (6) and communal farmers (5).Significantly (p < 0.05) more commercial farmers than communal farmers attributed the spread of wattle to water (91,7% and 53.8, respectively), livestock (41.7% and 10.8%, respectively) and birds (25% and 0, respectively) (Figure 2).Furthermore, many commercial farmers (58%) believed that the occurrence of unplanned fires stimulated wattle seed germination in the landscape (Figure 2).Some communal farmers (49%) believed that wattle spread in old croplands resulted from land not being worked to produce crops (Figure 1).

Local people's knowledge and perceptions of ecosystem disservices of wattle species
In general, a majority of communal farmers considered the current levels of wattle invasion in the landscape as excessive as reflected by their preference of wattle stands in different landscapes (Table 3).Significantly more communal respondents than commercial farmers preferred to have wattle removed from the various parts of the landscape, including grazing areas (p = 0.003), streams (p < 0.001), old croplands (p < 0.001) and from around homesteads (p < 0.001) (Table 3).Respondents associated wattle with many negative impacts (Table 4) on these parts of the landscape, as it resulted in ecosystem degradation, reducing the benefits gained directly or indirectly from the ecosystems (Table 4).As a result, almost half (46%) of respondents in communal areas wanted wattle to be completely removed from grazing areas, whereas 42% of communal as well as commercial respondents wanted it to be reduced and maintained at an acceptable level (Table 3).Amongst the perceived negative impacts on grazing areas, displacement of grass along with palatable species was mentioned by more than 90% and 80% of communal and commercial farmers, respectively (Table 4).The loss of palatable grazing species was  perceived to have negative effects on livestock, thus reducing benefits gained from them.According to 42% of commercial and 65% of communal respondents, the main negative impacts on livestock production came in the form of loss of body condition (Table 4).The majority of livestock owners mentioned that loss of body condition resulted in low production and reduced economic value (Table 4).Furthermore, a substantial proportion of communal farmers (40%) and commercial farmers (50%) perceived that wattle impeded monitoring of livestock in grazing areas, hence increased incidents of livestock theft.Furthermore, wattle provides easy cover to small predators (such as jackals) that prey on small stock such as goats and sheep (Table 4).
Another key disservice of wattle invasion, mentioned by 80% and 75% of communal and commercial farmers, respectively, was that wattle has occupied most parts of the landscape and takes up productive lands, more specifically the grazing areas and old croplands (Table 4).Wattle invasion near homesteads was associated with various negative perceptions, especially with increased crime.Attacks on and robberies of women and children, and house burglaries, were mentioned because wattle stands provided refuge for the perpetrators (Table 4).A small proportion of communal farmers noted that wattle invasion near homesteads tended to cause damage to building foundations through the expansion of the roots, in some cases destroying homes (Table 4).Approximately two-thirds (66%) of commercial farmers and one-third (39%) of communal ones mentioned that wattle reduced water flow from rivers and streams.Many respondents mentioned that some rivers and streams that used be perennial have since become dry because wattle takes up a lot of water (Table 4).
Significantly more commercial farmers (17%) than communal ones (2%) associated wattle invasion in the landscape with an increased need for buying livestock feed (Table 4).Because of shortages of forage, some livestock owners reported that they had to buy more feed to supplement grazing.Our analysis of the effects of wealth quartile on the perceived disservices of wattles revealed no significant differences (χ 2 = 20.54,df = 27; p = 0.807).

Local people's knowledge and perceptions of ecosystem services of wattle species
Many communal farmers believed that the ecosystem disservices outweighed the positive benefits (ecosystem services) from wattle due to current levels of invasion and rate of spread.The majority of farmers mentioned that they used wattle because of its high abundance and lack of alternatives.One respondent who sold fuelwood from wattle to other residents mentioned that he mainly sold wattle to control its spread, not just because he wanted to make profit.
Table 3. Land users' preferred level of wattle abundance in different parts of the landscape (Data are percentage of all households that suggested particular preference).Each respondent was asked if they would prefer to have more wattle on different landscapes, if wattle was maintained at current levels, reduced, or completely removed from each landscape.NB: All respondents were asked to give an opinion, based on the current levels of wattle invasion, on each landscape -so totals may not add up to 100%.Nine benefits were reported, the most common being fuelwood and fencing poles (Table 5).However, the number of benefits mentioned differed between the communal farmers (8) and commercial farmers (6).Significantly (p = 0.01) more communal farmers (98%) than commercial ones (83%) collected wattle for fuelwood (Table 5).And a larger proportion of commercial farmers (67%) mentioned the use of wattle for poles compared to 49% of communal stakeholders.Respondents mentioned that wattle is a valuable resource for fuelwood, and they didn't have to go far to collect it.Some communal farmers mentioned that wattle displaces indigenous trees, resulting in them relying on wattle for fencing poles and other uses.Use of wattle for poles was mainly limited to droppers (a light vertical lath in a fenceusually in between bigger poles), as farmers mentioned that wattle poles tend to rot quickly (in particular silver wattle).
Other wattle uses mentioned by a minority of communal farmers included construction material (29%), kraaling material (11%), and medicinal use (11%) (Table 5).The use of wattle for construction ranged from making small structures such as sledges for carrying materials, chicken coops, and the construction of mud houses.The most common use for medicinal purposes was the use of bark to treat stomach aches and asthmarelated illnesses in children and, at times, for the treatment of diarrhoea in young goats.Significantly more commercial farmers (50%, p < 0.001) mentioned that wattle trees in certain parts of the landscape provide shelter for livestock during harsh climatic conditions, with only (9%) of communal farmers sharing similar views.
Our analysis of the effects of wealth quartile on the perceived benefits of wattles revealed no significant differences (χ 2 = 18.7; df = 27; p = 0.881).

Discussion
Several researchers have argued that for better understanding of IAPs invasions before control programmes are implemented, the full range of socio-ecological and economic effects should be considered and evaluated (Forsyth et al. 2012;Wise et al. 2012;Zengeya et al. 2017;Novoa et al. 2018;Castro-Díez et al. 2019;van Wilgen et al. 2020).Some invasive alien trees provide benefits to local communities and the economy (De Neergaard et al. 2005;Kull et al. 2011;Ngorima and Shackleton 2019).However, while some IAPs (such as Acacia spp.) provide services, they also induce costs with negative implications locally and at broader scales (Shackleton et al. 2007;van Wilgen et al. 2008van Wilgen et al. , 2016;;Wise et al. 2012;Yapi et al. 2018;Ngorima and Shackleton 2019).Therefore, there is a need to understand where is the best place in a landscape to maintain IAP stands (if at all), at what density, how best to contain them, and how to satisfy human livelihood needs while preventing or minimising ecological degradation (Gaertner et al. 2016).
The results of this study show that wattles have both concurrent positive benefits and negative impacts for both communal and commercial farmers.However, communal farmers perceived wattles to have more negative effects than benefits, and thus, wanted wattles to be removed from certain landscapes.This is likely to be partly due to the relatively high abundance of wattles in the communal landscapes and impacts on other essential ecosystem services on which they are highly dependent.In contrast, large-scale commercial farmers typically had very low abundance of wattles on their landholdings -thus, perceived wattles to have more benefits than negative effects and were happy to maintain a portion of their properties within wattle.

Local people's knowledge and perceptions of wattle invasion
Understanding local people's views and attitudes is critical in mitigating conflicts over management and control of IAPs (Bennett and van Sittert 2019;Shackleton et al. 2019b).Recent studies have shown (Ngorima and Shackleton 2019;Potgieter et al. 2019) that the potential use of a species, and abundance of invasion, are some of the dominating factors influencing stakeholder perceptions.Other factors include proximity to invasion, the species characteristics, and the social context (Novoa et al. 2018;Shackleton et al. 2019d).
The results show that all respondents knew about wattles and of their invasive nature, except for a few communal respondents.The high level of awareness was not unexpected given that wattles are the dominant tree species in the area which is historically grassland.Australian wattles have been a feature of the landscape for at least eight decades, according to the respondents.All communal farmers stated that wattle abundance had increased over the years, and despite the ongoing management efforts, wattles continue to spread at an increasing rate.This view is similar to the one shared by communal respondents in other parts of the Eastern Cape province (Shackleton et al. 2007;Ngorima and Shackleton 2019).Furthermore, this perception is consistent with the findings by Gouws and Shackleton (2019a), who mapped increases in wattle abundance at an annual rate of 0.11-0.21%over six decades in the northern parts of the Eastern Cape.Similarly Scorer et al. (2019) found that A. mearnsii abundance could increase by 10% per year on abandoned farmlands of the Eastern Cape.Both groups attributed the cause of spread to high seed production and their dispersal by various agents, including water, wind, birds, and livestock.Water, in particular, was viewed as the primary vector of wattle spread by commercial farmers as most wattle invasions occurred along riparian areas.Indeed, wattle species in South Africa are known to grow along watercourses and produce abundant, tiny seeds suited to dispersal by water (Le Maitre et al. 2011).
A large number of communal farmers wanted wattles to be removed from certain landscapes as they believed that their negative effects outweighed the benefits.Similar perceptions about wattles species have been shared by people in the rural communities around Matatiele, Maclear, and Mount Fletcher in the Eastern Cape (De Neergaard et al. 2005;Ngorima and Shackleton 2019) and elsewhere (Kull et al. 2011).In contrast, all commercial farmers suggested that they still want to keep some wattles on their properties.This perception could be explained by the relatively low occurrence of wattle species on most commercial farms and the lack of alternatives for the benefits that wattles provide, such as fuelwood, fencing poles, shade, and shelter for livestock.The more active land management in terms of stocking rates, grazing rotation, and planned burning contributes to the mostly low occurrence of wattles on commercial farms.The contrasting views between the two groups correspond well with the scenarios documented in the Shackleton et al. (2007) framework, which depicts a likely shift in views and attitudes towards a species as its abundance increases and adversely affects one or more livelihood options.

Local people's knowledge and perceptions of ecosystem disservices of wattles
A high and reportedly increasing abundance of wattle in the landscape induced several disservices.The most frequently mentioned ecosystem disservices included loss of water, loss of arable land, reduced forage availability, and taking up land.From a social perspective, issues around the safety and security of women and children who are more reportedly vulnerable to attacks by criminals hiding in dense wattle stands were a significant concern in communal areas.
This echoes sentiments expressed by other communities in the region (De Neergaard et al. 2005;Ngorima and Shackleton 2019).Another issue was that in some cases, where wattles grew near homesteads, they can cause damage to housing foundations.This has resulted in some families abandoning their homes in fear for their safety.The resulting cost of repairing damaged walls or moving to new areas could be too high for some families, especially in a place where most people live below the poverty line.Such issues have been highlighted by respondents in other communal areas (De Neergaard et al. 2005;Shackleton et al. 2007Shackleton et al. , 2015a;;Kull et al. 2011;Ngorima and Shackleton 2019).
Several challenges associated with the high abundance of wattle in arable land were more prominent in communal areas than commercial farms.Firstly, the high cost and labour time required to remove wattle proved challenging, particularly for poorly resourced, communal farmers to undertake such interventions.Secondly, the documented effects of wattle on soil properties -including soil nutrient status, structure, and moisture -limit crop growth and render the land less productive (Lorenzo et al. 2010;Gaertner et al. 2011;Le Maitre et al. 2011;Oelofse et al. 2016;Yapi et al. 2018).Consequently, more croplands are increasingly abandoned as wattles continue to spread in these areas.The reciprocal relationship between arable land abandonment and IAPs increases is receiving greater attention and has been documented in South African studies (Shackleton et al. 2014(Shackleton et al. , 2019;;Scorer et al. 2019;Gouws and Shackleton 2019b).Overall, the presence of wattles in cropland has negative implications for food security as shown in other similar contexts in South Africa (De Neergaard et al. 2005;Shackleton et al. 2007;Ngorima and Shackleton 2019).
Similarly, the occurrence of wattles in grazing areas also led to several disservices there, including reduced forage and reducing the extent of grazing land.The ecological impacts of IAPs on rangeland ecosystems have gained growing attention in recent studies and have been documented in South African case studies, both locally (Ndhlovu et al. 2011;Oelofse et al. 2016;Yapi et al. 2018;Gouws and Shackleton 2019a) and at national scale (van Wilgen et al. 2008;Yapi 2013).Another concern raised by livestock owners was that dense stand of wattles in grazing areas provides a habitat for livestock predators, which may attack small stock such as sheep and goats.These views were shared between the two farming groups, consistent with the findings of studies in other communal areas (De Neergaard et al. 2005;Shackleton et al. 2007Shackleton et al. , 2015a;;Kull et al. 2011;Ngorima and Shackleton 2019).
There were contrasting views amongst communal farmers regarding keeping a portion of wattles in grazing areas.This result suggests a lack of consensus amongst communal residents about whether wattles should be retained in the landscape.In the above context, it is important to note that in communal areas, what is referred to as grazing areas, often include abandoned cultivated land as found in another study conducted in the region (Ngorima and Shackleton 2019).More attention needs to be given to the prevalence of wattles in rangeland ecosystems and its associated disservices.
High water consumption by wattle species was perceived as a significant EDS by both groups.The Drakensberg mountain range is a critical water source area in which the Umzimvubu catchment is embedded (Nel et al. 2017), and has been experiencing drought in recent years.Water availability in many areas of South Africa has become a significant concern, and efforts to conserve water are increasingly shifting towards demand management (Turpie et al. 2008;Egoh et al. 2012;Nel et al. 2017;Shackleton et al. 2019f).Invasive alien trees use significantly more water than indigenous grasslands (Dye et al. 2001;Dye and Jarmain 2004;Everson et al. 2011;Le Maitre et al. 2015), thereby reducing surface run-off, streamflow, and subsequently water supply and security (Le Maitre et al. 1996Maitre et al. , 2016;;Dzikiti et al. 2013;Preston et al. 2018).Awareness of the link between IAPs and water resources has increased considerably since the establishment of the Working for Water programme in the 1990s (Le Maitre et al. 1996Maitre et al. , 2016;;Turpie et al. 2008;van Wilgen et al. 2012).This is likely to influence respondents' perceptions of IAPs either by shifting or solidifying current views.Another concern raised by respondents was that water is the leading cause of wattle spread and that occurrence of wattles riparian zones sometimes prevented other essential activities, such as livestock access to water.Hence, the majority of respondents wanted wattles to be removed near water sources.Respondents from rural communities around Matatiele, Mount Fletcher and Maclear, have shared similar views (De Neergaard et al. 2005;Shackleton et al. 2007;Ngorima and Shackleton 2019).In contrast, most commercial farmers wanted to maintain a portion of wattles along the streams and rivers, as they believed the roots prevent siltation by stabilising riverbanks.Indeed, as reported by Rowntree (1991) removal of woody species from river banks could result in mobilisation of sediments.
Although not mentioned as an EDS, a significant portion of commercial farmers raised concerns about fire as the cause of wattle spread.Fire is an integral component of the South African grassland biome and its functioning (Parr et al. 2014;Bengtsson et al. 2019).To a large extent, fire occurrence is driven by the availability of fuel.Therefore, large quantities of aboveground biomass enhanced by IAPs can increase the extent and severity of fires, thereby altering fire regimes (Le Maitre et al. 2011).This often negatively affects ecosystem processes such as water repellency, soil stabilisation, and soil erosion, with negative implications for ecosystem services (Le Maitre et al. 2011).As mentioned by respondents, changes in ecosystems resulting from high-intensity fires tend to promote future invasion by wattle species, as wattle seeds are stimulated by fire (Hendry and Van Staden 1982;Fuentes-Ramírez et al. 2011;Le Maitre et al. 2011).From a social perspective, communities living in heavily invaded landscapes face a greater risk of fire and more extensive damage if a fire does occur.As with the effects of IAPs on water resources, the link between IAPs and fire regimes is receiving greater attention (Stehle 2017;Kraaij et al. 2018) with a potential to influence peoples' views about IAPs.

Local people's knowledge and perceptions of ecosystem services of wattles
According to both groups, the key services from wattles were the provisioning services of fuelwood and fencing poles.Fuelwood as a source of energy was more critical in communal areas, as there was no electricity supply to households in the area at the time of data collection.Shelter for livestock (regulating service) was vital for livestock owners.Indeed, in an environment where trees are naturally scarce as sources of fuelwood and other relevant services, residents tend to rely on IAPs for these purposes (De Neergaard et al. 2005;Kull et al. 2011;Shackleton et al. 2015a;Ngorima and Shackleton 2019).This is the case in many other areas where, in the absence of indigenous woody species, IAPs, such as Australian acacias and Prosopis, are a vital fuelwood source (Mwangi and Swallow 2008;Shackleton et al. 2015a;Shackleton and Shackleton 2018;Potgieter et al. 2019).However, some services were more important for one farming group than the other.For instance, regulating services such as shade and cover for livestock during extreme weather conditions were more important for commercial farmers, whereas many communal farmers perceived wattle as more problematic in grazing areas than useful.This is not surprising considering the differing levels of invasion in commercial and communal systems.
The effects of provisioning ecosystem services are likely to catalyse different perceptions than ES from other categories, such as cultural, supporting, and regulating services (Potgieter et al. 2017).This is because provisioning services provide direct tangible benefits that people can easily value, including a monetary value.On the other hand, other ES categories such as regulating services and supporting services do not directly affect people but play an essential role in ecosystem functioning, and therefore are the bases for which provisioning services exist.Therefore, services that fall in those categories seldom influence people's perceptions of these services.
Respondents with a high level of education, as is the case with commercial farmers (Table 1), tend to be more environmentally aware (Potgieter et al. 2019) and are therefore likely to acknowledge other ES categories provided by wattles as well as disservices.
Furthermore, socio-economic factors such as average household income and land tenure have been found to determine the level at which households depend on locally available natural resources (Shackleton and Shackleton 2006;Cumming et al. 2014b;Hamann et al. 2015;Daw et al. 2016;Shumi et al. 2021).High direct reliance on local ecosystem services is typically associated with low average household income and rural communities -most of which have limited access to market economies (Cumming et al. 2014a;Hamann et al. 2015;Shumi et al. 2021).This is especially the case for already marginalised smallholder communal farmers who have limited options and continue to rely on IAPs and may not be able to compensate for the loss of ES in the absence of such species.Consequently, the removal or loss of IAPs in the system is likely to have negative implications for their livelihoods as the landscape-scale supply of essential ES will be lost with the removal of the species (Daw et al. 2016;Shumi et al. 2021).Conversely, wealthier households have less direct dependence on locally available natural resources -and are able to cover their basic needs by purchasing these services from the market (Cumming et al. 2014a;Hamann et al. 2015).

Conclusions
As is the case with many invasive alien tree species, wattles were introduced to the region because of the benefits they provide or due to lack of trees in a particular environment.After introduction, invasive alien trees have also been introduced in other areas and often become an essential resource for local people (Shackleton et al. 2007(Shackleton et al. , 2015(Shackleton et al. , 2017(Shackleton et al. , 2019;;Ngorima and Shackleton 2019).It is the provisioning services offered by IAPs and their ability to spread and become a problem that results in conflict of interests, requiring substantive stakeholder engagement in terms of management.This study has provided insights into the stakeholder knowledge of wattles, perceived ES and EDS provided by wattle species in the upper Umzimvubu catchment and highlights the potential conflicting perspectives underlying the management of these species.The results show that wattles have both positive benefits and negative impacts concurrently.However, our findings demonstrate that the impacts of and the level of dependency on IAPs are not consistent among different social groups, i.e. communal land users and commercial livestock farmers.In this case, despite high dependence on locally available natural resources, including IAPs, communal land users perceived wattles to have more negative effects than benefits.The negative impacts were exacerbated by the high abundance of wattles in certain parts of the landscape and their impacts on other essential ecosystem services including water and grazing resources.In contrast, large-scale commercial farmers typically had very low abundance of wattles on their landholdingstherefore, were more likely to perceive wattles to have more benefits than negative effects.
Thus, these results highlight the importance of considering an approach that recognises the services and disservices of IAPs, and for divergent stakeholder views and attitudes to be included in management visions and decisions (Gaertner et al. 2012;Urgenson et al. 2013;Dickie et al. 2014;Potgieter et al. 2019).An improved understanding of land users' beliefs and attitudes can help identify the potential for conflicts regarding management of IAPs and thereby facilitate consultative and participatory processes to bridge differing perspectives and needs (Novoa et al. 2018;Shackleton et al. 2019bShackleton et al. , 2019d)).Management programmes for IAPs such as wattles need to be flexible and adaptive to their contentious status and the roles of these species in spaces such as rangelands, riparian zones, and homesteads.
project on Integrated Management of invasive alien species in South Africa, the Council for Scientific and Industrial Research through their Parliamentary Grant Project Fund, Southern African Systems Analysis Centre, and the Meat Industry Trust.

Table 1 .
Demographic, education and economic data for the two land user groups: commercial and communal farmers and livestock ownership.HH = household, LSU = Large Stock Units.

Table 2 .
Land users' knowledge and perceptions of wattle in the Umzimvubu catchment (percentage of households answering in affirmative).

Table 4 .
Land users' views of the negative impacts of wattle (ecosystems disservices).Values are the percentage of each land user group that suggested a particular disservice.

Table 5 .
Land user views of the ecosystem services provided by wattle species to communal and commercial households (values are the percentage of each land user group that suggested a particular service).