Introduction

Invasive alien species are among the leading causes of biodiversity loss worldwide (Vilà and Hulme 2017) and are a global threat to human well-being (IPBES 2023). Despite this, investment into their prevention and management across the globe is mostly reactive and rarely reflects the magnitude of the problem (Cuthbert et al. 2022; Early et al. 2016). While evidence of environmental impact should be sufficient to prompt action, economic projections are more useful when motivating for funding and are generally better understood by policymakers (Courchamp et al. 2017; Pyŝek et al. 2020). Environmental impacts, however, are often not monetised and the extent of the economic impacts of invasive alien species thus remains largely unknown (Vaissière et al. 2022). Moreover, where cost information is available, estimates are typically disparate or based on assumptions that reduce their reliability (Cuthbert et al. 2020). Estimates of the economic impact of invasive alien species have been made for selected nations (e.g., Pimentel et al. 2000, 2005; van Wilgen et al. 2022a) and sectors (e.g., Paini et al. 2016), though standardised and synthesised costs were unavailable for most contexts (Latombe et al 2017; Diagne et al. 2020a). To address this, Diagne et al. (2020b) developed InvaCost: a global-scale data repository of the reported economic costs associated with invasive alien species.

InvaCost has facilitated a broad range of cost assessments for biological invasions, covering several processes, habitats, and taxonomic groups at varying temporal scales and from national to global levels (Ahmed et al. 2023). Global costs of biological invasions run into trillions of USD (Cuthbert et al. 2022), but a recent InvaCost assessment of Africa reported the total cost of invasions to be a minimum of 18 billion USD (2017 values) from 1970 and 2020, with costs largely skewed towards eastern, and especially southern Africa (Diagne et al. 2021). The assessment further highlighted major gaps in cost estimates and emphasised the need for national cost assessments for individual African nations (Diagne et al. 2021); a gap similarly identified by the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services’ recent global report on invasive alien species and their control (IPBES 2023). South Africa provides probably the best opportunity for such an assessment in Africa, as the country is considered relatively data-rich compared to other countries in the InvaCost database (Diagne et al 2021). Furthermore, South Africa has one of the highest levels of invasion in the world (Turbelin et al. 2017), is a world leader in biological invasions research (van Wilgen et al. 2020a), and regularly reports on the status of biological invasions and their management (van Wilgen and Wilson 2018; Zengeya and Wilson 2020, 2023).

A recent review identified 71 invasive alien species in South Africa with documented impacts (van Wilgen et al. 2022b). Predictions of increasing invasions and an accompanying rise in the severity and extent of impacts in the country are a cause for concern (Zengeya and Wilson 2020, 2023), particularly considering invasion debt (i.e., the potential increase in impacts arising from existing invasions in the absence of strategic interventions, which is further compounded as new invasive species arrive and spread; Rouget et al. 2016). As a means to combat these threats, a programme on invasive species management—the Working for Water programme (WfW)—was initiated in South Africa in 1995 (van Wilgen et al. 1998a, b). Money spent by WfW on the control of alien plant species has been captured since 1998, with recent assessments showing an average management expenditure of ZAR347 million per year (adjusted to 2022 values of ZAR) over the last two decades (van Wilgen et al. 2022a). However, an important secondary objective of WfW was job creation, and so investments were not made wholly to manage invasive plants. This secondary motive may have increased the necessary management expenditure considerably.

Although previous studies have greatly improved the estimate of the money spent on managing biological invasions in the country, there remains no synthesis of the costs associated with biological invasions across all taxa, sectors, and cost types (i.e., both the observed costs, and estimated costs of management and damage using modelling approaches). This study provides an overview of the monetary costs of biological invasions in South Africa, by exploring: i) the trends in observed and modelled monetary costs over time; ii) the relationship between damage costs and management spending; iii) the distribution of costs across taxa; iv) the gaps in economic data for the country; and v) the implications for management and policy.

Methods

Data collection

This study is based on data from three sources: i) the InvaCost database; ii) a literature search; and iii) requests to stakeholders.

InvaCost is a ‘living’ global-scale data repository of the reported monetary costs associated with invasive alien species that was developed using a large pool of literature sources (Diagne et al. 2020b). The database distinguishes costs on the basis of whether they are observed (i.e., actual incurred costs) or modelled (i.e., cost projections or estimates). Modelled costs are not costs that have actually been spent or incurred, but are estimates from studies of the potential costs of damage or management due to invasive alien species. Modelled costs could reflect historical, current, or future estimations of economic impacts and are referred to as ‘potential’ costs in the InvaCost database; Diagne et al. (2020b); they could over- or under-represent actual expenditure depending on model accuracy. Cost entries in the database are further characterised using a range of descriptors. For example, cost type includes management costs and damage costs [encompassing economic damages and losses caused by an invasion, such as the value of crop losses or the cost of damage repair Diagne et al. 2020b)]; and impacted sector includes agriculture, forestry, authorities, and the private sector. Each cost entry also includes taxonomic and geographic contexts. For a full explanation of descriptive columns adopted in InvaCost, see Diagne et al. (2020b). For this study, InvaCost version 4.1 (January 2022) was used (https://doi.org/10.6084/m9.figshare.12668570). Extracted data included only entries where the “Official country” was “South Africa” to ensure that only costs specific to the country were retained.

Literature searches were conducted following the methods recommended by Diagne et al. (2020b). Search strings were constructed around the main concepts “Invasive Alien Species” and “Economic Impact” with the specific focus on South Africa and its currency (ZAR); see Supplementary 1 for a full list of search terms used. Finally, we engaged with relevant stakeholders to obtain additional information. Specifically: i) requests were made to various conservation and management entities across South Africa to provide data on costs where these existed; ii) a request for data was sent out through the Aliens-L mailing list server of the invasive species specialist group (ISSG) (www.issg.org) (10 June 2022); and iii) a call for data was made at the National Symposium for Biological Invasions (6–8 July 2022). All sources identified by 31 August 2023 were incorporated into an InvaCost data entry template as per Diagne et al. (2020b) and merged into a single dataset for data processing.

Data processing

To avoid taxonomic inconsistencies, all species names captured in the database were checked against the Global Biodiversity Information Facility (GBIF 2023, 28 August 2023) using the software package ‘rgbif’ (Chamberlain et al. 2023) in R Studio (R core Team 2022). To ensure that only species alien to South Africa were retained in the database, species were checked against the list of species alien to South Africa as provided in the latest National Status Report on Biological Invasions in South Africa (Zengeya and Wilson 2023) and all cost entries for species native to South Africa were removed. The species list in the national status report uses a range of taxonomic backbones (SANBI and CIB 2023; https://doi.org/10.5281/zenodo.8217197) and so both the national status report species list and the list of species from the database in this study were standardised to GBIF taxonomy to identify entries pertaining to native species.

In order to standardise the costs, all costs at the value and currency in which they were captured in the database were converted into 2017 values of USD following the InvaCost methods outlined in Diagne et al. (2020b). We then converted the 2017 USD values to 2017 ZAR and adjusted these by an inflation factor of 1.258 (the latest available conversion rate at the time of writing, https://data.worldbank.org/indicator/FP.CPI.TOTL?end=2022&locations=ZA&start=1960) to inflate these to ZAR values at the end of 2022.

Some cost entries in the database spanned a number of years and so costs were annualised to ensure an individualised cost for a single year. Costs were annualised using the expandYearlyCosts function of the ‘invacost’ R package (Leroy et al. 2022) in which each entry (a single row of data) is ‘expanded’ by the number of years for which the cost was reported. For example, a cost entry of ZAR1 million reported to span ten years was split into ten rows of ZAR100 000 per year. Although this assumes a steady annual spending rate, it allows for earlier costs to be adjusted for inflation and for the approximation of costs over time. Any projections of costs beyond 2023 were not considered.

Costs in the dataset were then classified according to their reliability, either as “High” if they were published in peer-reviewed or official (e.g., government) documents that used reproducible methods, and “Low” in the absence of reproducible methods (Diagne et al. 2020b). High reliability costs can include non-specific entries for some columns because reliability is based on the source material (i.e., peer-reviewed/official documents or those with reproducible methods), rather than its completeness. For instance, the study by Paini et al. (2016) provides a reliable, overall economic impact estimate for multiple invaders across various environments but does not specify the cost of invasions per environment type. Despite this, the results are considered reliable due to the soundness of the methods used. A reliability class is assigned to each cost by the researcher who enters the data into the dataset, and then double-checked by an InvaCost core team member to ensure consistency. If changes to reliability were deemed necessary, the revised reliability class was captured in the “Method reliability refined” column of the database, and the reason and reviewer were noted. The refined reliability class was retained for the purpose of analysis and only “High” reliability costs were considered for our analyses unless otherwise specified.

The data were divided into two distinct datasets that assessed observed and modelled costs separately. Overlaps in data from different sources were identified by assessing cost entries in terms of their source (stakeholder or publication name), taxonomy (species or genus), and the year for which the cost was reported. For example, WfW reported costs for managing species that were also reported by stakeholders funded by WfW. In these cases, we retained only the costs reported by WfW to avoid double counting. See Supplementary 2 for a workflow on the procedure followed for data processing in this study.

Descriptive analyses

All cost analyses were completed using the ‘invacost’ R package (Leroy et al. 2022) and performed separately for the two distinct databases—the ‘observed cost dataset’ and the ‘modelled cost dataset’. Cost analyses included assessments of cost type (damage losses or management costs), costs over time, and costs for different taxa (for a full list of analyses and associated package functions see Supplementary 3).

To make interpretation easier, the species-specific costs obtained in this study were summed and presented in groups. These groups were either i) genera which require similar forms of management, such as Acacia, Eucalyptus, Pinus and Prosopis [note: all species in the genus Prosopis sensu lato that are of major concern in South Africa have been reclassified under the genus Neltuma by Hughes et al. (2022); this proposal has not been adopted in GBIF as of 15 April 2024, so for the purposes of this study, the genus name Prosopis is retained]; ii) generalised groups (e.g., fishes, insects, grasses, and vines); and iii) ‘Diverse plants’ or ‘Diverse animals’ (i.e., broad categories where the kingdom but not the taxa were specified by the source document), or ‘Diverse taxa’ where no or multiple kingdoms were specified (i.e., the costs could not be parsed per kingdom, but were rather given as a total across kingdoms).

Lastly, to identify gaps in reporting on ‘money spent’ we developed a list of prominent stakeholders directly involved in invasive alien species control in South Africa. Stakeholders were identified using all the sources captured before the data were checked for overlapping costs, and through consultation with experts. The intent was to identify the most important stakeholders involved in the management of biological invasions in South Africa. We determined whether costs were captured for each stakeholder in the original dataset and, where captured, estimated the ‘completeness’ of the recorded cost. Data were regarded as complete where there were comprehensive records for a stakeholder that covered both recent and historical costs, or as incomplete where data were obtained from some but not all required sources, where data only reflected costs for recent years, or where data from a single source were supplied for some but not all activities. We assigned estimates of completeness and cross-checked these against the levels of confidence assigned to cost estimates in the latest National Status Report on Biological Invasions in South Africa (Zengeya and Wilson 2023).

Results

Of 68 source documents identified, 42 sources were from the InvaCost database, 16 were from literature surveys, and 10 were from stakeholders. Fifty-seven were considered to be of high reliability and retained for analysis, of which the majority provided management costs (Table 1).

Table 1 Total observed and modelled costs, divided into management and damage costs. The values presented (ZAR billion in 2022 values) only include data from highly reliable sources, and span 1960–2023. Observed costs were those where actual expenditure or losses were recorded, while modelled costs refer to estimated or extrapolated costs. The number of source documents here should not be totalled as some sources provided estimates for multiple cost types
Full size table

Types of costs

The majority of management (96%) and damage (79%) costs were reported as modelled rather than observed. The observed costs for damage amounted to ZAR52.7 billion, while the observed cost for management was ZAR9.6 billion, i.e., management spending was just less than 20% of the observed damage costs (Table 1). The modelled cost of management amounted to ZAR231.8 billion, which was considerably more than the modelled damage costs (ZAR195.2 billion), i.e., it would cost more to manage invasions than the damage that they caused. However, the damage costs captured here are likely to be substantially underestimated due to a lack of data (see Fig. 1b), and the real cost of managing invasions is expected to be considerably less than the cost of their damage. The figures reported are for reliable sources only. If less reliable sources were also included, costs were greater, but not drastically so (the observed cost of management rose by 1%, observed damage cost by 31%, modelled costs for damage by 38%, and modelled management costs by 29%). For the rest of the analysis, only sources scored as high reliability are considered.

Fig. 1
figure 1

Estimates of the cost of biological invasions to South Africa over time, with (a) management costs and (b) damage costs. The graphs differentiate between observed costs and modelled costs and reflect only the years for which data were available. Average costs are indicated every five years as data points connecting the trend lines, and the number of estimates informing annual averages are represented by the scaled circles. Observed damage cost data were too disparate to be connected with trend lines

Full size image

Management costs were reported from as early as the 1960s, whereas damage costs were only reported from the 1990s onwards. There were very few estimates of observed damage (Table 1), and so trends in the cost of damage over time could not be reliably inferred (Fig. 1). Estimated average annual damage costs amounted to ZAR742 million for observed costs and ZAR2.8 billion for modelled costs. Observed management costs increased steadily over time, with a notable increase after 1995 (from ZAR10 million to > ZAR250 million per year) and remaining at a similar order of magnitude thereafter with a slight decline recently. Modelled management costs also increased over time and were consistently an order of magnitude higher than observed management costs. For example, the average for observed management costs from 1960 to present was ZAR150 million per year, whereas modelled management costs have averaged at ZAR3.6 billion per year.

The majority (85%) of observed management costs were reported as reactive spending (i.e., interventions to control established invasions), with only a small proportion reported for preventative efforts (such as quarantine and border control) and research (Table 2). The remaining observed management costs were reported for varied efforts or did not specify the purpose of the cost. Modelled management costs reflected only reactive management interventions, with no cost predictions for other types of effort. In terms of environment type, the majority of costs were reported from the terrestrial environment, with fewer costs being reported for aquatic (marine and freshwater) and semi-aquatic (for organisms that live in both aquatic and terrestrial environments) environments (Table 2). Notably, many sources did not report on the environment type, particularly for modelled damage costs.

Table 2 A breakdown of the costs (ZAR billion at 2022 values) of biological invasions between 1960 and 2023 in South Africa in terms of the purpose of the cost, environment type, and taxonomic kingdom. Costs are presented as a comparison between management and damage costs and per implementation type, i.e., observed costs (realised costs), or modelled costs (estimated or extrapolated costs)
Full size table

Breakdown of costs by taxon

Observed management spending was overwhelmingly to control alien plants (Tables 2 and 3). Less than 1% of money was spent controlling 11 animal taxa (Table 3 and Supplementary 4). For plants, observed management expenditure was reported for 198 taxa in 11 broad groups (Table 3), 179 of which were reported at the species level (Supplementary 4). The following taxonomic groups accounted for the largest portions of observed management costs (Table 3): ‘Diverse plants’ (i.e., the breakdown of cost per individual plant species was not specified by the source), ‘Acacias’ (69% of which was reported for Acacia mearnsii; Supplementary 4), and various ‘Shrubs/ Herbs’ (36% of which was directed towards Lantana camara; Supplementary 4). Further spending was directed towards ‘Cacti/ succulents’ (largely focused on seven Opuntia species; Supplementary 4), various distinct invasive tree groups including mesquite (Prosopis species), eucalypts (Eucalyptus species), and pines (Pinus species), and ‘Other invasive tree’ species including Solanum mauritianum and Melia azedarach.

Table 3 The division of costs (ZAR billion at 2022 values) among taxa, differentiating between management and damage costs that are either observed (i.e., realised costs) or modelled (i.e., extrapolated or estimated costs). For ease of interpretation, the costs are broken down into groups representing: i) genera that tend to be similarly managed; ii) broad groups; and iii) unspecified categories where a single cost reflects multiple taxa (not distinguished by the source) under a specific kingdom (‘Diverse plants’ or ‘Diverse animals’), or where a single cost reflects multiple taxa under multiple kingdoms or where no kingdom is indicated (‘Diverse taxa’). The number of taxa included in each taxonomic group is indicated in square brackets behind each total (see Supplementary 4 for a full taxonomic breakdown)
Full size table

Most (55%) of modelled management costs were due to management projections for the polyphagous shot hole borer (Euwallacea fornicatus) and its fungal symbiont (de Wit et al. 2022). Modelled management costs were otherwise largely reflected by ‘Diverse plants’ and prominent invasive tree genera such as Acacia, Eucalyptus, Pinus, and Prosopis (Table 3).

Very few damage costs were reported, and for relatively few taxa (four at the species level and two at the genus level; Supplementary 4). Observed damage costs were reported for plants only, and largely for ‘Diverse plants’ and ‘Acacias’, under which Acacia mearnsii accounted for 68% of the costs (Table 3). The remaining observed damage costs were attributed to a single aquatic plant invader, Azolla filiculoides (< ZAR1 million). Aquatic invaders were mostly represented by modelled damage costs and for two species in particular: Azolla filiculoides (ZAR4.9 billion) and Pontederia crassipes (ZAR0.4 billion) (Table 3; Supplementary 4). The majority of modelled damage costs, however, were not species-specific. Several management initiatives targeting specific plant taxa (van Wilgen et al. 2020b) and animal taxa (Davies et al. 2020) are underway in South Africa. However, detailed breakdowns of expenditures for individual species are seldom provided. Any such data, if available, are grouped under ‘Diverse taxa’ in Table 3, with the overall cost of the initiative captured in the stakeholder costs in Table 4.

Table 4 Estimates (ZAR billion at 2022 values) of spending obtained from prominent stakeholders or groups of stakeholders known to be directly involved in the management of invasive alien species in South Africa. Completeness of estimates were cross-checked those provided in the latest National Status Report on Biological Invasions in South Africa (Zengeya and Wilson 2023)
Full size table

Stakeholder data contribution

Estimates of the money spent managing biological invasions (i.e., observed management costs) were largely incomplete (Table 4), resulting in an underestimate of management expenses in South Africa. Estimates for only one stakeholder (the Working for Water programme, Table 4) were assumed to be close to the real value, and constitute 86% of the ZAR9.6 billion in observed management costs reported in this study. Various stakeholders obtain funding for managing invasive species from the WfW programme, and the funding that these stakeholders reported was in all likelihood already included in the totals reported WfW. In addition, the majority of the estimates provide an incomplete reflection of money spent because of the current gaps in cost reporting. There are no reliable data on spending by several additional significant stakeholder groupings. These include research efforts, projects run by NGOs and volunteer groups, natural resource management programmes other than WfW that are run by the Department of Forestry, Fisheries and the Environment, other government departments (such as those responsible for agriculture and for water), and the private sector.

Discussion

In reviewing cost estimates for South Africa a few key points emerged: 1) The observed cost of damage of biological invasions is five times greater than the money spent on their management (the significance of this ratio should be carefully interpreted, see below); 2) The management spending required to effectively address all biological invasions is estimated to be an order of magnitude greater than actual money spent to date [the effectiveness of current interventions cannot be assessed due to a lack of regular monitoring (van Wilgen et al. 2020b) and most invasions are increasing in extent (Henderson and Wilson 2017)]; 3) Almost all of the money is spent on a few invasive species, and losses due to damage are only documented for a handful of invasive species, predominantly terrestrial plants; and 4) There are large gaps in cost reporting that limit our ability to assess how costs are divided amongst taxonomic groups and to determine which taxa have been most costly to manage, meaning that the estimates of economic impact (in terms of management and damage costs), although improving, remain incomplete and are inadequate for providing a basis for sound prioritisation. We discussed these points in turn below.

Damages exceed management costs

As with similar assessments completed for other countries (e.g., Australia, Bradshaw et al. 2021; Brazil, Adelino et al. 2021; India, Bang et al. 2022), this study found that observed damage costs substantially exceeded management costs. This appears to be a global feature related to the on-going increase in biological invasions worldwide (Seebens et al. 2017) as well as to insufficient awareness and funding (Cuthbert et al. 2022). The ratio of damage to management costs is not, however, an informative metric of whether the amount spent was appropriate. In general, management is only a good investment if damage is reduced by more than the cost of management (i.e., there is a positive return on investment). Understanding the return on investment requires a comparison of the costs with and without management, which necessitates some estimate of the efficacy of management. Decisions to invest in management should ideally be made in the context of the total resources available, the competing demands (both invasion-related and otherwise), and an evaluation of what type of returns and risks—based on industry standards or benchmarks—are viewed as acceptable and important. However, drawing reliable conclusions on management effectiveness in South Africa was not possible due to the absence of formal monitoring of intervention outcomes (Zengeya and Wilson 2023). What is possible, is to compare the amount spent with that predicted to be needed to bring invasions under control, and to evaluate the amount spent in the context of the overall trends in invasions in the country.

Management spending is insufficient to bring invasions under control

In 1998 it was estimated that ZAR5.4 billion (net present value in 1998 values of ZAR) would need to be spent over twenty years to bring invasive plants under control, i.e. ZAR272 million per year, Versfeld et al. (1998). Expressed in 2022 values of ZAR, this would amount to roughly ZAR1.04 billion per year. The money spent managing biological invasions between 1995 and 2023 was ZAR320 million per year (cf. van Wilgen et al. 2022a), i.e. only 30% of Versfeld et al.’s estimate in 1998, and only 4% of our modelled management costs. Unsurprisingly, and despite the successful implementation of classical biological control in a few cases, invasive plants have seen substantial increases in their ranges over this period (Henderson and Wilson 2017).

A further concern was that most spending was reactive rather than preventative. Money spent on prevention often has much higher returns on investment in the long-term (Leung et al. 2002; Reaser et al. 2020; Cuthbert et al. 2022). We know that there are significant initiatives to prevent potentially harmful alien species from entering the country (Saccaggi et al. 2021; Zengeya and Wilson 2023), but data on their monetary costs were not available. The underrepresentation of preventative costs has been highlighted in other national studies (c.f. Ahmed et al 2023), including that for New Zealand (Bodey et al. 2022) which is considered a world leader in proactive bio-security efforts.

A large focus on a few species

A substantial majority of the reported costs obtained in this study were for plants (95%) and largely for a handful of prominent invasive trees and shrubs (44% of the observed management cost was spent on Acacia, Chromolaena odorata, Eucalyptus, Lantana camara, Prosopis, and Pinus). Most of these taxa have major impacts (sensu Zengeya et al. 2017). The reported money spent on animals was < 1% of the total management expenditure and included only five taxonomic groups (birds, crabs, fishes, insects, and molluscs). These limited costs are in part likely attributed to: i) the relatively (compared to other countries) low number of invasive alien animals in South Africa (Picker and Griffiths 2017); ii) the high proportion (80 out of 107 species) of plants that comprise the most harmful species in South Africa (Zengeya and Wilson 2020); and iii) the large contribution of reported costs from WfW, which primarily focusses on alien plant control in South Africa, and specifically on invasive trees that affect water provisioning (van Wilgen et al. 2020b).

Globally, the amount spent on an invasive species is not necessarily proportionate to their area of occupancy, or their magnitude of impact (Ricciardi and Cohen 2007; Hulme et al. 2013, see also Latombe et al. 2022 for an analysis for South Africa), and in many cases the impacts caused are likely highly skewed to a few very harmful taxa. To establish whether a given focus is appropriate, information is needed on current impacts and the potential efficacy of different management options.

Large data gaps

There were major gaps in the reporting of the money spent managing invasive taxa in South Africa. For example, costs are represented in this study for only three fish species [Cyprinus carpio (common carp), Micropterus dolomieu (smallmouth bass), and Oncorhynchus mykiss (rainbow trout)] but 11 fish species that are alien to the country are known to be managed (Davies et al 2020). In fact, Davies et al (2020) reports on management efforts targeting 28 animal species in South Africa, while cost data were reflected for only nine animal species in this study. While a few invasive species may cause most of the damages, the absence of data on other taxa reduces the confidence in our estimate.

The total damage costs presented in his study (ZAR52.7 billion in observed and ZAR195.2 billion in modelled costs) are almost certainly substantially underestimated; further compounded by the rarity of impacts being translated into monetary terms (van Wilgen et al. 2022b). Furthermore, a portion of the estimates were informed by historical costs that have very little bearing on current costs. For example, Azolla filiculoides represented 10% of observed damage costs, but the species has been under complete biological control for many years and as such the species is no longer considered to be causing significant damage (McConnachie et al. 2003). Other observed damage costs correspond to impacts by Sirex noctilio on pine plantations of forestry sectors throughout the country (Hurley et al. 2007). Acacia mearnsii was also among the few species contributing observed damage costs, principally through reductions of surface streamflow as well as reductions in grazing areas in a cost–benefit analysis (de Wit et al. 2001).

While management costs are relatively simple to record, records of management effectiveness and long-term monitoring of implemented management schemes are rare, making it difficult to determine what the management has actually achieved. In addition, estimating damage costs requires research that is often based on a range of assumptions regarding the monetary value of environmental impacts As a result there are very few estimates of damage, which contributes to an underestimate of impact for the country and limits our understanding of changes in the magnitude of impacts over time. For example, according to expert opinion, at least 303 alien species have some form of negative impact [ranging from moderate to massive, sensu Hawkins et al. (2015)] in South Africa (Zengeya et al. 2017). In this study, however, damage cost estimates were available only for six taxa (Acacia mearnsii, Azolla filiculoides, Pontederia crassipes, Robinia pseudoacacia, Pinus, and Prosopis). The need to contextualise costs, and the highly fragmented estimations of economic damage, reinforces the call for improved estimates of impact in South Africa (Zengeya and Wilson 2020) and globally (IPBES 2023).

We also know that estimates of money spent by the private sector are missing; reported costs are rarely broken down into species-specific estimates, overheads, or by purpose (reactive, preventative, or research); and government spending is not fully accounted for. These are the same gaps noted by van Wilgen et al. (2022a) and multiple other InvaCost studies (Ahmed et al. 2023). Indeed, this study is subject to several limitations of the InvaCost database and of cost reporting infrastructures in general. First, cost information is missing for the vast majority of biological invasions because they are inaccessible, non-monetizable, or unstudied. Gathering cost information requires targeting different stakeholders at the subnational level, as per our study, but is still subject to data gaps due to fragmented data recording practices and unresponsiveness of correspondents. We call for greater openness of cost data through structured national reporting. Second, a large proportion of cost information is not reported with useful levels of detail. For example, records of spending to manage biological invasions often do not specify the particular taxa, regions, or time periods for which costs were incurred. This limits our ability to make data-driven recommendations to improve management actions. Thirdly, InvaCost typically aggregates cost data,but summing the costs of damage and management, and comparing these totals does not provide an indication on the success of management initiatives or their efficiency. It is clear that improved reporting on the efficacy and success of management actions is needed in a structured way, such that the true efficiency of investment towards biological invasions can be unveiled. A global repository of management actions and efficiencies could improve conservation and sustainability outcomes in the long-term.

Conclusion

South Africa is one of the most highly invaded countries in the world. It is evident that biological invasions have been a substantial cost to the South African economy since 1960. Current levels of management expenditure in South Africa appear to be insufficient to offset damages and bring invasions down to sustainably manageable levels. The actual costs of the impacts of biological invasions are likely to be greater than what is reported here, and they are likely to rise in future as invasions and their impacts grow. Limited quantifications of impact, combined with the lack of transparency in cost reporting, made it difficult to determine the true economic impact of biological invasions in South Africa. Regular engagement between the private and government sectors to collate information on target sites, priority species, and the available budget for interventions will be required to improve estimates and facilitate the prioritisation of interventions.