We relied on cost data recorded in the "InvaCost" database, which is the most up-to-date, comprehensive, and harmonized compilation and description of economic cost estimates associated with biological invasions worldwide (Diagne et al. 2020b). "InvaCost" has been generated following a systematic, standardized methodology to collect information from scientific articles, grey literature, stakeholders and expert elicitation. Each source was checked for relevance and the cost information was collated and standardized to a common and up-to-date currency in the database (i.e. 2017 US dollars). Each cost entry was depicted by a range of descriptive fields pertaining to the original source (e.g. title, authors and publication year of the reporting document), spatial extent (e.g. location and spatial scale), temporal coverage (e.g. time range and period of estimation), estimation methodology (e.g. method reliability and acquisition method) and the nature of cost (e.g. type of cost and impacted sector). All methodological procedures and details for data search (e.g. literature review), collation (e.g. cost standardization), validation (e.g. method repeatability) and improvement (i.e. corrections and inputs) are described elsewhere (Diagne et al. 2020b, 2020c). This updatable and publicly available data resource provides an essential basis for worldwide research and policymaking targeting IAS (Diagne et al. 2020a).

To get the most complete and up-to-date dataset of the reported economic costs attributable to biological invasions in Africa for the last fifty years (1970–2020), we used the most recent version of the "InvaCost" database (version 3.0; Diagne et al. 2020c). This updated database integrates and refines cost information (9,823 cost entries; 64 descriptive fields) from two other repositories generated in the frame of the broader "InvaCost" initiative (Diagne et al. 2020a), and which include cost data collected from multiple sources and languages throughout the world (Angulo et al. 2021). Using this latest version of "InvaCost" allows us to limit potential gaps in existing literature as well as common language biases due to the exclusive consideration of English in research (Haddaway et al. 2015; Konno et al. 2020; Angulo et al. 2021). Using successive filters in the descriptive fields of the database (i.e. "Geographic region" and "Country" columns), we identified and then extracted all economic costs which were exclusively associated with African countries. Therefore, any cost entry that concerned non-African territories located within African regions (e.g. La Reunion Island) was not considered. We carefully checked the data to correct or remove any potential mistakes or duplicated cost entries. Our final database (hereafter called "starting dataset") consisted of 696 cost entries (Suppl. material 1).

We homogenized our "starting dataset" so that each cost entry – realized over a single year, a period of less than a year, or a cost reoccurring over a series of years – corresponds to a single-year estimate, which is repeated over the number of years during which the cost occurred. For this purpose, we used the "expandYearlyCosts" function from the "invacost" package (Leroy et al. 2020) in R version 4.0.2 (R Core Team 2019). This operation allowed us to expand each cost entry over its actual or estimated duration time, which was derived from the difference between the first year ("Probable starting year adjusted" column) and the last year ("Probable ending year adjusted" column) of the recorded cost. Consequently, this process removed any cost entries occurring over an unspecified time period in the database. Nonetheless, this step was necessary to ensure accurate estimations of the cumulative and mean annual costs of invasions over time. The expanded version of our "starting dataset" contained 4,259 cost entries (Suppl. material 1).

To ensure a realistic and conservative synthesis of cost estimates reported for Africa, we applied two successive filters to this "starting dataset" (Suppl. material 2). The filters used were based on the categories listed for a set of descriptive fields in the "starting dataset" (see Suppl. material 3 for a detailed description of the fields). First, we kept only “observed” costs (rather than “potential” costs, under the "Implementation" column); second, we retained only economic estimates classified as “high” reliability (rather than “low” reliability, under the "Method reliability" column). Subsequently, all cost estimates for the year 2020 were excluded since these estimates were “potential” and/or of “low” reliability. Our final dataset (hereafter referred to as the "conservative subset") contained 2,302 cost entries between 1970 and 2019 (Suppl. material 4).

We categorized the cost data according to different descriptive fields (hereafter called “descriptors”) in our datasets. First, we grouped countries into the five geographical regions defined by the United Nations geoscheme for Africa (available at https://unstats.un.org/): “Western Africa”, “Southern Africa”, “Northern Africa”, “Middle Africa”, and “Eastern Africa” (the latter also includes countries in the Indian Ocean) (Suppl. material 5). Second, we considered information on the typology of the costs ("Type of cost merged" column) that groups each cost estimate under “damage” (i.e. economic losses due to direct and/or indirect impacts of invaders, such as yield losses, damage repair, medical care, infrastructure alteration or income reduction); “management” (i.e. economic resources allocated to actions that aim at avoiding the invasion or dealing with more or less established invaders, such as prevention, control, research, eradication, education or mitigation policies); or “mixed” (i.e. when a single cost simultaneously includes both “damage” and “management” components) category (Suppl. material 6). Third, we determined which sectors were impacted by the reported costs (using information from the "Impacted sector" column); cost estimates that were not allocated exclusively to a single sector were classified under the “mixed” category. Fourth, economically harmful species were classified into different major ‘organism types’ based on information from the "Kingdom", "Phylum", "Class" and "Environment" columns: “Animalia” (i.e. insects, mammals, birds), “Plantae” (i.e. aquatic plants, terrestrial plants, semi-aquatic plants), and “Virus”. For each descriptor, cost estimates that could not be unambiguously and exclusively assigned to one category were labelled as “diverse/unspecified”.

Our purpose was to draw a complete, as well as a robust picture of the cost of biological invasions throughout the African continent. We used the following R packages - ggplot2 (v.3.3.2, Wickham 2011), rnaturalearthdata (v.0.1.0, South 2017) and networkD3 (v.0.4, Allaire et al. 2017) – to generate an array of graphical representations for each descriptor of interest.

First, we used the "starting dataset" to describe the full cost information that was available. To do this, we investigated how individual cost estimates and their source materials (i.e. peer-reviewed articles and grey literature) were distributed over time. We focused on both the number of cost estimates and the total costs accumulated between 1970 and 2020. The latter was obtained by summing all cost estimates provided in the "cost estimate per year 2017 exchange rate" column of the expanded version of the "starting dataset" (Suppl. material 1). We systematically distinguished the proportions of the cost estimates that were of “high” versus “low” reliability, as well as those that were actually realized (i.e. “observed”) or just merely predicted (“potential”).

Second, we used the "conservative subset" to investigate how the cost amounts were distributed across geographic regions, types of costs, impacted sectors and taxonomic groups for the period 1970–2019. Finally, we investigated the trend of costs over time using two strategies.

The first strategy included an estimation of both the cumulative costs incurred between 1970 and 2020 (i.e. the sum of all cost estimates provided in the "cost estimate per year 2017 exchange rate" column of the expanded subset; Suppl. material 4) and the mean cost amount for each decade over the same period (i.e. obtained by dividing the total cost of each decade by ten years).

The second strategy consisted of modelling the long-term trends in economic costs of invasions by fitting models of annual costs as a function of time. Indeed, a reliable estimation of the average annual costs over time should take into account (i) the dynamic nature of costs, (ii) the time lags between the real occurrence of the costs and their reporting in the literature (called ‘publication delay’ hereafter), (iii) the heteroscedastic and temporally auto-correlated nature of cost data, and (iv) the effects of potential outliers in the cost estimates. For this purpose, we implemented the "costTrendOverTime" function ("invacost" package; Leroy et al. 2020) on the log₁₀-transformed cost estimates per year, which allowed modelling the trend of costs over time with a range of linear and non-linear modelling techniques while enabling a comparison of the respective outputs of all models generated. As statistical intricacies inherent to econometric data did not allow for a priori identification of the most relevant modelling technique to apply, we relied on ‘ordinary least squares regressions’ (linear, quadratic), ‘robust regressions’ (linear, quadratic - R package "robustbase", Maechler 2020), ‘multiple additive regression splines’ (MARS, R package earth, Milborrow 2017), ‘generalised additive models’ (GAM, R package "mgcv", Wood and Wood 2015) and ‘quantile regressions’ (quantiles 0.1, 0.5, 0.9, R package "quantreg", Koenker 2019). To optimize model performance, all models were calibrated following a robust linear regression using cost data as the response variable and time as a predictor, which allowed to identify obvious outliers in the years of cost occurrence. To account for potential data incompleteness due to the ‘publication delay’, we excluded from model calibration all cost estimates from 2014 onwards because they constituted obvious outliers with a sudden drop of two orders of magnitude. We confirmed these outliers by investigating robust regressions calibrated on all data, which had set the weights of years above 2013 near to zero (Suppl. material 7). Model discussion was based on the assessment of the predictive performance across models (Root-mean-square deviation, RMSE) as well as the goodness-of-fit measure (variance explained). Moreover, combining these diverse modelling procedures offers strong support for the observed temporal trends and provides consistent model outcomes. As this approach is highly data-demanding, we only applied it to the African continent without disentangling types of costs, regions, sectors or taxonomic groups.

During the 1970–2020 period, economic costs associated with biological invasions in Africa were obtained separately for 33 countries (i.e. 4 from Middle Africa, 3 from Northern Africa, 3 from Southern Africa, 10 from Western Africa, and 13 from Eastern Africa; see Suppl. material 5 for further details). The expanded dataset contained 4,259 cost estimates collected from 103 source documents from both the grey (n = 39) and scientific peer-reviewed (n = 64) documents (Suppl. material 1). Except for sixteen documents that were written in French, all reporting documents used English as the primary language. This shows a clear language bias despite all efforts made for collecting cost information reported in 15 languages in the updated "InvaCost" database (version 3.0; Diagne et al. 2020c). We also showed that since the 1970s, the number of both the cost estimates and source documents steadily increased over the years, along with the total estimated cost amounts (Figure 1). This is despite a slight decline in the number of cost estimates over the last decade, which might be the result of a time lag between the occurrence of the most recent costs and when they were reported in the literature (Figure 1).

About 86% of the cost entries (n = 3,653) collated were only incurred in Southern Africa (Table 1; Figure 2). Far behind, Eastern and Western Africa were the most represented regions, with 287 (7%) and 155 (3%) cost estimates, respectively. These patterns are influenced by a small number of countries with cost entries in each region (Suppl. material 5). Within Southern Africa, South Africa reported the majority of costs (together, the two other countries within this region, Lesotho and Swaziland, were only associated with 33 of the 3,653 cost entries recorded); more than 60% of the costs recorded in Eastern Africa were associated with three of the ten reporting countries (i.e. Kenya, n = 71; Uganda, n = 48; Tanzania, n = 53); and costs recorded in Western Africa mostly concern Benin (n = 78). The other regions harbored fewer than 15 cost entries, with Middle Africa reporting the smallest number of cost data (n = 6). Cost estimates associated simultaneously with two or more countries belonging to (at least) two distinct regions (i.e. “diverse/unspecified” category) consisted of 146 cost entries (3%) in the "starting dataset".

Except for Southern Africa and “diverse/unspecified” regions, more than two thirds of the recorded cost estimates were considered as having been empirically observed in each region (Figure 2; Suppl. material 8). Conversely for Southern Africa and “diverse/unspecified” regions, respectively about 42% and 66% of the reported data comprised potential costs. Given that Southern Africa is the most represented region in our dataset, this means that a substantial portion of the cost estimates recorded throughout the continent (n = 1,807 out of 4,259) were derived from extrapolation or modelling approaches rather than true observations (Figure 2; Suppl. material 8). Finally, the reported cost data mostly exhibited a high degree of method reliability (Figure 2; Suppl. material 8). Indeed, the proportion of cost entries resulting from highly reliable cost estimations range between 75% (for Northern Africa) and 98% (for Eastern Africa), suggesting that most cost estimates were obtained from relevant estimation methodologies (Figure 2; Suppl. material 8).

Considering all cost entries in our "starting dataset", the accumulated cost of IAS in Africa reached a total of US$ 78.9 billion between 1970 and 2020 (see Table 1 for a detailed cost breakdown by region, taxa, sector, and type of cost).

Biological invasions were estimated to cost a minimum of US$ 18.2 billion in Africa over the period 1970–2019 (Figure 3; Table 1). These conservative costs were not equally distributed across regions, between types of costs, or among sectors and taxa (Table 1).

Our findings undoubtedly illustrate that invasions incur substantial costs to national African economies, most of them being vulnerable and already weak (Lekunze 2020). The reported financial burden accumulated to a conservative total of approximately US$ 18.9 billion (annual average of US$ 303 million) between 1970 and 2019, reaching an estimated annual average of US$ 2.6–8.6 billion in 2019. However, these costs could seem relatively low compared with those from other continents such as North America (Crystal-Ornelas et al., submitted in the current issue), Europe (Haubrock et al., submitted in the current issue) or Asia (Liu et al., submitted in the current issue). On the one hand, this discrepancy likely reflects the strong geographical imbalance in research intensity and financial capacities (Early et al. 2016; Sooryamoorthy 2018) rather than the actual spatial distribution of the costs of invasions. Also, invasion-associated issues may not be perceived as a priority for many African countries where investments in many primary structural needs (e.g. roads, infrastructures, fight against extreme poverty, and building sustainable education and health systems) are still greatly needed (African Development Bank Group 2018; Adamjy et al. 2020). This may logically translate into reduced academic studies and operational programs on biological invasions. Accordingly, IAS were dramatically understudied in Africa compared with other parts of the world (Pysek et al. 2008) – with a notable exception being South Africa (van Wilgen et al. 2020). Moreover, wealthier or developed regions are also those with higher documented invasions and associated impacts (Bellard et al. 2016; Essl et al. 2020). In addition, the significant difference in both values of the money and price levels between areas (e.g. labor costs for similar management actions are likely cheaper in most African countries when compared with those in Europe or North America), might be contributing to increment the observed discrepancy in reported costs between Africa and other regions. Indeed, relevant monetary comparisons at macroeconomic scale require reliance on indicators such as the purchasing power parity (but see Gosh 2018), which reflects the relative purchasing power of different currencies between countries and over time. However, such reliable comparisons are still prevented by very limited information on this indicator for most countries and/or years (Diagne et al. 2020b). We therefore have to also acknowledge that some African countries invest a substantial amount of resources towards the appropriate management of invaders – as evidenced by the increasing successful control of invasive alien plants in several African countries such as South Africa and Namibia (Stafford et al. 2017). The cost estimates presented here are substantial and obviously detrimental for the African continent. An eloquent illustration comes from comparing our estimated costs with the African Union’s budget (https://au.int/). In 2019, the expected minimum cost of invasions was more than three times higher than the entire budget available for this continental organization (i.e. US$ 681.5 million in 2019). Therefore, we can safely assume that our conservative estimate of invasion costs largely exceeds the actual funding capacities of the largest regional organizations that support socio-economic development in African countries. Moreover, the highest average value estimated for 2019 (US$ 8.6 billion) is greater than the individual gross domestic products of the seventeen less developed countries across the whole continent.

Worryingly, we found that the economic costs of IAS in Africa are steadily increasing over time without any signs of slowing down, reflecting the continuous increase in the number of IAS worldwide (Seebens et al. 2017). A set of complementary reasons may explain this temporal pattern, and/or why we should not expect any deceleration in invasion costs in the years to come. First, there is a growing awareness of the impacts of invaders as well as a burgeoning interest in reporting their economic impacts along with an associated increase in management actions during recent years (Dana et al. 2013; Simberloff et al. 2013; Diagne et al. 2021). Scrutinizing our dataset reveals that while the first monetized impacts of IAS in Africa dates back to the 1970s, the first document providing IAS costs was published in 1991 (Suppl. material 4). All cost estimates recorded between 1970 and 1985 stemmed from only three sources which reported costs for South Africa and Indian Ocean islands, suggesting that the research interest in other African regions has been growing rapidly over the past few decades. Second, the ongoing globalization and climate change synergistically accelerate the opportunities and rate of species invasions almost everywhere, and Africa should be no exception (Seebens et al. 2015; Faulkner et al. 2020). Third, Africa has been shown to be among the key areas at risk for future invasions by at least 86 of 100 of the world’s worst invasive species, with most of these invasions likely to cause severe socio-economic impacts (Faulkner et al. 2020). The role of the socio-economic changes faced by most African areas is undebatable in this particular context. In particular, Africa is currently experiencing a rapid rate of urbanization that is only second to Asia, with an urban population that may at least triple between 2010 and 2050 to reach 1.339 billion people (Matamanda and Nel 2020). Evolutionary socio-ecological features associated with this urbanization process can promote invasion success (Klotz and Kühn 2010; Sinka et al. 2020). For example, the dense and various networks of exchanges of goods and people can create repeated opportunities for the introduction of a wide range of exotic species and a shift towards biotic and abiotic conditions can greatly favor opportunistic, adaptable and prolific species. In that sense, empirical evidence supporting this process was recently provided for different invasive taxa, including rodents in Western African countries (Garba et al. 2014; Hima et al. 2019) and cultivated ornamental plants in South Africa (Potgieter et al. 2020). Unfortunately, these examples constitute only a few among several others which demonstrates the changing context-related spread of harmful invaders throughout the continent (Early et al. 2016). The increasing costs reported here sound alarming, yet there are several reasons which can explain why these costs are likely much higher than we estimated.

A number of logistical, methodological and cost-intrinsic factors may have prevented the capture of the complete diversity – and thus the full amount – of costs. Costs can remain hidden and/or underestimated due to (i) the unclear status of some invasive species (Jarić et al. 2019), (ii) inaccessible source materials (e.g. grey literature; Adams et al. 2017), and (iii) methodological (e.g. inadequate extrapolations; see Jackson (2015) for a detailed synthesis) as well as ethical issues (e.g. monetary perception of ecosystem services; Meinard et al. 2016) that impair the evaluation process (Bradshaw et al. 2016; Hoffmann and Broadhurst 2016). For instance, costs from well-known economically harmful invaders could have been overlooked simply because they failed to be captured when building the "InvaCost" database (Diagne et al. 2020b, 2021). Moreover, costs are inherently complex and heterogeneous. As a consequence, misconceptions from the lack of reporting consistency in invasion science (Colautti and MacIsaac 2004; Richardson et al. 2020) likely lead to overlooking some cost estimates (Dana et al. 2013). Furthermore, we made highly conservative choices when generating our "conservative subset" in order to ensure reliable cost assessments, which led to consider only 2,302 out of 4,259 cost entries from the "starting dataset" (e.g. for some countries, such as Morocco and Angola, all costs were unreliable or potential, and were therefore discarded following our filtering procedure). More broadly, the skewed cost distribution (see below) revealed taxonomic, sectoral and geographic gaps that may contribute to our underestimations of the actual economic burden of IAS in Africa.

We showed that economic costs are widely but not evenly distributed across regions. Indeed, most cost estimates were associated with a single country (i.e. South Africa), which is internationally recognized as a pioneering and frontline country for research and management in invasion science (van Wilgen et al. 2020). It has been shown that South Africa comprised about two-thirds of the quantified research effort in the field across the African continent (Pysek et al. 2008). A similar unevenness has been found in relation to aquatic invasion costs, where South Africa dominated costs reported on the African continent (Cuthbert et al. 2021). The rich history of species introductions, higher economic capacity (compared with most African countries) and long tradition of large-scale conservation actions in this country may also contribute to this trend (Foxcroft et al. 2020; van Wilgen et al. 2020). Another reason for the higher cost estimates for South Africa comes from the fact that South African studies often rely on extrapolation-based approaches to provide economic estimates of IAS impacts (van Wilgen et al. 2020). Yet, these potential cost data were filtered out of our "conservative subset", explaining why the total cost for Southern Africa significantly decreased after filtering to reach an amount comparable to costs from Eastern Africa (Table 1). The high costs reported for Eastern Africa may be – at least partially – linked to the research activity of the Centre for Agricultural Bioscience International (CABI; www.cabi.org) and work done by the United Nations Environmental Program (UNEP) which have their regional centers located in Nairobi (Kenya).

Moreover, we may expect higher costs for the other regions than those reported here. For instance, Northern Africa has 13 cost entries recorded for only five species, while 157 species are listed in the GISD for this region. Also, Western African countries are historically and contemporarily threatened by a broad variety of biological invaders which is beyond insects and plants that were mostly reported for this region. Indeed, the succession of large international seaports along coastal cities (e.g. Abidjan, Cotonou, Lagos, Dakar) and the parallel development of the extensively urbanizing corridor from Côte d’Ivoire to Nigeria (i.e. the so-called Abidjan-Lagos corridor) may greatly facilitate the introduction of several vertebrate and aquatic invertebrate invaders (Habitat 2014; Bellard et al. 2016; Hima et al. 2019). Consequently, we advocate for increasing research effort towards the economic costs of biological invasions, mainly in the understudied regions where the costs are likely to be much higher than those currently reported.

Across the African continent, most of the reported costs were mainly driven by very few taxa, among which the costliest included three insect pests: the spotted stem borer (C. partellus), introduced in Eastern Africa in the 1930s, is suspected to be the most serious pest of maize and sorghum in Eastern and Southern Africa (Yonow et al. 2017); the fall armyworm (S. frugiperda) has now been reported in 45 African countries since its first report in Western Africa in late 2016, and is a known voracious consumer of more than 80 crop species of strong nutritional and socio-economic utility (CABI 2020); the tomato leafminer (T. absoluta) has now invaded 41 of the 54 African countries since its introduction in Northern Africa in 2007 (Rwomushana et al. 2019). Given the broad distribution (beyond the limited spatial coverage of each of these species in the database) and biological characteristics of these invasive pests, we can safely assume that their economic impacts largely exceed the monetary costs reported here (see Eschen et al. 2021 for a recent extrapolation attempt, using information obtained from the literature and stakeholder consultations). In this study, T. absoluta and S. frugiperda were still among the five costliest IAS, together with the invasive plants Eichhornia crassipes, Lantana camara and Prosopis juliflora. Typically, the over-representation of agriculture in the reported costs may reflect the direct influence of economic priorities and societal realities in political and research agendas. Indeed, building sustainable agriculture for food security is a priority for most African countries and their economies sometimes strongly rely on food production (Pratt et al. 2017; Wiggins et al. 2010). Given that Africa is highly vulnerable to invasions by exotic pests (Early et al. 2016; Paini et al. 2016), it would seem logical that local authorities invest more on research in the agricultural sector, especially given the very limited economic resources of these countries to fight against invaders (Early et al. 2016; Faulkner et al. 2020).

Focusing solely on major and well documented (and often mediatized) agricultural threats may have an ‘umbrella’ effect on other less visible but harmful invaders for which the costs may be unsuspected or neglected. Indeed, only a small spectrum of species (about 15%) from those recognized as invading Africa in the GISD were reported here. This strongly corroborates a previous assumption that only a small portion of invaders have been economically analyzed (Aukema et al. 2011). Besides, many of the species recorded in our dataset can have a broader range of economic impacts. An eloquent example of this is provided by rodent species (e.g. Rattus spp. and Mus musculus) for which only management costs were reported here. Yet, invasive rodents are responsible for significant damage costs to humans (e.g. medical care due to zoonotic infections, losses from consumption of stored food stocks, destruction of infrastructures and electric supply networks) (Drummond 2001; Han et al. 2015), as recently illustrated in different parts of Africa (Leirs et al. 2010; Dossou et al. 2020).

Therefore, it is evident that research intensity is closely connected with societal and economic realities in African countries. Hence, strong collaborations should be established and/or amplified between scientists, authorities, various sectoral stakeholders as well as local communities to understand and deal with the multidimensional issues raised by biological invasions.

Our results clearly highlight that IAS are a significant economic burden in Africa and the costs of these invasions are largely driven by damage induced by invaders. Monetary estimates associated with managing invasions were scarce and the amounts spent were essentially restricted to South Africa and North Africa. This pattern reflects a missed opportunity, since one of the rare examples we have for the entire African continent (i.e. the biological control of the cassava mealybug) suggests a benefit-cost ratio of management of 200 at minimum (Zeddies et al. 2001). If the lower investment in management is real (and not only under-reported), we hypothesize that this lower investment in management could possibly reflect a lack of awareness and/or insufficient capacities and means from national authorities and decision-makers facing invasions. Yet, invaders represent a significant shortfall for low income countries. In addition, this enormous financial toll represents only part of all the impacts incurred from invasions, which are also associated with major ecological and health issues (Kumschick et al. 2014; Ogden et al. 2019). Our findings should therefore be interpreted as an urgent call for considering invasion management as a major piece of sustainable development in these developing countries (Larson et al. 2011; Shackleton et al. 2017), in parallel with many of the Sustainable Development Goals defined by the United Nations and which serve as political, socio-economic and ethical guidelines globally (Sach et al. 2019).

We argue that efficient strategies towards management require cross-disciplinary and cross-sectoral efforts within and between scientists, decision-makers, stakeholders and civil society (Courchamp et al. 2017; Vaz et al. 2017; Richardson et al. 2020). Indeed, if research is necessary to produce knowledge about origin, impacts and spread of invasive species, a supportive political environment is critical to develop and implement long-term policies in Africa (Evans et al. 2018; Adamjy et al. 2020). Moreover, it has been shown that insufficient appreciation of socio-political context, non-existent or perfunctory public and community engagement, as well as unidirectional communications were associated with conflictual invasive species management (Crowley et al. 2017). Since an invasive species can be viewed as detrimental, neutral, or even beneficial in society, people who benefit from IAS may differ from those who suffer the costs (Estévez et al. 2015; Novoa et al. 2016b; Adamjy et al. 2020). As such, applying principles and concepts of sustainability science to invasion research and management should represent a key opportunity within the African context (Gasparatos et al. 2017; Tortell 2020). In addition, scientists and stakeholders need to engage in a joint paradigm for the concerted implementation of context-adapted policies and concerted implementation of management measures at relevant scales (Novoa et al. 2018).

The adoption and implementation of biosecurity measures appear particularly relevant for African countries where economic capacities are often limited. This is particularly true since many invaders introduced from other continents are also spreading within Africa in unpredictable directions (Faulkner et al. 2017; Keller and Kumschick 2017). The ultimate objective should be to act against invaders before they are introduced or become widely established, since controlling widespread invasions is often impossible or may require a high amount of resources. Furthermore, these actions should be applied at regional scales to balance expenditures and improve efficiency of actions (Faulkner et al. 2020). To date, such examples of regional cooperation are still scarce across the continent and the few attempts are restricted to South Africa (e.g. Shackleton et al. 2017). Our findings stress the need for integrating and/or reinforcing the place of biological invasions in the official agendas of African regional organizations.