We analysed 90 field management actions and proposals of IAS received or implemented by the Regional Environmental Administration of Andalusia in 2004 to 2018 (Suppl. material 1: Table S1). Andalusia (87,268 km²) is the southernmost administrative region in Spain. The region has a population of ca 8.4 million inhabitants (Instituto Nacional de Estadística 2018) and harbours a wide variety of inland, coastal and marine habitats including 340 protected areas accounting for 49% of the total administrative area (CMAOT 2017). From east to west, Andalusia is characterised by an extensive coast (945 km, along both the Mediterranean Sea and the Atlantic Ocean) and several mountainous ranges with a maximum altitude in Sierra Nevada National Park (Mulhacén Peak: 3,481 m a.s.l.). The dominant climate is Mediterranean, with dry, hot summers, but arid, cold steppe conditions occur in the south-east (Kottek et al. 2006).

The actions encompassed a variety of taxonomic groups, life forms, and habitats (Fig. 1). Plant taxa represented 50% of the action plans analysed (45 out of 90), most of them (37) were terrestrial plants, whereas animals included 24 vertebrate and 21 invertebrate taxa. Regarding the habitat type, action plans included both aquatic and terrestrial habitats with dominance of inland waters and coastal habitats (Fig. 1). Some of the actions involved the same species in different localities with varying characteristics or different elimination methods. Therefore, the total number of actions analysed (N = 90) was higher than the total number of taxa managed (59 taxa: 28 plants and 31 animals).

Figure 1.

a Taxonomic groups represented in 90 action plans evaluated in this study. The taxonomic classification is based on BOLD (Barcode of Life Data system), a cloud-based data storage and analysis platform developed at the Centre for Biodiversity Genomics in Canada (http://www.boldsystems.org) b major habitat types represented in these action plans, following the EUNIS classification. Numbers inside the pie chart indicate the number of actions for each.

Action proposals and implemented actions aimed at eradication, control or containment of IAS in the field were made by collectives (NGO, professional associations), managers, scientists, or public institutions. Part of the proposals were selected and implemented as the result of coordinated decisions made by regional decision-makers, local authorities, specialised technicians, and rangers and based on documented and expert knowledge.

We visited managed localities annually to check if initially defined goals were accomplished. Eradication was considered achieved when no new individual was detected for five years. Control was considered achieved when the IAS abundance (e.g. plant mean coverage or captures per unit effort) decreased at least 90% after the action and maintained at least at 75% of the initial abundance for a minimum of 3 years. Containment was reached when the treated area was not reinvaded after the action. Accordingly, each action was classified into the following categories: (i) implemented actions that did not achieve the goals or specific objectives initially defined (n = 22) (hereinafter ‘unsuccessful actions’), (ii) implemented actions that achieved the goals or specific objectives initially defined (n = 24) (hereinafter ‘successful actions’), or (iii) proposals for management actions that were not implemented (n = 45) (hereinafter, ‘not-implemented proposals’).

After the execution of the action, we gathered information on the causes responsible for not-implemented proposals and for unsuccessful actions based on discussions with the different participants responsible for the execution of the action. With all the information, we generated a raw list of items associated with rejection or failure of each action. Then, the raw list of items was refined (e.g., redundancies removed) and transformed into an easy-to-understand checklist of 40 items organised in five blocks, namely ‘basic prerequisites’ (items #1–12), ‘IAS and invasive population features’ (items #13–20), ‘administrative features’ (#21–26), ‘methodology effectiveness, efficiency and impacts’ (#27–34), and ‘native ecosystem features and social perception’ (items #35–40) (Table 1). This checklist was applicable to any action or proposal independent of its specific objective (prevention, eradication, containment, or control), the taxonomic group or the habitat type. To validate the extent to which the checklist serves to discern between feasible and unfeasible actions, the checklist was systematically used to assess all the not-implemented proposals (n = 44) and unsuccessful actions (n = 22) previously used for the raw list preparation and also successful actions (n = 24). For assessments, the proposed methodology was defined by the elimination technique (biological, mechanical, chemical), the time or season of application, the number of applications, the final concentration (in case of a biocide), the frequency of monitoring and rounds of control, and methodological adaptations to minimise the impact and to promote ecosystem recovery. In the case of not-implemented proposals, we only evaluated the prerequisites (first 12 items in Table 1) except for one case that did not show any deficiency in these prerequisites. Finally, we built up a database including the action, whether the goal was achieved or not and deficient items for each action.

First, we assessed the relative importance of each checklist item for determining the failure of unsuccessful actions or rejection of not-implemented proposals by calculating the frequency of deficiency of each item (number of times that item i showed a deficiency * 100 / total number of actions of category j), being j either not-implemented proposals (j = 44) or unsuccessful actions (j = 22). Second, to evaluate to what extent the checklist discerns between feasible (successful, j = 24 in the previous equation) and unfeasible (unsuccessful) actions, both the frequency of deficiency of each item and the amount of items with deficiency were compared between successful and unsuccessful actions. We compared the total number of items with deficiency, considering both the number of prerequisites (items #1–12) and the rest of items (items #13–40) separately and all together. Since the data did not follow a normal distribution, pairwise Mann-Whitney U tests (Zar 1996) were used. Significant differences were considered when p < 0.05. The software Past® version 3.15 (Hammer 2001) was used.

Mistakes or unexpected outcomes constitute the basis for individual, professional, and organisational learning (Clark 2002), and their analysis may provide useful information to improve management practice. Management failures may be considered fiascos prone to be swept under the carpet. However, excepting for negligence or wilful ineptitude, failures may be simply mismatches between expectations and outcomes or the result of contingencies, uncertainties, or limitations of existing knowledge (Argyris and Schön 1978; Simberloff 2003). In this study, the analysis of not-implemented proposals and unsuccessful actions served as the basis for defining what items were related to feasibility of IAS management actions. These items included basic prerequisites and different topics related to the IAS biology, the administrative requirements, the methodology used to remove the IAS and the characteristics of the native ecosystem. This checklist is the basis to avoid the same mistakes in future actions.

The feasibility analysis of IAS management actions has received little attention compared to risk (Dana et al. 2014) and priority analysis (Nielsen and Fei 2015; Kerr et al. 2016; Courtois et al. 2018). While feasibility analysis is based on criteria related to the outcome of management actions, priorities setting seeks to identify where, how, on what, and when we should act first (Wilson et al. 2009). A step-by-step assessment of feasibility and priority-setting as the one we present in this paper can help to differentiate those unfeasible actions from those that, being feasible, are not prioritized, for example, due to a transitory lack of resources (Simberloff 2003). Accordingly, decision-making in biological invasion management should be considered as a three-step process: invasion risk analysis (e.g., Vilà et al. 2018; Copp et al. 2005; D’hondt et al. 2015), feasibility analysis, and, finally, priority-setting (Fig. 4).

Figure 4.

Basic steps in decision-making for managing biological invasions. The checklist proposed in this study focuses on step 2.

The consideration of up to 40 sources of deficiencies highlights the underlying complexity associated with the decision making for IAS management. The relatively high number of items may be the consequence of the number and heterogeneity of actions and proposals assessed but also of specific circumstances of management at the regional scale (e.g., administrative features).

The application of methodologies that are not consistent with the IAS management objective is a prerequisite that was commonly overlooked in unsuccessful actions (up to 54.5% of cases analysed), probably because the initial objective was too ambitious (e.g., eradication of invasive crayfish in rivers; Dana et al. 2010) or simply because the impact thresholds was rarely defined in control actions (Panetta and Gooden 2017). Consequently, it is difficult to establish the necessary removal rate or even to conclude whether the action was a success or a failure.

The absence of funding during the time frame to accomplish the management goals showed the highest occurrence in unsuccessful actions. This is a consequence of the current approach of conservation in the area of study, where IAS conservation actions are financed by biennial programmes whose long term durability is not guaranteed. Funding is a critical factor (Simberloff 2003) but may be a temporary obstacle as the planner or decision-maker can often search for different funding sources to implement feasible actions. Therefore, we decided not to include this item as a prerequisite to avoid stopping the feasibility assessment of an action at early stages.

Technical viability, social acceptance, legality, assumable impact, the possibility of restoration or the availability of specialised personnel have been considered as key factors related to feasibility of actions in several reports (Simberloff 2003; Pacific Invasives Initiative 2011; Dana et al. 2016), but the relative importance of each factor has not been previously assessed. IAS biology and life cycle are crucial aspects to define effective removal methods. For instance, annual dicotyledons and grasses may request very different biocides. The relevance of these aspects was considered in items #17, 18, 30, and 33 (Table 1). Besides, our analysis demonstrated that also non-biological aspects such as administrative features linked to the management action have often determined the likelihood of goal achievement. For example, management of IAS populations that are located inside private properties can involve additional difficulties when implementing an action plan. This was the case of Aedes albopictus in Málaga (Suppl. material 1: Table S1), which currently colonises a high number of private properties (e.g., saucers under flowerpots, and other small containers). In such cases, it is unfeasible to enter house by house to control potential breeding sites or colonies. Very often, part of the invasive population traverses private properties (e.g., Colocasia esculenta in Dos Hermanas, Seville or Ailanthus altissima in Aracena, Huelva) (Suppl. material 1: Table S1). In Andalusia, management of IAS in private properties requires of signed agreements between the public administration and landowners are necessary. Thus, the absence of will by any of the parties prevents a feasible, effective management of the invading population. Our work also revealed limitations in the efficacy of the existing approved methodologies, which often may lead to inadmissible impacts or to social rejection. This was the case of chemical control of Oenothera drummondii in coastal dunes of Huelva (García-de-Lomas et al. 2016). The analysis supports the need of investing in adequate (in terms of ecological indicators selected and design), adaptive, long-term monitoring (Lindenmayer and Likens 2009). Although directly related to goal achievements, the incorporation of long-term monitoring may involve an added challenge, as it requires extending the duration of the projects beyond possible changes of government. Again, long-term IAS management actions do not fit well with current short-term funding sources (Blossey 1999). The use of the proposed checklist prior to the implementation of IAS management proposals will encourage the definition of a monitoring programme in advance with an array of indicators consistent with a previously defined conservation goal and specific objectives (Lovett et al. 2007).

The significant differences found in the number, frequency, and identity of deficient items between successful and unsuccessful actions suggest that the present checklist discerns reasonably well between feasible and unfeasible actions. The use of a checklist prior to implementation of management proposals is of high interest to decrease the number of unsuccessful actions worldwide. The presence of basic prerequisites in some unsuccessful actions supports a systematic assessment of feasibility before action. The basic prerequisites (i.e., 12 simple items) we listed may be seem obvious, however, we decided to include such items in the checklist for three different reasons: (i) the analysis of feasibility starts at the planning stage and pre-requisites are essential to decide whether or not to implement an action proposal; (ii) planning may be done by people from different disciplines or different level of expertise on IAS, therefore, items that may seem obvious for some decision-makers may be overlooked by others; and (iii) the assessment of prerequisites is a quick step in decision-making in comparison with the major implications that the implementation of actions can have. For example, in the present study area up to four native species were confused with IAS, something that it is not rare among practitioners (Bardsley and Edwards-Jones 2006). Other rejected proposals included species whose origins of introduction remain uncertain. Such cases (e.g., Alopochen aegyptiacus, Alpheus ponteridae) may represent expansions at range edges without an apparent human intervention or in response to climate change (Gutiérrez 2003; Lindström et al. 2013).

In successful actions, the occurrence of deficiencies in two items suggests that certain items may or may not provoke failure depending on different circumstances. The valuation of a greater number of cases from different regions could help to distinguish items that are unambiguously related to non-feasibility (as the basic prerequisites seem to be) from others that may or may not motivate the action failure depending on additional factors (e.g., the planner experience). In this sense, conducting pilot tests or research projects on the use of novel methodologies are needed as a basis for improving the management of IAS.

To our knowledge, the present checklist is the most comprehensive ever done to date, as it includes a broad range of items integrating an interdisciplinary scope, useful to evaluate management of biological invasions in different habitats, involving taxonomic groups, and specific objectives. Therefore, the present checklist could be potentially used to detect weak points of IAS management actions before implementation in different parts of the world.