Developing policy-relevant biodiversity indicators: lessons learnt from case studies in Africa

There is an increasing need for monitoring schemes that help understand the evolution of the global biodiversity crisis and propose solutions for the future. Indicators, including temporal baselines, are crucial to measure the change in biodiversity over time, to evaluate progress towards its conservation and sustainable use and to set conservation priorities. They help design and monitor national and regional policies on biodiversity; they also feed into national reporting on international agreements such as the Convention on Biological Diversity and the Sustainable Development Goals. We analyse the methodological approach of five small African projects resulting from a call to promote indicator development, improve monitoring capacity and strengthen the science-policy interface in the field of biodiversity. We compared their approach to existing guidance provided by the international community, specifically the Biodiversity Indicators Partnership. To this end, we assess whether internationally recommended steps are effectively applied to national/local biodiversity monitoring in selected developing countries. We also present lessons learnt from workshop interactions between partners involved in these projects. Through our pilot projects we identified data availability and data accessibility, together with the involvement of stakeholders, as critical steps in indicator development. Moreover, there is a need for a better awareness and a wider application of the indicator concept itself. Hence, training of key actors both in the policy and science spheres is needed to operationalize indicators and ensure their continuity and sustainability. We hope that these case studies and lessons learnt can stimulate and support countries in the Global South to formulate policy-relevant biodiversity indicators.


Introduction
Earth is experiencing a mass extinction event (Ceballos et al 2017, Tilman et al 2017. To address the current biodiversity crisis, understand its evolution and propose solutions, monitoring schemes are needed. Baselines are crucial to understand temporal changes in biodiversity, to evaluate progress towards its conservation and sustainable use and to set conservation priorities (Mihoub et al 2017). They help design and monitor regional and national biodiversity policies, feed into national reporting on international agreements such as the Convention on Biological Diversity (CBD) and the Sustainable Development Goals (Bubb et al 2010, Joppa et al 2016, and provide input for global reports such as WWF's Living Planet Report, IPBES assessments or CBD's Global Biodiversity Outlook. Jones et al (2011) define monitoring as a process that includes collection of primary biodiversity data, feeding of data into indicators, and public dissemination of their spatial and temporal trends. Indicators can be defined as a 'measure based on verifiable data that conveys information about more than just itself', hence they are purpose-dependent (Biodiversity Indicators Partnership (BIP) 2011). Global efforts-e.g. by the Group of Earth Observations Biodiversity Observation Network (GEO BON), the Biodiversity Indicators Partnership (BIP) and the ad hoc Technical Expert Group on Indicators for the Strategic Plan for Biodiversity 2011-2020-identify and develop indicators for assessing progress towards the CBD Strategic Plan and its Aichi Targets, intended to guide the efforts of the international community to address the biodiversity crisis. Although these indicators could be used both at the global level and by parties and international organisations, recent studies show that the disaggregation of global indicators for use at the national level is rare, one of the reasons being that they are often intended for different types of users and purposes (Bubb 2013, Convention on Biological Diversity (CBD) 2017, Geijzendorffer et al 2017, Han et al 2017). Bubb (2013) strongly argues for aggregating nationallevel indicators to create a global index, instead of disaggregating global indicators to the national level: indicators should primarily be designed for national needs and be tailored to nationally adopted targets.
Despite significant efforts to develop biodiversity indicators, global indicators still fall short of an accurate and comprehensive picture of biodiversity, and are often underused in decision-making. Biodiversity data richness is skewed towards the poles, while species richness and conservation relevance is greatest in the tropics (Collen et al 2008, Wilson et al 2016. Biodiversity data are lacking in the tropics because (1) data is either scarce or of poor quality due to poor research efforts and (2) available data may not be accessible for lack of collaboration between institutes or because data are not adequately presented, published or centralized. Unfortunately, there is a lack of collaboration and communication at the science-policy interface between the data holders and the policymakers who need to use the information. The lack of cross-sectoral or inter-ministerial collaboration is a major barrier for using and mainstreaming biodiversity information (overview: Stephenson et al 2017). Moreover, a lot of technical barriers prevent efficient use of data, e.g. with regard to data-sharing platforms or availability of software or internet for e.g. remote sensing applications (de Klerk and Buchanan 2017). Despite efforts to disaggregate global indicators for use at national level (Biodiversity Indicators Partnership (BIP) 2011), major gaps persist in the Global South, beyond just data quality and availability: relevance to national context, capacity, funding, infrastructure, technologies, and governance (Han et al 2014). Also, the willingness to generate indicator data varies between countries (Bubb 2013, Han et al 2014. Finally, awareness of indicators is insufficient, limiting their use . This results in biodiversity not being adequately taken into account in decision-making and national planning and reporting and hinders countries in monitoring their biodiversity for their National Biodiversity Strategies and Action Plans (NBSAPs) (Collen et al 2008). This can then lead to a negative feedback loop: a lack of policy-relevant scientific research leads to inappropriate policies because these are not sufficiently science-based and vice versa.
To improve capacity for indicator-based monitoring and stimulate collaboration between scientists and policymakers, the CEBioS programme, funded by the Belgian Development Cooperation, developed a competitive call for projects in line with the Measuring, Reporting and Verification (MRV) approach mostly known from carbon management and forestry (Vanhove et al 2017, CEBioS (Capacities for Biodiversity and Sustainable Development) 2018). As with MRV within the typically multilevel framework of carbon management, we expect the integration of information to be key to success but also a challenge for MRV of biodiversity and biodiversity policy (see Korhonen-Kurki et al 2013).
The five African case studies resulting from this call aimed to develop biodiversity indicators to support reporting on their NBSAPs. While the countries in question do conduct monitoring of biodiversity, biodiversity policies or ecosystem services, it is generally focused on specific services or indicator taxa and does not routinely feed into national policy-relevant indicators. Examples include a study on the conservation importance of sacred forests in Central Benin (Ceperley et al 2010), and a recent project on monitoring anthropogenic impact in the littoral zone of Burundian Lake Tanganyika (VLIR-UOS 2018).
We discuss lessons learnt from project workshops. Moreover, the quality of the indicators and the methodological approach of these projects are analysed. We compared it with guidance provided by the international community, specifically the BIP (Bubb et al 2010), to assess whether recommended steps are effectively applied to national/local biodiversity monitoring in developing countries. We hope that these case studies and lessons learnt, and the comparison between countries, can encourage and support other African countries to develop policy-relevant biodiversity indicators.

Methods
The case studies A competitive call was launched for year-long projects aiming to improve monitoring capacity and strengthen the science-policy interface in the field of biodiversity in selected partner countries of the Belgian development cooperation (Vanhove et al 2017, CEBioS (Capacities for Biodiversity and Sustainable Development) 2018). In line with the abovementioned recommendation of Bubb (2013) to focus on indicator development at the national level, it aimed to develop and test an approach, for national or sub-national monitoring and reporting (Measuring, Reporting, Verification) of biodiversity and biodiversity policy, that is better tailored to countries in the Global South, specifically Africa. The call required submission of detailed proposals, ensuring genuine demand from partners in the South, and supporting local institutions to stimulate local development of capacity needed for long-term monitoring (see also Collen et al 2008).
The call asked applicants to develop, apply or improve one or more biodiversity indicators that should serve to set a baseline for one or several NBSAP objectives and/or to measure the status of such objectives. While national-level indicators were preferred, projects at a sub-national level were also eligible. The steps promoted under this call to convert policy-relevant biodiversity data into trends that are communicable to decision makers are described in figure 1 and table 3. A jury of in-house and external experts selected four out of nine projects submitted. A fifth selected project was terminated halfway for administrative reasons, but its results are included here. Table 1 and Vanhove et al (2017) give an overview of topics and the type of partners involved.
All four selected countries-Benin, Burundi, Democratic Republic of the Congo and Moroccorecently updated their NBSAP and intend to develop indicators for reporting. To bridge the gap between data collection by academics and data use by decisionmakers, a 'tandem' approach was chosen, in which a scientific institution (university, government research centre) collaborates with a National Focal Point for the CBD or for the Clearing House Mechanism.
Training, collective restitution and exchange of best practices in workshops Two representatives from the CEBioS programme and two from each project-one scientist and one decision-maker-met twice to share best practices and methodological concerns. An opening workshop in Brussels, Belgium (September 2015), provided opportunities to discuss and address expectations, priorities, methodological concerns and requests from the partner countries. Participants received training on elements of MRV and on subjects relevant to the topics of their projects. Sessions addressed the Global Biodiversity Information Facility (GBIF), indicator development, Geographic Information Systems, economic valuation of ecosystem services, and ethnobiology. During the closing workshop organized in October 2016 in Cotonou, Benin, participants discussed best practices, lessons learnt, conclusions and perspectives after the projects ended. One session also focussed on communicating with decision-makers, and co-produced policy briefs (Akouehou et al 2016, Akpona et al 2016, Mayundo et al 2016, Nzigidahera and Habonimana 2016, which were disseminated through follow-up awareness projects in 2018. The Beninese teams hosting this workshop offered sessions on additional topics relevant to MRV in general and the participants' specific projects, including species distribution modelling and community-based conservation. The capacity building components of the call are summarized in table 2, together with the main challenges they aim to address. Lessons learnt from the projects, on the basis of workshop interactions with project participants are summarized in the Discussion.

Comparison between projects' methodologies and international guidelines
The five projects represent different stages of biodiversity indicator development (figure 2). They differ in choice and prioritization of indicators, scale, Figure 1. Approach promoted in the call for projects: the case studies are commonly developed by scientists and decisionmakers to fill the gap between biodiversity data and their use for policies, strategies and action plans. trends in the coverage, state, representativeness and efficiency of approaches linked to protected areas, as well as other local policy interventions.
National Table 2. Correspondence between the MRV approach used in the climate change field, the Biodiversity Indicator Development Framework and the MRV approach used in the present call for projects. Capacity building components of the call are also summarized, together with the main issues for MRV for biodiversity they aim to address.  'Verify (V) by periodically subjecting the reported information to some form of review or analysis or independent assessment to establish completeness and reliability.' • 10-test and refine indicators • Steps 1-5 can also be seen as verification when they follow step 10 Review the developed indicators, and adapt them if appropriate

• No permanence and quality review of developed indicators
Call for indicator follow-up projects in 2019 development and use of indicators, and level at which data is collected (local or national). The quality of the indicators is assessed using criteria identified by the BIP for 'successful' indicators, which are key factors in determining whether an identified indicator is taken up and produced over time (Biodiversity Indicators Partnership BIP 2011). The methodological diversity between projects is analysed using the Biodiversity Indicator Development Framework (BIDF) developed by the BIP (figure 2), whose mandate is derived primarily from the CBD and decisions taken by its Parties. As shown in table 2, indicators are at the heart of the MRV approach used in the climate change field, in the call for projects addressed in this paper, and the BIDF and strong similarities exist between these three approaches, which oriented our choice of using this internationally recognized framework for our analysis.
The BIDF is promoted globally as methodology to develop successful biodiversity indicators and is 'intended to be used as a guide to better understand what is involved in producing biodiversity indicators and to help plan the most appropriate development process for each situation' (Biodiversity Indicators Partnership BIP 2017). It describes ten key steps divided into three themes to support indicator development.
The first four steps under the 'purpose' section refer to actions needed for selecting successful indicators. The next four steps are essential to generate the indicators and fall under the 'production' section. The last theme-'permanence'-comprises two steps for ensuring indicator continuity and sustainability (Biodiversity Indicators Partnership BIP 2011). Although most project partners had heard of this framework, most of them only implemented it partly. We analyse whether and how each step was taken into account in each project.

Results
Between two and four indicators were developed per project, sometimes associated to a set of sub-indicators (see supplementary material, available online at stacks. iop.org/ERL/14/035002/mmedia), that address different NBSAP objectives and the corresponding Aichi targets (table 3). The most common themes are traditional knowledge, habitat loss reduction, protected area expansion and the conservation of ecosystems and ecosystem services. The involvement of different people at each step reflects the level of stakeholder engagement in the project.
The analysis of the quality of the developed indicators against the BIP criteria for successful indicators, shows that indicators are rarely 'based on available data', 'championed' and 'used' (table 4) (see supplementary material for the analysis of each indicator).
Examples of indicators that were fully developed are provided, along with their interpretation, in figure 3. The interpretation refers to the relevant national target or the key-messages for decision making.
An overview of the flow of each of the five projects compared against each step of the BIDF is provided in table 5, as well as the contribution of the CEBioS capacity building to the process. Not all steps are applied by each project, and the order in which they are implemented differ among projects.

Discussion
Developing indicators: a plea for more awareness of the indicator concept and for stakeholder involvement The case studies demonstrate that, even to professional scientists and policymakers working on biodiversity monitoring, the concept of 'indicator' is insufficiently understood and applied in a standardized or operational way. While the small number of funded projects, and thus the modest number of indicators in this study, is a limitation, the focus was on the process of piloting an indicator-based MRV approach rather than on the resulting indicators. For example, in view of the above-mentioned challenge of integrating information across different levels, our participants found it instructive to bring together teams collecting data locally with those focusing on country-wide It is divided into three themes: purpose (1-4, in red), production (5-8, in purple) and permanence (9-10, in green).   -Two people involved in the project coordination and implementation indicator development. Also, the needs expressed even in these small projects already aptly illustrate that the working conditions and availability of skilled people limit the ability to uptake capacity building and take advantage of funding for biodiversity monitoring in developing countries (Danielsen et al 2005). We cannot agree more with Jones et al (2011) that the 'constant reinvention of indicators is costly in terms of time, cost, and public engagement'. To strengthen capacity for indicator use, indicator development initiatives and South-South cooperation should be better promoted (Convention on Biological Diversity (CBD) 2010). Exchanging good practices between countries with similar ecosystems or similar contextual challenges (such as governance or the lack of funding and capacity) enabled the workshop participants to acquire and use more interoperable and relevant methodologies.
Steps 5-9 of the BIDF, referring to the 'Production' and 'Permanence' sections, are implemented by    Table 5. Summary of the inclusion of the BIDF key steps in each project and the contribution of the CEBioS capacity building to the process. The framework is separated into three themes (see figure 2): purpose (actions needed for selecting successful indicators), production (essential stages for indicator development), permanence (mechanisms for ensuring indicator continuity and sustainability) (Biodiversity Indicators Partnership BIP 2011). Workshop with key informants to analyse the main ecosystems for which the trends are important for the country (after step 2).
Identification but no consultation of stakeholders interested in traditional knowledge (local communities as knowledge holders) and users of data (scientists and competent authorities).
Identification of the stakeholders that could be interested by the indicator, after steps 2 and 5 were implemented. However, they were not consulted due to a lack of financial resources.

Identify management objectives and targets
Selection of 5 priority NBSAP objectives.
-Objectives selected before the other steps, while writing the project.
Choice of the national objectives that address traditional knowledge (because of expertise of the researcher in this field) and for which data are available or can be collected.
The selection of the objective was a first step to answer the following criteria: focus of the call, available data and expertise of the implementing research institute.

Determine key questions and indicator use
For each priority objective, key questions were stated.
-Key questions determined during the workshop.
Three key questions unilaterally developed by the researchers to address the selected objective. Unilaterally developed by the researchers based on research subject and data to be collected. Five indicators were proposed to address key Selection of a set of indicators based on available data, and to respond to the selected national objective A3.

Calculate indicators
Calculation methodology developed in the reference sheet of each indicator (see step 9 and supplementary material).
Vulnerability index calculation, diversity of used species, mapping of rare medicinal plants.
A set of indicators to report on (1) size, state and vulnerability of ecosystems and natural habitats; (2) trends in the influence of unsustainable agricultural practices on species typical to a natural forest habitat;(3) trends in distribution, state and sustainability of ecosystem services (see supplementary material for all indicators).
The number of NTFPs identified and their use by local and indigenous communities.
Trends in coverage, condition, representativeness and effectiveness of protected areas and other local actions, represented by a set of indicators (see supplementary material).
Capacity building contribution: The closing workshop organized by CEBioS and gathering all project representatives included training and brainstorming sessions about the best way to communicate indicators, leading to the co-production of policy briefs and the development of their dissemination strategies. Follow-up awareness projects were supported by CEBioS to disseminate key-results and produced policy briefs.

Communicate and interpret indicators
Information detailed in the reference sheet of each indicator (see step 9) about: (1) representation type suggested, Capacity building contribution: The closing workshop organized by CEBioS and gathering all project representatives included a general reflexion about the best way to develop monitoring and reporting systems for the projects in particular, and for biodiversity indicators in general. The main outcomes of the discussions are summarized in this paper.

Develop monitoring and reporting systems
Operationalization of the selected indicators through indicator reference sheets to guide and support indicator development and their ongoing production (see supplementary material for an example). Training of key actors for collecting data. A policy brief summarizes the selected indicators and the content of the reference sheets, and recommends their inclusion in existing monitoring processes.
A policy brief gives recommendations for the creation of a national structure for future reporting about traditional knowledge, for the follow-up of the value chains and the inclusion of all concerned actors.
A database was created for the continuous updating of data. It contains all figures gathered in the form of time series and graphs for three categories: species, ecosystems, protected areas. Maps are also centralized. The activity of continuously updating the database was integrated in existing national biodiversity monitoring. The indicators will be included in future biodiversity reports.
A reference sheet for the indicator was developed using the BIP fact sheet model.
An information sheet for the indicator was developed using the BIP fact sheet model.

Test and refine indicators with stakeholders
Application of the methodology in the indicator reference sheets to collect data on indicators -Validation workshop with stakeholders.
--all projects, which can be related to the fact that they correspond to the 'M' (measure or monitor) and 'R' (report) aspects of the capacity building process (see table 2). Through these steps, indicators are produced and communicated, but do not necessarily fulfil the requirements for a 'successful' indicator. Most projects failed to meet the criteria 'based on available data', 'championed' and 'used' (table 4); possible reasons are further investigated in the discussion. Indicators should also be purpose-dependent; hence their development or selection should start with identifying the issue or policy need concerned (Biodiversity Indicators Partnership (BIP) 2011, figure 2). Therefore, the first BIDF steps, corresponding to the 'Purpose' area, increase stakeholder involvementand consequently ownership-and sustainability of the indicators. In Benin, the existing framework for stakeholder dialogue on biodiversity issues was an asset for project A to prioritize objectives and indicators in a participatory, multidisciplinary approach. It enabled the development of a strategy for gathering available and new data. These stakeholders must be involved throughout the process as they are crucial for data mobilisation, including localizing and accessing existing information, and assessing data reliability. They are key in prioritizing indicators, as the lack of data, funds and capacity may mean a limited number of indicators can actually be used in NBSAP reporting, as illustrated in the prioritization methodology of project A. As a result, three of the developed indicators under project A were included in the Beninese NBSAP and will be used in national reporting. Likewise, it is the only project to have developed indicators that are positively assessed for the 'championed' criterion (see supplementary material). Conversely, in projects C and E, stakeholders were involved at a point when the indicators were already defined, thereby limiting their effective use for reporting. This again illustrates the importance of the order of steps and the involvement of stakeholders from different sectors and backgrounds. This approach reflects field reality, data availability and policy needs and promotes cost-efficiency, investment in positive outcomes, and validity of indicators (e.g.

Creative solutions for data availability and accessibility
When asked about the key steps identified in their MRV approach, responses vary between the project teams. The above-mentioned emphasis on prioritization of targets and indicators was echoed by projects A, D and E. Conversely, project C underlined the importance of gathering quantitative, reliable and comprehensive data, while project B also stressed that the development of a methodology for data processing is crucial. This reflects the weight that many projects give to acquiring new data. It also reflects substantial gaps in (existing) data availability and accessibility in the South (Chambers et al 2017), as demonstrated by the poor results for the 'based on available data' BIP criterium for most developed indicators (table 4). Moreover, data may be of insufficient quality (weak sampling design and methodology), badly processed or interpreted, presented in a user-unfriendly format or insufficiently harmonized, rendering replication and comparison difficult and time-consuming (Costello et al 2013, Stephenson et al 2017. Therefore, scientific capacity-resources or skills-for data collection and processing is essential and should be continuously supported, as under this CEBioS programme. The lack of data for certain indicators forced some projects towards indicators for which data exist. For example, for the indicators aiming at measuring the effectiveness of protected areas management of project E, many parameters were required to ensure a scientifically sound indicator that is responsive to the issue of interest-two main BIP criteria for 'successful' indicators. However, data on total species richness, habitat cover or total vegetation cover were missing (see supplementary material), although they are crucial to assess the ecological performance of these areas, in terms of representation and maintenance of key biodiversity features and hence ecosystem services (Gaston et al 2008). Indicators had to be built using parameters for which quantitative data are available: trends in the coverage of protected areas and some selected conservation actions. The same happened for project C, where the focus on marshes and wetlands was driven by the availability of data; savannahs and Lake Tanganyika and surroundings were excluded because of data scarcity. It is hence clear that, often, feasible indicators will not align with the (often topdown) priority needs identified by authorities or other stakeholders.
Considerable information gaps may already be resolved through existing data collection efforts (Geijzendorffer et al 2016) and sources of biodiversity data, such as herbaria or other collections (Greve et al 2016), grey literature or environmental impact assessments (Hugé et al 2017), and tapping into currently underused information providers. Citizen scientists (including indigenous communities), observations by tourists (Pimm et al 2015), and paraecologists can greatly improve the flow of biodiversity information to all users, from local stakeholders to the academic realm (Schmiedel et al 2016. Establishing a baseline based on historical data allows to detect subsequent trends, as was the case in projects A, C and E. Therefore, for the subsequent MRV call, which focused on the DR Congo only, we decided to capitalize on existing data, rather than collect new information . There are a number of global initiatives to provide access to data (e.g. World Database on Protected Areas, Integrated Biodiversity Assessment Tool or the GBIF) but these resources are often poorly known and not easily discoverable or available in the global South (in particular where English is not widely used). Apart from language, capacity for database management, software availability and internet connection quality are barriers for the use of such databases or tools at the national level. Despite being promoted during the opening capacity building workshop, only one project (project E) used such databases. Efforts should be made to harmonize these existing databases and communicate about them and their relevance/applicability for national monitoring and to lower access thresholds by promoting technology transfer through easy and user-friendly tools accessible to a wide range of users.
The BON (Biodiversity Observation Networks) concept developed by the GEO BON network promotes networks which may be regional, thematic or national; they engage in harmonized and interoperable biodiversity monitoring and in making biodiversity data and data products publicly available (Group on Earth Observation (GEO BON) 2017a). Such networks should be developed keeping in mind the policy priorities of the countries and international bodies involved, so that their observations contribute to management, conservation and sustainable use of biodiversity and ecosystem services (see Walters and Scholes 2017). There are no such BONs in Africa so far and we would strongly recommend their creation.
Opportunities at the science-policy interface and the challenge of permanence Unequivocally, all project participants stressed the need to repeat monitoring over time, becoming a permanent process. Assessments of biodiversity and ecosystem services are often too costly to be sustained once project funding ends in developing countries (Danielsen et al 2003;Schmeller et al 2017) and elsewhere (Watson and Novelly 2004). Hence, they do not become established monitoring, nor do they inform on policy-relevant trends. This may be a challenge for our projects funded for a single year. However, the influence of local research funding should not be underestimated, nor should the importance of funding for data mobilization over funding for research leading to peer-reviewed output (Meyer et al 2015). A solution may be locally-based monitoring. This has been demonstrated to be sustainable in developing countries after foreign funding ceases, although external input for capacity building or data analysis remains beneficial (Danielsen et al 2005). Several recommendations that also emerged from our MRV workshops, such as institutionalization across different levels of governance, and involvement of local communities, are known to contribute to the sustainability of monitoring schemes (Danielsen et al 2005).
Producing indicator reference sheets following the BIP model, as most projects did, is an important tool to guide and support the development and use of indicators. However, to ensure permanence, monitoring should be included in governance and management structures, (national) plans and strategies. Monitoring results need to be integrated and centralized in a harmonized database to encourage continuous collection and processing of data, and its development should involve both researchers collecting and aggregating data, and policymakers. The lack of an efficient and well-established national coordinating structure that is responsible for coordinating national biodiversity indicators (Bubb 2013) was emphasized by all participants (see table 4: 'Championed by an institution responsible for the indicator' and consequently the 'used' BIP criterion). Most indicator development projects in these countries are sporadically supported by international donors for short periods and no local institution is responsible for the indicator's continued production and communication. For example, the indicator 'Percentage of forest land annually converted to other categories of land use' developed by project A (figure 3(a)) was selected during the indicator prioritisation process but was not integrated into annual monitoring and evaluation systems of the Beninese forestry department. During the last 12 years, this department was 'housed' within four different ministries. Consequently, the indicator was not standardized and data collection was neither systematic nor mandatory, which led to a data gap from 2006 to 2016. More awareness and lobbying to integrate this indicator systematically into the monitoring and evaluation system of the ministry responsible for forests should improve data availability in the future. Similarly, whereas the quality of technical capacities seemed less of a problem, administrative capacity and good governance and management related to MRV was shown to be more problematic for most REDD+countries, requiring capacity building for a wide range of actors (Ochieng et al 2016).
Continuously updating biodiversity indicators obviously presents a financial challenge that poses a threat to the permanence of these monitoring initiatives. However, networking and gap-filling between existing initiatives is much more cost-efficient than setting up new structures. Agreements between government agencies, NGOs, academic institutions and even the private sector can fulfil many roles related to the collection of data, calculation of indicators, and their communication to users (Scholes et al 2008, Bubb 2013. Whether new entities are created or not, scattered, unharmonized data and lack of collaboration across governance levels have been identified as major obstacles to national MRV systems in a carbon context, where information flow across levels is crucial (Korhonen-Kurki et al 2013).
The capacity development model elaborated by CEBioS stimulates and facilitates collaboration between scientists and policymakers , hence between entities that often evolve and work independently. It can be seen as an action research component: 'a learning process between stakeholders to understand the problems at stake, each other's roles, the policies and mandates and the access to information' (Janssens de Bisthoven 2015). This contributes to 'decompartmentalisation' of different types of actors and of different levels , which is essential for mainstreaming biodiversity information into decision-making (Stephenson et al 2017). Participants positively evaluated their interactions, and the better understanding of each other's expectations and role in the process. They also identified additional positive outcomes, such as the joint exploration of available data: their potential, where to find them and who can work with them. The involvement of scientific institutes ensured the use of quantitative data, collected in a scientifically sound way. Therefore, participants felt that the chosen priorities and the ensuing co-produced policy briefs calling policymakers to action had received scientific legitimacy.
Policy briefs may serve as knowledge-brokering tools or 'digests' to address issues of biodiversity and other topics related to sustainable development. Ideally, key results and recommendations should be presented orally, accompanied by the policy brief. Designing policy briefs requires an understanding of the audience's needs, and a personal engagement between policymakers and scientists (Balian et al 2016). Policy briefs are indeed a co-produced output of joint prioritization exercises of scientists and policymakers (Sutherland et al 2011). Such a joint project may reconcile academic and policy motivations (Young et al 2014), ensure shared ownership, and make the application of results in the policy realm more likely (Sutherland et al 2011). It is therefore unsurprising that the 'tandem' approach of our capacity development concept proved particularly rewarding to participants, as it provides a basis for long-term dialogue at the science-policy interface. In this way, our capacity building contributes to two of the 'core capabilities' that, according to the European Centre for Development Policy Management, come with effective capacity: the capability to relate, and to achieve coherence . Therefore, it would be commendable for funding agencies to stimulate this (Neßhöver et al 2013) by specifically and systematically allocating funds to the identification and formulation of monitoring projects together with stakeholders from the policy-side. Project E is an example that seems very relevant to NBSAP reporting and monitoring but that could not apply any of the 'purpose' steps due to the lack of funding, hence it could not involve stakeholders in the process, risking that Moroccan national reporting on protected areas will not consider its results.
Participants stressed that indicators should not be the only focus of messages to policymakers. Other useful communication tools that they identified are the economic valuation of ecosystem services or the use of scenarios, projections and models to simulate the impact of future policies (e.g. Visconti et al 2016, Kubiszewski et al 2017). For example, regarding the use of scenarios related to biodiversity and ecosystem services in policy planning in Africa, there is a clear lack of capacity and of inter-African knowledge transfer (Biggs et al 2018). Another reason to not only focus on indicators is the notion that progress towards certain Aichi Targets is not purely hampered by a lack of quantifiable data, but rather by institutional challenges, namely the mismatch with existing commitments. This is the case e.g. for targets related to access and benefit sharing (Hagerman and Pelai 2016). In this respect, projects B and D explicitly focused on traditional knowledge, and hence bridged the local level of data acquisition in this field with national indicator development. Such subnational projects have indeed been proposed, for MRV in a REDD+context, to serve as case studies for a more participatory approach that also considers social impact (Korhonen-Kurki et al 2013). The poor coverage of the BIP criteria (table 4) by the indicators developed under project B and C also reflects that indicators are not always the best suited tools to address specific questions. The weak 'scientific validity' of the indicators indeed refers to the lack of widely accepted scientific knowledge and understanding of traditional knowledge (see also Hernández-Morcillo et al 2013, for an overview of the difficulties surrounding indicators for cultural ecosystem services and the importance of involving stakeholders). These projects call for authorities to set up structures to access scientifically sound information on the sustainable exploitation of plants and to play a role in training in and conservation of related traditional knowledge. This illustrates the importance of better communication between scientists and policymakers in this field. The approach taken by Burundi, where a formal agreement between the scientific community and traditional practitioners was reached regarding access and benefit sharing, may serve as a pioneering example of best practice (Janssens de Bisthoven et al 2017).

Conclusions and perspectives
Within the methodology for indicator development in the South, through our pilot projects we identified data availability and, probably even more important, data accessibility, together with the involvement of stakeholders, as critical steps. Indeed, while we agree with Wilson et al (2016) that certain countries in the South, and even more their researchers, are underrepresented in research and policy, our case studies demonstrate that numerous and diverse scientific (monitoring) activities exist, but are unfortunately poorly accessible, visible or harmonized. Moreover, at a more basic level, there is a need for a better understanding and a wider application of the indicator concept itself.
Various recent capacity development initiatives are being developed to assist governments in developing more effective and timely biodiversity monitoring and policy responses, e.g. 'BON in a Box' (Proença et al 2017, Group on Earth Observation (GEO BON) 2017b). Our MRV pilot programme demonstrated the need to step up capacity development, to initiate an indicator-based approach and to sustain long-term monitoring. This fits into the trend of focusing on policy-relevant biodiversity monitoring when setting up new capacity building programmes in the Global South. Of the main existing schemes of biodiversity data collection sensu Proença et al (2017), our MRV projects seem closest to the 'intensive monitoring programmes', with capacity building focusing on professional development. In this scheme, Schmeller et al (2017) highlight the importance of transferring technical expertise and strengthening institutions. It was clear from the onset that our participants valued training, especially in ground-truthing, economic valuation of ecosystem services, the indicator concept in itself, and a variety of information technologies to collect, access and share data . Given limited resources to initiate capacity building in biodiversity monitoring, identifying local and topical priorities relevant to biodiversity policy instruments together with national governments, is a way to increase impact and sustainability (Henle et al 2013. With our MRV approach, we subscribe to this approach. We recommend that in future capacity development efforts, the perspectives that we identified from an African viewpoint, in particular the demonstrated potential of collaboration between scientists and policymakers, and the South-South collaboration and South ownership, be promoted. Adopting this approach would necessitate changes in information flow, incentives and structures across multiple scales at the institutional level and including policy and funding agencies. Therefore, we propose that setting up a biodiversity MRV system could foster wider transformational changes in biodiversity governance, as has been suggested in a REDD +context (Korhonen-Kurki et al 2013).

Acknowledgments
Sincere gratitude goes to A De Kesel, S Dessein, A Heughebaert, T Huyse, P Lejeune, F Malaisse and N Witters for their participation and contributions to the 2015 MRV opening workshop and to M Agarad, H de Koeijer, H Keunen, Y Loufa, F Muhashy Habiyaremye, M-L Susini Ondafe, S Van den Bossche, V Pinton, E Verheyen and K Vrancken for their input in developing and executing the MRV program. This work was supported by the Belgian Directorate-General for Development Cooperation and Humanitarian Aid (CEBioS program). Two anonymous referees are also gratefully acknowledged for their input.