NbS Triple Win Toolkit: Biodiversity Indicators in Context 31 iv) More generally, there needs to be enough understanding of the system that appropriate reference models and activities can be planned – this can be challenging in some environments where knowledge and resources are limited (also see Limitations and Challenges below). Limitations, Challenges, and Areas for Further Discussion Suitability of aggregating across projects Reference models and a consistent restoration scale allow standardisation across projects, and so facilitate aggregation of the indicator. However, this approach to aggregation does have limitations. Firstly, there will still be subjectivity and variation across projects in how the reference model is defined and in how the restoration scale is interpreted. If this proves problematic, more specific guidance to those reviewing project plans could help improve consistency. Secondly, ecological restoration is likely to be more feasible in some locationsand ecosystems than others due to technical and logistical challenges. For example, per hectare costs of coral reef restoration are typically substantially higher than for mangrove, saltmarsh, or seagrass restoration35. The benefits of smaller scale restoration in high biodiversity value ecosystems might therefore not be fully reflected by an aggregated area indicator. Disaggregating the indicator by ecosystem type could help to communicate these benefits more clearly and avoid incentivising restoration focusing on only a restricted set of ecosystems. Feasibility The proposed indicator may need refining to ensure practical feasibility. In particular, it could be that in practice very few projects implement ecological restoration designed to reach at least 4 stars across the six categories of ecosystem attributes. Potential alternatives to this could be discussed (e.g. the need to achieve at least two ‘levels’ worth of improvement on the SER 5-star scale). Discussion Point: Prior to finalising the indicator, discussion with programme managers and those involved in carrying out projectswill be important to ensure the indicator reflects true biodiversity benefits without being unattainable. Data and Knowledge To count towards the indicator, projects must have a reference model that describes relevant ecosystem attributes in the natural or near-natural state. The SER provide guidance on approaches to defining reference models, although in practice this might still be challenging in ecosystems where knowledge and/or the ability to collect new data is limited. Projects also need to plan for monitoring the effects of restoration activities on ecosystem attributes. Again, this may not be straightforward for logistical or technical reasons. In marine ecosystems in particular it might be difficult to collect sufficient data on the status of the ecosystem and its threats to be able to reasonably plan and achieve a sufficient level of restoration, although the vast majority of marine-related NbS projects occur in coastal ecosystems36 where the required data collection is feasible. The counter argument to these challenges is that if a project is unable to describe a natural state to work towards or monitor how the project is affecting key ecosystem attributes, it would be difficult to justify claiming a biodiversity benefit.In practice, a degree of pragmatism will be needed in this respect. Discussion Point: Prior to finalising the indicator, discussionwith programme managers and those implementing projects will be important to ensure the indicator fairly accounts for project-specific challenges without overly compromising the need for suitablereference models and monitoring.