Investments in ecosystem-based adaptation (EbA), which usually entails building ecological infrastructure, will generate benefits for society for decades. EbA projects funded by governments or donors typically run for 5 years, or at the most 10 years, and consequently require detailed exit strategies  to ensure effective maintenance of the infrastructure beyond their closure dates. The financial and operational arrangements for such maintenance will vary according to the social, economic and biophysical contexts of a particular site. The nuances and details of these contexts that are important for designing an effective exit strategy are, invariably, not available at the start of an EbA project. As a result, it’s crucial that EbA project managers are actively collecting information from early on in the project to refine the exit strategy presented in the project document. As new information comes to light over the course of the project, the exit strategy will need to go through numerous iterations, with the ultimate end goal being a set of plans that can be easily implemented after the grant funding period has ended. Examples of lessons learned from the three EbA South countries, namely the Republic of Seychelles, Islamic Republic of Mauritania and Nepal, are presented below.
In the case of the Seychelles, mangrove, marshes and riverine vegetation were restored through planting of tree seedlings at 10 sites on three different islands. At the outset of the project, it was not anticipated that major investments of time and resources would be required after the seedlings had been planted. The assumption was that the ecosystem would be self-sustaining as long as direct damage to trees, like harvesting for fish traps and other construction purposes, did not occur. The risk of tree harvesting at the project sites was deemed negligible and consequently an exit strategy was not one of the main foci at the outset of the project. Unanticipated long-term threats did, however, emerge in the last year of the project. At one site, landscaping and maintenance contractors cut down several hundred of the project’s tree seedlings during road maintenance work, despite the seedlings being more than ten metres away from the roadside. At other sites, alien weedy plants, mangrove crabs and giant tortoises damaged the project’s tree seedlings more extensively than had been expected. The exit strategy for the project required enhanced communication protocols between government and the Landscape and Waste Management Agency to prevent damage to the trees by roadside maintenance teams. It also required long-term agreements with government and non-governmental organisations to safeguard the project’s tree seedlings until the trees are large enough to withstand pressures from alien plants, crabs and tortoises.
In the case of Mauritania, where degraded dryland ecosystems were restored by planting indigenous trees, the extent of adaptive management that was required during the project implementation was not anticipated. For example, local politics required changes in project sites, with some of the new sites requiring fencing to protect tree seedlings from livestock. The shift in project sites also meant that planting of tree seedlings continued right up to the point of closure of the project instead of a year before closure. The exit strategy for the project consequently needed to be revised to include a long-term commitment from government to maintain the project’s fences and a short-term commitment to water seedlings for several months after project closure until they were established. It is likely that government will follow through on these commitments because the project’s sites were on state-owned land and the government recognised the value of the EbA investment in the land. In retrospect, it would have been preferable to invest the time and resources well in advance of the closure of the project to obtain a written agreement with the government, specifying the anticipated activities and costs over a period of decades. This exercise would also have raised awareness of how the EbA benefits will accrue for future generations and need to be managed by government accordingly.
In the case of Nepal, where tree seedlings were planted on both private and communal land, the project budget was insufficient to provide follow-up maintenance services such as weeding and fertilizing to maximise seeding survivorship. Private landowners and communities had been selected for participation in the project based on their interest in the project and their commitment to invest in the maintenance of the seedlings. As the project was implemented, it became evident that the amount of resources that landowners and communities invested in the seedlings varied considerably. This variation was far larger than had been anticipated at the outset of the project. In retrospect it would have been appropriate to undertake more in-depth socio-economic surveys at each site to assess the likelihood of a particular individual or community investing in the long-term maintenance of the tree seedlings. Two extremes of the continuum highlight this point.
A successful example of a situation where the seedlings were well maintained was a Nepali woman who owned a piece of land that had been planted with hundreds of the project’s tree seedlings. This woman routinely weeded the seedlings and fertilized them with goat droppings. She regularly encouraged her son and sister-in-law to travel from the city to her rural plot on the weekends to assist her with the weeding and fertilizing. Her objective was to leave a green legacy for her grandchildren, and to convert eroding fallow land into a verdant forest of valuable trees which provide fodder, medicines, fruit and timber, and which retain an organic-rich topsoil. It should also be noted that her neighbours were entitled to graze their cattle on her fallow land and in effect occupy it if she did not demonstrate an economically viable activity on the land. She was therefore also clearly motivated to maintain the trees because it protected her land asset.
An example of where tree seedlings were not maintained was a community that had agreed to have thousands of tree seedlings planted on a piece of communal land. The tree seedlings at this site were difficult to find in amongst the grass and weeds. At the outset of the project the community had assured the project manager that they were fully committed to having a multi-use forest on the land in the long-term. It took several years for the project manager to work out why the community was not following through on their initial commitment of actively investing in maintenance of the tree seedlings. What emerged, as relationships were built with the community, was that the security of the land was at risk. A wealthy private landowner had donated the land to the community, allegedly to avoid paying taxes on the land. For unknown reasons, the community had, during the course of the project, developed suspicions that when the land value increased through time, as the trees matured, the landowner would find a legal route to reverse the donation and reclaim the land for himself. It was consequently a calculated economic decision for the community to hold back on the long-term investment in tree seedling maintenance.
It should, however, be noted that no matter how well the project manager develops the EbA project’s implementation plan and iterates the project’s exit strategy, there will always be a continuum of success across sites. Unlike other infrastructure projects, where specific targets of quality and quantity can be reached (e.g. x litres of water stored behind a non-cracked dam wall of y metres height), EbA interventions will invariably have considerable variation across sites as a result of socio-economic and biophysical factors beyond the control of the project manager. As a result, an EbA project investing in interventions spanning tens or hundreds of hectares should be viewed as a large experiment, in which project resources are allocated across multiple sites, with the full understanding that not all sites will prove to be viable locations for long-term ecological infrastructure. Crucially, the reasons for failure at certain sites within a project should be well documented to ensure upscaling of EbA in the future is done at appropriate sites.
 Ecological infrastructure is the solid mass of biomass (roots, stems, leaves, lichen, fungi, fauna) and soil carbon that make up an ecosystem and provide a host of ecosystem services to society (e.g. provision of clean water, fruits, fibre, medicine, timber, fodder).
 An exit strategy provides details on how the ecological infrastructure will be maintained through time. Importantly, funding for the maintenance is required. It is preferable that such funding is acquired through domestic sources because the new ecological infrastructure is a national asset. However, it is noted that further international funds may be necessary for maintaining the assets until such time that the local government is fully sensitised and aware of the value of these assets.