NbS are defined as actions to protect, sustainably manage, and restore natural or modified ecosystems, that address societal challenges effectively and adaptively, simultaneously providing human well-being and biodiversity benefits (Cohen-Shacham et al., 2019; Seddon et al., 2021). For example, protecting and restoring forests can help to reduce the impacts of floods and landslides, while nature-based agricultural techniques such as the use of leguminous cover crops can improve the water-holding capacity of the soil, helping to combat droughts. We include modelling studies that assess the potential benefits of NbS that have not yet been implemented, and, similarly, we include assessments of the benefits delivered by existing ecosystems, because this evidence is a useful proxy for the benefits that would be delivered through protecting, restoring, or managing the ecosystems through future NbS actions. We use this approach because our aim is to gather evidence on the potential future effectiveness of scaling up the deployment of NbS in Bangladesh, not only on the benefits currently being delivered by NbS that are already implemented.

NbS should be designed and implemented with the full engagement and consent of Indigenous Peoples and local communities, and must sustainably provide one or more benefits for people whilst causing no loss of biodiversity or ecological integrity (or preferably a gain) compared to the pre-intervention state (Seddon et al., 2021). It was rarely possible to determine whether all these criteria were met based on the information given in the papers, hence some interventions may not qualify as ‘solutions’, as per the NbSI guidelines and the IUCN standard (NbSI, 2020; IUCN, 2020). However, we recorded any relevant information on biodiversity and social impacts, even where it seemed possible that mixed or negative impacts might have occurred, because this is important in highlighting lessons for NbS design in future. We considered the likely baseline or counterfactual scenario, i.e. what would have happened in the absence of the intervention. We therefore included interventions such as agro-ecological farming methods if the most likely alternative was a continuation of more damaging practices, and sustainable fishery management if the alternative was a continuation of over-harvesting. We use the term biodiversity in its broadest sense, with a positive outcome for biodiversity indicating a move towards an appropriate mix and abundance of habitats and species for each location, acknowledging that some ecologically valuable habitats have naturally low species diversity.

The strength of NbS is that they can simultaneously address multiple challenges. Outcomes can arise either from changes to ecosystems to support or increase the provision of ecosystem services or through the process of implementing the NbS, such as through training, employment, and empowerment. In this way, NbS can address climate and natural hazards at the same time as contributing to other sustainable development goals. They can reduce vulnerability to climate change and natural hazards by reducing exposure to impacts (e.g., forests protecting against floods), reducing sensitivity to impacts (e.g., by diversifying livelihood options) and increasing the capacity to adapt to change (e.g., by empowering communities and individuals) (Thiault et al., 2021). In addition, GHG mitigation reduces hazards by limiting the magnitude of climate impacts, and biodiversity underpins the adaptive capacity and healthy functioning of the ecosystem, for example by maintaining genetic diversity that could confer resilience to future pests and disease. We drew up a list of relevant NbS outcomes adapted from Roe et al. (2021), and identified how they address climate and natural hazards and contribute to the SDGs (Table 1). These are hereafter referred to as the ‘target outcomes’ for the review. Further details are provided in the Supplementary Information.

The systematic review was based on the methodology of Chausson et al. (2020), which used the scoping elements listed in the top row of Table 2. In order to restrict their global search to a manageable number of articles, Chausson et al. (2020) searched academic articles only, and excluded the categories shown in the middle row of Table 2. These criteria retrieved only two papers for Bangladesh. For this in-depth country-level study we removed all these exclusion criteria, as shown in the bottom row of Table 2, and we also searched for evidence from books, conference proceedings and grey literature.

The search string (see Supplementary Information) included recognized intervention terms (e.g., nature-based solution, ecosystem-based adaptation, agroforestry), the people or sector benefiting from NbS (e.g., local communities, policymakers, food systems), the challenge targeted (e.g., climate change, flood, drought, landslide) and the outcome (e.g., food security, adaptation, mitigation, protection, resilience). The string was based on that used by Chausson et al. (2020) but after review by the co-authors based in Bangladesh we included one additional local term (“floating gardens”). We also added an extra step in which we simply searched for “Nature-based solutions” (and related terms) and “Bangladesh”. This was to check that we had not inadvertently screened out any relevant studies due to the complex and specific search terms used for the main search. We searched only for studies of NbS in Bangladesh.

We searched Web of Science and Scopus on May 7, 2020 for articles, reviews, conference proceedings, reports or book chapters that matched these terms. We excluded duplicates using EndNote, and screened first titles and then abstracts to eliminate sources that clearly did not contain any evidence of NbS effectiveness for delivering the target outcomes. As the aim of this review was to conduct an in-depth analysis of the strongest evidence, we then performed a further screening round, selecting only the sources that explicitly referred to evidence on the effectiveness of NbS for delivering the target outcomes in the abstract. The sources excluded at this stage included many general texts or reviews about climate change adaptation or mitigation, some of which did not explicitly refer to NbS in the abstract, most of which appeared to consist largely of secondary information taken from other studies. However, it is possible that some of these sources may contain some primary evidence on NbS effectiveness. Finally, some additional studies were excluded at the full-text screening stage, either because they were duplicate studies or because they were not relevant.

Many NbS projects in Bangladesh are not included in peer-reviewed journals. It was therefore important to analyze grey literature on these interventions. Because non-academic search engines return very large numbers of hits, most of which are not relevant, we used the knowledge of local experts and networks to narrow down the search to the most relevant resources. From our personal knowledge and experience of the implementation of NbS in Bangladesh, together with discussion with members of the ‘NbS Bangladesh Network’, a community of researchers, practitioners and policymakers (www.nbsbangladesh.info), we identified four major projects implemented in Bangladesh over the past 22 years. These projects focused on community-based natural resource management, ecosystem-based adaptation and biodiversity conservation, and covered multiple sites in the Sundarbans, the coast, the Chittagong Hills and the Haor wetlands. We searched for reports from these projects through online sources and personal contacts, and finally selected three final reports (one in two volumes) (DoE, 2015; IUCN Bangladesh, 2016, MACH-II, 2007a; MACH-II, 2007b) and one performance report (Winrock International, 2018). From these we identified 24 interventions that qualify as NbS.

Many other projects followed ecosystem-based approaches in Bangladesh, but documents with adequate evidence were not available. This reflects challenges with the grey literature evidence base. Documents often state outcomes (mainly in qualitative terms) without fully describing the methodology used to determine the outcome, so it is not possible to assess the robustness of the information, and often they do not follow a consistent impact assessment methodology throughout the whole project period.

There were 22 records of positive outcomes for GHG reduction and one mixed outcome. Sixteen were from forests and six from conservation agriculture.

Protecting and restoring mangrove forests is particularly important as they trap carbon-rich sediment amongst their roots, as well as storing carbon in biomass. Carbon storage was estimated at 219 tC/ha (Rahman et al., 2017) to 257 tC/ha, of which 63% was belowground in the soil and roots (Abdullah-Al-Mamun et al., 2017). However, the global average for oceanic mangroves was estimated as 400 tC/ha, suggesting that mangroves in Bangladesh could be relatively degraded (Chow, 2018). Mangroves were estimated to sequester carbon four times faster than mature land-based forests, offsetting 1.5% of Bangladesh’s fossil fuel carbon emissions in 2014 or 10% from 1997 according to different estimates (Table 5). Conversion to aquaculture was highlighted as a threat, as it requires excavating at least 2 m of sediment, which can release 70 tC/ha of carbon (Chow, 2018).

We found fewer studies on carbon storage and sequestration in native terrestrial forests, which include the mixed evergreen forests in the Chittagong Hill Tracts and the much smaller fragments of deciduous Sal forests (Shorea robusta) in central Bangladesh. Progress towards protecting and restoring forests for carbon benefits via REDD+ faces governance challenges in Bangladesh, especially in the Chittagong Hill Tracts (Richards and Hussain, 2019). Low income households are highly dependent on forests for fuelwood and timber, and over-harvesting is causing severe forest degradation which reduces carbon storage (Yong Shin et al., 2007). For example, one study found that plots reforested with species that were valued as wood fuel showed much lower levels of carbon storage than those using other species, due to harvesting by local people. It was suggested that providing secure land tenure and clean energy options for indigenous communities could enable more sustainable management of forest resources with benefits for carbon storage (Yong Shin et al., 2007).

Forests managed for production are only categorized as nature-based solutions if they provide biodiversity benefits (compared to business-as-usual) and are co-implemented with local communities. Carbon storage will be offset by emissions from wood extracted for fuel or short-lived products such as paper. Estimates of carbon storage are lower than for natural forests, ranging from 200 tC/ha for roadside social forestry plantations (Rahman et al., 2015) to 118 tC/ha for homegardens (Nath et al., 2015) and 31–37 tC/ha for agroforestry (Hanif et al., 2015). However homegardens in Bangladesh stored more carbon than those in India, which could be due to their higher tree density and diversity (Nath et al., 2015).

Conservation agriculture can play an important part in reducing emissions by improving soil health and soil carbon. This enables inputs of mineral fertilizers to be reduced and thus also cuts nitrous oxide emissions (Islam et al., 2011; Alam et al., 2020), and it may reduce the need for pumped irrigation, saving carbon emissions from fuel use (Begum et al., 2018). However, switching to 100% organic fertilizers may not be the best strategy as although it increases soil carbon, it may reduce yields and increase methane emissions (Begum et al., 2018). Also, cow dung and crop residues are in demand as fuels (Begum et al., 2018). Integrated methods that combine organic fertilizers with reduced tillage, increased residue retention and lower use of synthetic fertilizers can successfully maintain yields whilst cutting costs and emissions (Islam et al., 2011; Begum et al., 2018).

Non-puddled transplanting of rice is an important option for reducing methane emissions. It involves strip tillage (tilling only the strips to be planted rather than the whole field) followed by planting into saturated soil, rather than the usual practice of planting into a ploughed and flooded field (Bell et al., 2019). This cuts the total lifecycle greenhouse gas emissions of rice production by between 16 and 31% depending on the level of crop residue (straw) retention (Bell et al., 2019). Residue increases soil carbon storage and crop yield, but this is offset by higher methane emissions as the incorporation of residue into the soil stimulates the activity of methane-producing bacteria (Alam et al., 2019).

Bangladesh is one of the most vulnerable countries in the world to coastal hazards (Shi et al., 2016). A large proportion of the population live in low-lying coastal areas on a funnel-shaped delta with a high tidal range, putting them at high risk from cyclones and storm surges (Das et al., 2010; Rahman et al., 2019) which cause extensive wind and flood damage. In addition, coastal areas are often lost to erosion, and suffer from salinization of groundwater and agricultural land due to seawater intrusion and frequent flooding. We found 24 outcomes for coastal protection: 22 positive, 1 mixed and 1 unclear. Most (15) were for mangrove protection and restoration.

Mangroves have provided a natural barrier to coastal hazards for centuries. They are estimated to protect 1.1 million to 3.5 million people in Bangladesh from coastal flooding during cyclones (Akber et al., 2018), avoiding damage worth at least US$1.56 billion per year on average (Menéndez et al., 2020). Villages protected by mangroves had only about half of the monetary loss from flood and wind damage associated with cyclone Sidr, compared to other villages (Akber et al., 2018). Even a 100 m deep coastal shelterbelt of healthy mangroves can reduce storm surge velocity by up to 92%, protecting embankments from costly damage (Dasgupta et al., 2019), and a double shelterbelt of mangrove and Casuarina trees, 200–300 m in depth, can reduce storm surge height by up to 22% and velocity by up to 49% (Das et al., 2010).

The Forest Department of Bangladesh started planting mangroves along the whole coastline in 1966, initially aiming to boost protection against cyclones and storm surges, and later to stabilize newly accreted (char) land so that it can be used for agriculture (Iftekhar and Takama, 2008). Planting with mangroves greatly increased the ratio of accretion to erosion, and cut the risk of erosion loss by a third (Chow, 2018). Long term community-based adaptation projects have involved local communities in planting mangroves and re-vegetating sand dunes to protect against coastal flooding and erosion (DoE, 2015; Winrock International, 2018; Table 5).

Despite the success of these initiatives, both existing and replanted mangroves have been extensively cleared for housing, infrastructure, agriculture and aquaculture, and degraded due to over-extraction of timber and fuelwood. Timber harvesting from mangroves was banned in 1991; this reduced the short-term benefits for local people but was expected to improve coastal protection, fish production, biodiversity and carbon storage (Abdullah-Al-Mamun et al., 2017). However such bans are not always effective, due to corruption and weak enforcement (Iftekhar and Takama, 2008; Abdullah-Al-Mamun et al., 2017; Rahman et al., 2018). Mangroves are also threatened by climate change impacts including fresh water flow reduction, sea level rise, salinity increase and storm damage (Ahammad et al., 2013; Abdullah-Al-Mamun et al., 2017; Chow, 2018; Rahman et al., 2018). Restoration should use species that are well adapted to current and future local conditions, and replanting may be necessary after a few years because there can be a high failure rate (Dasgupta et al., 2019).

Bangladesh lacks coral reefs, but one study showed that artificially created oyster reefs could help to reduce coastal erosion, trap suspended sediment, and support saltmarsh expansion (Chowdhury et al., 2019). However, their potential may be limited due to the high turbidity of Bangladesh’s coastal waters.

Floods, landslides, and soil erosion are severe problems in Bangladesh. In the Chittagong Hills, steep slopes, heavy rainfall, poor soil and intensive cultivation for cash crops is leading to soil erosion of over 100 t/ha/year, loss of soil nutrients, falling crop yields, landslides, and sedimentation of reservoirs (Rasul, 2009). This is exacerbated by changes in traditional slash-and-burn cultivation (‘jhum’), as population pressures have shortened the fallow period during which secondary forest usually regenerates from 6 to 7 years to just 3–4 years (Gafur et al., 2003; Nath et al., 2005; Rasul, 2009). We found 16 examples of how NbS can address these hazards, and one unclear outcome.

Protecting and restoring forests plays a key role. Soil erosion was found to be three to four times lower and peak flow seven times lower in a forested catchment than a cleared catchment (Gafur et al., 2003), and modelling indicates that trees with deep tap roots could prevent landslides on slopes less than 70^° (Islam and Rahman, 2019). Agroforestry can reverse soil loss, as the soil-formation rate can exceed the erosion rate (Rasul, 2009). Contour planting (planting trees across slopes) can be particularly effective, and can increase yields, income and food security (e.g., in the MACH project, MACH-II, 2007a,b). In the Haor Basin, community-based adaptation initiatives are restoring freshwater swamp forests and planting trees on embankments to protect villages on ‘kanda’ (raised land that becomes islands when the seasonal wetlands flood) from erosion due to wind-driven wave action (DoE, 2015; IUCN Bangladesh, 2016; Winrock International, 2018). Villagers also grow low plants and heap piles of rotten water hyacinth around homesteads to prevent soil erosion during floods (Reid and Alam, 2017).

Cyclones and other tropical storms cause damage from high winds as well as flooding. Trees can help to protect against this, although they can also cause damage if they are blown over (Nath et al., 2015) and, conversely, forests and mangroves can themselves be damaged in storms (Dasgupta et al., 2019). We found six positive outcomes of NbS for addressing wind damage.

Protection by mangroves in the Sundarbans reserve significantly reduced the cost of repairing houses due to combined wind and flood damage (Akber et al., 2018), and coastal shelterbelts of Casuarina trees were also perceived to be effective in reducing wind speed and increasing the size of sand dunes (Miah et al., 2013). In homegardens, a survey found that taller trees are grown at the boundary where they provide protection against wind but do not shade the other plants, and will not fall on the house if blown over (Nath et al., 2015). On exposed beach and hillside locations, species such as coconuts were preferred due to their strong stems, low weight, and light canopy, which reduced the wind load on trees and reduced damage if they fall, while species such as Acacia were chosen on mudflats because of their strong and spreading root systems.

Only two of the studies retrieved in our review referred to NbS for protecting against heat, and neither provided robust evidence. In the Haor basin, villagers value the forest for providing shade and cooling the air, and they are also planting more fruit trees at their homesteads for shade during heatwaves (Reid and Alam, 2017). However, native trees such as mango and jackfruit are susceptible to hailstorm damage, so they are being replaced with fast-growing and storm-resistant non-native timber trees such as teak, eucalyptus and acacia (Reid and Alam, 2017), so this action might not be classified as a NbS as there could be adverse biodiversity impacts. On the coast, local people planted salt-tolerant mangrove trees on land degraded by a saline storm surge where all the other trees had died, to provide protection from summer heat (Imam et al., 2016). The evidence in these studies was based on the perception of the villagers who carried out the intervention.

Although Bangladesh experiences extremely heavy rainfall in the monsoon season, droughts and water shortages are a growing problem in the dry season (Sayed et al., 2020), due to climate change, over-abstraction of water for human use, soil degradation, water pollution and salinization.

Ten interventions were associated with outcomes for water security: 8 positive and 2 unclear. Four of the positive outcomes were from conservation agriculture, where improving soil structure and retaining crop residue can reduce evaporation and enhance soil water infiltration and storage, reducing the need for irrigation (Alam et al., 2017; Bell et al., 2019; Islam et al., 2019; Sayed et al., 2020). For example, these techniques increased the water productivity of wheat by 60% (Bell et al., 2019) and increased irrigation water productivity by 25% (Islam et al., 2019), increasing the resilience of farmers to unpredictable rainfall patterns.

Rainwater harvesting mainly involves engineered options such as installation of rooftop tanks. However, a study in the Barind Tract described how re-excavation of silted-up ponds and beels (permanent wetland waterbodies) allowed rainwater storage and infiltration to recharge a depleted aquifer, so that the community now has water all year round rather than only for 4–5 months of the year (Rahaman et al., 2019).

Four interventions addressed water pollution, which seriously affects both surface water and groundwater in Bangladesh. We found evidence that constructed wetlands could be used to help tackle the problem of groundwater pollution by arsenic, which affects over 1.5 million wells used by 35 million people in Bangladesh. Experiments indicate that water from wells could be passed through a series of wetlands containing river sand and planted with bulrush (Typha latifolia), which would reduce arsenic concentrations to the WHO safe limit (Schwindaman et al., 2014). Two studies mentioned the potential role of water hyacinth for removing water pollution in inland wetlands, but neither provided robust data (MACH-II, 2007a; MACH-II, 2007b; Reid and Alam, 2017).

Food security accounted for 43 of the 140 positive outcomes identified in the review, and nine mixed, negative, neutral or unclear outcomes. Conservation agriculture was the most frequently cited NbS, but others included agroforestry, homegardens, wild food in protected forests, and the role of protected and restored mangroves and inland wetlands in supporting fisheries.

Smallholders face severe challenges in Bangladesh, as soil fertility is declining, rainfall is becoming more unpredictable and the cost of agrochemicals, irrigation and fuel is increasing. Conservation agriculture has been extensively studied as a way of improving soil health and water storage capacity, with a focus on adapting the approach for the rice-based rotations that predominate in Bangladesh. The key benefit for food security is that conservation agriculture enables inputs of mineral fertilizers and irrigation to be reduced without loss of yield (Islam et al., 2011; Alam et al., 2020), and this in turn increases profitability (Aravindakshan et al., 2015; Gathala et al., 2016). The papers stated that the key is to apply integrated science-based approaches that combine conservation agriculture techniques with reduced (but not zero) use of synthetic fertilizers, to maintain yields whilst cutting costs (Islam et al., 2011; Begum et al., 2018).

For example, a 10-year program involving over 6,000 farmers found that strip planting increased yields by up to 28% and cut cultivation costs by 75%, labor requirements by 50%, irrigation water requirements by 11–33% and fuel costs by up to 85%, increasing profits by 47–560% for different crops (Bell et al., 2019) (Table 5). Researchers on this program worked with farmers and equipment supplies to develop a lightweight reduced tillage planting machine; they estimated that if this was used by 2.5% of farmers in Bangladesh it would generate US$21–38 million per year from increased yield and reduced production costs.

Similarly, integrated pest management, which combines non-chemical methods of pest control (such as manual removal of pest eggs, and provision of perches for birds) with reduced application of pesticides at the economic threshold level (at which the value of the crop destroyed exceeds the cost of controlling the pest), can increase yields and save money compared to conventional heavy pesticide use, while reducing pollution and health impacts from pesticide poisoning (Alam et al., 2016).

Agroforestry offers a method of increasing food, fuel and timber production despite shrinking available land and a growing population, by making use of vertical space (Hanif et al., 2015). It can help to reduce the negative environmental impacts of food production such as soil erosion, water pollution and (if native tree species are used) biodiversity loss, and can help to restore degraded land. The first two Community Forests in Bangladesh were established on deforested land in the Chittagong Hill Tracts that the Forest Department had already tried to replant several times but without success, due to lack of maintenance and encroachment for farming by local people. The land was given to local landless or almost landless people, who first planted eucalypts and fruit trees, and later shifted to native forest species. Sale of tree products improved their incomes and enabled them to diversify to other income sources, relieving pressure on local natural forests, but as they become less dependent on forest resources it remains to be seen whether this reduces their incentive to conserve and replant the community forest (Mohammed et al., 2016). Agroforestry could also potentially help to restore land that has been degraded by saline intrusion, although there is a risk that it could exacerbate the problem by concentrating salts in the root zone (Wicke et al., 2013).

Homegardens are widespread and play a vital role in food security in Bangladesh. One study found that applying a year-round rotating system of different vegetables and fruit trees, chosen through a participatory approach using local traditional and indigenous knowledge, could more than double annual production, improve nutrition, increase household income, alleviate poverty and provide employment and empowerment for female family members (Ferdous et al., 2016). In dense urban areas, rooftop gardens can play a key role; 35% of properties surveyed in one part of Dhaka had rooftop gardens providing a wide range of fruit and vegetables (Zinia and McShane, 2018).

Floating gardens are a traditional approach to cultivation in some wetland areas, and have more recently been introduced to the Haor basin. Mats of floating wetland vegetation (mainly water hyacinth, which is an invasive non-native species) are used as the base for vegetable gardens in the wet season when farmland is inundated, and the rotting mat is then used to fertilize the soil in the dry season when the water recedes. This can help families to produce food all year round, reducing the need for unsustainable harvesting of wild resources to sustain livelihoods (Irfanullah et al., 2008).

Wild food resources are important to many low-income households. People living closer to the Sundarbans mangrove forest have higher levels of dietary diversity, partly due to direct consumption of bushmeat, although the authors noted that bushmeat harvesting for subsistence often coexists with commercial bushmeat trades that may threaten endangered species (Baudron et al., 2019).

Forest ecosystem services can also contribute positively to food production in nearby cropland. A survey of 275 households in the Chittagong Hill Tracts found that households closer to forests were more likely to have homegardens and own livestock, and had higher fruit consumption and higher dietary diversity as a result. The nearby forest was thought to support crop production in homegardens through maintenance of soil fertility, micro-climate regulation, and pollination (Baudron et al., 2019). The benefits of these existing forests are a proxy for the benefits that would continue to be delivered through protecting and/or restoring the forests in future. However, protecting or restoring ecosystems such as forests and mangroves can reduce the area available for food production, potentially leaving a ‘food gap’ where an area is unable to produce enough food to meet local demand (Hoque et al., 2020). To address this, it has been suggested that, for example, coastal areas targeted for NbS to provide hazard protection could receive food subsidies from surplus food produced in other areas further inland (Hoque et al., 2020).

Sustainable fishery resource management has long been practiced by communities in the freshwater wetlands of Bangladesh. Community-based projects have brought all the remote beels (permanent pools within the seasonal wetland) under one management practice, allowing the community to harvest fish to a sustainable level under certain conditions determined by biodiversity conservation goals (IUCN Bangladesh, 2016). This supports poor people’s rights and access to resources, as well as boosting food and nutrition security, employment and biodiversity. In the Tanguar Haor project, endorsement of a ‘core zone and buffer zone’ approach by the government was a significant achievement which established the rights of the poor to access fisheries across the wetland, and reduced illegal harvesting (IUCN Bangladesh, 2016). In addition, the projects build the capacity of the community by developing income generating and community management skills, raising awareness of wetland resource issues and understanding the value of biodiversity conservation (IUCN Bangladesh, 2016). Community-managed fish sanctuaries have also been created in deeper parts of the wetland, where fishing and collection of other aquatic resources are restricted or totally prohibited. In Hakaluki Haor, fish sanctuaries were established in two beels by using katha (bamboo and tree branches) to create breeding grounds and food supply for the fish and protect from illegal fishermen. This helped to increase fish catches from 171 kg/ha in 2013–14 to 277 kg/ha in 2015–16, while enriching biodiversity, employment and tourism (Winrock International, 2018; Table 5). The MACH project also supported local resource management organizations in re-stocking nearly 1.2 million fish of 15 native species, and restoring critical wetland habitats. This enriched fish production and biodiversity, more than doubling fish catches, and increasing fish consumption in local villages by about 45%, with landless people benefiting as much as larger landowners (MACH-II, 2007a).

River protection, restoration and sustainable management can sustain and enhance essential services including provision of fish and fresh water. For example, the River Halda is the only river in the world where major Indian carp species can spawn (Kabir et al., 2015). Fish eggs and fry are gathered and sold to underpin aquaculture across Bangladesh. However, the river is under threat from over-fishing, pollution and other human activities. A survey found that local people were willing to contribute their time and money to help conserve the river, although their willingness-to-conserve was less than the value of the services it provided (Kabir et al., 2015). The authors recommended that community-based management of rivers could be supported by a Payment for Ecosystem Services approach.

Forests are a vital source of wood, timber and other products for local people in Bangladesh, although over-exploitation has led to widespread forest degradation (Yong Shin et al., 2007). There were 19 outcomes related to this, of which three were negative, all from a study of the impacts of forest protection in the Chittagong Hills (Miah et al., 2014). This found that local households generated an average revenue of Bangladesh Taka 13,473 per year from gathering timber, firewood, bamboo, medicinal plants, bushmeat and nuts, as well as being dependent on these goods for their own use. Dependence on the forest for medicinal, religious and food purposes was felt to be non-negotiable, while some people were prepared to forego extraction of timber, bamboo and vegetables if appropriate cash compensation or alternative livelihoods were provided. The authors recommended that providing secure land tenure for the indigenous communities could help to establish more sustainable management of the forest resources, which were perceived to be severely degraded due to over-extraction.

Many NbS interventions attempt to manage this conflict by implementing participatory sustainable management approaches. One example is the social forestry program, which was launched by the Bangladesh Forest Department in 1989 to support forest restoration, agroforestry, village woodlots, and roadside plantations (Rahman et al., 2015). Most of the roads in south-western Bangladesh are lined by social forestry plantations that are protected and managed by local landless and land-poor people, who are allowed to gather fuelwood and receive 40% of the proceeds after felling (Rahman et al., 2015). Similarly, in Hakaluki Haor, the ‘green belts’ of indigenous tree species established on embankments provide fuel wood and local economic benefits as well as reducing flood risk (DoE, 2015). Local communities are also allowed to gather branches for fuelwood from restored swamp forest when it is sufficiently mature, but there are agreements not to fell the trees (MACH-II, 2007a; IUCN Bangladesh, 2016).

Conflicts between different beneficiaries are illustrated by a case study of Nijhum Dwip, an uninhabited mudflat island that was planted with mangroves by the Forest Department and was later declared a National Park, due to its value for migratory birds (Iftekhar and Takama, 2008). However, around 700 households who had been displaced from their homes due to river erosion gradually settled on the island, leading to illegal encroachment and extraction of timber from the forest. Around a quarter of people interviewed on the island were highly dependent on forest resources, especially women and low-income households, and the forest also attracted tourists who boosted the local economy. Yet many residents wanted to convert the forest to agriculture, aquaculture or commercial forestry. Despite this, and in apparent contradiction, almost all thought that the forest should be better protected and that new mangroves should be planted. It was suggested that adaptive co-management involving local people could help to resolve these conflicts and trade-offs through sustainable use and equitable distribution of benefits.

We found only six cultural outcomes in the review, three related to the value of protecting the Sundarbans mangroves for aesthetic value, recreation and cultural heritage and spiritual inspiration (Rahman et al., 2018). Cultural heritage was also an important benefit of protecting the River Halda, which inspired local festivals, characteristic Sampan boats and the Halda Fada songs (Kabir et al., 2015). However, trade-offs were found for a shelterbelt of casuarina trees, which increased the aesthetic value of the area and its attractiveness for tourism, but also acted as a location for increased anti-social behavior such as theft (Miah et al., 2013). Finally, the CREL (Climate Resilient Ecosystems and Livelihoods) project (Supplementary Table S5) was estimated to deliver US$53 million of tourism and cultural services (Winrock International, 2018).

There were 52 outcomes reported for socio-economic benefits: 50 positive, one negative and one mixed. These included local economic benefits (15), employment (8), tourism (8), rights, empowerment and inequality (8 positive, 1 mixed), education and training (6), and social cohesion, governance and engagement (5 positive, 1 negative).

Local economic benefits include increased profits from implementation of conservation agriculture or agroforestry, income from use of sustainably harvested natural resources (including social forestry), and support for livelihoods provided through community-based management programs. Tourism also provides benefits for local economies, through opportunities for eco-tourism or enhanced attractiveness of tourist destinations. For example, a coastal shelterbelt was perceived to have increased tourist visits to a beach, due to provision of shade and increased attractiveness (Miah et al., 2013). Employment benefits came from jobs created through habitat restoration and protection, or livelihoods sustained through better management of resources. Out of the 154 interventions, 42 explicitly targeted poverty reduction although only 12 of these provided clear evidence that poverty had been reduced as a result of the intervention, while one reported no effect and five had unclear evidence. There was one mixed outcome for a social forestry initiative where poor people were excluded because the land was illegally occupied by local elites (Muhammed et al., 2008).

Out of the 50 positive socio-economic outcomes, 33 came from the community-based initiatives described in the grey literature reports (Supplementary Table S5). These were generally initiated by government departments with support from international development or conservation NGOs. They aimed to conserve biodiversity and provide sustainable livelihoods for local people, through empowering and engaging the community to manage their wetland and forest resources sustainably, with a strong focus on reducing poverty and inequality (DoE, 2015). Community organizations were established, and local people were offered training in natural resource management and conservation techniques and supported to develop diversified livelihood options, including through skills training and establishment of micro-credit loan schemes. For example, the CREL project provided training and support for enterprise development to 60,000 households, of which 51,400 reported that they had adopted more resilient agricultural practices, and 38,500 were estimated to have enhanced their incomes by a total of over US$5 million. Over 8,000 poor women (73% of livelihood beneficiaries) were empowered through financial training, helping them to improve access to services and credit, increase asset ownership and play a greater role in decision-making. CREL also helped 45 community management organizations to develop governance and financial capacity, with 35 of these (79%) reporting that they were able to become recognized local implementing partners for government and donor programs, and 23 (52%) becoming able to generate sustainable income from charging visitors to enter their protected areas (Winrock International, 2018).

These capacity building activities were accompanied by community-led restoration and protection of the ecosystems on which the communities depend, such as restoring swamp forests, excavating dried up wetlands and re-stocking with fish, with community guards protecting the restored forests from illegal exploitation. Co-management approaches were established, along with mechanisms for sharing the benefits from natural resources, such as by distributing 60% equally to households involved in the initiative, 25% to the community-based organization, and 15% to the government as revenue (IUCN Bangladesh, 2016).

Community management can help to build social cohesion and provide opportunities for engagement, education and cultural enrichment. For example, community organizations established two bird sanctuaries in Tanguar Haor, and employed community guards to protect them from illegal hunting, providing opportunities for eco-tourism development. A nature club was also established to engage young people in conservation and raise awareness on the importance of maintaining the ecological integrity of the wetland (IUCN Bangladesh, 2016). In the CBA-ECA project, local people started to feel a sense of pride in having a bird sanctuary in their neighborhood, and this empowered them to resist illegal hunters (DoE, 2015).

Several studies noted the importance of protecting intact ecosystems in order to sustain the flow of benefits on which many households depend, especially the most vulnerable. For example, the Sundarbans mangroves were estimated to provide public goods worth US$ 1,135 per ha each year, greater than the net economic return from shrimp farming at US$ 713 per ha (Rahman et al., 2018). A land use model of the Lower Meghna River Estuary estimated that loss of forests and mangroves due to urban expansion resulted in loss of ecosystem services worth US$118 million from 1988 to 2018, and continuing with business as usual or prioritizing economic development will lead to further losses of US$41 or US$16 million respectively, while protection and restoration will deliver an additional US$131 million of benefits (Hoque et al., 2020).

NbS should, by definition, support and preferably enhance ecosystems and their biodiversity. However, ecological outcomes were only explicitly reported for 28 interventions (all positive), and only four of these had robust evidence. Often the benefits were reported in terms of species richness or presence of iconic species, and in anecdotal terms with no clear methodology or baseline.

We aimed to screen out any interventions for which there was no obvious pathway for delivering ecological benefits, which would not be defined as NbS, but in several cases this was not clear. For example, social forestry plantations were found to contain 36 tree species from 17 families, but 94% of the biomass was from just four fast-growing timber species (Rahman et al., 2015). Similarly, it was reported that homegardens have the same tree species diversity as natural forests, but the species composition is different to natural forests, with a bias towards fruit, nut and ornamental trees, some of which are not native species (Bardhan et al., 2012). The ecological outcomes of these interventions would depend on the most likely alternative use of the land.

Nevertheless, we did find evidence of the role of NbS in supporting the biodiversity of the unique and threatened forests, mangroves and wetlands in Bangladesh. Many of these were related to the community-based management projects, which protect threatened habitats and species while also providing jobs and eco-tourism opportunities (Supplementary Table S5). For example, protection and restoration of Baikka Beel within Hail Haor resulted in an increase in wintering water bird populations from about 300 birds of 16 species in 2004 to 7,200 birds of 35 species in 2007 (MACH-II, 2007b). Similarly, rapid assessments in Nuniarchhara mangrove forest found 24 wildlife species in 2011–2013 (18 birds, 2 mammals, 1 reptile, 3 amphibians) compared to 14 bird species in 2007–2010, although it is not clear whether the earlier assessment looked for non-bird species. The report also notes that seven species of kingfishers were observed in more recent years, which indicates abundance of native fish, as well as some rare species such as the fishing cat and monkey, and Purple Swamphen (Kalim) bird.

The literature identifies a long tradition of research and implementation of participatory and pro-poor approaches to natural resource governance in Bangladesh, including harnessing local and traditional knowledge (Alam et al., 2016; Ferdous et al., 2016; Dasgupta et al., 2019), targeting interventions towards landless or land-poor households (Muhammed et al., 2008; Miah et al., 2014; Rahman et al., 2015; Ferdous et al., 2016), and working with local communities to protect and manage resources (Rahman et al., 2015). It was suggested that participatory co-management could help to resolve trade-offs and conflicts between beneficiaries or between different outcomes (Iftekhar and Takama, 2008).

The grey literature describes long-term projects (Supplementary Table S5) that aim to build capacity for community-based management of natural resources, by establishing community organizations and supporting vulnerable people to adopt diversified and sustainable livelihoods (Socio-Economic Outcomes). They show how participatory approaches are crucial for engaging and motivating the community to protect and sustainably manage natural resources (DoE, 2015). Adopting a co-management approach for managing ecologically important wetlands and forests during these initiatives was said to trigger a cultural and policy shift for the government of Bangladesh, from a top-down protection approach towards working with local communities to address the underlying drivers of over-exploitation of natural resources (Winrock International, 2018). Although the benefits of these approaches are self-reported by the project implementers, they hold considerable promise for supporting livelihoods and empowering the vulnerable while protecting critically endangered habitats and species. However, the study methods and outcomes were not always clearly reported.

We found one example of an approach that was implemented without sufficient participation. Attempts to train landless indigenous people to establish contour hedgerows to help them cultivate degraded forest land in the Chittagong Hill Tracts failed, because they found the system too complex and labor-intensive, it used unfamiliar hedgerow species and did not leave enough available land to meet their needs for subsistence cereal crops (Nath et al., 2005). International guidance has now been developed to help define best practice for participatory NbS (IUCN, 2020), including standard reporting criteria, which should help to improve future outcomes.

A common theme in the literature is that poor governance and corruption frequently undermines the effectiveness of interventions (Iftekhar and Takama, 2008; Abdullah-Al-Mamun et al., 2017). For example, a participatory forestry initiative in Tangail Forest Division showed a bias towards allocating land to local elites who had illegally encroached into forest areas, rather than landless people as intended, due to corruption of forest officials (Muhammed et al., 2008). Rahman et al. (2015) recommend stricter implementation of operational rules, strengthening of institutions for regular monitoring, and increased authority to implement sanctions against violators, to enhance the outcomes of roadside social forestry.

The cost of corruption has been estimated in a study of the Chittagong Hill Tracts, where smallholders must bribe local forestry officials in order to get a license to sell timber from agroforestry, as well as paying bribes at all the checkpoints established to (in theory) control illegal felling. Without paying the bribes of Taka 150 per cubic foot of timber, roughly 20% of the market price, the profitability of agroforestry would double. Setting up local collectives would enable smallholders to get a fairer price for agroforestry products by weakening the market dominance of large traders and middlemen (Rasul and Thapa, 2007). Strong, transparent and fair institutions that focus on empowering the vulnerable can ensure that the benefits of NbS flow to those most in need.

Although NbS can offer more cost-effective solutions than alternatives in the long term, when all public and private costs and benefits have been taken into account, governments and other funding agencies may need to provide practical and financial support to enable the transition to NbS in a way that meets local needs. For example, farmer training and knowledge of pests and beneficial insects is crucial to application of integrated pest management (Alam et al., 2016), and farmers on the plains of north-west Bangladesh were more likely to adopt conservation tillage if they had access to an agricultural extension office to provide unbiased advice and training (Aravindakshan et al., 2015).

Similarly, in the Chittagong Hills, subsistence farmers who could benefit from agroforestry to stabilize and regenerate the eroding soil (Section 3.3.3) face short term barriers including a high cost of borrowing, small size of land holdings, and pressure to produce sufficient crops to feed their families. Agroforestry does not produce economic returns for the first 5 years for fruit, or 10–12 years for timber, while shifting ‘jhum’ agriculture produces crops in just a few months (Rasul and Thapa, 2006; Rasul, 2009). Also, most farmers do not have secure land tenure, as the forests were nationalized during colonial times, and this discourages long term investments and prevents access to credit for covering initial costs. As a result, farmers continue to practice slash-and-burn cultivation on common land, avoiding the costs of nutrient depletion and soil erosion in the short term by shifting to new locations, but undermining their livelihoods in the long term. This implies that governments and other funding agencies need to enable a shift to more sustainable farming practices by providing financial incentives (such as Payment for Ecosystem Services), access to credit, secure land tenure or inheritable land use rights, and practical training and support (Rasul and Thapa, 2006; Rasul, 2009).

There is an opportunity to capitalize on recent interest in NbS, both globally and in Bangladesh, and to promote evidence-informed policy and practice to influence nature conservation and climate resilience. The evidence that we have compiled on the effectiveness of NbS interventions, such as the creation of coastal green belts with mangroves over the last 56 years, protection of World Heritage and Ramsar Site the Sundarbans, and sustainable management of wetlands over the last 22 years, is a good starting point, but we also recommend strengthening this evidence base through a more systematic approach to monitoring, evaluating, and reporting the process and outcomes of future NbS projects. Nevertheless, our analysis can help Bangladesh to effectively incorporate nature conservation and ecosystem-based approaches in implementing its current plans in the short-term (e.g., 8 ^th Five-Year Plan 2020–2025), medium-term (e.g., Perspective Plan of Bangladesh 2021–2041), and long-term (e.g., Bangladesh Delta Plan 2100). The evidence on NbS effectiveness that we have compiled can also help the country to take pragmatic steps in implementing the NDC and National Adaptation Plan (NAP) (under preparation) as well as any plans developed in response to the Post-2020 Global Biodiversity Framework to guide biodiversity conservation through 2050.

As the world is trying to focus on post-pandemic recovery, the International Labour Organization has advocated the potential for NbS to boost economic recovery by increasing green employment opportunities (WWF and ILO, 2020). Under its Nature-based Recovery Initiative, IUCN has been working with its Members and partners to create evidence to support governments to invest at least 10% of overall investments in nature and to ensure that economic investment in the post-COVID era doesn’t cause further harm to nature and livelihoods (IUCN, 2021). Bangladesh is in a good position to harness this opportunity, given its long experience of implementing NbS interventions as community-based management of natural resources, community-based adaptation, and co-management of protected areas. The economic recovery potentials of different NbS interventions in a wide range of ecosystems of the country could be investigated and incorporated in the national COVID recovery plans.

Our world is urbanizing exponentially, and 68% or 7 billion people could live in cities and towns by 2050 (WEF, 2020), yet experience of urban NbS in Bangladesh is relatively limited. Some recent initiatives bring together policies and practices on ecosystem-based approaches in urban areas, such as the Global Commission on Adaptation (www.gca.org), the Network Nature (www.networknature.eu) of the European Union, and the BiodiverCities by 2030 initiative of the World Economic Forum and the Government of Colombia (www.weforum.org). In Bangladesh, with highly vulnerable coastal towns and increasing climate-induced displacements, urban local government institutions and development partners should make NbS an integral part of urban development strategies and plans (Irfanullah, 2021b). NbS interventions in urban settings should restore and manage urban ecosystems and biodiversity, help to address conflicts over natural resources, and ensure social equity within the expanding urban slums. Local institutions and communities should either lead or be appropriately and sufficiently involved in planning, executing, and monitoring NbS in towns and cities.

It is important to understand the scope, effectiveness and limitations of NbS to avoid any miscommunication, misuse and misinterpretation (Irfanullah, 2021a). Funding and implementing agencies and other stakeholders should abide by the available standards and guidelines when designing, implementing, and scaling up NbS initiatives. The IUCN Global Standard for NbS (IUCN, 2020) brings together experience from 100 countries. It guides stakeholders to co-define societal challenges so that they can co-design suitable NbS at an appropriate scale, and checks that NbS are economically feasible, provide sufficient biodiversity and human well-being benefits, involve all stakeholders equitably, and manage trade-offs. Similarly, Seddon et al. (2021) urge stakeholders to follow four guiding principles: NbS are not a replacement for the rapid decarbonization of the economy; they should involve a range of terrestrial, freshwater and marine ecosystems; they are designed, implemented, managed and monitored by or in partnership with Indigenous peoples and local communities; and they should provide measurable benefits for biodiversity.