This study uses a case study approach, using contrasting sites from Ethiopia and Mali to appraise restoration innovations implemented through various programs. Two distinct sites were chosen from each study case based on available datasets and documentation. In Ethiopia, the Yewol watershed in the South Wollo Zone and the Chifra and Yallo sites in the Afar regional state were considered. In Mali, the study encompassed the Bougouni and Koutiala districts in southern Sikasso (Figure 1). Though each case study is quite distinct, for example, in terms of average household farm sizes (Table 1), persistent problems of land degradation, soil erosion, limited smallholders’ access to sufficient resources to invest in agricultural intensification solutions, and lack of farm to landscape scale natural resource management are common in both cases. In Ethiopia, the study areas are in the highland sloppy catchment (>30%) with an elevation that ranges from 1,090 m to 3,170 m. Mali’s relatively flatland characterizes the study site (with slopes less than 5%) and elevation ranging from 350 m to 400 m. Agriculture dominates the land use system in both countries, with mixed-crop livestock as the primary farming system (Table 1). Before the intervention practices, the study areas represented degraded, drought-prone, and poor soil fertility conditions.

Major scientific websites such as ResearchGate, Google Scholar, and ScienceDirect were used to search literature on the sustainability gains of landscape management practices. The literature search did not impose a time limit but considered two case studies from Ethiopia and Mali, representing East Africa and West Africa, respectively, under contrasting conditions. Literature outside the two countries was considered on a limited case to show comparable advantages of implemented practices. Keywords used were ‘degraded landscape,’ ‘agricultural productivity,’ ‘sustainable development,’ ‘integrated practices,’ ‘sustainable landscape management,’ ‘farmers perception,’ ‘Ethiopia,’ and ‘Mali.’ Criteria for further screening of the search results were done based on criteria such as (i) the presence of innovation sites for validating and co-learning landscape management practices, (ii) community or end users’ participation in the landscape management process, (iii) the existence of practices at different scale (field/plot, farm, and landscape), (iv) land use and land cover changes, and finally (iv) outputs that signify agricultural water and land sustainability options. Most of the articles (over 96%) used in the literature were published in the last 10 years, a period that signifies the establishment of the selected innovation sites in both countries.

In the rainfed highland of northern Ethiopia, a good example of an innovation site is the Yewol mountain landscape’s rehabilitation initiated in 2013. Landscape intervention practices were implemented jointly through public works of the Productive Safety Net Program (PSNP), where rural communities played an active role in the implementation process (Child et al., 2021). Activities were co-designed and driven by the needs of communities and local partners, and through local government support, approximately 7,500 hectares of degraded landscape were restored. In the co-designing process, integrated solutions of locally adaptive practices such as soil and water conservation, fertility management, improved crop and fodder varieties, contour ditches, and agroforestry practices were implemented through investing in public works. Another example of an innovation site is in the dry lowlands of Afar Regional State, where two learning sites were established in 2015 to benefit more than 200 agro-pastoral households. The region is a drought-prone area with low rainfall and high temperatures. The lowlands suffer from recurrent droughts and flash floods emerging from adjacent highlands. Flood-spreading weirs, an engineering structure aimed to divert and spread flash floods and create recession farming, were identified as innovative solutions to promote agro-pastoral intensification through food and fodder varieties in a purely pastoral farming system.

In Mali, under the Africa Research In Sustainable Intensification for the Next Generation (Africa RISING), innovation sites were established in the districts of Bougouni and Koutiala of the Sikasso region. The Africa RISING program commenced in early 2011 to showcase different technologies and practices through the inclusive engagement of government extension agents, NGOs, and R4D partners (Africa RISING, 2018). Reducing soil loss, enhancing water utilization, and integrating soil fertility management were a few of the expected outputs from the innovation sites (ICRISAT, 2022).

Evidence-based data monitoring and landscape management assessment help to demonstrate how socio-economic and environmental benefits are improving beyond productivity gains. We selected the Sustainable Intensification Assessment Framework (SIAF) with its comprehensive approach encompassing five domains, namely productivity, economic, environmental, social, and human condition, and featuring scales ranging from plot-level to landscape-level, all facilitated by simplified yet robust indicators (Musumba et al., 2017). This framework was chosen for its ability to provide a holistic and nuanced understanding of sustainable intensification, allowing us to assess the effective and multifaceted impacts of integrated landscape management practices across diverse contexts. Selected indicators and metrics for each domain were considered essential and present assessment methods to show broader implications and interactions of sustainability at multiple scales. Sustainability indicator outputs were adopted from previous works (Snapp et al., 2010; Zurek et al., 2015; Frelat et al., 2016). Table 2 presents the adopted approach for this study. Results were compared against the outputs obtained for each indicator in the traditional or non-intervention practice.

Before intervention practices, yields of major staple crops such as wheat and barley in the Yewol watershed were stagnant at 0.7 tons/ha and 0.8 tons/ha, respectively. Crop yields increased significantly after introducing in-situ and ex-situ management practices such as hillside terraces, Fanya juu, and contour ridge tillage. Farmers could diversify cropping systems by adding other crops such as faba bean and potato into their farming practices. Compared with untreated plots, implementing management practices reduced soil loss in treatment plots (Table 3). In places where crop covers and green manuring were integrated, plot-level soil loss was reduced below 10 tons/ha (Desta et al., 2021). At the farm level, the introduction of new crop varieties (such as faba beans and potato) integrated with the in-situ and ex-situ management practices enabled a reduction of runoff and soil loss by 27 and 37%, respectively. Faba beans are nitrogen fixing and enhance soil fertility and if incorporated regularly into diets, enhance household nutrition. They are also relatively resilient to varied rainfall. By increasing the cultivable land by 44%, household income was raised by selling potatoes and agroforestry products (Yaekob et al., 2020). As shown in Table 3, farmers were able to obtain an income of $140 per season from the sale of potatoes and $115 from agroforestry products. This is an opportunity for Yewol watershed communities to benefit from the implementation of ILM practices. At the watershed level, significant improvements in sediment load reduction (by 75%) and groundwater recharge were observed (Gumma et al., 2021). The vegetation covering the entire landscape increased from 13 to 29%. A five-fold increase in irrigable land from 200 ha to 970 ha (Kane-Potaka, 2015) enabled improvements in the food security status of the Yewol community (additional benefits are presented in Table 3 and in ICRISAT, 2018).

In the dry lowlands of the Afar regional state, participatory flood-based farming using water spreading weirs (WSW) helped convert degraded dry rangelands to be productive and support diversified crop and fodder production. As highlighted in Table 3, the major staple crop (maize) grain yield increased by 3.5 to 5.8 tons/ha. Before the intervention, farmers were getting biomass yields of less than 7 tons/ha, with the use of WSW biomass of maize increased by over twofold (13.4 to 20 tons/ha). Additionally, the 80% increase in fodder biomass provided feed for calves and milking cows, reducing calf mortality from 35–40% to 5–8% in the dry season (Getnet et al., 2020). This is an important livelihood benefit to the livestock community of Chifra and Yallo communities in the Afar regional state. At the landscape level, increments in cultivable land (from nil to 44%), vegetation cover (from 13 to 29%), and reductions of degraded rangelands from 87% to just 28% were the environmental benefits gained by introducing flood recession farming (Gumma et al., 2021). Improved land quality (soil moisture and nutrients) benefitted the community by enabling the introduction of new dryland crops and fodder varieties into the flood-based farming system (Amede et al., 2023). The primary forage types utilized by farmers include desho grass (Pennisetum glaucifolium), vetch (Vicia sativa), and Napier (Pennisetum purpureum; Adie and Blummel, 2019). Additionally, pigeon pea and sesbania species were introduced to increase the amount and quality of available forage. Apart from utilizing them for feed, farmers utilized improved forage species for soil and water conservation (desho grass), fencing, and as a windbreak (sesbania and leucaena; Mijena, 2022). With fodder trees and shrubs, soil fertility was improved, and landscape level erosion was controlled. Additional benefits include the provision of firewood and timber. Challenges are related to the shortage of land, lack of capital to purchase improved forage seed, and poor extension services, among others (Zereu and Lijalem, 2016; Melese et al., 2019). A gradual change in land allocation from communal use covered by grasses and shrubs to a fenced agro-pastoral farming system enabled community leaders to negotiate with communities to protect crops from livestock (ICRISAT, 2021). Further, with water-spreading weirs, crop and fodder production could be scaled to over 1.27 million hectares in different seasons (Gumma et al., 2021). Such innovations co-validated at the learning sites are one entry option to support the government policy to expand dryland irrigated agriculture, for example, to the dry lowlands of the Somali and Oromia states (OICR, 2018).

In southern Mali, contour bunding technology (CBT) is the most widely adopted management practice from farm to landscape. Improvements in crop productivity, erosion control, and increased household incomes are a few of the reported benefits (Table 3). Regarding crop productivity, the introduction of contour ridge tillage in farmers’ fields increased yield of sorghum and maize by 33, and 63%, respectively. Compared to non-treated fields, the biomass of sorghum and maize increased by 42 and 64% in the Koutiala and Bougouni innovation sites, respectively. The overall advantage of CBT technology was reflected at the household level by providing improved incomes. For example, the net benefit generated by sorghum production with contour bunding was about 24% higher than that generated without contour bunds (Birhanu et al., 2019). Improvements in water productivity and soil moisture at different soil layers, reductions in landscape runoff coefficients, and soil loss rates are some of the benefits of ILM practices (detailed data is presented in Table 3). The majority of communities living in the watershed (>78%) witnessed higher income by selling sorghum and maize crops planted on farms treated with contour ridges. Integrated landscape management practices combining trees and crops have helped manage agricultural resources and benefit from ecosystem services (Dembele et al., 2021). The most common tree species used for erosion control and dry-season livestock feed include Leucaena leucocephala and Gliricidia Sepium. Additionally, several tree species, such as Vitellaria paradoxa, Parkia biglobosa, Adansonia digitata, and Tamarindus indica, are left in farmers’ fields to grow with crops to serve as a shed and improve soil fertility. In addition to the farm level and household benefits, CBT was evaluated at the watershed scale, with measurements performed on water table levels in 2017 and 2018. As shown in Figure 2, toward the end of the rainy season in October 2017, recorded water levels below the ground surface at the watershed outlet, CBT plots, and non-CBT plots were 1.52 m, 3.11 m, and 5.08 m, respectively. Improvements were observed after a year in 2018, where the respective recordings were 0.86 m, 2.44 m, and 4.33 m. CBT implementation and recharging of the subsurface soil moisture have implications for facilitating better water uptake for crops grown near CBT ridges, especially during the growing season in Mali, where dry spells are common. A recent study from the innovation sites by Sanogo et al. (2023) highlighted that the highest rates of erosion (7.8 to 17.53 t/ha/year) were recorded from landscapes without CBT and hence resulted in significant deficiency (p < 0.05) of soil nutrients in most untreated agricultural fields.

The analytical study using sustainability indicators enabled us to understand gender roles and decision-making power at the household level. In male-headed households, men have a high level of land ownership and hence control income on productions made from farms with ILM practices. However, women in female-headed households indicated high decision-making power concerning income. In most households (60%), household heads and adult males are primarily responsible for implementing ILM practices. Communities living in the innovation areas were beneficial in providing feedback regarding other management practices implemented at the landscape level. Socio-economic data from 112 farm households in the nine innovation sites in Bougouni and Koutiala enabled us to understand that water productivity was consistently higher in farms treated with ILM practices (Table 3). For example, solar based irrigation systems enabled local communities to increase household vegetable consumption by 55% and household incomes by 40% (Birhanu et al., 2023). Other studies conducted in the innovation sites at the plot level further enabled the understanding and usefulness of farm-based intervention practices, such as vegetable sack gardens to increase household income and nutrition (Badolo et al., 2022; Govoeyi et al., 2022; Tignegre et al., 2022).

Though the idea of having a standard operation unit, such as innovation sites, was common in the studied cases, the collective effort in rehabilitating degraded landscapes and co-validation of management at different scales were guided by specific contexts in the study cases. In Ethiopia, for example, integrated watershed management was considered a key area of intervention by the government to address developmental challenges related to land degradation. According to the World Bank report, four strategic operations were necessary to achieve these: (i) soil restoration, (ii) management of rainwater, (iii) land use and vegetation management, and (iv) diversifying livelihood options (World Bank, 2019). Thus, these different operations are implemented in a watershed approach through the community’s multi-stakeholder collaboration and collective actions. The government’s ambition was to deliver intervention practices in 5,000 community watersheds that cover 2.5 million hectares of land (9.3% of 27 million ha degraded area) until 2023.

In terms of collaboration the Yewol watershed community contributed 60 labor days per year at a household level, and, through the lead of the PSNP public work and technical support provided by different local organizations, the entire watershed (7,500 hectares) was terraced, and 2000 households have been benefitting with more food, higher incomes, and a revitalized environment.

In Mali, the implementation of landscape management practices was promoted by the Institut d’Economie Rurale (IER) and Centre de Coopération Internationale en Recherche Agronomique pour le Développement (CIRAD; Birhanu et al., 2019). CBT was a landscape management practice introduced to improve agricultural productivity and erosion control. The CBT technology gained quick acceptance and was adopted by the Malian cotton company Compagnie Malienne de Development des Textiles (CMDT) to increase cotton productivity in southern Mali (Birhanu et al., 2022). The rollout of ILM practices considered watersheds a landscape comprising several micro-watersheds (Gumma et al., 2021; Amede et al., 2023; Sanogo et al., 2023). Practices such as terracing and earthen bunds are usually constructed in the upstream areas. The middle and lower parts of the landscape are for crops (e.g., sorghum and faba beans) and several soil fertility measures (intercropping and cereal legume rotations). Training activities are facilitated through extension agents and lead farmers at the watershed level (Birhanu et al., 2022; Amede et al., 2023). In most cases, married women are essential contributors to the agricultural labor force (Grosz-Ngaté, 2000; Doka et al., 2014). In contrast, for Malian households, decision-making on ILM practices (such as contour bunds) and economic benefits are controlled by the head of the household, who is usually an adult man (Birhanu et al., 2022). Documentation on gender dynamics is limited. The ILM practices were mainly conducted in several cases to capture excessive rainfall and maximize infiltration into the soil. A recent study by Abera et al. (2023) indicated that gender and social inclusion are rarely considered in landscape restoration studies in Ethiopia. Most studies and interventions are heavily focused on the biophysical and environmental aspects of restoration. Despite the importance of gender and social inclusion in supporting positive decision-making and outcomes (Leisher et al., 2016), scientific evidence on gender power dynamics and the role of social capital is lacking (Wossen et al., 2015; Abera et al., 2023). In some instances, the ILM approach is embedded in existing local coordination mechanisms, but the participation of women groups is limited. The study by Umutoni et al. (2016) in southern Mali highlighted that decentralized management of natural resources is dominated by a small group of individuals, often community leaders and elites. Women are marginalized and are not encouraged to participate in decision-making actively.

The co-learning from innovation sites enabled an understanding of the stability of crops under climate change conditions. In Mali, for example, by establishing innovation sites for technology co-validation and showcasing, it was possible to generate high-quality data for decision-making. Akinseye et al. (2023) utilized data collected from innovation sites in Mali from 2014 to 2021 to predict sorghum yield under changing climate scenarios. Accordingly, they reported that the yield of early maturing sorghum varieties will decline by 2.5 to 4% during 2010–2039, and a further decline by up to 5.7% would be expected in the mid-century scenario (2040–2069). In this case, helpful policy guidance was provided to adopt medium-maturing varieties that exhibit a high-yielding potential with a mean simulated yield of up to 12.2% in the near future. Other studies conducted at different scales highlighted how co-validation of technologies and participatory research at the innovation sites brought significant improvements in the management of soil, water, and integrated crop-livestock systems (Birhanu et al., 2019; Traore and Birhanu, 2019; Birhanu et al., 2022; Guindo et al., 2022). In Ethiopia, the experience obtained in the Yewol watershed paved the way to scale out integrated landscape management practices and community-led experiences to over 20 other adjacent watersheds.

Evidence-based data monitoring and sustainability assessment helped demonstrate how landscape management practices improved the socio-economic and environmental benefits beyond productivity gains. Different sets of data collected in the domains of productivity, environment, social, economic, and human well-being enable the quantification of the gains in sustainability indicators to promote restoration practices and to comply with international climate change treaties such as the Paris Agreement in reducing GHG emissions. From 2020 to 2030, for example, the government of Mali has determined a mitigation scenario of reducing GHG emissions to mitigate national carbon emissions and build community resilience. By diversifying livelihood options, increasing soil fertility, reducing soil loss, and addressing rural energy security needs using renewable energy sources, the ambitious plan of reducing GHG emissions by 29% from agriculture, 31% from energy, and 21% from forest and land use change could be achieved.

Several technical options exist to improve the sustainability of landscape management and reverse degradation. In most cases however, institutional, socio-cultural, economic, and policy barriers hinder adoption at a scale (Thomas et al., 2018). In the case study sites of Ethiopia and Mali, for example, the absence of quicker gains from restoration efforts discourages local communities from being fully committed. Additionally, NGOs and the private sector are less engaged because of initial investment risks. Other challenges are related to uncoordinated and sector-oriented development actions such as land tenure insecurity, inadequate data availability, limited technical support, and weak research-extension linkage (Haregeweyn et al., 2015; Fleskens, 2019).

In most African countries, policies promoting ILM exist; the challenge lies in pilot project-based projects with inadequate strategies for upscaling by the respective governments (Chirwa et al., 2015). Sometimes, the situation is global as well. In Sweden, for example, ILM practices are hindered by institutional and regulatory barriers and the timely availability of sufficient funds (Dawson et al., 2017). In Ecuador, the barriers to large-scale adoption of ILM (for example, Agroforestry) are rooted in the misalignment between small-scale farmers and beneficiaries at large (Buck et al., 2020). Mismatch of political cycles and the absence of proper restoration objectives were identified as scaling challenges in the ILM practices in Ecuador (Wiegant et al., 2020). In Ethiopia and Mali, successful restoration practices are documented from specific localities by specific studies with little luck in collecting long-term research data that supports designing and targeting sustainable action plans in the face of current and future climate challenges (Abera et al., 2023; Birhanu et al., 2023).

Competing interest in natural resource use and the absence of guiding bylaws in natural resource use management is another challenge worth mentioning (for example, Umutoni and Ayantunde, 2018). Promoting and broader scaling of ILM requires collective effort and commitment that includes government extension systems, NGOs, communities, and others interested in investing in management practices and combating long-term impacts (Dawson et al., 2017; Plieninger et al., 2020). For a successful scaling strategy, efforts are required to protect existing ecosystems and rehabilitate degraded areas. Additionally, investments in soil health, good agronomic practices, timely funding requirements, and monitoring practices are crucial (Mansuy et al., 2020; Peng et al., 2022; Abera et al., 2023). Often, there needs to be more balanced investments in the different sustainability options.

In most cases, extension service agents need upskilling and capacity strengthening for example through the farmer field schools (Fleskens, 2019), to enable them to utilize digital technologies to facilitate data monitoring and guidance of implementation practices. The study has been limited to two restoration innovation sites in Ethiopia and Mali. We recommend the inclusion of sustainability and livelihood gains from other SSA regions for more inclusive decision guidance in integrated landscape management practices.