Food Sec. DOI 10.1007/s12571-016-0633-3

REVIEW

Increasing productivity and improving livelihoods in aquatic agricultural systems: a review of interventions O.M. Joffre 1 & S.A. Castine 2 & M.J. Phillips 2 & S. Senaratna Sellamuttu 3 & D. Chandrabalan 4 & P. Cohen 2,5

Received: 27 August 2015 / Accepted: 23 November 2016 # Springer Science+Business Media Dordrecht and International Society for Plant Pathology 2017

Abstract The doubling of global food demand by 2050 is driving resurgence in interventions for agricultural intensification. Globally, 700 million people are dependent on floodplain or coastal systems. Increased productivity in these aquatic agricultural systems is important for meeting current and future food demand. Agricultural intensification in aquatic agricultural systems has contributed to increased agricultural production, yet these increases have not necessarily resulted in broader development outcomes for those most in need. Here we review studies of interventions that have sought to improve productivity in aquatic agricultural systems in Bangladesh, Cambodia and Zambia. We review evidence of development outcomes from these interventions and the particular role of participatory approaches in intervention design and deployment. There was evidence of increases in productivity in 20 of the 31 studies reviewed. Yet, productivity was only measured beyond the life of the intervention in one case, income

and food security improvements were rarely quantified, and the social distribution of benefits rarely described. Participatory approaches were employed in 15 studies, and there was some evidence that development outcomes were more substantial than in cases that were less participatory. To explore the impact of participatory approaches further, we examined five empirical cases. Review and empirical cases provide preliminary evidence suggesting participatory approaches contribute to ensuring agriculture and aquaculture interventions into aquatic agricultural systems may better fit local contexts, are sustained longer, and are more able to deliver development benefits to those most in need. A worthy focus of future research would be comparison between outcomes achieved from interventions with differing levels of participation, and the social differentiation of outcomes. Keywords Aquatic agricultural systems . Productivity . Food security . Nutrition . Income . Participatory approaches

Electronic supplementary material The online version of this article (doi:10.1007/s12571-016-0633-3) contains supplementary material, which is available to authorized users.

Introduction * O.M. Joffre [email protected]

1

WorldFish, WorldFish Greater Mekong Office, #35, Street 71 Sangkat Boeung Keng Kang 1, Phnom Penh, Cambodia

2

WorldFish, Jalan Batu Maung, Batu Maung, 11960 Bayan Lepas, Penang, Malaysia

3

International Water Management Institute (IWMI), Southeast Asia Regional Office, Vientiane, Laos

4

Bioversity International, 43400 Serdang, Selangor Darul Ehsan, Malaysia

5

Australian Research Council Centre of Excellence for Coral Reef Studies, James Cook University, Townsville 4811, Australia

Sustainable Development Goals (SDGs) highlight poverty eradication as one of the biggest global challenges today. Despite significant progress in addressing the needs of the world’s poorest at the beginning of the twenty-first century, approximately 1.2 billion people still live in extreme poverty, and 800 million do not have access to adequate food.1 The link between improved agricultural food systems and the alleviation of extreme poverty, hunger and malnutrition in the developing world is increasingly recognized and reflected in 1

UN MDG1fast facts (Sept 2013): While the proportion of undernourished people globally decreased from 23.2% in 1990–1992 to 14.9% in 2010–2012, this still leaves 870 million people—one in eight worldwide—going hungry.

Joffre O.M. et al.

the SDGs (CGIAR 2015). Aquatic agricultural systems, defined as floodplain or coastal systems in which the annual dynamics of fresh- or brackish water contribute significantly to the production of crops, livestock and fisheries, support the livelihoods of over 700 million people in developing countries (CGIAR 2012). There are five common components of aquatic agricultural production systems; fisheries, aquaculture, field cropping systems, horticulture and livestock. These components are intimately connected by the flow of nutrients, water and ecosystem services. Overlaying and influencing these components are the human dimensions of aquatic agricultural systems that are driven by complex social structures, with different groups of people, with differing values and priorities, having different access to opportunities, information, agricultural inputs and markets (Fig. 1). The social and ecological complexity of aquatic agricultural systems presents a substantial challenge in implementing and delivering interventions that ‘fit’ local contexts, and lead to improvements to productivity and equitable development outcomes. In recent decades fishing and farming practices have intensified in pursuit of short term increases in productivity, however, as a result this has led to environmental degradation, loss of ecosystem function, and ultimately diminishing productivity in many aquatic agricultural systems (Foley et al. 2005; Datta et al. 2012). Within the next 20 years, driven by rapid population growth2 and given the fact that food demand is set to increase by as much as 70% (Pretty et al. 2011), reliance on aquatic agricultural systems for food production is likely to increase. The expected increase in future food demand has sparked a resurgent investment into agricultural intensification, taking into consideration the valuable lessons that came out of the Green Revolution (Pingali 2012). The Green Revolution (1940s – 1960s) led to substantial improvements in the productivity of major staple crops in developing countries. Yet, despite these improvements, poverty and food insecurity persist. Five main reasons are put forward to explain the failure for the gaps between improved productivity and delivery of development outcomes. First, new technologies were developed for the relatively less marginal areas where productivity improvements were easily achieved. In marginal, rainfed areas where poverty rates are the highest, new technologies from the Green Revolution were not suitable. Second, the benefits did not reach those who needed them most, households not having sufficient resources and skills to innovate. Third, new technologies were disseminated but they often bypassed marginalized groups due to, for example, unequal access to land, poor value chain development and only access to credit or research and policies that support large farms to the detriment of smallholders (Hazell 2003). Women farmers 2

Current UN projections indicate that world population could increase by 2.25 billion people from today’s levels, reaching 9.15 billion by 2050 (Alexandratos and Bruinsma 2012).

benefited less from agriculture innovation, with innovation centered on male farmers (Doss 1999; Paris 1998; McIntyre et al. 2009). Fourth, despite the overall gain in intake of calories, the nutritional benefits have been uneven, with persisting micronutrient malnutrition due to decreased dietary diversity (Webb 2009; Cagauan 1995). Fifth, increases in productivity resulted in a substantial environmental impact, and in many cases, extending beyond the area cultivated (Pingali and Rosegrant 1994). Despite these insights from the Green Revolution into the failures of productivity improvement to deliver equitable or substantial development benefits, there remains a critical need for research to support the development of solutions to sustainably increase productivity3 in marginal farming areas, and that lead to the equitable distribution of benefits amongst the social groups who are most in need (Pingali 2012). An increasing body of evidence suggests that this will be best achieved using a participatory approach to agricultural (Scoones and Thompson 1994; Lilja et al. 2001; Buhler et al. 2002; Hoffmann et al. 2007) and natural resource management interventions (Lilja et al. 2001; Johnson et al. 2004). Evidence emerging from these studies suggests that participatory action research helps to ensure that technologies are tailored to the people using them, and that farmers’ ability to learn, adapt and innovate, is improved through their participation in experimenting within their systems. Despite these benefits of participatory approaches in agricultural research and development, to date there has been no systematic review of the outcomes from applying elements of participatory approaches in aquatic agricultural systems. This review aims to not only identify the productivity gains and associated development outcomes resulting from innovations in aquatic agricultural systems, but also to understand what elements of participatory approaches are used in the design and deployment of technologies. We provide an early exploration of how participatory approaches may be linked to outcomes. The paper reviews peer-reviewed and grey literature on reported productivity interventions in aquatic agricultural systems in three developing countries; Bangladesh, Cambodia and Zambia, where large proportions of the population are reliant on the productivity of alluvial flood plains. The focus is on aquatic agricultural systems because of their inherent productivity, but with the apparent paradox that these systems cannot sustain the needs or larger numbers of poor people now reliant on that productivity. We supplement our review of the published literature with analysis of published and unpublished evidence from case studies of ongoing interventions, 3

Although the words Byield^ and Bproductivity^ are sometimes used interchangeably in the literature, yield refers to the volume of food produced per unit of area and productivity refers to the volume of food produced per unit of input. Depending on the methods used to calculate productivity, these inputs may include all or some of the following: seed, fertilizer, water, land, energy, finance or labor.

Increasing productivity and improving livelihoods in aquatic agricultural systems

Fig. 1 Aquatic and agriculture based food production systems and the interactions between their components

within each country where the authors have direct experience. The review addresses the following three key research questions: & &

&

To what extent, and how have interventions improved productivity in aquatic agricultural systems? What development outcomes4 have been realized from these interventions, in terms of income, nutrition and food security outcomes, particularly for the poor and marginalized? What elements of participatory approaches have influenced development outcomes from interventions broadly, and from productivity increases specifically?

In section 2 of this paper we present the methodology used to search for, and select, the specific cases for the review. Section 3 presents the analysis of the outcomes of each intervention on the productivity of aquatic agricultural systems and the extent to which the intervention delivers desirable development outcomes in terms of income, food and nutrition security. In the same section, we review how the population 4 Here we consider outcomes as Bthe likely achieved short-term and mediumterm effects of one or more interventions outputs’ (OECD 2010) and distinguish which impact can be observed on a longer term after a development intervention and is most likely a result of the combination of outcomes and other factors.

targeted by the selected intervention was defined, before answering the third research question and present results from five selected case studies. In the fourth and last section, we discuss the current knowledge on the influence of interventions on productivity and the complementarity of conventional and participatory approaches before identifying the knowledge gaps and improvements needed to sustainably increase productivity in AAS to benefit those most in need (section 4).

Method Through a literature review we examined the development and deployment of technological interventions in aquatic agricultural systems and their effect on system productivity and development outcomes. We use the term Binnovations^ to encompass traditional technological innovations, such as improved tools and inputs, but also consider non-technological innovations to encompass application of new farming management practices, policy developments, new forms of social organization or collective action, and mechanisms that foster learning at individual or collective levels. We focused on three development outcomes based on the AAS CGIAR Research Program (WorldFish 2011), to assess the innovations’ outcomes: improved income, improved nutrition and improved food security of poor rural households (Table 1) in the aquatic

Joffre O.M. et al. Table 1 Indicators of productivity and development outcomes

Productivity

Indicators Yield (absolute value ton or kg, relative change %, per area unit ton ha−1, kg ha−1) Water productivity (kg ha−1 mm−1), refers to water use in surface irrigation (including rainfall and irrigation) Total Factor Productivity (TFP) index, is defined conceptually as the ratio of an index of total output to an index of all factor inputs used to produce this output

Development Outcomes Improved income

Direct indicators: household income, net profit, net income or gross margin (US$ per household) Indirect indicators of income: US$ per area unit (ha), % change of net margin per area unit (ha)

Improved nutrition

Dietary variety per household (number of field crops cultivated by household); Consumption of calories, protein and micronutrient-rich food per household Nutritional value of crops (glycemic index of rice varieties) Micronutrient intake per household (vitamin A and calcium) Prevalence rates of stunting, wasting and underweight among children

Improved food security

Food consumption per household (percentage of change and absolute value of intake of food in kg per household) Percentage of food produced and consumed per household, food production per household (kg per household), harvesting frequency. This indicator relates to the food provision delivered by the production system to the producing household.

agricultural systems of Bangladesh, Cambodia and Zambia. We selected cases that spanned regions (South Asia, South East Asia and Africa) in order to examine outcomes across very different historical, ecological and social contexts within which a wide array of productivity interventions have been tested and deployed. We added cases from these regions that were a part of the AAS CGIAR Research Program. Selection of productivity interventions The review process was inspired by a recent scoping review implemented by Béné et al. (2016) and the review methodology was adapted from Arksey and O’Malley (2005) and Levac et al. (2010). The scoping review process was broken down into 2 steps: i) systematic review and selection of the documents (detailed below); and ii) extraction of data (detailed below). The selection process started with an inventory of peerreviewed journal articles and project reports of aquatic agricultural productivity compiled for each of the three countries following an online literature search (Web of Science and Google Scholar). We used combinations of the following search terms in Web of Science and Google Scholar: Bproductivity, yield, household, homestead, food security, nutrition, income, agriculture, aquaculture, integration, natural resource management, food production, poverty, fish, crops, livestock^ with BBangladesh, Cambodia, Zambia^ (Castine et al. 2013). The reference lists of pertinent articles within the inventory (i.e. those that were identified in the initial

search) were also screened for relevant publications and added to the process (Levac et al. 2010). The inclusion/exclusion criteria were: documents in English language, geographic focus (Bangladesh, Cambodia and Zambia floodplains), documents that described an innovation intended to improve productivity of agriculture and/or aquatic systems, and quantified changes in productivity as a result of the innovation. To conduct this assessment, we screened the title, abstract and conclusion and considered relevant documents that described and directly reported productivity interventions in aquatic agricultural systems in a report, journal or book format (i.e., media reports were excluded). After a first screening, documents were selected from the inventory by tabulating the key findings from each study and generating a score based on the following criteria: i) a clear description of the intervention; ii) documenting a change in productivity; iii) reporting qualitative or quantitative changes in income; iv) reporting qualitative or quantitative changes in nutrition; and v) reporting qualitative or quantitative changes in food security. Studies that scored at least 2 on a scale of 1 to 5 were selected. In addition, to evaluate the quality of the documents, we adapted the individual scoring method used by Béné et al. (2016), ESRC (2003) and Petticrew and Roberts (2006); we detailed nine screening questions or criteria (Table 2). Each document was scored as follows; a 1 was given if the document fulfilled the criterion, 0.5 if the criterion was partially fulfilled and 0 if the criterion was not met. Documents with low quality (average score below <0.5 for validity, rigor and reliability) were rejected.

Increasing productivity and improving livelihoods in aquatic agricultural systems Table 2 Criteria used to assess the quality of selected documents (from Béné et al. 2016; ESRC (2003) and Petticrew and Roberts (2006))

Indicators

Criteria

Validity

Are the findings substantiated by the data and has consideration been given to limitations of the methods that may have affected the results? Was the methodology adequate for the research question?

Rigor

Is the context or setting adequately described? Is (are) the research question(s) clear? Is the method used appropriate to answer the research question(s)? Is the method applied correctly? Is there evidence that the data collection was rigorously conducted to ensure confidence in the findings?

Reliability

Is the data analysis rigorously conducted to ensure confidence in the findings? Is the methodology adequately described to ensure confidence?

In most instances, more than one article reported on the same intervention, so these were discussed together, as a single case study. From the selected study, the participatory approach methodology was further screened following the technique described later. Given our interest in the impact of participatory processes, we supplemented our review with five cases from the CGIAR AAS Research Program from Bangladesh, Cambodia and Zambia. Cases were nominated by CGIAR researchers who were experts in focal countries and who suggested that participatory processes for these cases were well described and that there was early evidence of productivity and development outcomes. Experts provided five additional reports for analysis that met the criteria, but had not been found in our online search. Semi-structured interviews were conducted with 14 researchers and field officers involved in the intervention. The interviews included descriptions of scope and objective of the intervention, the targeted population, of the innovation and the process for selecting direct beneficiaries. During the interviews, indicators of the participatory approach were determined and levels of community engagement were collected based on the list presented in Table 3. We classified all papers identified for review, and the five cases, according to the production system targeted: i) aquatic, ii) agricultural and iii) integrated aquatic agricultural production systems. We reviewed evidence of development outcomes and examined the use and impact of participatory approaches. Productivity improvement and development outcomes There were a variety of indicators used to analyze the change in productivity, income, and food and nutrition security (Table 1). This list was populated from the selected literature. A change in productivity is quantified by increased yield in absolute value or by its relative increase compared to baseline or control site and by comparing water use efficiency. The effect of the intervention on income is assessed by a change

in household income or indirectly by estimating the increase in income per area unit. Indicators for nutritional status include dietary diversity, micronutrient intake, nutritional value of crops, food consumption per household and prevalence of stunting, wasting and underweight. Improved food security is assessed by the absolute value or relative increase in food provision delivered by the production system. In the selected publication, those indicators are usually monitored and compared with control groups or baseline to assess the outcomes of the intervention. Participatory approaches The inventory of papers and reports from the literature and case studies from the AAS Research Program was subsequently screened again to identify studies which adopted key elements of a participatory approach. From the subset of AAS Research Program, case studies, which used key elements of participatory approaches, we identified the primary researchers involved in the study and interviewed them to gain a better understanding of the projects’ implementation and participatory approaches (Table 3). Only studies from Bangladesh and Cambodia were selected because case studies from Zambia were limited and did not include sufficient information regarding participatory approaches. While the AAS program adopts a participatory action research (PAR) approach with a focus on experimental learning and empowerment of stakeholders and iterative processes of planning, acting and reflecting (Dugan et al. 2013), in our review we investigated productivity case studies that adopted any key elements of a participatory approach (not necessarily the entire PAR process). Since the AAS program was transitioning to include other bilateral projects that may not have used PAR in their original approach, we felt it was important to leave the criteria broad enough to capture other types of participatory approaches in addition to PAR. To capture elements of participation we looked at different dimensions of the participatory approach, namely how the

Joffre O.M. et al. Table 3 Participatory approach indicators and levels of community engagement throughout the research process (adapted from Biggs 1989; Chambers 2007; Apgar and Douthwaite 2013; Neef and Neubert 2011)

Criteria

Indicator of adopting elements/components of participatory approaches

Involvement of community in research process

Was the community involved in the research process and from which stage? Role of community members?

Use of participatory planning tool

Role of researchers? Was a visioning tool used? Was a community action plan developed with the community? Is there evidence that poor and marginalized communities were involved in action planning?

Type, frequency, and intensity of interaction

How often did community members come together to reflect and discuss the process? Who came together? Have there been situations in which you had to adapt and revise your plan? How was progress measured and did both community and researchers learn during the process?

Researcher’s participation

What was the level of participation/engagement of the researcher in research design and implementation? What was the level of community participation in design, implementation and results analysis?

Inclusiveness

Who identified the problem, needs and priorities? Which actors were represented in decision-making and how? Who were the target groups and did they include both men and women? Did they include poor and more vulnerable communities?

community was engaged, the characterization of target group(s); tools and methods used to plan the research, and the interaction between researcher and communities (Table 3). The criteria used to assess participatory approaches employed in each of the case studies were selected to broadly capture i) the community’s involvement in the research process such as identification of the needs and priorities, research implementation, observation and analysis of the intervention and reflection on the results; ii) the use of a participatory planning tool; iii) type, frequency, and intensity of interaction (derived from Neef and Neubert (2011), Apgar and Douthwaite (2013) and Ratner et al. (2013)). BInclusiveness^ was estimated based on interviews with researchers and field officers for the case studies and on indication of target groups, such as women and the poor, in the different selected studies (Lambrou 2001). Based on those indicators, the degree of community participation in research activities varied from limited (positivist theoretical research), to increasingly higher levels of participation in the case of consultative, on-farm testing, evaluation, collaborative planning and partnership types (Lambrou 2001).

Results The web search returned a total of 387 peer reviewed journal articles and book chapters, of which 34 were selected for review. During the selection process three documents incurred scores below 0.5 and were therefore discarded due to

inadequate methods. Average score for validity, rigor, and reliability of the selected documents per category of interventions are found in Table 4. The remaining 31 selected documents presented low to high scores for the three criteria assessment system and were classified in three main categories: aquatic, agricultural and integrated production systems. Most of the literature found analyzed interventions in Cambodia (n = 10) and Bangladesh (n = 19), while studies from Zambia were scarce (n = 2). One reference included a comparative analysis of a similar innovation in both Cambodia and Bangladesh and counted only once in the total number of selected references. Productivity interventions and development outcomes The 31 selected cases presented innovations suitable for a range of aquatic agricultural system components, including fisheries and aquaculture, livestock, horticulture and crops as well as innovations suitable for integrated aquacultureagriculture (IAA) systems. The majority of these studies were from Bangladesh and Cambodia with only two from Zambia. Within the 31 case studies reviewed, 20 reported an increase in yield realized through improved management of existing systems, introduction of new cultivars, integrated farming systems or collective management of inland fishery resources, suggesting widespread opportunity for productivity improvement across aquatic agricultural systems. Observed increases in yield varied from less than 10% to more than 61% across the different types of interventions.

Increasing productivity and improving livelihoods in aquatic agricultural systems Table 4 Quality of the selected literature per type of intervention

Type of intervention

Number of documents

Validity

Rigor

Reliability

Quality

6

0.87

0.95

1

High

5

0.75

0.92

0.85

Moderate

Rice Horticulture

7 2

0.78 0.5

0.85 0.8

0.85 0.5

High Low

Livestock

2

0.87

0.95

0.87

High

4

0.75

0.85

0.87

Moderate

Aquatic production systems Aquaculture Fisheries Agricultural production systems

Integrated systems Rice-fish Ghers

3

0.75

0.83

0.93

Moderate

Integrated Aquaculture Agriculture (IAA)

2

1

0.85

1

High

Quality is ‘high’ when the three score are >0.75; ‘moderate’ at least one score ≤ 0.75 and low when 2 scores ≤0.50

Aquatic production systems The selected studies reported a range of interventions into aquatic food production systems, including capture fisheries and aquaculture, and intermediate production systems, bridging fishing and farming.

Aquaculture Aquaculture is growing rapidly in Bangladesh, but less so in Cambodia and is limited in Zambia. Productivity-improving interventions in Bangladesh aquaculture systems were reported in a range of systems from small homestead ponds to commercially-oriented aquaculture farms. We found five interventions in Bangladesh which aimed to increase productivity either through enhanced stocking of aquaculture fish species (Karim et al. 2011; Kadir et al. 2006; Roos et al. 2007), improved pond management practices or integration of agricultural production linked to homestead ponds (Azim et al. 2004) or production of fingerlings (Barman and Little 2011). Yield increases of 25% and 59% were reported in homestead ponds using new pond management techniques (Azim et al. 2004), by producing fish fingerlings in ponds (Barman and Little 2011), or addition of small indigenous fish species (SIS) with high nutritional value for household consumption (Roos et al. 2007). Increasing nutrient inputs achieved an increase yield of 2 tons per hectare in a carp and tilapia polyculture area of Bangladesh where tilapia is accepted by farmers, but this increase in productivity requires a significantly higher input cost (Karim et al. 2011). Additionally, 650 kg of vegetables were produced along the pond’s dikes (Karim et al. 2011). Various improved Bangladesh-based pond management innovations leading to an increase in fish production between 90 and 940 kg ha−1 compared to existing practices were reviewed (Thompson et al. 2006).

Improved pond management innovation resulted in an increase in gross margin of between US$ 150 and 200 per household in peri-urban areas (Karim et al. 2011) or increased net returns of US$ 40 to US$ 445 per household per annum compared to the control group (Thompson et al. 2006). Introducing SIS in homestead ponds with carps increased household income by 13 to 24% (Kadir et al. 2006). However, in other studies it was noted that increased productivity does not necessarily lead to higher income for farmers (Azim et al. 2004). Household food provision can be increased in the case of fingerling production (Barman and Little 2011) and with regular harvesting of SIS and has the potential to improve nutrition security, as SIS are rich in vitamin A, iron, zinc and calcium (Roos et al. 2007). The quality of the documents selected is ranked high, with well-designed studies and robust methodology in which possible biases are discussed. Fisheries Seasonal floodplain and perennial water body fisheries cover 20% of the country territory in Zambia (Mukanda 1998), 30% in Cambodia (So and Touch 2011) and up to 88% of the arable land in Bangladesh (BBS 2006). In Cambodia, inland fisheries provide the main source of animal protein and are an important livelihood for local populations, making substantial contributions to national food security (Baran et al. 2001). Interventions in fisheries have attempted to better integrate productivity-improving measures with measures focused on conservation of natural resources. Interventions include turning management over to community groups or individuals (Viseth et al. 2008; Nao 2009; Thompson et al. 2003; Amilhat et al. 2009), implementing fishing quotas (Thompson et al. 2003), and stocking fingerlings (Mustafa and Brooks 2009; Thilsted and Wahab 2014). The selected documents were of moderate quality due to the methodological challenge of assessing and monitoring change in fish catch, and impact at the household level in systems with high

Joffre O.M. et al.

variability, and open systems that in some case are influenced by many other factors external to the food production system. The use of Community Fish Refuges (CFRs) was reported as an intervention used in Cambodia. This innovation relies on community engagement to protect dry season ponds (water bodies that retain water throughout the year) from fishing and to use them to provide a fish refuge or sanctuary during the dry season (Joffre et al. 2012). During the flooding season, fish broodstock and fingerlings from the refuge pond can migrate to flooded rice fields, canals and streams, which then also become fishing grounds. The rice-field fisheries are open access and contribute to a large extent to the livelihood portfolio and food security of communities in the Cambodian floodplain (Gregory et al. 1996). The Cambodian Government supports CFRs through its fisheries and aquaculture development plan by setting the ambitious target of 1200 CFRs operational by 2019 and thus channeling development aid toward this particular intervention. In nine locations, CFRs were implemented and led to 20 to 50% increases in annual fish catch landed in fishing communities (Viseth et al. 2008; Nao 2009). In Cambodia, CFR innovation was reported to increase household income from US$ 18 to US$ 924 per fishing season (Nao 2009). The outcome on nutrition was not quantified. Household food security improved significantly, with 40 to 90% of the fish catch consumed by the household (Viseth et al. 2008; Joffre et al. 2012). Results from managed flooded rice field fisheries at the farm level were highly variable with reported changes in fish catch from −11% to +127%, depending on the technical intervention (Amilhat et al. 2009). In Bangladesh, stakeholder groups developed management rules to prevent unsustainable fishing practices and fish catch increased by 70% to 100%, while community management of fisheries could improve household income by 37% with a net benefit per hectare of US$ 214 to US$ 541 ha−1 (Thompson et al. 2003). However, neither study provides an estimate of the increase in revenue for the fishers dependent on these resources, nor was the impact on fish consumption and nutritional status of households reported. Additional stocking of carp fingerlings into fishery systems increased the productivity of adjoining rice fields by 300 to 700 kg ha−1 (Mustafa and Brooks 2009). In a recent innovation, small nutritious fish were stocked into floodplain fisheries with qualitative evidence that production and availability of nutrient-rich small fish was enhanced (Thilsted and Wahab 2014). Agricultural food production systems Rice Rice interventions in Bangladesh and Cambodia include the use of both high-yielding varieties (HYV) and stress tolerant varieties and the alteration of rice field management techniques to include System of Rice Intensification (SRI) techniques. This cluster of documents presents an average

high quality score even if some documents have a lower rigor and reliability score, with lack of discussion about methodological limitations. The introduction of HYV of rice in Bangladesh in the late 1960s spearheaded the Green Revolution with a recorded average yield increase from 1.7 to 3.5 metric tons ha−1 between 1970 and 2001(Hossain et al. 2006). The rice varieties were selected for their higher yield, shorter growth cycles and tolerance to a range of biotic and abiotic stressors (salinity, drought and submergence) (Hossain et al. 2006). This technology is usually paired with a modification of the rice production system and rice field management that comprises the improvement of irrigation and drainage systems, use of commercial fertilizers and pesticides and other inputs. In the coastal zone of Bangladesh, improving management of rice production systems was made possible by the introduction of salttolerant varieties, modifying the planting calendar to limit risk of abiotic stress due to cold weather and efficient water management. This package increased yield by 2.3% per annum between 1970 and 2000 (Mondal et al. 2010). Shankar et al. (2004, 2005) modelled the impact of HYV intensification for rice field fisheries and proposed new management practice to mitigate trade-off between those two important components of Bangladesh food production systems i.e. rice and fish. The System of Rice Intensification (SRI) introduced in Cambodia is based on a set of principles applied to rice culture. It includes early transplanting of rice seedlings at lower density than normal to allow for manual weeding and to expose the rice plant to more resources (light and nutrients). It also uses organic fertilizers and alternate flooding and drying of the plot, which requires good water management capacity but reduces the amount of water used. This combination provides optimal conditions for the rice seedling to grow a deep and dense root system. The SRI technique increased water productivity by 0.95 kg ha−1 mm−1 compared to conventional techniques (Suryavanshi et al. 2013). Farm trials in Cambodia showed an increase in the yield of 27% to 61% compared to conventional techniques (Ly et al. 2012; Anthofer 2004) with a lower labor demand of 10 men per day per ha−1 during the uprooting and transplanting period, the most critical laborbottleneck period (Anthofer 2004). However, Dobermann (2004) showed that SRI is overall more labor intensive than the conventional technique and the debate as to whether or not this makes the innovation unsuitable for resource-poor farmers is still open. In Bangladesh between 1987 and 2000, the use of HYV rice, together with improved management practices and the spread of mechanization and inputs, led to a 30% and 83% increase in the economic return of dry and wet season crops, respectively, compared to traditional rice cultivation (Hossain et al. 2006). The effect on household nutritional status is questionable, however, with HYV rice having a higher Glycemic Index and lower protein content than traditional varieties

Increasing productivity and improving livelihoods in aquatic agricultural systems

(Howlader and Biswas 2009). The spread of HYV rice culture increases household food security (Hossain et al. 2006) but reduces dietary diversity following expansion of the rice cultivation area at the expense of pulses and other minor cereals (Oakley and Momsen 2005). In Cambodia, the adoption of the SRI technique increased farmers’ income by 48%, reaching US$ 153 ± 8 per households and per year (CEDAC 2008). Increased profit margins, from US$ 120 per ha to US$ 209 per ha, were reported, with larger rice supplies estimated to last between 2.2 and 4.6 months for households that converted 21% and 42% of the rice land into SRI, respectively (Anthofer 2004). Livestock In spite of over 310,000 rural households owning cattle in Zambia and cattle playing a significant role in food security and income generation (The World Bank and UKAID 2011), neither documented research on, nor evidence of, cattle husbandry and disease prevention were found. Cattle in rural Cambodia are the most important asset for rural households (Stür et al. 2002). They provide meat and milk, and manure for fertilizer and cooking and are used as draught animals. However, their productivity is low. They traditionally feed on fallow land and crop residues of low nutritional value. The intervention called Bforage bank^, aims at improving small-scale livestock productivity by cultivating plots of grass and legumes near the homestead and thus provide nutrient rich forage for local livestock (Maxwell et al. 2012). There was no clear evidence linking forage banks to improved productivity but qualitative benefits, including weight gain, ensuring calf birthing earlier in the season and securing cattle feed provisions during the lean period were observed. Time spent collecting forage for livestock was reduced, benefitting women and especially children, who were previously skipping school to collect forage for livestock (Maxwell et al. 2012). Husbandry techniques, targeting disease prevention, proved effective in Cambodia where no cases of foot and mouth disease were recorded following their implementation including training on disease risk prevention and vaccination. This led indirectly to an increase in productivity, unlike in villages where farmers only received the vaccine (Young et al. 2014). Providing access to improved husbandry techniques involving disease prevention and forage production for 645 households in Cambodia increased annual farm income from US$ 1360 to US$ 3340 and saved time for farmers compared to simple vaccination programs. This study did not investigate linkages between the intervention and improved household food security and nutritional status. Both studies were of high quality, with discussions of methodological limitations. Horticulture Dambo aquatic agricultural systems consist of seasonally flooded natural shallow wetlands used for

recession agriculture (Ndiyoi et al. 2009). They cover between 10 and 15% of Zambia’s land area (Sampa 2007). The cultivation of a legume (Glircidia sepium) with high nitrogen content and used as an organic fertilizer was tested for increased soil fertility for the following dry season crop. Yield from cabbage culture improved almost threefold from 19 tons per ha in the control treatment to 51 tons per ha using the legume as organic fertilizer (Kuntashula et al. 2006). Using organic fertilizers in dambo systems produced a return of US$ 220 and US$ 300 in two cases (Sampa 2007). Gross margins increased by almost US$400 in farms that cultivated on average 0.13 ha of cabbage on their dambo plot (Kuntashula et al. 2006). This innovation increased food supply during the lean season (Sampa 2007,) but empirical evidence of improved household food security is lacking. Even if those studies document interesting productivity interventions, the quality of the research is low, with issues regarding the validity and the reliability of the results in one of the two selected studies. Other horticulture interventions commonly found in Bangladesh were included in integrated aquatic and agricultural food systems, when vegetable production was associated with aquaculture. Integrated aquatic and agricultural food production systems The studies review various integrated systems found in Bangladesh and Cambodia. They include integrated rice and fish culture, aquaculture pond with vegetable production on the dikes in gher systems or the Multiple Purpose Farms (MPF) in Cambodia that combine rice, fish, fruit trees and horticulture within the same farm. Rice-fish systems Integrated rice-fish systems are commonly found in Bangladesh and to a lesser extent in Cambodia. They are implemented in rainfed and irrigated systems and the fish component may either occur naturally with very little management or may be managed with aquaculture inputs such as fingerlings and fish feed. In the aquaculture-style system, rice and fish are either integrated all year round or only during one season in concurrent or alternate systems, respectively (Halwart and Gupta 2004). In an alternate system, producing a fish crop before rice culture increased rice yield by 12% compared to the monoculture system (Ahmed and Garnett 2011) while an increase of 425 kg ha−1 in rice yield was reportedly common for concurrent rice-fish systems (Gupta et al. 1998). Fish productivity depends on multiple factors such as pond depth, feeding and water quality management. Reported yields in integrated systems in Bangladesh varied between 116 and 605 kg ha−1 in concurrent systems and up to 1108 kg ha−1 in alternate rice fish culture (Ahmed and Garnett 2011). The quality of the documents selected was moderate due to the lack of discussion regarding methodological limitations that may have

Joffre O.M. et al.

affected the results and gaps in the methodological framework in non-peer reviewed documents. Households adopting this innovation reported a 56% to 61% ha−1 increase in net return (Gupta et al. 1998), and a US$ 125 increase in annual net income while increased fish consumption provided dietary benefits (Ahmed and Garnett 2011). Gher systems In Bangladesh’s coastal areas, integrated rice aquaculture systems are known as gher systems (Belton et al. 2011), where rice and brackish water aquaculture (mostly shrimp culture), alternate during the year according to variation in water salinity. When the freshwater period is too short to allow rice culture, the systems shift to year-round brackish water aquaculture. Better pond management and production diversification with fish culture and vegetable cultivation on dikes can increase their productivity. Reportedly the intervention increased shrimp yield to between 600 and 1500 kg ha−1 compared to traditional ghers (100–600 kg ha−1), and further to 986 to 2209 kg ha−1 of fish when produced with additional vegetable production (Ahmed et al. 2010; Karim et al. 2011 and Wahab et al. 2012). Gher innovation is likely to increase farmers’ income but there is no clear empirical evidence with wide variability of economic results. Net profit per hectare, for a typical gher sized 0.06 to 1 ha, showed large variation, from US$ 141 to US$ 3145 ha−1 year−1 in improved ghers compared to US$ 651 ± 1065 per ha−1 year−1 in traditional systems (Alam et al. 2007; Karim et al. 2011). The increase in income was correlated with landholding size. Smallholders, with under 0.2 ha harvested fewer benefits (US$ 476 per household per year) than farmers owning more than 0.4 ha of land and earning US$ 1132 per ha per year. The poorest farmers were excluded due to relatively high production costs but the innovation provided employment opportunities for the poor and women (Ahmed et al. 2010). Effects on food security and household nutrition are also unclear. Qualitative evidence showed that household food security can be at risk in the case of crops lost to flood or conversion of the rice area to dikes and canals, limiting households’ rice production (Ahmed et al. 2010). At the same time, no significant difference in fish intake was observed when increasing pond productivity by more than 3 tons ha−1 (Karim et al. 2011). Modification of household food supply and nutrition were not explained. This group of documents were ranked of moderate quality due to the lack of discussion regarding the methodology used for on-farm data recording. Integrated aquaculture agriculture (IAA) In Bangladesh, Integrated Aquaculture Agriculture (IAA) farms are broadly defined as the concurrent or sequential linkages between two or more farming activities, of which at least one is aquaculture (Edwards 1993). Farmers receiving three-year participatory

training in IAA had a Total Factor Productivity (TFP) index of 4% to 10% which was higher than at the beginning of the project, and a TFP index of 8.9 to 10.6%, higher than farmers not receiving any training (Murshed-E-Jahan and Pemsl 2011). In Cambodia, intercropping of legumes and cassava increased the total biomass by 100% compared to cassava monoculture (Borin and Frankow-Lindberg 2005). Farmers implementing IAA innovation recorded a 22% increase of the net income (Murshed-E-Jahan and Pemsl 2011). This innovation increased household fish consumption by 21% (Murshed-E-Jahan and Pemsl 2011), with more than 40% of the fish produced consumed by households. The number of documents in this group (n = 2) resulted from the selection process in which a number of reports were discarded due to their low quality. Target groups For productivity improving interventions, target groups should be well defined in order to identify and respond to their specific needs and be adapted to their specific capacities. However, within the selected literature, references to groups targeted during technology testing or deployment are scarce. When mentioned, they are characterized by specific assets, usually land size to indicate wealth status (Murshed-E-Jahan et al. 2008; Kuntashula et al. (2006), or only ‘poor households’(Maxwell et al. 2012), without defining or clarifying what ‘poor households’ means in the context of the intervention. This lack of precision is also found in the Cambodian CFRs, where the potential target group was defined as fishers in rice fields, including part-time, full time and seasonal fishers, an open access resource where about 7.2 million people fish (Hortle 2007). This definition includes a high proportion of the population without specifying if there were differences within this large group in access to the intervention or outcomes provided by the intervention. By contrast, interventions in Bangladesh floodplain fisheries targeted the poorest households highly dependent on fisheries during both the participatory approach and early phase of the intervention design (Thompson et al. 2003; Mustafa and Brooks 2009). The targeted population was then better identified and the short term outcomes and long term impact could be better assessed. Another group of studies of productivity-improving interventions discussed a posteriori the enabling and constraining factors specific to poor, smallholder or marginal farmers. For example, although the SRI innovation is supposed to be best suited to resource-poor farmers with limited input use (Azim et al. 2004), research results narrow the target group to households with high labor availability, access to irrigation and drainage and small farms specialized in cultivation of rice plots near the homestead (Ly et al. 2012). Access to the gher

Increasing productivity and improving livelihoods in aquatic agricultural systems

system in Bangladesh is limited by the capital investment required and risk related to culture (Ahmed et al. 2010). In those studies, the intervention was not designed for a specific target population, but analysis of the intervention provided indications for future deployment, while the innovation could have been designed to meet the capacity of specific target groups. A less common trend found in the reviewed literature is to conclude on potential target groups after the intervention. For example, the SIS aquaculture innovation targets smallholders with land size below 1 ha (Murshed-E-Jahan et al. 2008) and the authors only mention women as their potential target group because the innovation is transferable to a homestead pond which is accessible to women. A similar conclusion was drawn regarding fingerling production in homestead ponds, an innovation suited for smallholders and poorest households with limited costs and land requirement (Barman and Little 2011). However, none of these studies described a specific approach to include the target groups during design or deployment of the innovation. Target groups are generally poorly documented in those interventions, beside basic wealth status. A definition of the target group and specific method to ensure the inclusion of the poor and marginalized groups at the beginning of the intervention was usually missing. Influence of participatory approaches on productivity interventions Insight from literature search Only seven studies on productivity interventions explored the relationship between the adoption of participatory approaches and development outcomes (Table 5). Other studies only mention tools and elements of the participatory approach within their research. Participatory Rural Appraisal (PRA) tools were used in four cases for assessing intervention results such as the introduction of new HYV rice (Hossain et al. 2007), assessing rice-fish systems (Oakley and Momsen 2005), selection of participants (Karim et al. 2011) or engaging with a community for collective action (Joffre and Sheriff 2011). Interventions at the community level were found in relation to fisheries (Thompson et al. 2003; Mustafa and Brooks 2009). Here, interventions were specifically designed around a community based management approach, while action research supported the design of interventions with communities. Researchers, NGOs, government agencies and beneficiaries (fishers, landowners) interacted in a participatory process to design technical interventions and new governance rules for the access and exploitation of the resource. This engagement of different stakeholders supported the inclusion of specific target groups (women and poor) within the intervention. However, the role of fishers was limited to design and implementation and they were not included in the monitoring or the

research process, led in this case by NGOs and researchers. Feedback to communities was not documented, and process of changes, adjustment in the technical intervention and how governance rules changed across the period of the monitoring were not analyzed by researchers. In other community based or collective actions (Viseth et al. 2008; Joffre and Sheriff 2011), the participation of farmers or the community remained vague. Fishers and fish farmers were consulted for understanding bio-physical or the local institutional context using a participatory rural appraisal method. But the contribution of fishers and farmers to the design of the actual intervention or the research remained limited to the contextualization of the intervention. The flow of information between researchers and participants was therefore top-down, with researchers considered ‘experts’ and farmers as a source of information to contextualize a specific technology – thus ‘adopters’ of knowledge and technology designed by ‘experts.’ In the case of interventions targeting individual households, the participatory tools and methods used by researchers were not described in detail. For example, Karim et al. (2011) mentioned the use of participatory approaches such as BFarmer Field Schools^, but the methods for implementing BFarmer Field Schools^ and the participation of farmers or other stakeholders in the intervention were largely undocumented. Five publications reviewed indicated higher levels of participation, integration of the farmers in the trial design (Kuntashula et al. 2006), and use of participatory tools such as iterative cycles of action, reflection and feedback in IAA evaluation (Murshed-E-Jahan and Pemsl 2011), fish polyculture (Karim et al. 2011), and livestock raising (Maxwell et al. 2012; Young et al. 2014). Here farmers were considered as experimenters (Karim et al. 2011; Murshed-E-Jahan and Pemsl 2011) and experts (Kuntashula et al. 2006) - providing knowledge to researchers not only about the context but also the technical design and monitoring of the innovation. For example, the participatory livestock extension program in Cambodia focused on knowledge transfer for cattle husbandry (Young et al. 2014) and forage production (Maxwell et al. 2012). But detailed information on participatory approaches during the research was scarce in these studies and the role of farmers was only described as experts and experimenters, providing information to the researcher on the type of foraging (Maxwell et al. 2012) and experimenting with extension services and new husbandry techniques (Young et al. 2014). The nature of the feedback and contribution of users to innovation design was not mentioned, so it is unclear which aspects within the participatory process were critical in achieving these results. Meanwhile, the importance of the participatory approach in some of those studies was highlighted in the discussion. Karim et al. (2011) mention the importance of trial design workshops to create dialogue between different stakeholders,

Joffre O.M. et al. Table 5 Evidence of participatory approaches used in assessing or testing new AAS innovation in the selected literature

System

Aquatic Production system Aquaculture

Fisheries

Evidence of participatory approaches and their influence on development outcomes

Participatory trials and PRA-based participants’ selection helped to engage community and included poor and non-poor. Community planning helped to develop social learning (Karim et al. 2011). Aquaculture extension with group approach and farmers involved in action research was more likely to succeed (Thompson et al. 2006). Engagement of community to protect and manage fish refuge ponds (Viseth et al. 2008). Participatory planning with community, government agencies and NGOs but monitoring and research conducted by researchers. Inclusion of the poor and women (Thompson et al. 2003; Mustafa and Brooks 2009)

Agriculture System Rice

Use of participatory tools to stratify sample in wealth and gender groups and assessment of the adoption of HYV rice helped understanding of the relationship between innovation adoption and poverty (Hossain et al. 2007). Understanding gender role in selection of HYV rice, using a gender disaggregated sample (Oakley and Momsen 2005).

Livestock

Adoption and outcomes of this approach led to time saving for children and women (Maxwell et al. 2012). Using participatory field research with farmers in addition to formal training resulted in greater knowledge, better husbandry techniques and lower cattle mortality, saved time for children, men and women and increased income (Young et al. 2014)

Horticulture

Participatory trial with farmers designing the experiments (Kuntashula et al. 2006), provided new soil fertility and pest management techniques that increased yield and were affordable for resource-poor farmers.

Integrated system Rice-fish

Limited use of participatory approach (PRA) in the case of collective fish culture in rice fields (Joffre and Sheriff 2011)

Gher System

Participatory hands-on techniques to deploy the innovation (Karim et al. 2011)

Integrated Aquaculture Agriculture

Using a participatory training approach over 3 years with small scale farmers on IAA systems led to 21% income increase and 21% higher fish consumption compared to non-project farms (Murshed-E-Jahan and Pemsl 2011)

and improve social relationships and social learning. Improvement of social capital was discussed by Murshed-EJahan and Pemsl (2011), the Participatory Adaptive Learning approach facilitating linkages between extensions services and farmers. The approach had a direct impact on farmers’ behavior, with more farmers taking a leading role in their communities’ organization and in sharing knowledge with other farmers, supporting the diffusion of the innovation. The participatory extension program helped to involve the early adopter of the innovation, support community engagement and the development of a broader learning community. Finally, Thompson et al. (2006) compared different approaches to aquaculture extension, from traditional technology transfer where farmers are adopters of technological packages provided by projects, to a group approach and adaptive research development. Results of the comparative studies showed that 2 to 6 years after training the (linear) transfer of technology using demonstration ponds was less successful in terms of adoption and economic results compared to a group approach and adaptive research development. However, the

analysis did not detail the different types of exchange and knowledge transfer within the different approaches. From the selected literature, the discussion on participatory approaches adopted in aquatic agricultural system productivity interventions was limited. Few examples described the approaches sufficiently, and in most cases, elements of participatory approaches were used to contextualize the research or selected target groups. When participation of farmers was included as experts and/or experimenters, the outcomes of the participatory approaches seemed to be on social capital, ownership of the intervention and contribution to social learning of the target groups. However, the level of detail with regard to participation and methods to measure the influence of such approaches on the outcomes of the intervention remained limited. Case studies Five case studies of innovations in Bangladesh and Cambodia involved increasing productivity of one or several AAS

Increasing productivity and improving livelihoods in aquatic agricultural systems

components. These case studies used participatory approaches to different degrees; from limited participation of farmers receiving pre-defined training on aquaculture innovation to PAR in horticulture and homestead pond innovation in one specific project (Table 6). The type of community engagement, tool used, frequency and quality of interactions between researchers and communities also varied among our case studies. An example of an approach with limited participatory elements was the Aquaculture for Income and Nutrition (AIN) project. This project was classified as Bpassive information sharing^ within the different grades of participation defined by Lambrou (2001). In this project the target group were provided with assets and there were informal modes of information sharing between farmers and researchers. Both researchers and farmers contributed to data collection in order to monitor the innovation’s implementation, however, the way in which research was undertaken in this case was distinct from a PAR approach because there had been no farmer participation in the choice, design and distribution of the intervention. Farmers’ involvement was limited to data collection and their participation in data analysis and interpretation results was not included in the innovation process. A model more closely aligned to PAR was when interactions between researchers and communities was ‘collaborative’ and farmers participated in recording data, discussing the results and collectively determining actions to optimize or adapt interventions, at least to a certain level. Cereal Systems Initiative for South-Asia project (CSISA) and Aquaculture and Nutrition Extension Project (ANEP) project have elements of ‘collaborative’ and ‘evaluation’ grades of participation (Table 6). Those projects used participatory tools to identify populations’ needs. These tools included Rapid Rural Appraisal in the CSISA project and Participatory Market Chain Analysis in the ANEP project. The tools focused on specific production systems (i.e. in these cases aquaculture) and were not applied to understanding household’s food production systems more broadly. These tools facilitated the engagement of communities during the early stages of project implementation but for a limited time period (i.e. usually shorter than one week). Participatory planning with the community was limited to the crop cycle in the case of CSISA and no long-term vision of the project goals and outcomes was developed with the community. The type of household targeted by the intervention was determined by criteria that were set by the research project. In the case of CSISA, criteria meant there was a focus on those with land assets and a particular focus on women. In addition to collaborative and quantitative data collection efforts, anecdotal evidence was collected from farmers. These reflections about the CSISA project, demonstrated that the explicit efforts to involve women in these collaborative research efforts had led to further outcomes at the community level. In particular, farmers expressed

the belief that women’s involvement had shifted opinions within the community about women’s capacity to actively earn income and to apply innovative agricultural and aquaculture techniques (CSISA 2014). The next level or grade of participation that these projects met was ‘collaborative’. The Rice Field Fisheries Enhancement Project (RFFEP) falls under Bcollaborative’ grade of participation because more elements of the participatory approach were used during engagement, in particular the collective (i.e. researcher and community) definition of problems and design of interventions (although this didn’t occur immediately at project inception in this case). In this project, and at this level of participation, community members were considered experimenters and co-researchers, rather than recipients of a technology and subjects of research. After one year of project implementation, the RFFEP project team deemed it necessary to develop a common vision of the project objectives amongst the different stakeholders. The vision and action plan, was developed with 40 Community Fish Refuges (CFRs) operating communities in Cambodia. The action plan informed the design of new activities such as vegetable cultivation around the CFRs and the collection of additional funds within the community to improve water management infrastructure. These newly developed activities responded to local needs and were identified, planned and implemented by communities using participatory tools (Miratori 2015). The collective identification of intervention priorities, and the subsequent six-monthly collective and critical reflections on progress facilitated common learning and fine tuning of the intervention in order to respond to emerging challenges. This project led to productivity increases of 45% in the CFRs, and a higher estimated fish biomass in the ponds. Two years after the project had started fish catch among the 40 CFRs increased on average by 7%, from 7.1 kg to 7.6 kg of fish caught per household per week (RRFEP 2015). These benefits were experienced by all wealth groups within the community, including those that had been categorized as being poor.It was evident that these increases were contributing mainly to higher household income as fish consumption per household was stable at around 3 kg per week and there was no clear evidence of increased consumption of micro-nutrient rich fish (RRFEP 2015). As part of collaborative design and delivery of some of the AAS CRP interventions, a series of visioning tools were employed with target groups in order to develop consensus regarding their goals, and to determine next steps to achieve these goals. Subsequently, AAS CRP project teams facilitated regular iterative cycles of critical reflection and action (Aktar and Faruque 2015). In this case, the grade of participation corresponded to ‘partnership’ (Table 6). In these interventions there was an in-depth analysis of the context, and the identification and prioritization of the needs of different target groups. This included using gender disaggregated visioning and action

Cambodia

Farmers are source of information and experimenters Researchers are experts and capacity builders

Identification of priority using RRA tools (FGDs) Data collection by farmers but analysis by researchers Reflection with farmers in selected cases

Collaborative planning

Partnership

Evaluation / Collaborative planning

Passive information sharing

Evaluation / Collaborative planning

Grade of participation*

*Seven Bgrades^ of participation: (1) positivist theoretical research (the least inclusive type of approaches), (2) passive information sharing (farmers are informed of the processes and outcomes of the research), (3) consultative stage (farmers are consulted and their needs may be included in the research design), (4) on-farm testing (researchers continue to dominate the research process, but farmers’ expertise is recognized), (5) evaluation (farmers are involved in assessing the process and results of the research), (6) collaborative planning (scientists join hands with farmers in defining problems and in designing the research process), and (7) partnership (scientists and farmers engage in a long-term mutual learning and research process). (Adapted from Lambrou 2001)

Fisheries in RFFEP project Community Fish Refuge

Men and women are targeted. No No specific mechanism to involve women and neither poor nor marginal farmers in the project: only requirement is to own a pond. Land size requirement (0.2–0.6 ha) No excludes poorest. Specific requirement for including women but within HH, − not directly

Long term visioning Women’s opinions taken into Constant feedback and (5 years) and account from identification reflection led by yearly action plan of priorities to implementation. farmers Poor and marginalized are identified and included. In-depth analysis of context and diversity of social groups and norms. Selection of participant-based on Reflecting on the results Visioning for 4 years voluntary participation. Non every 6 months 6- months action exclusion of the poorest plan and landless

Bangladesh Farmers are adopters Researchers are experts

Engagement of the community throughout the entire learning cycle

Aquaculture in AIN project Homestead ponds & Carp polyculture with SIS

Short term (crop cycle) business plan among market actors

Type, frequency, and intensity of interaction

Selection of participants based on Limited to the production land size and gender criteria cycle and technical (homestead ponds include only aspects and within a women target groups in gher selected number of systems include both men farmers (within and women) Adaptive trials component)

Inclusiveness

Identification of priorities using a Participatory Market Chain Analysis Data collection by farmers but analysis by researchers and limited reflection

Use of Participatory Planning Tool

Aquaculture and Integrated Bangladesh Farmers are source of systems in ANEP project information and Carp polyculture with SIS; experimenters Stocking large fingerlings Researchers are experts and feeding technique; Improved Rice-fish culture; vegetable production Aquaculture and Horticulture Bangladesh Farmers are partners in ASS CRP Researchers are capacity Homestead ponds. builders and Carp polyculture with SIS facilitators of learning and vegetable seed selection

Involvement of community in learning cycle

Identification of priority No using Rapid Rural Appraisal (RRA) tools (FGDs) Data collection by farmers but analysis by researchers. Reflection with farmers in selected cases No involvement of farmers No throughout the learning cycle

Researcher and Community Member involvement

Aquaculture and Integrated Bangladesh Farmers are source systems in CSISA projects of information and Homestead ponds experimenters Gher systems Researchers are experts and capacity builders

Country

Indicators of participatory approaches and criteria for target groups for different technological interventions

Type of intervention

Table 6

Joffre O.M. et al.

Increasing productivity and improving livelihoods in aquatic agricultural systems

planning to allow interventions to be more tailored to the needs of specific target groups; for example, intervention of homestead pond aquaculture that targeted women in poor households (Douthwaite et al. 2015). The intervention’s aim was to improve the productivity of land that women control - which is generally near their homes. According to ‘outcome evidencing’ conducted by the CGIAR, the uptake of this homestead innovation by 80% of the households within the project area was possible through a network of on-farm trials and demonstration plots, of which several were led by women. In addition, these interventions were more holistic in the sense that they were more sensitive to all components of the farming system. For example, in Southwestern Bangladesh, women prioritized the development of a method to select vegetable varieties for homestead gardening or design specific technologies for homestead aquaculture that did not interfere with other pond uses, such as bathing and washing. In addition to the collective cycles of critical reflection, there was a continuous reflexive interaction for long-term mutual learning between researcher and a selected panel of farmers. Together with researchers, these farmers would record and discuss experiences and outcomes with innovations in order to fine-tune them (such as in the case of selection of okra varieties) (Cole et al. 2014). The above analysis of case studies showed early evidence of interventions that both increased productivity and delivered desirable development outcomes. Moreover, analysis of the case studies confirmed the results of in-depth studies of innovations such as CSISA, AAS CRP or RFFEP projects, showing that involving communities in the research and reflection process from conception of ideas to innovation design and implementation, can result in greater impact on development outcomes (Thompson et al. 2006; Murshed-E-Jahan and Pemsl 2011; Young et al. 2014). It suggests that while designing, testing and deploying productivity improving interventions, participatory approaches can help to contextualize the intervention and promote the ability to adapt interventions to respond to the differing needs of people within a target population.

Discussion How and to what extent have interventions improved productivity in AAS? Global food demands will more than double by 2050 (Tilman et al. 2011). Increasing productivity of current food production systems is essential to meet this demand, and there has been a long history of fisheries, livestock, field crop and horticulture interventions to this end. However, investments to date have failed to result in sustainable increases in productivity, or in realizing development gains for those most in need

(Dethier and Effenberger 2012; Godfray et al. 2010). It is critical that as we address the future need for food, we do so with a clear view of where our efforts have succeeded, and which characteristics of the interventions have helped lead to that success. We identified 20 of 31 interventions that increased the productivity of aquatic agricultural systems, with yield increases of between 10 and 61%. Yet, we found that in all but one case (Hossain et al. 2006), measurement of productivity ceased when the project finished. We found that accounts of field crops, integrated systems and aquaculture and fisheries systems interventions were relatively common (n = 29), compared to livestock interventions (n = 2, in Cambodia only). Further literature on livestock interventions may have been inadvertently excluded due to search terms. Our selection excluded initiatives that may have been abandoned in earlier stages due to poor results. Therefore, our cases represent more successful interventions, but we suggest that there may also be much to learn from an analysis of less successful initiatives. We recognize that the articles captured in our review represent the common bias of positive reporting. These do, nonetheless, provide a sound basis for us to unpack these positive outcomes, but a deeper exploration of no change or negative consequences would require an ex-ante impact assessment that, while beyond the scope of this review, would be a worthwhile next step. In this review, change in productivity is reported as an increase of production or yield. Little is known about productivity regarding nutrient use efficiency in aquatic agricultural systems. Broad scale studies estimate that the yield gap could be significantly reduced with slight changes of nutrient use in agricultural systems (Tilman et al. 2011; Tittonell and Giller 2013). In a recent study, Mueller et al. (2013) showed that production of 17 major crops can be increased by 45–70% with improved water and nutrient management. Input use and efficiency of application is critical in addressing the food security challenge but this is not well addressed by research regarding productivity interventions in aquatic agricultural systems. In research to improve productivity of aquaticagricultural production systems, interactions and feedback of the intervention at the landscape level are rarely addressed. Measuring the productivity of a targeted component of a food production system is sometimes not sufficient or adequate to estimate the outcome on food security. Poor and small-scale farmers often diversified their food security portfolio, using a combination of natural resources, collection and production systems. Enhancing the productivity of one component of the food production system, for example a rice field, can alter the productivity of natural systems or other production systems. For example, introduction of HYV rice in floodplains and its impact on fisheries is not mentioned nor discussed in those interventions, while trade-

Joffre O.M. et al.

offs exist (Shankar et al. 2004; Shankar et al. 2005) and can affect the population dependent on those natural systems. Often the contribution of those elements on productivity of the system and food security are downplayed or not measured while they can be significant in the livelihood portfolio and for food security of the poorest (Gregory et al. 1996). This dependence on natural systems and diversified food security portfolios is especially true in aquatic agricultural systems. Food security includes not only the provision of food but also regular access to food (World Food Summit 1996) and depends on patterns of distribution within and between populations. To assess if a productivity-improving intervention will contribute to food security for the poor, estimation of productivity gains is not sufficient, rather monitoring the changes of the nutritional status, food intake at an individual level, seasonality and changes in income and income use at the household level are necessary. To what extent do productivity interventions in AAS increase food security and income of the poor? Intensification of, and investment in, food production systems do not necessarily result in improved livelihoods and the well-being of poor populations (Fan and Hazell 2001; Pingali 2012). Measuring and reporting on improvements in income is relatively simple and in the present study 16 of the 31 cases reviewed demonstrated an increase in income per household. Although it was not clear in all cases whether the benefits were accrued by the poorest and most marginalized households. The remaining 15 studies did not monitor such outcomes, so any changes in income were not detected. Where changes in income were detected there may also have been corresponding increase in costs. Further, our research did not provide a deeper understanding of who was controlling and benefiting from extra income within households. This is particularly evident when indirect indicators of income, such as income per surface area, are used instead of income per household per season. For example in the gher system there was a shift away from the staple crop, to the commercial crops, shrimp and fish. This led to a shift in household economics i.e., from having high levels of self-sufficiency to becoming more integrated in the cash economy. However, for the poor subsistence fishers who once caught fish in rice fields, access to essential rice field fisheries would cease. These complex interactions and their poverty alleviation and food security implications for the whole community are rarely adequately explored and require in-depth examination. In comparison to productivity, and to a lesser extent income, our review found there to be relatively little

reported on changes of food security and nutrition. Out of 31 selected studies, only one quantified the nutritional benefit of the intervention (Roos et al. 2007), while six demonstrated an increase in household food provision. A further nine presented qualitative evidence of an increase in food provision but often the assumption was made that increased productivity correlates with increases in household food intake (Oakley and Momsen 2005). Béné et al. (2016) reviewed literature concerning the contribution of fisheries and aquaculture to food security and nutrition. The review clearly identified a consistent body of literature showing the importance of finfish for nutrition and its potential to contribute to micro-nutrient deficiency, but only a few examples were based on productivity intervention (see Kawarazuka and Béné 2010 for a review). Thus the knowledge gap still persists as to how fish can contribute to the diet of the poor (Béné et al. 2016). The scientific community is oriented towards a debate along biological sustainability and economic efficiency of fisheries and aquaculture and less on their contribution to food security and nutrition (Béné et al. 2015). Masset et al. (2012) reviewed 7000 studies on agricultural interventions but could find only 23 demonstrating links between agricultural interventions and changes in nutritional status. The complexity of, and sometimes inadequately defined, methodologies to measure food intake and nutritional impact limit the evidence demonstrating links between agricultural interventions and improved nutrition. Studies including disaggregated consumption data, indicators of dietary diversity and child height and weight are needed to better understand links between agricultural interventions and nutrition (Masset et al. 2012). Demonstrating that a new intervention can increase the productivity of a given food system does not necessarily result in the population taking up the intervention. The review conducted by Loevinsohn and Sumberg (2013) looking for evidence of adoption of technologies that increased agricultural productivity in low and lower medium income countries found that assessment of productivity change, as a result of technology adoption, was imperfect in most of the research screened because of lack of robust methodology, poor definition of the term ‘adoption’, lack of clear definition of productivity and of adequate measures of change in productivity associated with the innovation. The authors concluded that these breaches undermined the ability to learn and elaborate evidence-based guidance on the conditions required for productivity gains from most of the screened reports and articles. Our review revealed that neither the characteristics nor the capacity of the target population to acquire and use the intervention is often investigated. Only in a few cases, indications of specific characteristics constraining adoption by the poorest are reported, such as in-pond aquaculture, gher systems and interventions that require specific

Increasing productivity and improving livelihoods in aquatic agricultural systems

skills and knowledge such as in rice-fish systems (Azim et al. 2004; Ahmed and Garnett 2011). Households, with sufficient resources and skills to make a technological shift, capture the new technologies (Haque et al. 2010; Ahmed and Garnett 2011), hence preventing the poorest and marginalized from benefiting. Access to inputs by marginalized groups, unequal access to knowledge and development project targeting groups that are most likely to succeed, partially explain the exclusion of the poorest (Pingali 2012). Reducing poverty and food insecurity for the poorest requires a better distribution of the benefits of improving productivity interventions. This can only be achieved by understanding the context in which the poorest households operate. Both lack of evidence linking increased productivity to household food security and limited understanding of interventions’ sustainability and equity call for the design of an approach that integrates an improved monitoring and evaluation framework. This framework should systematically include nutrition indicators, food security and income changes whenever a new technology, management practice or learning are introduced. The approach should be carefully tailored to address sustainability and equity in the intervention design and deployment and it should be clear from the outset who the target audience is for the particular intervention. What benefits can participatory approaches bring? Analysis of specific productivity-improving interventions reveals a lack of contextualization to adequately respond to local needs and priorities (Anthofer 2004; Azim et al. 2004; Ahmed and Garnett 2011). Use of participatory approaches is often limited to the early stages of the project, during engagement with the community and study of the local context, using tools such as PRA or RRA to assess the nuances and constraints of specific production systems. Some studies show evidence of engaging communities in the research and design processes (Kuntashula et al. 2006; Thompson et al. 2003; Mustafa and Brooks 2009) while others engage communities for an iterative reflection cycle during the evaluation process (Young et al. 2014; Murshed-E-Jahan and Pemsl 2011; Thompson et al. 2006). However most of those studies do not assess longer term improvements. Only Thompson et al. (2006) demonstrated a sustainable increase in productivity and income five years after the end of the project, suggesting that the use of participatory approaches in this intervention resulted in more sustained adoption and longer lasting yield improvements and development outcomes. Collaborative and partnership research approaches not only help farmers achieve higher productivity (Thompson et al. 2006; Murshed-E-Jahan and Pemsl 2011; Young et al. 2014), but also improve their confidence and enable

new skills (Cole et al. 2014; Douthwaite et al. 2015; CSISA 2014). The research community and development partners learn from the community, jointly identifying issues, initiating action to address those issues, monitoring the results of actions and the processes used and then reflecting upon the results to determine future action (Kuntashula et al. 2006; Murshed-E-Jahan and Pemsl 2011, Miratori 2015, Cole et al. 2014; Douthwaite et al. 2015). However, implementing a participatory approach is resource-demanding, with high and frequent levels of engagement from both the target population and the researcher. It demands that: the researcher engages in a high level of community immersion with frequent visits; requires a longer process than a rapid assessment; does not necessarily correspond to project resources or time frames; and requires specific skills for researchers and field officers (Rhoades and Nazarea 2006). Government involvement is essential so that monitoring can continue after the project ends. Delivering quality data for analysis requires frequent researchers’ supervision and cannot be scaled out without significant human resource investment. With limited resources, it is important for researchers to decide which objectives and phases within the project could be used to improve research outputs for the benefit of the targeted population (Neef and Neubert 2011). Evidence from this review, the reviews of others (Kuntashula et al. 2006; Thompson et al. 2003; Mustafa and Brooks 2009; Young et al. 2014; Murshed-E-Jahan and Pemsl 2011; Thompson et al. 2006) and case studies (RFFEP, AAS CRP) suggests that involving target populations in the design of interventions can yield more desirable development outcomes by responding to context specificities and local needs. It is in this regard that agricultural research for increasing productivity should aim and should also include social and governance elements (Ratner 2012). A deeper understanding and consideration of the informal set of norms, traditions, social networks and power relationships that characterize a target population is required, and in addition their access and uses of resources and how they make decisions to truly innovate and address productivity challenges. This also means moving beyond simply including men and women as stakeholder groups, but understanding gender norms and power relations within the particular social and cultural context in the AAS program sites and improving social relations (Dugan et al. 2013). Within a sound monitoring and evaluation framework, participatory approaches have the potential to transform food production systems and inform research in development systems in other regions.

Joffre O.M. et al.

Conclusion It is widely acknowledged that productivity interventions will be required to meet the growing global food demand. This review has started to unpack the elements that make up interventions that succeed in increasing productivity whilst enhancing sustainability and driving development outcomes, with an emphasis on aquatic agricultural production systems. Yet, given the level of past and projected investments in agricultural interventions, robust supporting evidence of development outcomes is surprisingly scant. Appropriate methods to adequately measure not only productivity, but also development outcomes are critical for demonstrating impact to community members and researchers and to help scale and disseminate technologies. There is emerging evidence that participatory approaches achieve greater, more sustainable and more equitable development outcomes. Further evidence building is required, however, to gain a better understanding of how best to integrate conventional research and participatory approaches to achieve not only improved productivity but better development outcomes and better targeting of the poor and marginalized. Acknowledgments The authors would like to thank research and field officers from CGIAR Research Program on Aquatic Agricultural Systems (AAS) in the Khulna Hub office (Bangladesh) and Bangladesh WorldFish Office, Zambia WorldFish Office and Mekong WorldFish Regional Office. The contributions of Kim Miratori, Alan Brooks, Ahmed Orko Nur, Golam Faruque, Karim Mandjurul, Sarwer Rayhan Hayat, Kevin Kamp, Mohammad Mokarrom Hossain, Murshed-EJahan, Khondker, Andrew Ward and Maravanyika Tendayi are particularly appreciated. The authors wish to thank Simon Attwood for his comments and feedback during the early parts of this research. The preparation of this paper was supported by funding from the CGIAR Research Programs on Aquatic Agricultural Systems (AAS). Compliance with ethical standards Conflicts of interest The authors declare no conflict of interest.

References Apgar, M., Douthwaite, B. (2013). Participatory Action Research in the CGIAR Research Program on Aquatic Agricultural Systems. CGIAR Research Program on Aquatic Agricultural Systems. Penang, Malaysia. Program Brief: AAS-2013-27. Ahmed, N., Allison, E. H., & Muir, J. F. (2010). Rice fields to prawn farms: a blue revolution in Southwest Bangladesh? Aquaculture International, 18, 555–574. Ahmed, N., & Garnett, S. T. (2011). Integrated rice-fish farming in Bangladesh: meeting the challenges of food security. Food Security, (3), 81–92. Aktar, S., & Faruque, G. (2015). Engaging community in RinD. Visioning, Action Planning & Issue Prioritization. Southern Bangladesh Polder Zone. World Fish Internal. Report. 19 pp.

Alexandratos, N., & Bruinsma, J. (2012). World agriculture towards 2030/2050: the 2012 revision. ESA Working paper No. 12–03. Rome, FAO. Alam, M. J., Islam, M. L., Saha, S. B., Tuong, T. P., & Joffre, O. (2007). Improving the productivity of the rice-shrimp system in the southwest coastal region of Bangladesh. In C. T. Hoanh, B. W. Szuster, K. Suan-Pheng, A. M. Ismail, & A. D. Noble (Eds.), Tropical deltas and coastal zones: food production, communities and environment at the land-water Interface. Oxfordshire: CAB International. Amilhat, E., Lorenzen, K., Morales, E. J., Yakupitiyage, A., & Little, D. C. (2009). Fisheries production in southeast Asian farmer managed aquatic systems ( FMAS ) II. Diversity of aquatic resources and management impacts on catch rates. Aquaculture, 298, 57–63. Anthofer, J. (2004). The potential of the system of rice intensification (SRI) for poverty reduction in Cambodia. Conference on International Agricultural Research for Development, October 5– 6, 2004. Berlin: Deutscher Tropentag. Arksey, H., & O’Malley, L. (2005). Scoping studies: towards a methodological framework. International Journal of Social Research Methodology, 8(1), 19–32. Azim, M. E., Rahaman, M. M., Wahab, M. A., Asaeda, T., Little, D. C., & Verdegem, M. C. J. (2004). Periphyton-based pond polyculture system: a bioeconomic comparison of on-farm and on-station trials. Aquaculture, 242, 381–396. Barman, B. K., & Little, D. C. (2011). Use of hapas to produce Nile tilapia (Oreochromis niloticus L.) seed in household foodfish ponds: a participatory trial with small-scale farming households in Northwest Bangladesh. Aquaculture, 317, 214–222. Baran, E., Zalinge, N. V., & Ngor, P. (2001). Floods, floodplains and fish production in the Mekong Basin: present and past trends. In A. Ali et al. (Eds.), Proceedings of the Second Asian Wetlands Symposium (pp. 920–932), 27–30 August 2001 (p. 1116pp). Penang, Malaysia: Penerbit Universiti Sains Malaysia, Pulau Pinang, Malaysia. BBS (2006). Report on labor force survey, 2005. Bangladesh Bureau of Statistics. Dhaka: Ministry of Planning, government of the People’s Republic of Bangladesh. Belton, B., Karim, M., Thilsted, S., Murshed-E-Jahan, K., Collins, W., & Phillips, M. (2011). Review of aquaculture and fish consumption in Bangladesh. Studies and Reviews, 2011, 53. Béné, C., Barange, M., Subasinghe, R., Pinstrup-Andersen, P., Merino, G., Hemre, G.-I., & Williams, M. (2015). Feeding 9 billion by 2050 – putting fish back on the menu. Food Secur., 7, 261–274. Béné, C., Arthur, R., Norbury, H., Allison, E. H., Beveridge, M., Bush, S., Campling, L., Leschen, W., Little, D., Squires, D., Thilsted, S. H., Troell, M., & Williams, M. (2016). Contribution of fisheries and aquaculture to food security and poverty reduction: assessing the current evidence. World Development, 79, 177–196. Biggs, S. (1989). Resource-poor farmer participation in research: a synthesis of experiences from nine national agricultural research systems. OFCOR comparative study paper. The Hague: International Service for National Agricultural Research (ISNAR). Borin, K., & Frankow-Lindberg, B. E. (2005). Effects of legumescassava intercropping on cassava forage and biomass production. Journal of Sustainable Agriculture, 27, 139–151. Buhler, W., Morse, S., Arthur, E., Bolton, S., & Mann, J. (2002). Science, agriculture, and research: a compromised participation? London: Earthscan. Cagauan, A. (1995). In P. Pingali & P. Roger (Eds.), Impact of pesticides on farmer health and the Rice environment (pp. 203–248). Boston: Kluwer. Castine, S.A., Sellamuttu, S.S., Cohen, P., Chandrabalan, D., & Phillips, M. (2013). Increasing productivity and improving livelihoods in aquatic agricultural systems: a review of interventions. CGIAR Research Program on Aquatic Agricultural Systems. Penang, Malaysia. Working Paper: AAS-2013-30.

Increasing productivity and improving livelihoods in aquatic agricultural systems CEDAC. (2008). Evaluation study: Adoption and non-adoption of system of rice intensification (SRI) in Cambodia. Cambodia Center for Study and Development in Agriculture. Chambers, R. (2007). From PRA to PLA and pluralism: practice and theory. IDS working paper 286. CGIAR. (2012). CGIAR research program on aquatic agricultural systems: Program proposal. Penang, Malaysia. The World Fish Center. AAS-2012-07. CGIAR. (2015). CGIAR Strategy and Results. Framework 2016–2030. 52 pp. Cole, S.M., Kantor P., Sarapura S., & Rajaratnam S. (2014) Gendertransformative approaches to address inequalities in food, nutrition and economic outcomes in aquatic agricultural systems. Penang, Malaysia: CGIAR Research Program on Aquatic Agricultural Systems. Working Paper: AAS-2014-42. CSISA. (2014). Life–Changing Stories of Successful Women Farmers. Cereal Systems Initiatives for South Asia in Bangladesh (CSISA-BD). 15 pp Datta, D., Chattopadhyay, R. N., & Guha, P. (2012). Community-based mangrove management: a review on status and sustainability. Journal of Environmental Management, 107, 84–95. Dethier, J.-J., & Effenberger, A. (2012). Agriculture and development: a brief review of the literature. Economic Systems, 36, 175–205. Dobermann, A. (2004). A critical assessment of the system of rice intensification (SRI). Agricultural Systems, 79, 261–281. Doss, C. (1999). Twenty-Five Years of Research on Women Farmers in Africa: Lessons and Implications for Agricultural Research Institutions (International Maize and Wheat Improvement Center, D.F., Mexico). Douthwaite, B., Apgar, J. M., Schwarz, A., McDougall, C., Attwood, S., Senaratna Sellamuttu, S., & Clayton, T. (Eds.) (2015). Research in development: Learning from the CGIAR Research Program on Aquatic Agricultural Systems. Penang, Malaysia: CGIAR Research Program on Aquatic Agricultural Systems. Working Paper: AAS-2015-16. Dugan, P., Apgar, M., & Douthwaite, B. (2013). Research in Development: the approach of AAS. AAS Working Paper. Penang: World Fish. Edwards, P. (1993). Environmental issues in integrated agricultureaquaculture and wastewater-fed fish culture systems. In R. S .V. Pullin, H. Rosenthal, & J. L. MacClean (Eds.), Environment and aquaculture in developing countries. ICLARM Conf. Proc. 31, pp. 139–170. ESRC (2003). Fit for purpose? Assessing research quality for evidence based policy and practice. UK Centre for Evidence Based Policy and Practice, Working Paper 11. Fan, S., & Hazell, P. (2001). Returns to public investments in the lessfavored areas of India and China. American Journal of Agricultural Economics, 83, 1217–1222. Foley, J. A., DeFries, R., Asner, G. P., Barford, C., Bonan, G., Carpenter, S. R., Chapin, S. F., Coe, M. T., Daily, G. C., Gibbs, H. K., Helkowski, J. H., Holloway, T., Howard, E. A., Kucharik, C. J., Monfreda, C., Patz, J. A., Prentice, I. C., Ramankutty, N., & Snyder, P. K. (2005). Global consequences of land use. Science, 309, 570–573. Godfray, H. C. J., Beddington, J. R., Crute, I. R., Haddad, L., Lawrence, D., Muir, J. F., Pretty, J., Robinson, S., Thomas, S. M., & Toulmin, C. (2010). Food security: the challenge of feeding 9 billion people. Science, 327, 812–818. Gregory, R., Guttman, H., & Kekputherith, T. (1996). Poor in all but fish. AIT Aquaculture working paper No. 4, Cambodia. Bangkok: Asian Institute of Technology. Gupta, M.V., Sollows, J.D., Mazid, M.A., Rahman, A., Hussain, M.G., Dey, M.M. (1998). Integrating aquaculture with rice farming in Bangladesh: feasibility and economic viability, its adoption and impact, vol. 55. International Center for Living Aquatic Resources Management (ICLARM) Tech. Rep, 90 pp. Halwart, M., & Gupta M.V. (2004). Culture of fish in rice fields. FAO and World Fish Center. 83 pp

Haque, M. M., Little, D. C., Barman, B. K., & Wahab, M. A. (2010). The adoption process of ricefield-based fish seed production in Northwest Bangladesh: an understanding through quantitative and qualitative investigation. The Journal of Agricultural Education and Extension, 16, 161–177. Hazell, P. (2003). Agricultural Research and Poverty Reduction: Some Issues and Evidence. In S. Mathur & D. Pachico (Eds.), Centro Internacional de Agricultura Tropical (CIAT) (pp. 43–58). Colombia: Cali. Hoffmann, V., Probst, K., & Christinck, A. (2007). Farmers as researchers: how can collaborative advantages be created in participatory research and technology development? Agriculture and Human Values, 24, 355–368. Hortle, K.G. (2007). Consumption and yield of fish and other aquatic animals from the lower Mekong basin. MRC Technical Paper No. 16. Vientiane, Lao PDR: Mekong River Commission. Hossain, M., Bose, M., & Mustafi, B. A. A. (2006). Adoption and productivity impact of modern rice varieties in Bangladesh. The Developing Economies, XLIV-2, 149–166. Hossain, M., Lewis, D., Bose, M. L., & Chowdhury, A. (2007). Rice research, technological progress, and poverty: the Bangladesh case. In M. R. M. D. Adato (Ed.), Agricultural research, livelihoods, and poverty: studies of economic and social impacts in six countries (pp. 56–102). Baltimore and Washington: Johns Hopkins University Press and International Food Policy Research Institute. Howlader, M.Z.H., & Biswas, S.K. (2009). Screening for nutritionally rich and low glycemic index Bangladeshi rice varieties. National Food Policy Capacity Strengthening Programme. IFAD. (2010). Rural poverty report. Rome, Italy: IFAD. Joffre, O., & Sheriff, N. (2011). Conditions for collective action: Understanding factors supporting and constraining communitybased fish culture in Bangladesh, Cambodia and Vietnam. World Fish Center Studies and Reviews 2011–21 (p. 46pp). Penang, Malaysia: The World Fish Center. Joffre, O., Kosal, M., Kura, Y., Sereywath, P., & Thuok, N. (2012). Community fish refuges in Cambodia: lessons learned. Phnom Penh, Cambodia: The World Fish Center. Johnson, N., Lilja, N., Ashby, J., & Garcia, J. A. (2004). The practice of participatory research and gender analysis in natural resource management. Natural Resources Forum, 28, 189–200. Kadir, A., Kundu, R. S., Milstein, A., & Wahab, M. A. (2006). Effects of silver carp and small indigenous species on pond ecology and carp polycultures in Bangladesh. Aquaculture, 261, 1065–1076. Karim, M., Little, D. C., Kabir, M. S., Verdegem, M. J. C., Telfer, T. & Wahab, M. A. (2011). Enhancing benefits from polycultures including tilapia (Oreochromis niloticus) within integrated pond-dike systems: A participatory trial with households of varying socioeconomic level in rural and pen-urban areas of Bangladesh. Aquaculture, 314, 225–235. Kawarazuka, N., & Béné, C. (2010). Linking small-scale fisheries and aquaculture to household nutritional security: a review of the literature. Food. Security, 2(4), 343–357. Kuntashula, E., Sileshi, G., Mafongoya, P. L., & Banda, J. (2006). Farmer participatory evaluation of the potential for organic vegetable production in the wetlands of Zambia. Agriculture, 35, 299–305. Lambrou, Y. (2001). A typology: Participatory research and gender analysis in natural resource management research. Working document No. 15. Cali, Colombia: CGIAR Participatory Research and Gender Analysis Program, CIAT (Centro Internacional de Agricultura Tropical). Levac, D., Colquhoun, H., & O’Brien, K. (2010). Scoping studies: advancing the methodology. Implementation Science, 5(1), 1–9. Lilja, N., J.A. Ashby, & L. Sperling. (2001). Assessing the impact of participatory research and gender analysis. Cali, Colombia: Participatory Research and Gender Analysis Program, CIAT (Centro Internacional de Agricultura Tropical).

Joffre O.M. et al. Loevinsohn, M., & Sumberg, J. (2013). Under what circumstances and conditions does adoption of technology result in increased agricultural productivity? A Systematic Review Prepared for the Department for International Development. Ly, P., Jensen, L. S., Bruun, T. B., Rutz, D., & de Neergaard, A. (2012). The system of rice intensification: adapted practices, reported outcomes and their relevance in Cambodia. Agricultural Systems, 113, 16–27. Masset, E., Haddad, L., Cornelius, A., & Isaza-Castro, J. (2012). Effectiveness of agricultural interventions that aim to improve nutritional status of children: systematic review. British Medical Journal, 344, d8222. Maxwell, T. W., You, S., Boratana, U., Leakhna, P., & Reid, J. (2012). The social and other impacts of a cattle/crop innovation in Cambodia. Agricultural Systems, 107, 83–91. McIntyre, B.D., Herren, H.R., Wakhungu, J., Watson, R.T. (2009). Agriculture at a Crossroads: The Global Report. International Assessment of Agricultural Knowledge, Science and Technology for Development, Washington, DC. Miratori K. 2015. Good governance of rice field fishery management. Penang, Malaysia: World Fish. Program Brief: 2015–19. Molden, D. (2007). Water for food, water for life. London: Earthscan. Mondal, M.K., Tuong, T.P., Sharifullah, A.K.M., & Sattar, M.A. (2010). Water supply and demand for dry-season rice in the coastal polders of Bangladesh. In Tropical Deltas and Coastal Zones: Food Production, Communities and Environment at the Land-Water Interface. C.T. Hoanh (Eds.). CAB International. Mueller, N. D., Gerber, J. S., Johnston, M., Ray, D. K., Ramankutty, N., & Foley, J. A. (2013). Closing yield gaps through nutrient and water management. Nature, 409. doi:10.1038/nature11420. Mukanda, N. (1998). Wetland classification for agricultural development in eastern and Southern Africa: the Zambian case, in FAO (ed) Wetland characterization and classification for sustainable development, Proceedings of a sub-regional consultation, 3–6 December 1997, Harare, Zimbabwe. Murshed-E-Jahan, K., Beveridge, M. C. M., & Brooks, A. C. (2008). Impact of long-term training and extension support on small-scale carp polyculture farms of Bangladesh. Journal of the World Aquaculture Society, 39, 441–453. Murshed-E-Jahan, K., & Pemsl, D. E. (2011). The impact of integrated aquaculture-agriculture on small-scale farm sustainability and farmers' livelihoods: experience from Bangladesh. Agricultural Systems, 104, 392–402. Mustafa, M. G., & Brooks, A. C. (2009). A comparative study of two seasonal floodplain aquaculture systems in Bangladesh. Water Policy, 11(S1), 69. Nao, T. (2009). Community fish refuge husbandry in lowland agricultural ecosystem (p. 211). Phnom Penh: Build Bright University.Ph.D. thesis Ndiyoi, M., Sampa, J.B., Thawe, P., & Wood, A. (2009). Striking Balance: maintaining seasonla dambowetlands in Malawi and Zambia. In wetland International. 2009. Planting trees to eat fish: Field experiences in wetlands and poverty reduction. Wetlands International, Wageningen, The Netherlands. Neef, A., & Neubert, D. (2011). Stakeholder participation in agricultural research projects: a conceptual framework for reflection and decision-making. Agriculture and Human Values, 28(2), 179–194. Oakley, E., & Momsen, J. H. (2005). Gender and agrobiodiverstiy: a case study from Bangladesh. The Geographical Journal, 171, 195–208. OECD.. (2010). Glossary of Key Terms in Evaluation and Results Based Management. Paris, TR. (1998). Impact of Rice Research, Eds Pingali PL, Hossain M (International Rice Research Institute), 241–262. Los Banos, Phillipines. Petticrew, M., & Roberts, H. (2006). Systematic reviews in the social sciences: A practical guide. Blackwell Publishing. Pingali, P. L., & Rosegrant, M. W. (1994). Confronting the environmental consequences of the green revolution in Asia. Washington, DC: International Food Policy Research Institute.

Pingali, P. L. (2012). Green revolution: impacts, limits, and the path ahead. Proceedings of the National Academy of Sciences of the United States of America, 109, 12302–12308. Pretty, J., Toulmin, C., & Williams, S. (2011). Sustainable intensification in African agriculture. International Journal of Agricultural Sustainability, 9, 5–24. Ratner, B.D. (2012). Collaborative Governance Assessment. CGIAR Research Program on Aquatic Agricultural Systems. Penang, Malaysia. Guidance Note AAS-2012-27. Ratner, B. D., Cohen, P., Barman, B., Mam, K., Nagoli, J., & Allison, E. H. (2013). Governance of aquatic agricultural Systems : analyzing representation, power, and accountability. Ecology and Society, 18(4). Rhoades, R. E., & Nazarea, V. (2006). Reconciling local and global agendas in sustainable development: participatory research with indigenous Andean communities. Journal of Mountain Science, 3(4), 334–346. RFFEP (2015). Rice Field Fisheries Enhancement Project (RFFEP). Catch & Consumption survey trends over 3 years. Draft report. World Fish. 33p. Roos, N., Wahab, M. A., Hossain, M. A. R., & Thilsted, S. H. (2007). Linking human nutrition and fisheries: incorporating micronutrientdense, small indigenous fish species in carp polyculture production in Bangladesh. Food and Nutrition Bulletin, 28, 280–293. Sampa, J. (2007). Sustainable dambo cultivation. Report of Striking a Balance Project. Scoones, I., & Thompson, J. (1994). Beyond farmer first: rural people’s knowledge, agricultural research, and extension practice. London: Intermediate Technology Publications. Shankar, B., Halls, A., & Barr, J. (2004). Rice versus fish revisited: on the integrated management of floodplain resources in Bangladesh. Natural Resources Forum, 28, 91–101. Shankar, B., Halls, A., & Barr, J. (2005). The effects of surface water abstraction for rice irrigation on floodplain fish production in Bangladesh. International Journal of Water, 3, 61–83. So N., and Touch, B. (2011). Fisheries resources in Cambodia: Implications for food security, human nutrition and conservation. Proceedings for the International Conference on Asian Food Security (ICAF 2011), 10–12 August 2011, Singapore. Stür, W. W., Horne, P. M., Gabunada, F. A., Phengsavanh, P., & Kerridge, P. C. (2002). Forage options for smallholder crop-animal systems in Southeast Asia: working with farmers to find solutions. Agricultural Systems, 71, 75–98. Suryavanshi, P., Singh, Y. V., Prasanna, R., Bhatia, A., & Shivay, Y. S. (2013). Pattern of methane emission and water productivity under different methods of rice crop establishment. Paddy and Water Environment, 11, 321–329. The World Bank & UKAID. (2011). What would it take for Zambia's beef and dairy industries to achieve their potential? The World Fish Center. (2011). CGIAR Research Program on Aquatic Agricultural Systems: Program brief. Penang, Malaysia. http://www. worldfishcenter.org/resource_centre/WF_2934.pd. Thilsted, S.H., & Wahab, M.A. (2014). Increased production of small fish in wetland combats micronutrient deficiencies in Bangladesh. CGIAR Research Program on Aquatic Agricultural Systems. Penang, Malaysia. Policy Brief: AA-2014-10. Thompson, P. M., Sultana, P., & Islam, N. (2003). Lessons from community based management of floodplain fisheries in Bangladesh. Journal of Environmental Management, 69(3), 307–321. Thompson, P. M., Firoz Khan, A., & Sultana, P. (2006). Comparison of aquaculture extension impacts in Bangladesh. Aquaculture Economics and Management, 10(1), 15–31. Tilman, D., Blazer, C., Hill, J., & Befort, B. L. (2011). Global food demand and the sustainable intensification of agriculture. Proceedings of the National Academy of Sciences of the United States of America, 108(50), 20260–20264. Tittonell, P., & Giller, K. E. (2013). When yield gaps are poverty traps: the paradigm of ecological intensification in African smallholder agriculture. Field Crops Research, 143, 76–90.

Increasing productivity and improving livelihoods in aquatic agricultural systems Viseth, H., Leap, H., Savry, C., Thon, H., & Doi, M. (2008). Propagation of rice-field fish resources for rural communities through establishment of dry season fish refuges in Cambodia. Freshwater aquaculture improvement and extension project (FAIEX). Phnom Penh, Cambodia: Fishery Administration. Wahab, M. A., Ahmad-Al-Nahid, S., Ahmed, N., Haque, M. M., & Karim, M. (2012). Current status and prospects of farming the giant river prawn Macrobrachium rosenbergii (de man) in Bangladesh. Aquaculture Research, 43, 970–983. Webb, P.J.R. (2009). Fiat Panis: For a World Without Hunger. Eiselen H (eds), 410–434 Hampp Media/Balance Publications, Stuttgart World Food Summit. (1996). Rome declaration on World Food Security Young, J. R., O’Reilly, R., Ashley, K., Suon, S., Leoung, I. V., Windsor, P., & Bush, R. D. (2014). Impacts on rural livelihoods in Cambodia following adoption of best practice health and husbandry interventions by smallholder cattle farmers. Transboundary and Emerging Diseases, 61(1), 11–24.

Olivier Joffre completed his PhD in coastal aquaculture planning at Wa g e n i n g e n U n i v e r s i t y (The Netherlands). His PhD research focuses on developing new tools and methods that support the development of sustainable aquaculture in coastal areas by combining economic, social and biological perspectives in order to understand farmer’s decision-making. His dissertation investigated new approaches combining participatory tools, agent based modelling and aquaculture farming systems analysis to better plan coastal aquaculture, using the Mekong Delta (Vietnam) as a case study. Olivier is now employed as Post-Doc on a Wageningen University and World Fish led project to support the Aquaculture Innovation System research in the Mekong Delta (Vietnam) and initiate an innovation platform on aquaculture.

Sarah Castine completed an industry-based PhD in aquaculture wastewater treatment and a Post-Doctoral Fellowship on international development, based in rural Bangladesh. The aim of the Post-Doc was to improve human health through optimising the species of fish cultured in small-scale pond aquaculture. Underpinning Sarah’s PhD and Post-Doctoral work was the need to produce more food while simultaneously reducing environmental impact and improving human health. Sarah is currently based in Tasmania where she is consulting on a range of aquaculture and municipal waste outfall projects.

Michael Phillips received a PhD in Aquaculture and Fish Behavior in 1982 from the University of Stirling, UK, and has been involved in research and development on various aspects of Asian aquaculture and aquatic systems since 1985, with occasional short periods of work in the Americas, Africa, and the Pacific. Before joining World Fish in 2009, Dr. Phillips was Research and Development Manager at the Intergovernmental Network of Aquaculture Centres in AsiaPacific (NACA) in Bangkok, Thailand. Since joining World Fish, he has been involved in sustainable aquaculture research in Africa, Asia and the Pacific. He recently contributed to two seminal publications: the landmark BBlue Frontiers^ study on the global ecological footprint of aquaculture, conducted in partnership with Conservation International and a publication on aquaculture in 2050, with the World Resources Institute. With World Fish colleagues and private partners, he researches on business innovations and financing for investment into SME aquaculture enterprises. He is currently the Director of the Sustainable Aquaculture Program of World Fish.

Sonali Senaratna Sellamuttu is Head of IWMI’s Southeast Asia Regional Office in Lao P D R . Sh e s e r v e d a s th e Acting Theme Leader for IWMI’s theme on G o v e r n a n c e , Ge n d e r an d Poverty Reduction (May 2014 – March 2015). Since joining IWMI in 2006 she has been responsible for leading and managing a number of large, multidisciplinary projects and has worked in Southeast Asia, South Asia and Africa. Sonali was a member of the AAS CRP Strategic Leadership Group and the IWMI AAS focal point. She was the IWMI Representative on the Ramsar Convention’s Scientific and Technical Review Panel (STRP) for 2013–2015, in addition to being the Lead for the STRP Working Group on Wetlands and Poverty Eradication (2013–2015). She was recently appointed Co-Chair for the Intergovernmental Platform on Biodiversity and Ecosystem Services (IPBES) Asia-Pacific Regional Assessment. Sonali’s main research areas include; natural resource management & governance, sustainable livelihoods and poverty reduction related issues especially in the context of agricultural and aquatic systems (including wetlands). Sonali has a multidisciplinary background with a BSc (Honors) in Biology, an MSc in ‘Ecosystems Analysis and Governance’ – both from the University of Warwick (UK) and a doctoral degree from Imperial

Joffre O.M. et al. College London (UK) that investigated the sustainability of livelihood dynamics in a rural community in Sri Lanka.

Dorothy Chandrabalan prior to joining Bioversity International, w o r k e d w i t h Wa g e n i n g e n University as a research assistant while pursuing her master’s degree in plant biotechnology for establishing protocols for the successful cryopreservation and micropropagation of recalcitrant native tropical fruit species. Dorothy is an experienced research professional with Bioversity International, an international research organization undertaking research for development in agricultural and tree biodiversity. Since 2004, Dorothy has been engaged in numerous capacities for international and regional research projects across Asia which seek to improve

livelihoods while ensuring environment and food security from the sustainable use of agriculture biodiversity. She has specialized research experience in plant biotechnology, with almost five years working in the laboratory on cryopreservation, micropropagation and molecular marker technology. Her current research interest lies at the intersection of gender and agrobiodiversity/natural resources use.