Innovative approach to build a “no regret” framework for reinforcing agricultural water resilience under climate risks and change in Burkina Faso Open Access

Despite several efforts undertaken at international, national and local levels, agricultural productivity is still relying on rainfall patterns in Burkina Faso (BF) (Hesse et al., 2013). The performance of agriculture in the country is frequently hindered by water scarcity due to climate risks, especially by recurrent droughts, flood and higher temperature (IPCC, 2014; Sarr, 2012; Barbier et al., 2009). In fact, rainfall declines and evaporation directly result in a decrease in water bodies at more than 2 m per year, whereas groundwater levels may fall down between 0.5 and 0.6 m during the dry season in the country (Ibrahim et al., 2015; Sandwidi, 2007). Demonfaucon (2011) mentioned that small reservoirs are more vulnerable to flood insofar, as many of them are destroyed whenever flood occurs (18 small reservoirs have been damaged by the 1994 and 2009 floods). The consequence is that yield decreased by 32.3 per cent in dry years and food insecurity spreads out over the whole country from year to year. Under future climate compulsion, agricultural productivity is therefore likely to fall down, amplifying food insecurity and migration in the area (IPCC, 2012; UNEP, 2012). Many studies reported that an increase in temperature, changes in rainfall regime and extreme event intensities and frequencies can cause decrease in agricultural water (AgWater) availability in Burkina Faso (Ibrahim et al., 2014; Henry et al., 2004). Subsequent consequences of that situation will include water shortage in water bodies and their overuse, changes in optimal farming lands and decrease in yields and people’s livelihood (Kemp-Benedict et al., 2011). These adverse and much debated drawbacks from climate risks and change (CR&C) suggest elaborating and implementing new frameworks for actions which will be able to reinforce Water Resources (WR) resilience. In addition, uncertainties in rainfall in Burkina Faso (Ibrahim et al., 2014) command not only enhancing the current strategies, policies and processes for preventing and managing climate effects but also guarantee sustained adaptation. At this end, “no regret” options that yield socio-economic benefits even in absence of expected climate effects (Hallegatte, 2009) are likely to help achieve this goal.

Using an innovative approach adapted to the Sahelian context, this paper aims at proposing, for the very first time for Burkina Faso, a “no regret” referential framework for actions to improve water policies which could ensure sustained AgWater resilience under CR&C.

Burkina Faso is a landlocked country in West Africa that covers an area of about 274,000 km² and whose 86 per cent of inhabitants live thanks to agriculture. The climate is characterized by a short rainy season from June to October and a long dry season from November to May (Ibrahim et al., 2014). The hydroclimatic conditions show significant spatial rainfall variability and an increase in temperature trends. The mean annual rainfall ranges from 1,100 mm in the south to 500 mm in the north. Temperature during dry seasons may be beyond 40°C. The country is entirely covered by three cross-border basins, including the Volta basin covering 63 per cent of the total area of the country. The latter includes Nakanbe sub-basin which has served as the reference field for this study. The AgWater main sources are rainwater, rivers, water bodies (ponds, lowlands, lakes, dams and reservoirs) and groundwater. Increase in temperature, changes in rainfall regime and extreme event intensities and frequencies are likely to cause much decrease in water availability (Kaboré et al., 2015). Sandwidi (2007) reported that the evaporation rate is 2 m/year in water bodies, whereas groundwater levels may fall down between 0.5 and 0.6 m during dry seasons.

The current AgWater management strategy is part of BF’s Integrated Water Resources Management (IWRM), which has been adopted as the water resources governance model in BF since 2001. Its institutional framework includes:

• Ministry of Agriculture, Water and Hydraulic work (MEAH) responsible for the inter-ministerial coordination and water resources policy making;

• National Water Committee, that is supposed being the backbone consultative framework of all stakeholders;

• the General Directorate of Water Resources (DGRE) in charge of coordinating the technical aspects pertaining to water resources management;

• the five basin water agencies (Cascades, Black Volta, Nakanbe, Gourma and Liptako) responsible for developing the water resources master plan and water management plans, respectively, at the basin and sub-basin levels; and

• Local Committees for Water (CLE) that are responsible for managing water resources at local level.

Though existence of this water policy framework, water security is still questionable, as facing climate risks and managing water resources for the benefits of the poor becomes a challenge increasingly.

The ClimProspect model is the basic methodological tool that has served to elaborate the AgWater indicative resilience framework. It has been designed by “Institut d’Application et de Vulgarisation des Sciences (IAVS)”. As an approach to measure social-ecological system resilience, the tool used allows a systemic view on social-ecological interactions, which according to Béné et al. (2014) appears necessary to understand the links between human systems, ecosystems, shocks and trends. It incorporates absorptive coping capacity, adaptive capacity and transformational resilience as well. In addition, it takes into account rainfall uncertainties in the Sahel in the formulation of resilience development options through “no regret” solutions. The ClimProspect-based method is a systematic, participatory and integrated approach. It considers a system with its main components. It includes diverse stakeholders at different scales – horizontal and vertical – in vulnerability assessments and in drawing resilience measures. It is an integrated approach as it takes into account within the same analysis both climate and disaster risks and their socioecological impacts on the target system. The main purpose of participatory assessments is to identify efficient resilience options that are feasible, practical and effective to the benefit of communities and policy makers as well. Figure 1 shows the simple design of the model adapted to the paper purpose. It is implemented through several steps grouped into three main components, which are vulnerability mapping (Steps 1-4), adaptation strategies and planning adaptation (Steps 5-6) and monitoring (Step 7).

The first step of vulnerability mapping in the ClimProspect model is to define the key units of the targeted system that are likely to be significantly affected by climate CR&C. The approach used considers a system as being an ensemble of components/units with their interacting relations, rather than a unique unit of analysis. Doing so, the analysis allows for a complexity of the system and its functioning. Thus, AgWater (the target system) has been subdivided into four units so-called “AgWater components” exposed to CR&C. The climate risks that are susceptible to significantly impact each unit of the system are droughts, flood and heat weaves. In fact, the literature on climate shows that these three risks are those which cause the highest negative effects on AgWater in Burkina Faso (Mertz et al., 2012; Barbier et al., 2009). These risks were also identified by stakeholders during interviews. Thus, AgWater vulnerability to climate will be assessed referring to these three climate shocks.

The second step consists of assessing direct and indirect impacts on each unit of the target system. Practically, it is the question of drawing an impact chain for each defined unit, as presented in Figure 2. To identify a given impact, an iterative exercise is required as many times as it is useful. First, one identifies a set of immediate drawbacks related to environmental, economic, social, technological, human, political and institutional aspects. Second, the significant impacts are selected.

The third one is related to assessing impacts of future climate on the units of the target system. To do that, the model proposes three qualitative scenarios which are based on the outputs of the different climate models (IPCC, 2007). The scenarios that allow for climate uncertainties are the following:

• Scenario S1: It forecasts a drier and warmer climate than the current climate characterized by an increasing trend of the temperature, a decreasing trend of the rainfall, a permanent drought and frequent occurrence of high floods.

• Scenario S2: Under this scenario, a significant change and warmer climate compared to the current climate is expected. This climate is fundamentally characterized by a rising of temperature and a significant increase in the frequency and intensity of climatic shocks (droughts, floods and heat waves).

• Scenario S3: It refers to a wetter and warmer climate compared to the current climate, but overall characterized by the wet periods, in fact in this scenario, rainfall returns to near-normal as it was before the droughts of 1970s that occurred in the Sahelian region periods.

The fourth step consists of identifying vulnerability factors. There are many different definitions of vulnerability (Nazari et al., 2015). ClimProspect model considers that “vulnerability” of a given system is derived from characteristics of the system itself and characteristics of the environment where the system evolves. Thus, two categories of vulnerability factors are considered by the model. The first one includes vulnerability factors based on specifications or characteristics of the system itself. The second one is related to the context where the system is evolving. These types of factors are the contextual vulnerability factors, which comprise environmental, social, human, technological, economic, institutional and political considerations. The main question here to identify these vulnerability factors is:

Adaptation and resilience of socio-ecological systems, including AgWater, have been studied by many authors (Gunderson, 2000; Walker et al., 2004; Smit and Wandel, 2006; Folke, 2006; Enfors and Gordon, 2007; Gordon et al., 2010; Miller et al., 2010). Using the ClimProspect approach, based on the vulnerability factors, resilience phase consists of three main steps as follow:

“to identify development options (the fifth step of the model) which can reduce the vulnerability of the system”. The decision support tools used in practice to formulate the resilience set of solutions are the vulnerability classes that include four classes of vulnerability. There are:

1. the vulnerability class known as V1 that is linked to early warning;

2. the vulnerability class so-called V2 that is linked to response;

3. the vulnerability class so-called V3 that is related to recovery; and

4. the vulnerability class V4 that is associated to structural vulnerability.

Practically, the ClimProspect-based approach is an integrated analysis. It takes into account the complexity of the meaning of adaptation and transformational resilience, as defined by O’Connell et al. (2015), among others. In other words, adaptation, resilience and transformation are viewed as a set of closely related concepts that go toward sustainable living environment. For instance, long-term resilience which includes sustainable development is based on transformational resilience in the sense of social system (reformed governance arrangement including improved institutions and water policy). As mentioned above, the method includes physical, institutional, technical, social and human aspects and not just in the sense of the ability of a system to absorb disturbance and reorganize so as to retain its “identity” of maintaining the same function, structure and feedbacks (O’Connell et al., 2015). The analysis also allows for adaptive capacity. Actually, “no regret” term takes all of these into account (sustainability; better the former state, adaptive and transformative capacity to climate risks). Therefore, the resilience pillars P1-P5, as shown by Figure 3, are a set of integrated solutions to overcome climate vulnerability under V1-V5, respectively.

The conceptual framework above is based on two main components. The first one refers to the current climate and includes four pillars, which are warning, response, recovery and chronic resilience. As to its second component, it integrates a set of solutions for long-term resilience.

The last two steps of the ClimProspect approach are planning development options (now, short, mid and long terms) and defining indicators for monitoring vulnerability of the system. The aim of this paper is beyond the last two steps. It is worth noticing that the approach suggests planning the resilience options before implementing them.

In this paper, based on ClimProspect approach, AgWater units which are likely to be significantly impacted by current and future climate risks are the following:

• Water supply: Here, it refers to the amount of water available for farming production. It includes rainwater, Bagre Dam and Mogtedo Reservoir and the ephemeral Nakanbe River strictosensus (Nakanbe ss). Bagre Dam, with an initial storage of 1,700mm³, is for both hydro-agricultural and hydropower functions. As for Mogtedo Reservoir, which has a storage capacity of 6.5 mm³ (Sally et al., 2011), is likely to be used for irrigation only. The two water bodies are fed by the Nakanbe River. The Nakanbe River is shared with Ghana, a neighboring country.

• Water quality: It refers to sedimentation and siltation in the water sources aforementioned. Considering that erosion brings sediments and silts to the water sources that is, apart from reducing water quantity, one of the basic causes of water deterioration over time, stakeholders compromised to consider “sedimentation and silting” as being a water quality issue. Doing that, they argue that they took the root of the problem and not the head, as usually done. Therefore, “sedimentation and silting” has been considered in this paper as being a water quality issue. It is worth noting that though water quality issue (e.g. chemical products) exists in the country, it is not much considered to be the main problem in the present.

• Water demand: It refers to the crop water needs, including major crops in growing season (rice and cowpea) and crops in dry seasons (tomatoes and onions).

• The managerial frameworks: It includes MEAH, DGRE, Nakanbe Basin Board and CLE.

Sociological and biophysical data were collected for assessing vulnerability and elaborating the indicative framework. The former data were collected through structured interviews among key stakeholders participating in IWRM in BF. These are, inter alia, policy makers, non-governmental organizations (NGOs), farmers’ organizations and researchers. A simplified random sample method was made to interview 25 key stakeholders. A pre-testing of questionnaire was carried out for the validity and reliability with regard to stakeholder’s stakes on climate issues. Based on that, the final questionnaire was put in line with stakeholder interests regarding climate change and water resources development issues. The collected information was related to the water vulnerability analysis focused on AgWater under the current and future climate variability. Scale and field of intervention for various stakeholders were taken into account for the interviews. They include the international level (development partners), the sub-regional scale (Authority of Crossborder Rivers), the national level (Ministry, Water Agencies and Directorate and environmental services) and the community level (NGOs, users’ organizations). Field observations allowed to compare their statement regarding their practices (ground truth). The former data were collected from the extension services. They concern the hydro-meteorological (temperature, rainfall, runoff, sunshine, evapotranspiration and relative humidity) and agricultural data (crop area, production and yield). These services include the Directorate of National Meteorology, DGRE and Volta Basin Authority Observatory. In-depth, literature has been helpful in identifying the best and relevant practices and lessons learned regarding water resources management, particularly AgWater management in BF or in similar contexts. In addition, both literature review and specialists helped get crop coefficients and types of soils around the dams.

To illustrate quantitatively these climate effects on AgWater, the three qualitative scenarios aforesaid have been combined with analog forecast. Climate analogs consider using as a future climate for a given area the past or the current climate data (temporal analogy) or the data from another area (spatial analogy). Information from such a benchmark can help have both the future vision and the possible responses to arouse learning in exchanging knowledge (Kopf et al., 2008. On the basis of that, we might better understand a climate for a given region in a long term. In other words, the current climate issues for the compared area will be the one for the target field. However, the socio-cultural, economic and policy consideration should be taken into account and modulated. For this purpose, the three climatic zones of BF have been considered. In this study, both spatial and temporal climate analogs are used, as shown in Table I. Referring, on the one hand, to an increase in temperature between 0.5 and 2°C, and on the other hand, to a decrease in approximately 30 per cent in rainfall (Kaboré et al., 2015), the following assumptions were formulated:

• Ouagadougou area, over 61 years of observation (1952-2012), has been assimilated to the Nakanbe Basin (Soudano-Sahelian zone).

• Dori area in the Sahel zone has been selected for the spatial scenario (spatial analogue).

• The period ranging from 1981 to 2010, at Ouagadougou, was taken as the normal climate over 30 years for the Nakanbe Basin (reference climate S0).

• For the future climate, under Scenario S1, Dori climate data over 1971-2000 were considered for the Nakanbe Basin.

• In scenario S2, Dori climate data over the period of 1961-1990 were projected for the Nakanbe Basin.

The sunshine period took into account the same period aforementioned. Wind speeds over 1981-2010 at Ouagadougou level were kept identical in the different scenarios. This exercise helps quantify the impacts of CC on rice and onion water needs under different climate scenarios.

Several tools were used to analyze the data. The content analysis method was used to analyze data from interviews. Within the CROPWAT software version 8.0, the crop water requirements under different climate scenarios were estimated. However, CROPWAT software computing crop water requirements per decade, monthly values, used in this paper, have been calculated using the Excel program. The expert knowledge method was used to collect and analyze the opinions from experts working in the AgWater fields. In addition, the check list has been applied to ensure that each option (solution) met the previous criteria to wit: be likely to be implemented in each scenario (S1, S2 and S3), be likely to provide socioeconomic benefits with farmers and be acceptable by stakeholders (farmers and policy makers).

Table II describes the direct and indirect effects of climate risks on agricultural water. The most obvious effects include a reduction or loss of water supply, and increase in sediment concentration in the water sources and inefficiency of IWRM. The analysis shows that the three climate risks considered have similar socioeconomic impacts on water supply, demand and quality. These socioeconomic impacts include water-related conflicts among users particularly in dry seasons, food shortages and insecurity and increase in rural exodus and migration toward coastal neighboring countries such as Côte d’Ivoire and Ghana.

Based on the ClimProspect scenarios, CC impacts on AgWater have been assessed. The analysis showed that under the Scenario S1, the main expected impacts are a rapid and irreversible depletion of AgWater; permanent unmet water demand for crops, intensive farmer migration toward rare water bodies and permanent conflicts for resource control; loss of agricultural production and incomes; high pressure on governance frameworks leading to their decline; and IWRM and famine followed by uncontrollable migration toward coastal countries. Under the Scenario S2, impacts will be, inter alia, unbalance between water supply and demand for agricultural production; recurrent restrictions on water supply; water-related conflict intensification; inefficiency of water managerial frameworks in facing CC impacts; food shortages; and massive rural exodus. The Scenario S3 will be characterized by normal rainfalls similar to 1950s; destroying irrigation infrastructures in growing seasons; AgWater unavailability in dry seasons due to the destruction of the irrigation dams by heavy floods; and lower yields in dry seasons. Figures 4 and 5 present the estimated water demand for rice (May-October) and onion (September-December) at the Mogtedo irrigated unit under the Scenarios S1 and S2. Under the drier Scenario S1, the water needs are 1,155.9 mm for rice and about 504 mm for onion. That corresponds to an increase rate of 38.85 and 12.36 per cent, respectively, compared to the reference period S0 (832.5 mm for onion and 448.9 mm for rice). As for Scenario S2, crop water needs increase up to 31.45 and 10.94 per cent for rice and onion, respectively, compared to the reference period S0.

The impacts reported in Table II above are recurrent impacts that clearly give evidence to better the ongoing processes for climate risk prevention and management. For instance:

• Impacts such as imbalance between water supply and demand refer to deficits in early warning.

• Reduction in water stored and supply for irrigation, food production, amplification of the rural exodus, conflicts and food crisis call for responses.

• Loss of food and incomes and food shortage refer to recovery.

• Substantial changes in water demand, increase in silting in the water sources due to soil erosion, and amplification of stakeholder solicitations to decision makers are closely related to structural vulnerability.

These impacts show that AgWater in Nakanbé Basin is vulnerable to CR&C. Therefore, AgWater vulnerability factors have been ranked in four vulnerability classes so-called V, such as V1, V2, V3 and V4, related to early warning, responses, recovery and structural vulnerability, respectively, as follows:

• V1: It is deficits in early warning systems linked to water availability and prevention against conflicts.

• V2: It is lack of alternative water sources to the Nakanbe River, Mogtedo Reservoir and Bagre Dam to supply crops in dry seasons; deficits in the AgWater allocation mechanisms in dry seasons; deficits in water-related conflict management strategies; inadequacies in the mechanisms to face food shortages.

• V3: It is deficits in the recovery mechanisms and models in response to climate risks.

• V4: It is deficits in water source planning to respond to climate risks; poor agricultural practices around the water bodies and the river; deficits in mainstreaming climate risks into the design, implementation and management of the water sources; land degradation in the Nakanbe basin; inefficient water saving strategies and techniques with regard to heat waves and droughts; deficits in integrating climate risks in designing and implementing of the AgWater management frameworks; and deficits in mainstreaming climate risks into the water policy and regulations.

Based on the above-mentioned vulnerability factors, no regret options are identified and selected. To be selected as no regret options, these criteria are defined and applied to each option identified. Those fitting all the criteria have been considered in this paper as no regret AgWater resilience options under CR&C. These main criteria are the following: the resilience option will be susceptible to be implemented in each of the four climate conditions (current climate and the three CC scenarios from ClimPropect model), be likely to have socioeconomic impacts even in absence of CC and be acceptable by stakeholders. Based on vulnerability analysis and the conceptual framework, agricultural water resilience (AgWR) has been considered as a function of five key pillars:

where P1 is early warning set of solutions; P2 is response set of solutions; P3 is recovery set of solutions, P4 is structural resilience set of solutions and P5 is a long-term resilience set of solutions. The elements of these key pillars are specified below.

The measures for reducing cyclical vulnerability are described as a set of solutions by pillars P1, P2, P3 and P4 as follows:

• P1: It is promoting improved system to monitor weather; supporting optimized networks to enhance early warning and ecological resilience to secure water supply and use; creation of the national and local expertise to develop early warning; elaborating consensual models to inform, manage and disseminate early warning; elaborating consensual tools to train farmers to use early warning efficiently through learning by doing (gardens schools); and promoting community-driven alert.

• P2: It is promoting a reliable information system on water balance; elaborating practical crisis information exchange models; making water allocation models available for the Nakanbe River and Bagre and Mogtedo Dams; promoting alternative water sources to face water shortage; strengthening capacities of water policy makers and planers to be able to build, implement and monitor responses to climate risks; supporting farmers training to face climate risks; and promotion of community-driven response.

• P3: It is promoting reliable devices to well-informed climate assurance (automatic weather, rain gauge and groundwater monitoring stations to combine meteorological and land cover data); promoting a national multi-hazard early warning system for AgWater; promoting farmer training on climate insurance; elaborating information exchange models during recovery periods; smart facilities for having access to AgWater; reinforcing agricultural input banks; promoting “cash for work” programs to help users to pay for climate assurance; promotion of alternative water sources on farm through creating awareness of farmers and policy-takers; supporting continuous training of farmers in soils and water conservation techniques on farm; and promoting community-driven recovery.

The set of solutions to achieve such a steady state of AgWater under current climate influence is described below in pillar P4.

• P4: It is promoting national and local multi-hazard climate information and weather forecast systems; promoting AgWater allocation models; making available a database on the best technologies and practices; promoting technology transfer models; elaborating models to allow for sedimentation when building small reservoirs; promoting colloquial toolkits to train farmers in use of weather forecasts; promoting user-driven multi-hazard resilience; reinforcing farmers’ organization leadership through practice exchanges; elaborating modules and training materials on AgWater and climate risks in primary and secondary school programs; promoting low-cost, eco-innovative technologies to harvest water for food production; elaborating toolkits to mainstream climate knowledge in policy and management plans for AgWater; continuous building capacities of water policy makers and planers on smart climate risk management; providing modern technical equipment for the water managerial frameworks; supporting partnerships among researchers-policy makers and farmers’ organizations through knowledge exchange platforms; integrating explicitly socioeconomic and environmental information in water resources database; promoting rural platforms to exchange knowledge on climate risks; encouraging collaborative platform among DGRE – banks and farmer organizations to support farmers having access to smart technologies and credits; initiate climate resilience national tax to help fund resilience initiatives; identifying and disseminating smart endogenous technologies through farmer and policy-maker networks; and identifying consensual set of indicators for monitoring both climate and disaster risks in water policy and practices.

A set of development options for reducing future vulnerability has been identified too as presented by the pillar P5.

• P5: It is building a climate change resilience foundation that brings research, technology and development together; promoting prospective water research and policy which fully integrates socio-economic aspects; making available a regional climate modeling system; building a climate and disaster risk expertise center; elaborating a knowledge management and transfer consensual tool; elaborating a reference guideline for mainstreaming CC into water management frameworks and policies; elaborating technical guidelines for assessing AgWater and environment vulnerability to CC; devices to prevent extreme events (droughts, floods and higher evaporation) for extending climate warning; new resilient tree species seeds and crops; promoting a national remote sensing network and geomantic tools for improving and monitoring agro-ecological systems; national funding agency for climate resilience giving a priority to science and technology; promoting a mutual fund to reinforce farmers’ recovery from damages due to climate hazards; elaborating two consensual (national and local levels) models for assessing costs and benefits of climate resilience; vulgarizing “no regret and flexible” agro-ecological models; and elaborating consensual scientific method including indicators for monitoring AgWater resilience to monitor future climate risks.

However, reducing the key main structural vulnerability factor of each AgWater unit exposed to climate and disaster risks (CC&DR) constitutes the foundation of sustained resilience. In fact, it is such vulnerability that compromises the early warning, response and recovery needs. A set of key factors constitutes an entry point for resilience of the overall AgWater under CC&DR. Therefore, a set of a major vulnerable factors (V) has been identified as follows:

V: It is low mainstreaming of the climate risks in the design and management of the dam, reservoir and river; land degradation in the Nakanbe Basin; inefficiency of water-saving techniques to respond to heat waves and droughts; and deficits in integrating climate risks in making and running water managerial frameworks and policies for food production.

The resilience set of solutions (R) related to it is the following:

R: It is the best mainstreaming of climate risks in agricultural water infrastructure design and management; land cover restoration in the Nakanbe Basin; vulgarization of the smart water saving techniques and strategies for food production; and better mainstreaming of CC&DR within the water managerial frameworks and policies.

The persistence of climate effects on agricultural water calls for rethinking the strategies and policies, especially with regard to climate uncertainties. In this sense, “no regret” solutions that provide socioeconomic benefits, even in absence of expected climate effects (Hallegatte, 2009), are identified to overcome this complex issue. The proposed solutions could be looked at in two different ways: enhancing knowledge and supporting decision-making. With regard to the former, first, the findings established a current and future climate impact mapping. This mapping shows a convergence of socioeconomic impacts concerning water supply, quality and demand. It implies that the process aiming to enhance AgWater resilience to CC&DR in Burkina Faso should prioritize an integrated approach. Second, in BF, AgWater resilience to CC&DR is required to improve farmer living conditions. Yet, despite the current and expected climate impacts on AgWater, there is still not a consensual and robust framework for addressing vulnerability to a changing climate. In such a context, the proposed resilience framework, particularly those related to structural resilience, is a new one whereupon further thoughts, strategies and policies could be built for AgWater security. It covers all phases of disaster risk reduction and the long-term resilience as well. The paper also proposes the new way of which resilience issues should be undertaken. It clearly highlights that climate and disaster risks overlapped in BF, and as such, they should be dealt with at the same time for effectiveness of actions. Another output of this paper is the construction of an integrated and coherent tool that can help to foster synergy among stakeholders in water, agriculture and forestry sectors. As pointed out by Mamun (2013), co-management contributes to reinforcing ecological system resilience.

In terms of decision-making, the approach brought together national level (where policies are made) and local level (where they are implemented). In practice, the implementation of the five pillars requires adjustments at the institutional level and a new interface between research institutes, policy institutions and farmers’ organizations. Specifically, the implementation of Pillar P1 set of solutions will be useful in reducing the impacts significantly, as it can help prepare farmers to face the risks. This is also highlighted by UNEP (2012) and Roncoli et al. (2002) for the Sahel. However, in BF, producing and disseminating reliable early warning and climate information and seasonal forecasts require the modernization of the weather monitoring network by promoting automatic network and building up a national expertise. The solutions of Pillar P2 can help address damages effectively, singularly in emergency context. The Pillar P3 set of solutions could reduce the current vulnerability while serving as a basis for building a recovery framework especially at the local level. Pillar P4 should stabilize the system, and Pillar P5 should overcome the long-term effects. It appeared clearly that the vulnerability factors and solutions identified are dominated by non-structural factors and options. These types of solutions are popular “investment” to get high support with a little financial expense. Thus, they can easily also be supported by farmers’ organizations. Identifying such solutions can be explained by the fact that up to now, in BF, efforts for reducing AgWater vulnerability have been limited to building infrastructure and other structural options (dams, reservoirs and wells).

However, reinforcing sustainable AgWater resilience should primarily be focused on reducing structural vulnerability (Pillar P4), as it is this kind of resilience that compromises the early warning, response and recovery needs and prevents long-term effects. Here, some of these kinds of solutions are discussed to show their relevance in the context of a changing climate with uncertainties, especially in relation to the future of rainfall. They relate to mainstreaming CC in water infrastructure design, environment integrity, water productivity, access to information, anticipative water allocation models, capacity building and prioritizing research in decision-making.

Developing infrastructures for AgWater security should avoid the deficits of the ongoing water mechanisms by explicitly integrating climate issue. The practical options involve promotion of smart water and land conservation techniques and strategies, especially in the Soudano-sahelian and Sahelian zones. The former includes promoting underground ecological microdams, micro-ecological basins and flexible rainfall and run-off harvesting techniques, whereas the latter refers to agroforestry associated with water-saving techniques. Some advantages of such options include water availability and reliability on farm; increase in groundwater amount by infiltration, flood control, erosion and pollution limitation; and reducing costs. They can sustainably prevent watershed degradation and improve soil fertility. Bouma et al. (2015) and FAO (2013) noticed the relevance of such options, as they increase food productivity and farmer’s incomes.

Improving water productivity on-farm includes promoting proven technologies for water saving, low-carbon techniques while increasing agricultural productivity. These techniques and strategies are diverse. They incorporate multi-hazard information for seasonal forecasts, underground broadcaster irrigation, alternate irrigation channel, micro-irrigation, wireless sensor networks, water-efficient cropping patterns, watering at less hot time, early crop varieties with high yields and the use of films and plastic tunnels (Morris and Barron, 2014; Mehta et al., 2013; Mirza et al., 2013; Atta et al., 2011). For instance, in semi-arid areas, the combination of such techniques with user and planer trainings can save AgWater up to 90 per cent and increase yields, as experimented in more than 45 countries (Blanke et al., 2007; Kulkarni, 2011).

Sadoff and Muller (2009) noted that information is essential, and ability to use it at the appropriate time is more helpful. In this regard, developing local multi-hazard information systems should provide actionable and useful warnings and season forecasts. Practically, policy makers and AgWater users should have access to end-user-tailored climate information on a need-to-know basis and how to use it. It can help plan irrigation efficiently and other agricultural activities including selection of seeds and crops, weeding, tillage and sowing (Roudier et al., 2014; Furman et al., 2014). Compared to traditional techniques, Kulkarni (2011) supported that seasonal forecasts have allowed saving 50 per cent of water, increasing food production and scaled-up returns on investment.

Among the well-known water allocation mechanisms, a user-based allocation should be promoted in BF. In fact, it can be managed easily by users themselves with a regulation from water management agencies. Other advantages include being participatory and preventing water-related conflicts, sustainable, feasible and politically acceptable, as it provides advantages both for users and policy makers. In addition, by raising water productivity, it can also boost farmers’ participation in IWRM. Notwithstanding, its efficiency can be limited by users with poor socioeconomic conditions. To this end, Place and Dewees (1999) supported that economic mechanisms such as incentives and subsidies can raise awareness for adapting this best practice. In fact, it will be wrong to expect changes in poor farmer practices if they do not easily access to information, credits and technologies. This is also pointed out by Mwakalila (2014) in two districts in Tanzania. However, unsuccessful initiatives may be useful to draw lessons and take challenges up.

In addition, training, knowledge exchange programs and interfaces among climate information producers, popularizers and users will be helpful to develop successful joint initiatives, as underscored by Wossen et al. (2013) in arid zone in Ethiopia. Khodran et al. (2013) asserted that where capacity building is successful, stakeholders are generally able to provide sustainable services saving water up to 90 per cent. This is also noticed by Adger et al. (2011) and WMO and GWP (2006).

Beyond that, culture of resilience should be promoted by integrating CC in basic education. Thus, mainstreaming CC&DR into legislation on water could be an entry point for integrating CC into development for sustainable food security. In fact, the mere mention of CC in water policies and legislation can change the way of acting, as also noticed by Suhardiman et al. (2014). However, an effective management is the one that includes rigorous research and technology. This is also mentioned by Klopper et al. (2006) in managing seasonal forecasts. Therefore, research should be a main component of such a resilience policy.

Though many efforts have been made, climate impacts on water resources impede the achievement of food security in BF. Such recurrent impacts question the way decision makers and development agencies act to overcome climate effects. Thus, the purpose of this paper was to develop a “no regret” framework by using an innovative approach that will strongly be able to attenuate the current and future climate impacts on AgWater and subsequently food production in BF. This framework is a new resilience indicative framework that encompasses current and future climate risk resilience. In practice, this “no regret” resilience tool is structured into five pillars with a set of solutions related to early warning, response, recovery, chronic resilience and long-term resilience. The proposed resilience tool calls for an adjustment of ongoing water policies in BF. Such adjustments could include reinforcing national and local water framework capacity to deal with climate risks. Solutions encompass both those which are directly related to WR and those beyond, including agriculture and land use. Indeed, combating effectively climate harmful effects on WR requires cross options, including human well-being, and goes beyond the water sector only. In addition, these solutions could fit the political aim that seems to attract communities’ confidence and so should be supported easily by the government. In fact, citizens could show strong appreciation for the political leader. Concerning scientific consideration, the proposed intervention framework has been based on an integrated approach that will be useful for future research on climate risks and change. The proposed tool should be seen as a toolkit and a guide to sustainably support the current undertakings. Each solution should be tailored to each community to guaranty its success. The future work consists of planning these resilience options based on costs-benefits analysis with regard to the country’s socioeconomic and human context.

The authors are grateful to the West Africa Economic and Monetary Union (UEMOA) for giving a PhD research Excellence Fellowship to the lead author. They would like to thank National Meteorological and hydrological Services of Burkina Faso and Volta Basin Authority for providing the hydroclimatic and hydrological data for free of charge. The authors are also grateful to Dr Bruno Barbier from CIRAD for his help and assistance. They are also grateful to stakeholders, especially the experts, decision makers and the farmers’ communities, to take great interest in this research. The study was carried out in Burkina Faso, where the first author was hosted by the Institute d’Application et de Vulgarssation des Sciences (IAVS).