Promue comme la technique d'irrigation en théorie la plus efficiente, et mise à contribution pour régler les grandes crises d'eau, le goutte à goutte se diffuse très rapidement de par le monde. La perspective d'une technique étudiée en station expérimentale, sans utilisateurs, et en définitive idéalisée, nous a d'emblée interpellé. Rejetant une vision normative du goutte à goutte (ce qu'il devrait être/faire), nous l'avons étudié en pratique, en interactions avec une multitude d'acteurs (ce qu'est le goutte à goutte et ce qu'il fait et pour qui). Notre objectif est d'analyser comment la pratique de cette innovation technique a changé la technologie et les trajectoires socioprofessionnelles des acteurs, et l'impact sur les efficiences d'irrigation. Nous montrons comment les acteurs locaux ont pris le contrôle de l'innovation, et des responsabilités dans des domaines que l'on pensait réservés aux ingénieurs dans la production, l'utilisation et la diffusion de l'innovation. Ces initiatives locales croisent et renforcent les programmes étatiques de développement du goutte à goutte. La technologie a séduit un grand nombre d'acteurs non-conventionnels qui ont changé la technologie, mais ont changé aussi eux-mêmes. La diffusion réussie du goutte à goutte assure leur promotion socioprofessionnelle, et à leur tour ils vont attirer d'autres utilisateurs dans le monde du goutte à goutte. C'est là où se résume la force du goutte à goutte comme objet technique autour duquel de puissants réseaux sociotechniques se sont construits. Le bricolage a permis la conception de systèmes plus adaptés aux conditions locales, mais aussi un apprentissage de la technique pour apprivoiser le changement. D'une technologie importée accessible à une minorité de grands agriculteurs, une pluralité de systèmes de goutte à goutte de proximité sont désormais accessibles au plus grand nombre. Enfin, les performances d'irrigation de ces systèmes sont hétérogènes, certains agriculteurs irriguant 3-4 fois plus que les volumes nécessaires aux besoins de la culture. Ces performances peuvent être expliquées par des pratiques d'irrigation favorisant le confort hydrique de la plante, traduisant ainsi les logiques des acteurs. Aujourd'hui, « l'économie d'eau » est un objectif irréalisable, qui n'est prioritaire pour aucun des acteurs, y compris l'Etat-promoteur. Nous concluons qu'il serait intéressant de mobiliser les acteurs non-conventionnels pour en faire des alliés pour intégrer le concept d'économie d'eau sur le terrain auprès des utilisateurs. / Abstract Promoted as the most efficient irrigation technology, able to contribute to solving the major current water crises, drip irrigation spreads quickly throughout the world. The perspective of an idealized technology with a theoretical efficiency studied in experimental stations without users struck us immediately. Rejecting a normative view of drip irrigation (what it should be/do), we studied it in practice, interacting with a multitude of actors focusing on what it does and for whom. Our objective is to analyse how the practice of this technical innovation has changed the technology and the socio-professional pathways of different actors involved in drip irrigation, and analyse the impact on actual irrigation efficiencies. We show how local actors took control of the innovation taking responsibility in spheres once thought reserved for engineers in the production, use and dissemination of innovation. These local initiatives crossed and strengthened state development programs promoting drip irrigation. The technology attracted a large number of non-conventional actors who changed the technology, but also changed themselves. The successful diffusion of drip irrigation ensures their socio-professional promotion, and they in turn will attract more users to the world of drip irrigation. This constitutes the strength of drip irrigation, which is a technical object around which powerful socio-technical networks were built. Bricolage helped design systems better suited to local conditions, but was also a learning process to incorporate change. From an imported technology accessible to a minority of large farmers, a plurality of drip systems is now accessible for a wide diversity of farmers. Finally, the performance of irrigation systems is heterogeneous, some farmers irrigating 3-4 times the volume necessary for crop water requirements. This performance can be explained by irrigation practices favouring a water comfort to crops, reflecting the logic of the actors. Today, "water saving" is an unattainable goal, as it is not a priority for any of the actors, including the state. We conclude that it would be interesting to engage non-conventional actors as allies in order to promote the concept of water saving on the field with users.

Policy makers adopt irrigated agriculture for food security, since irrigation doubles crop production. Therefore, the development of large irrigation systems has a long history in many places worldwide. Although large-scale irrigation schemes play an important role in improving food security, many schemes, especially in Africa, do not yield the expected outcomes. This is related to poor water management, which is generally due to a lack of effective evaluation and monitoring. The objective of this study, therefore, is to propose a new methodology to assess, evaluate and monitor large-scale irrigation systems. Information on irrigation indicators is needed to enable the evaluation of irrigation performance. The evaluation is the first and the most significant step in providing information about how it is performing. After reviewing extensive literature, a list of indicators related to the performance of irrigation, rainwater supply and productivity is suggested. The irrigation efficiency indicators Relative Irrigation Supply (RIS) and Relative Water Supply (RWS) are selected. Potential rainwater supply to crops can be tested based on the Moisture Availability Index (MAI) and the Ratio of Moisture Availability (RMA). Water productivity can be assessed by Crop Yield (Y) and Water Use Efficiency (WUE). However, the central problem facing large-scale irrigation schemes is always the lack of data, which calls for the development of a new method of data acquisition that allows evaluation and monitoring. Remote Sensing (RS) technology makes it possible to retrieve data across large areas. Two different approaches via RS, the Normalized Difference Vegetation Index (NDVI) and Actual Evapotranspiration (ETa), can be utilized for monitoring. The well-known Vegetation Condition Index (VCI), derived from the NDVI, is modified (MVCI) to allow a qualitative spatio-temporal assessment of irrigation efficiency. MVCI takes into account crop response to water availability, while ETa indicates whether water is used as intended. Furthermore, the assessment of the possible hydrological impact of the irrigation system should be considered in the evaluation and monitoring process. The Sudanese Gezira Scheme of 8,000 square kilometers in the Nile Basin, where performance evaluation and monitoring are absent or poorly conducted, is no exception. This research takes the large-scale irrigation of the Gezira Scheme as a case study, as it is the largest scheme, not only in the Nile Basin but also in the world, under single management. The first long-term historical evaluation of the scheme is conducted for the period 1961–2012 rather than only on a short-time scale as is the common practice. An increase in RIS and RWS values from 1.40 and 1.70 to 2.23 and 2.60, respectively, since the 1993/94 season shows decreasing irrigation efficiency. MAI and RMA for summer crops indicate a promising rainfall contribution to irrigation in July and August. The Gezira Scheme achieves low yield and WUE in comparison to many irrigation schemes of the globe. Low productivity is mainly due to poor distribution and irrigation mismanagement. This is indicated by the 15-year MVCI spatio-temporal analysis, which shows that the northern part of the scheme experiences characteristic drought during the summer crop season. Although MVCI can be considered a monitoring tool, the index does not deduct the soil water content, and water could be wasted and available in other ways (e.g. water depressions). Spatio-temporal information for ETa is required to better quantify water depletion and establish links between land use and water allocation. However, several RS models have been developed for estimating ETa. Thus, improving the understanding of performance of such models in arid climates, as well as large-scale irrigation schemes, is taken into account in this study. Four different models based on the energy balance method, the Surface Energy Balance Algorithm for Land (SEBAL), Mapping EvapoTranspiration at High Resolution with Internalized Calibration (METRIC™), Simplified Surface Energy Balance (SSEB) and MOD16 ET are applied in order to determine the optimal approach for obtaining ETa. Outputs from these models are compared to actual water balance (WB) estimates during the 2004/05 season at field scale. Several statistical measures are evaluated, and a score is given for each model in order to select the best-performing model. Based on ranking criteria, SSEB gives the best performance and is seen as a suitable operational ETa model for the scheme. SSEB subsequently is applied for summer and winter crop seasons for the period 2000–2014. Unfortunately, one of the limitations faced in the current research is the absence of validation data on a regional scale. Therefore, the assessment focuses on spatial distribution and trends rather than absolute values. As with the MVCI distribution, the seasonal ETa for the Gezira Scheme is higher in the southern and central parts than in the northern part. This confirms the robustness of the developed MVCI. To avoid using absolute values of ETa, the ratio of ETa from agricultural areas (ETagr) to the total evapotranspiration (ET) from the scheme (ETsum) is calculated. The ETagr/ETsum ratio shows a descending trend over recent years, indicating that the water is available but not being utilized for agricultural production. This study shows that SSEB is also useful for identifying the location of water losses on a daily basis. Around 80 channels are identified as having leakage problems for the 2013/14 crop season. Such information is very useful for reducing losses at the scheme. In addition, Rainwater Harvesting (WH) is addressed and found to be applicable as an alternative solution for accounting for rainfall in irrigation. It is seen that these management scenarios could save water and increase the overall efficiency of the scheme. It is possible to save 68 million cubic meters of water per year when the overall irrigation efficiency of the scheme is improved by only 1%. A level of efficiency of 75% is predicted from the proposed management scenarios, which could save about 2.6 billion cubic meters of water per year. In conclusion, the present study has developed an innovative method of identifying the problems of large-scale schemes as well as proposing management scenarios to enhance irrigation water management practice. Improved agricultural water management in terms of crop, water and land management can increase food production, thereby alleviating poverty and hunger in an environmentally sustainable manner.

Policy makers adopt irrigated agriculture for food security, since irrigation doubles crop production. Therefore, the development of large irrigation systems has a long history in many places worldwide. Although large-scale irrigation schemes play an important role in improving food security, many schemes, especially in Africa, do not yield the expected outcomes. This is related to poor water management, which is generally due to a lack of effective evaluation and monitoring. The objective of this study, therefore, is to propose a new methodology to assess, evaluate and monitor large-scale irrigation systems. Information on irrigation indicators is needed to enable the evaluation of irrigation performance. The evaluation is the first and the most significant step in providing information about how it is performing. After reviewing extensive literature, a list of indicators related to the performance of irrigation, rainwater supply and productivity is suggested. The irrigation efficiency indicators Relative Irrigation Supply (RIS) and Relative Water Supply (RWS) are selected. Potential rainwater supply to crops can be tested based on the Moisture Availability Index (MAI) and the Ratio of Moisture Availability (RMA). Water productivity can be assessed by Crop Yield (Y) and Water Use Efficiency (WUE). However, the central problem facing large-scale irrigation schemes is always the lack of data, which calls for the development of a new method of data acquisition that allows evaluation and monitoring. Remote Sensing (RS) technology makes it possible to retrieve data across large areas. Two different approaches via RS, the Normalized Difference Vegetation Index (NDVI) and Actual Evapotranspiration (ETa), can be utilized for monitoring. The well-known Vegetation Condition Index (VCI), derived from the NDVI, is modified (MVCI) to allow a qualitative spatio-temporal assessment of irrigation efficiency. MVCI takes into account crop response to water availability, while ETa indicates whether water is used as intended. Furthermore, the assessment of the possible hydrological impact of the irrigation system should be considered in the evaluation and monitoring process. The Sudanese Gezira Scheme of 8,000 square kilometers in the Nile Basin, where performance evaluation and monitoring are absent or poorly conducted, is no exception. This research takes the large-scale irrigation of the Gezira Scheme as a case study, as it is the largest scheme, not only in the Nile Basin but also in the world, under single management. The first long-term historical evaluation of the scheme is conducted for the period 1961–2012 rather than only on a short-time scale as is the common practice. An increase in RIS and RWS values from 1.40 and 1.70 to 2.23 and 2.60, respectively, since the 1993/94 season shows decreasing irrigation efficiency. MAI and RMA for summer crops indicate a promising rainfall contribution to irrigation in July and August. The Gezira Scheme achieves low yield and WUE in comparison to many irrigation schemes of the globe. Low productivity is mainly due to poor distribution and irrigation mismanagement. This is indicated by the 15-year MVCI spatio-temporal analysis, which shows that the northern part of the scheme experiences characteristic drought during the summer crop season. Although MVCI can be considered a monitoring tool, the index does not deduct the soil water content, and water could be wasted and available in other ways (e.g. water depressions). Spatio-temporal information for ETa is required to better quantify water depletion and establish links between land use and water allocation. However, several RS models have been developed for estimating ETa. Thus, improving the understanding of performance of such models in arid climates, as well as large-scale irrigation schemes, is taken into account in this study. Four different models based on the energy balance method, the Surface Energy Balance Algorithm for Land (SEBAL), Mapping EvapoTranspiration at High Resolution with Internalized Calibration (METRIC™), Simplified Surface Energy Balance (SSEB) and MOD16 ET are applied in order to determine the optimal approach for obtaining ETa. Outputs from these models are compared to actual water balance (WB) estimates during the 2004/05 season at field scale. Several statistical measures are evaluated, and a score is given for each model in order to select the best-performing model. Based on ranking criteria, SSEB gives the best performance and is seen as a suitable operational ETa model for the scheme. SSEB subsequently is applied for summer and winter crop seasons for the period 2000–2014. Unfortunately, one of the limitations faced in the current research is the absence of validation data on a regional scale. Therefore, the assessment focuses on spatial distribution and trends rather than absolute values. As with the MVCI distribution, the seasonal ETa for the Gezira Scheme is higher in the southern and central parts than in the northern part. This confirms the robustness of the developed MVCI. To avoid using absolute values of ETa, the ratio of ETa from agricultural areas (ETagr) to the total evapotranspiration (ET) from the scheme (ETsum) is calculated. The ETagr/ETsum ratio shows a descending trend over recent years, indicating that the water is available but not being utilized for agricultural production. This study shows that SSEB is also useful for identifying the location of water losses on a daily basis. Around 80 channels are identified as having leakage problems for the 2013/14 crop season. Such information is very useful for reducing losses at the scheme. In addition, Rainwater Harvesting (WH) is addressed and found to be applicable as an alternative solution for accounting for rainfall in irrigation. It is seen that these management scenarios could save water and increase the overall efficiency of the scheme. It is possible to save 68 million cubic meters of water per year when the overall irrigation efficiency of the scheme is improved by only 1%. A level of efficiency of 75% is predicted from the proposed management scenarios, which could save about 2.6 billion cubic meters of water per year. In conclusion, the present study has developed an innovative method of identifying the problems of large-scale schemes as well as proposing management scenarios to enhance irrigation water management practice. Improved agricultural water management in terms of crop, water and land management can increase food production, thereby alleviating poverty and hunger in an environmentally sustainable manner.