J-M. Deumier, Institut Technique des Céréales et des Fourrages, Baziege, P. Leroy, Institut National de la Recherche Agronomique, Grignon, and P. Peyremorte, Société du Canal de Provence, Le Tholonet, Aix-en-Provence, France
The improvement of irrigated crop system management has become a priority mainly for economic and environmental context evolution reasons.
For agricultural crop systems, for which the irrigation cost is high in relation to the gross product, irrigation control is primarily a question of management of means: water resources, equipment, manpower. It therefore depends on strategic choices which must be considered well before the irrigation season:
· choice of cropping plan according to farm constraints: which crops, which crops to irrigate and which production and margin objectives;
· provisional irrigation programme, which is a set of decision-making rules concerning irrigation management; this programme has to be coherent with the strategy previously defined.
These strategic choices are determinant for the tactical decisions made during the irrigation season: irrigation management is usually limited to carrying out the forecast plan.
Three tools have been developed to provide answers to these questions:
· LORA is a software which enables farmers to achieve the optimal cropping plan on the irrigable area.
· IRMA is an irrigation simulator which allows farmers to test and to improve a provisional irrigation programme.
· IRRISA is a software designed to decide irrigation strategies; this method gives a timetable which allows farmers to schedule irrigation according to the strategy.
There are many reasons for devoting attention to the management of irrigated systems:
· the rapid evolution of economic conditions which leads farmers to modify their production systems,
· the need to respect the environment leads us to reflect on the consequences of the management of irrigated systems on the quantitative and qualitative control of the water resource.
The modalities of irrigation management are very different for high value and agricultural crops.
For high value crops, water cost is a small fraction of the gross product. The farms concerned are well supplied with irrigation equipment and the economic optimum for the crop is close to the technical optimum. Farmers' queries concern irrigation scheduling meeting the objective of satisfying the plants' water requirements without overirrigating.
The problems posed by the irrigation management of agricultural crops are very different. It is strongly influenced by the high cost of irrigation relative to the product. Thus, within the framework of collective arrangements, irrigation costs can be broken down as: $US 260 for the infrastructure, $US 120 for equipment and $US 80 working cost, that is a total of $US 460 per hectare. In reality costs can vary between $US 300 and $US 600 per hectare (Balas and Deumier, 1989). This represents between 20% and 30% of the gross product (including subsidies) with a high proportion (80%) due to infrastructure and equipment compared to running costs (20%). Because of this high cost, the capacity of the equipment is never structured to manage dry years, and irrigation systems on farms do not enable crop water requirements to be satisfied in all situations.
Furthermore, recent economic change has led to increased irrigation costs relative to the product and induced a change in the use of irrigation systems. For instance, a tendency to increase the irrigated area can be observed: the objective is to spread the fixed charges over a greater area whilst not aiming for maximum yield. This is not without risk in dry years and in relation to an uncertain water supply (shortage due to dry winters, in volume from lakes and in flow from water tables and rivers).
In this type of context, particularly frequent in the irrigated agricultural crop systems in France, the farmers' queries concern the overall management of their irrigated systems that is the choice of cropping plan on the irrigable area, consideration of a forecast irrigation plan and, finally, the management of irrigation.
STRATEGIC AND TACTICAL CHOICES FOR IRRIGATION MANAGEMENT OF AGRICULTURAL CROPS
It is possible to identify strategic decisions taken before the irrigation season (Peyremorte and Tron, 1989; Balas and Deumier, 1989) as follows:
· Choice of cropping plan: this choice is made for the irrigable area of the farm in the light of the water resources, equipment and manpower and takes into account the economic and regulatory context (product price, level of subsidy and the area ceiling, input costs (particularly of water) and set-aside). The questions are: Which crops and which crops to irrigate? For each crop what production and margin objective? At the same time the irrigation strategy is fixed in order to achieve the production objective, that is to say, allocation of volume and flows:
· by species, according to yield objectives and priorities between species,
· in time, taking into account the water provision aimed for, and possible rationing of crops.
· Consideration of a provisional irrigation programme. This programme is a set of decision rules concerning irrigation management that are coherent with the strategy previously define:
· What is the start rule for each species? What is the irrigation amount and repeat frequency during the season? How is rainfall taken into account? How and when is irrigation brough to an end?
· How are priorities managed between species and between soil types for the species?
The irrigation programme must respect the constraints imposed by the management of equipment and the water resource.
To solve these questions, the procedure for defining a strategy consists of three stages:
1. For each crop several provisional programmes are developed whose simulations allow us to obtain a range of values for the level of water requirement satisfaction. These programmes are constructed taking account of:· the agronomic parameters of each crop (water requirement, stress sensitivity, sensitive periods, etc.),
· the soils and climate characteristics and, particularly, climate variability.
2. A calculation is made for the different levels of water requirement satisfaction and, for each crop, of the irrigation water requirement (volume and flow) and the economic consequences.
3. A choice is made of the 'crop satisfaction of water requirement' combination which will produce the best economic result over a range of climatic scenarios.
It is, in fact, particularly important to take account of climate variability. One can work from an average climate year but this method is less satisfactory as it is unable to estimate the risk, for example, in a dry year. Another possibility consists in setting up a typology of the climate years: for instance, 'very dry year', 'dry year', 'average year'. It is then possible to choose the 'crop satisfaction of water requirement' combination which gives the best average economic result for the set of scenarios. This procedure allows us to assess the risk taken for each type of year.
These strategic choices are determinant vis-a-vis the tactical decisions made during the irrigation season: irrigation management is usually limited to carrying out the forecast plan while taking into account climate and cultural occurrences.
Farmers' queries are, in fact, very varied. It is sometimes necessary to deal with the overall problem (cropping plan, irrigation programme, irrigation scheduling). In other cases the cropping plan has already been established in the light of economic factors while respecting water constraints. In this case an irrigation programme has to be constructed and a management method adapted to the farmer's strategy has to be proposed.
In the following part of this paper there is a description of three decision support tools developed to provide answers to questions related to:
· choice of cropping plan and irrigation strategy: LORA, developed by the Institut Technique des Céréales et des Fourrages (ITCF) and the Institut National de la Recherche Agronomique (INRA);
· forecast irrigation programme: IRMA, developed by the two above-mentioned institutes;
· irrigation management method coherent with irrigation strategy: IRRISA, developed by the Société du Canal de Provence (SCP) in collaboration with ITCF and INRA.
TOOLS FOR THE MANAGEMENT OF IRRIGATED SYSTEMS
LORA, a decision support system to facilitate the choice of cropping-plan on the irrigable area
LORA is a decision support system for helping farmers and their advisers to develop a cropping plan on the irrigable area of the farm (Jacquin et al, 1993; Leroy and Jacquin, 1994).
The questions concern crop and area. Which crops should be irrigated? What level of their water requirement should be aimed at? Different elements need to be taken into consideration: soil and climate characteristics, the response to water of different crops, the relationship between needs and means (water resource and equipment capacity). The objective is to obtain the best economic return from these irrigation means for the whole of the irrigable area. Estimation of the risks inherent in climate variability and uncertainty regarding production prices or the water resource are also factors which are essential to the choice.
Having taken into account soil water capacity, climate and the technico-economic context of the farm (potential yields; product prices; subsidies; input costs, particularly of water), LORA seeks a cropping plan for irrigated and non-irrigated crops which:
· maximizes the margin on the irrigable area,
· is compatible with the irrigation set-up in terms of volume and ten-day capacity, linked to the available flow and the working time of the equipment,
· takes climate variability into account.
The climate and its variability are assessed from data obtained from a nearby meteorological station. The user selects a series of years from those available. Each one is treated as a possible weather scenario for the next season.
For each crop envisaged by the farmer and for each soil type on the irrigable area, the model considers different management options: a well-irrigated strategy, restrictive irrigation management, non-irrigated management.
At the first stage, LORA elaborates an irrigation schedule for each crop, soil and management choice for each climate scenario. It then assesses the effect of each schedule on yield and gross margin.
The soils are described as an available water capacity (AWC) and an easily available water capacity (EAWC) (Thévenet et al., 1987).
For well irrigated management, a schedule of water requirements is calculated for each selected climate scenario using a ten-day water balance. At intervals often days, taking rainfall into account, the irrigation requirement is calculated in order to cover maximum evapotranspiration (ETM)1 of the crop and to maintain a minimum level of water content in the soil at the end of the ten-day period (EAWC empty).
1 ETM is calculated every ten days from reference evapotranspiration (ETref) and crop coefficients (kg). These are set against a standard growth cycle which is predetermined for each crop (ETM = kcETref).
For the management of restrictive irrigation, the irrigations are calculated pro rata of a restriction coefficient (80% and 60% of the water depth calculated for well-irrigated management). For restrictive and dry management, water stress is calculated from a water balance of the total AWC; the actual evapotranspiration (ETA) is equal to ETM as long as the EAWC is not exhausted, and falls as soon as the remaining AWC is put to use (Brochet and Gerbier, 1975).
The potential yield of the crops, as defined by the farmer, is, by hypothesis, always attained with well-irrigated management. For restrictive and non-irrigated management, the consequences on yield are estimated using production functions: relative yield is calculated as a linear function of the ETA/ETM ratio on all or part of the crop cycles (Doorenbos and Kassam, 1980). Gross margins are then calculated in the light of hypotheses concerning sale prices, subsidies and the different input costs, notably of water.
FIGURE 1 - Structure of the software
At the second stage, a linear program searches for cropping plan solutions compatible with the water resource and irrigation capacity which maximize the overall gross margin. Two modalities are proposed to account for the series of climate years.
LORA first seeks an optimal solution for each climate scenario considered independently. This allows us to distinguish cropping plans particularly adapted to certain types of year and provides a ceiling curve of economic results for climate scenarios; these solutions are termed 'ideal' on Figure 2.
Subsequently, LORA seeks a single cropping plan in terms of irrigated and non-irrigated crops which, on average for the sum of climate scenarios envisaged, produces the best gross margin. However, the management of irrigation is adapted to each scenario respecting the constraints associated with the irrigation equipment.
The user can also test his own irrigated and non-irrigated cropping plan. The model evaluates it over all the climate scenarios and seeks the optimal management methods.
The use of the program comprises four stages (Figure 1): data input, agronomic calculations, search for solutions and analysis of results.
Regional databases collect the information needed for the agronomic calculations: meteorological information from a group of weather stations, characteristics of the principal soil types in the region, cycles for the crops grown in the region and the agronomic parameters necessary to calculate water requirement and the consequences of water stress (crop coefficients, production functions).
The recording of data takes place in two stages: (i) input of general data (choice of regional database and a nearby weather station, extent of irrigable area and choice of soil types); (ii) creation of a data set (choice of climate scenarios; choice of crops in the database; description of the irrigation installation characterized by volume, flow, operating time and water cost; product sale prices, potential yields and input costs per crop and per soil type; crop area limitations).
During the calculation of irrigation requirements, yields and margins the whole set of results can be visualized. This enables validation of these results from local references.
Following the search for solutions and the testing of the farmer's cropping-plan intentions, the analysis of results is an essential element in the dialogue with the farmer.
FIGURE 2 - Economic results of the different solutions
Results are supplied in the form of graphs and enable analysis of each solution:
· the cropping plan chosen on the irrigable area and by soil type,
· the volume of water consumed and the ten-day periods when irrigation capacity is saturated for each climate scenario,
· the management methods chosen by crop according to climate scenarios.
A synthesis of economic results (Figure 2) allows visualization of the gross margin per hectare on the irrigable area according to climate scenario for the solutions put forward by LORA and those tested by the farmer. Analysis of the relative positions of these curves allows us to assess the value and risk of each solution and to set up a dialogue with the farmer.
It is then possible to modify the hypotheses retained in the data set and to test their consequences against the results. Each data set and its results are stored. This allows us to compare them and to measure the sensitivity of each solution to the different parameters.
IRMA, a simulator for management of the irrigated area of the farm
The objective of this simulator is to facilitate the analysis and improvement in organization of the irrigation sites over the whole irrigated area of the farm. In technical and economic terms, it evaluates the relationship between the farmers' strategic choices (cropping plan and irrigation strategy per crop) with their tactical decisions during the season whilst keeping in view the conflicting demands on available resources (water, equipment and manpower) (Leroy et al., 1996).
In terms of its use, the objective of the tool is to lead the farmers to reflect on their action rationale and on the way in which they organize their work in their particular context of resources and constraints, and in the face of variable factors such as climate in particular. The simulator allows assessment of the coherence of this action rationale in the light of the farmers' objectives and strategy, and to seek improvements to it.
IRMA is based on a representation of the farmer's decision models. This model translates fee strategic choices. It takes fee form of a set of decision rules for fee irrigation scheduling of fee different crops and for managing resources throughout a season. This set of rules governs fee organization of fee irrigation sites taking into account fee crops irrigated, the physical and agronomic nature of fee fields, fee available volume and flow of water, fee structure of fee irrigation network, fee characteristics of fee equipment and fee work time required for its use.
FIGURE 3 - Working principle of IRMA
Following a series of interviews wife farmers and observation of irrigation seasons, we were led to structure these decision rules in fee following way:
· rules for constituting irrigation blocks, grouping together positions irrigated in a homogeneous way;
· rules for irrigation management by block to decide on start dates, irrigation amounts, repeat dates and stop dates; these rules, which are expressed as 'if condition then decision', mobilize a set of indicators relating to climate conditions, soil water state, the development stage of the crops, the actual or forecast resource state, etc.;
· rules for managing priorities between blocks; the irrigation season is broken down into periods during which the composition of the blocks and the order of priority between blocks does not vary;
· rules governing the use of water resources, flow and equipment on the various blocks according to their characteristics and the particular constraints imposed by the farm and the environment;
· rules for managing the farmer's working time.
This decision-making model is set against different climate scenarios. For each of them it simulates the progress of the season on a daily basis: it applies the rules to decide on irrigation and it uses sub-models to make the system state evolve (Figure 3). Thus, IRMA generates irrigation schedules for the totality of irrigated fields by applying the rules and respecting the constraints and the working conditions of the resources.
During the entire simulation, agronomic models evaluate the state of the indicators used in the decision rules. They equally allow us to assess the consequences of the irrigation schedules simulated on the soil water state, the satisfaction of crop water requirement and the effect on yields. Economic hypotheses concerning crop price, the cost of water and other inputs allow us to calculate gross margins per crop and for the cropping plan envisaged for each climate scenario.
The results obtained for the various climate scenarios can then be analysed from different points of view: organization of irrigation schedules, use of water resources and equipment, time given over to irrigation, yields and margins achieved, etc. It is also possible to envisage the consequences of modifications to the decision-making rules, economic data, irrigated cropping plan, water resources, irrigation equipment and the availability of human resources.
FIGURE 4 - Organigram of IRRISA
In addition to its value in facilitating decision-making for farmers, this tool is also a means for understanding and making explicit the management of irrigation at farm level.
This software derives from the 'bilaneaumètre' method (Peyremorte and Tron, 1986, 1987). Its aim is to:
· check the compatibility between water requirements and irrigation means, in particular equipment and supply, for a given cropping plan,
· provide a support tool for irrigation management (Figure 4).
Before the start of the irrigation season, to help farmers to build a strategy, a balance is sought between the needs of the irrigated cropping plan and the means calculated on a weekly basis.
The water requirement depends on the crop, its area and the yield objective. It has been hypothesized that the level of water requirement of crops depends on the yield objectives. Three need levels are identified:
· 'comfort' with ETM = kc ETref (Doorenbos and Pruitt, 1975),
· 'sufficiency' and 'shortage': the value of the crop coefficient kc is reduced in a non-uniform manner during the cycle and particularly at the least sensitive stages. The more moderate the yield objective, the greater the reduction.
The needs are calculated from the average ETref over ten years.
The means of supplying the plants depend on soil type and irrigation capacity.
For each soil type and each crop a soil contribution level is suggested to the user. This reserve will be partially consumed before the start of the irrigation season (MMD = maximum mobilizable before irrigation starts). The other part of the reserve will be used during the season to reduce peak requirements according to the rule described in Table 1 (deferred soil contribution). At present this rule does not depend on the soil type. In the future we will have to adapt it for various soil types.
TABLE 1 - Deferred soil contribution related to soil water deficit
Soil deficit in mm
Deferred soil contribution in mm/day (in addition to rainfall and irrigation)
Irrigation capacity is obtained weekly by multiplying equipment flow by duration of use within the limit of the available volume.
After describing his system (crops, areas, soil water capacity and irrigation means) the farmer makes a choice to establish a balance between requirements and needs:
· a rainfall level associated with its probability based on a frequency analysis over 25 years. Usually a dry year hypothesis is adopted, for example, occurrence of rainfall in two years out of ten,
· a water requirement chosen from the three types described: comfort, sufficiency, shortage,
· the strategy for using the soil water capacity: the user decides on the start level for irrigation (MMD) and IRRISA then presents a way of mobilizing the remaining water capacity to alleviate peak requirements.
If from these first hypotheses the means are insufficient to cover needs, the user can:
· accept the risk and in the future be more prudent in his choice of cropping plan,
· reduce the requirement by lowering the production objectives of certain crops.
In every case, after the test phase of the strategy, IRRISA produces a guide usable for the subsequent irrigation of each crop.
In the example given in Figure 5, the maximum deficit at the start is 50 mm (MMD) and the total soil contribution is 150 mm; 100 mm are therefore available during the season to complement rainfall and irrigation (deferred soil contribution).
FIGURE 5 - Example of guide for corn
During the irrigation campaign the guide is updated to include rainfall, ETref and irrigations carried out. The soil water state indicator 'MMD + rainfall + irrigation' (dotted curve) should remain between 'S kc.ETref + MMD' and 'S kc.ETref - soil contribution'.
In this way the user always knows the state of each crop and can decide on future irrigations in the light of forecast requirements.
USE OF THESE TOOLS AND CONCLUSION
LORA has been available in France since 1990. Since that time the principal modification has been to take into account the management of subsidies following changes in the Common Agricultural Policy.
Our first marketing strategy was to commercialize the tool via a computer program publisher. The limited market for this sort of decision-making product and the difficulty for the publisher in training potential users in the principles involved lead to the relative failure of this marketing strategy.
At present, LORA is distributed by ITCF to potential users by way of a contract including training on the tool and its principles.
LORA is used in three areas:
· Direct help in decision-making for farmers via their advisers.
· Study of irrigated systems on a regional basis: to analyse their diversity taking into account the different technical and economic contexts and to anticipate change due to modification of the economic and regulatory contexts.
· Training in water management at farm level for advisers.
The 'bilaneaumètre' method has been distributed for several years among farmers who purchase water from the SCP in south-eastern France (Tron, 1991-1995).
IRRISA, which was developed in 1995, was used by a group of advisers and farmers during the summer of 1995 (Boyer and Campagnaud, 1996). This software should be marketed in 1996 in 'farmer' and 'adviser' versions, the latter being designed to allow an adviser to manage a group of farmers.
IRMA is still at the research model stage. It was conceived and developed in the framework of a group of farmers operating in different crop, climate and water resource contexts. In a first step, it has been used to analyse the observed irrigation seasons on these farms. In a second step, we use it to answer farmers' questions about some crop irrigation strategies and irrigation programmes before the irrigation season (Leroy et al., 1996).
These three tools deal with the question of irrigation management at farm level. They give help to farmers and advisers in solving, at different levels, the question of the links between strategic decisions, tactical aspects and resources management: from choice of cropping plan to irrigation programmes and irrigation guides. We now have to think about advising methods, using these tools, that are adapted to the different context of irrigation of agricultural crops in France.
Balas, B. and Deumier, J.M. 1989. Gestion de l'eau au niveau de l'exploitation agricole. Colloque sur les Méthodes d'irrigation améliorées pour préserver et protéger les ressources en eau et augmenter le rendement des cultures. Nations Unies. 16 p.
Boyer, F. and Campagnaud, M. 1996. Synthèse de l'expérimentation sur le logiciel IRRISA. Chambre d'agriculture de Dordogne. 18 p.
Brochet, P. and Gerbier, N. 1975. L'évapotranspiration. Monographie de la Météorologie Nationale, no. 65, 95 p.
Doorenbos, J. and Pruitt, W.O. 1977. Crop water requirements. FAO Irrigation and Drainage Paper 24. FAO, Rome. 179 p.
Doorenbos, J. and Kassam, A.H. 1980. Réponse des rendements à l'eau. Bulletin FAO d'irrigation et de drainage no. 33. FAO, Rome. 235 p.
Jacquin, C., Deumier, J.M. and Leroy P. 1993. Lora et la gestion de l'eau dans l'exploitation agricole. Perspectives agricoles, no. 184, pp 73-82.
Leroy, P. and Jacquin, C. 1994. Un logiciel pour le choix d'assolement sur le périmètre irrigable d'une exploitation. 17ème Conférence Régionale Européenne sur les Irrigations et le Drainage ICID/CIID, Varna Bulgarie, 16-22 mai 1994. Vol 2, pp 61-72.
Leroy, P., Deumier, J.M. and Jacquin, C. 1996. Water management at farm level. Method and tool to support strategic decisions - In: Management of Limited Water Resources. Agro-economical Consequences. Final report of the EU contract no. 8001-CT91-0109.
Peyremorte, P. 1986. Simplification de l'estimation approximative des besoins d'arrosage: le «Bilaneaumètre». Eau et Aménagement de la Région provençale no. 43.
Peyremorte, P. and Tron, G. 1987. Stratégies de pilotage de l'irrigation, moyens d'aide à la décision. Colloque «Alimentation hydrique du maïs» AGPM-ITCF, Agen, December 1987.
Peyremorte, P. and Tron, G. 1989. Prédominance du contexte physique et humain pour le choix d'une stratégie de pilotage des irrigations en grandes cultures. Colloque international sur les méthodes améliorées d'irrigation Avignon 25-29 September 1989. 12 p.
Thévenet, G., Degoulet, A. and Deumier, J.M. 1987. Irrigation du maïs. Brochure AGPM-ITCF maïs no. 4, 40 pages.
Tron, G. 1991 to 1995. Comptes rendus annuels des suivis parcellaires de la Société du Canal de Provence.