Programme Poster session 1
Elaboration of a Learning and Testing Simulation Tool to improve water
The Crau Hay Cropping System
Author(s): A. Merot(1)(2), J-E Bergez(3), J-C Mailhol(4), C.
Isbérie(4), F. Charron(4), A. Capillon(1), J. Wery(1)
Merot A.(1,2), Bergez J-E.(3), Mailhol J-C.(4), Isbérie C.(4),
Charron F.(4), Capillon A.(1), Wery J.(1)
(1)UMR System, SupAgro, Montpellier, France.
Canal de Provence, Aix-en-Provence, France.
(3)UMR AGIR, INRA, Castanet-Tolosan, Franc
Keyword(s): Water management, hay cropping system, Learning and Testing
Simulation Tool, multifonctionality, border irrigation
Session: Poster session 1
to environmental impacts, conventional water management practices in the Crau plain must be modified in order to
ensure higher water productivity of the border irrigation system while keeping hay production and quality and lower
environmental impacts. Tools exist to analyse water management at different scales and for different management
questioning, but no simulation models have been developed to represent gravity irrigation management at cropping
system scale. In this paper, we present the method used to develop a computer-based Learning and Testing
Simulation Tool (LTST) at farm scale to evaluate the agronomic and environmental impacts of water management
practices and regulations in water distribution, in the context of multi-functionality of agriculture in the Crau territory.
The LTST integrates, in a simplified way, the complexity of the system considering the needs of farmers and
water managers as well as the functional relationships between water supply, hay production and water flows at field
level. It includes the following characteristics identified by experiments and farmers interviews: (i) border irrigation
with various durations of the irrigation events and various spatial orders for water distribution modified by the farmer;
(ii) multi-species grasslands highly sensitive to water deficit, (iii) strong interactions between irrigation and mowing. It
takes into account the calendars of water distribution and the water flow. It is based on existing knowledge, adapted
models and on new modules built from experiments and regional surveys data.
The LTST includes a dynamic
decisional model, a dynamic crop model, a hydraulic model and a graphic user interface. The decision model takes
into account the main constraints regarding irrigation and mowing, the decision sequences describing irrigation
planning and the interactions between irrigation and mowing. The dynamic crop model allows simulating plant and
soil functioning in relationship with water supply. The hydraulic model reproduces an irrigation event at border scale
depending on the biophysical and water distribution context. Finally the graphic user interface allows visualizing the
spatial organisation of the farm and its evolution during a simulation, particularly the spatial orders in which fields are
mown and borders are irrigated. The spatial heterogeneity is also included. In the LTST, there are 4 time scales
(year, irrigation season, growth period and day) and 4 spatial scales (irrigated area, irrigated sub-area corresponding
to a given water resource, field and border). The outputs of the LTST are environmental (water supply, drainage),
social (labour) and agronomic indicators (yields, water productivity and irrigation efficiency).
The structure of the
LTST was validated by the target users: water manager and agricultural advisors. This tool aimed at allowing them to
conduct strategic thinking on new irrigation schedules, new design of water channels or fields and new distribution
planning for a given water resource.