Page 80 - Modelisation du devenir des pesticides...

Chapitre 3 Evaluation du modèle STICS version pesticide sur différents sites
expérimentaux
exceed 0.1µgl-1. In the European Union, France is the country that uses the lar-
gest amounts of pesticide agricultural production. Pesticide concentrations excee-
ding the environmental quality standards recommended by the European Council
(2008/105/EC) have already been reported in groundwater and water bodies in in-
tensive farming production areas such as the Seine River basin. 25% of the ground-
water gages measured concentrations exceeding 0.1µgL in the Seine area in 2009. 2
-1
The PIREN Seine research program (http ://www.sisyphe.jussieu.fr/internet/piren/),
has been monitoring the diffuse contaminations from agriculture within the Seine
river basin have been considered for the past 20 years. 3 A database describing
the agricultural practices in homogeneous districts of 100 km² has been created
at the Seine River Basin scale. 4 A modeling chain based on the linking the crop
model STICS with the hydrological model MODCOU has been established to
6
5
simulate the occurrence of nitrate in the hydrosystem at the scale of the Seine
basin. 7,8 The crop model STICS is a generic crop model ables to simulate an-
5
nual and perennial crops such as cereals, 9–11 vineyards, and grassland. 12,13 The
model has been compared to other agronomic models in the context of national
and international projects. 14,15 In addition, pesticides sources were quantified in
small areas of the Seine basin to estimate the contribution of urban areas, 16,17
atmospheric fallout 18 and agricultural areas to rivers. 19,20 The STICS model was
chosen to integrate the pesticide transfer in the STICS/MODCOU modeling chain
in order i) to be consistent with previous works on nitrogen diffuse pollution mo-
deling , ii) to take advantage of the availability of the soil, crops and agricultural
databases over the whole Seine River Basin and iii) to take advantage of the recent
developments in the crop model including the simulation of foliar disease dynamics
in order to better define pesticide applications . This approach is consistent with
21
previous implementation of pesticide fate modules in crop models such as DAISY,
EPIC or CERES. Pesticide fate and transport models are assumed to be re-
22 23 24
levant tools for pesticide environmental risk assessment and pesticide registration.
The formalisms to describe pesticide processes in soils, i.e. sorption and degrada-
tion are relatively similar in all the pesticide transfer models. The main differences
between models are usually related to water transfer, as capacity models such as
PELMO, 25,26 PRZM 27 and PLM 28 or Richards type models such as PEARL, 29,30
MACRO, 31,32 LEACHP, 33 Wave, 34 AGRIFLUX 35 and RZWQM. 36 Most phy-
sically based models included a dual porosity module in order to take preferential
flow into account. 37,38 The implementation of pesticide fate processes in the crop
model STICS were derived from the models cited above. However, the difficulty of
obtaining the input parameters with sufficient accuracy for model evaluation and
the model parameterization strategy 39 make using mechanistic models a tedious
task. The methodology for model parameterization and evaluation used for this
work, is derived from guidance established by the FOrum for the Co-ordination of

68