Page 84 - Modelisation du devenir des pesticides...

Chapitre 3 Evaluation du modèle STICS version pesticide sur différents sites
expérimentaux
layer i, Mp deg(liq,ads) is the degraded mass of pesticide from liquid (liq) or linearly
sorbed (ads) phases, and DT50 is the half life (in day) of the pesticide.
The difference with the previous version of the pesticide module implemented
in the STICS model 20 is that, the new version includes the variation of Koc
with depths and the Arrhenius formulation for degradation. Moreover, the code
is able to manage as many pesticide and metabolites as required. In addition to
the classical STICS parameters, the pesticide module requires 6 parameters : the
pesticide linear sorption coefficient, the sorption and desorption coefficient for the
non-equilibrium phase, the half life, and the molar enthalpy and the solubility (to
determine the maximum mass of pesticide that water can contain).

3.2.3 Study sites
In order to assess the ability of the model to simulate the fate of various pes-
ticides in contrasted climate and soil, we used the data from three experimental
sites : Vredepeel in the Netherlands, and Kerlavic and Thiverval-Grignon in France.
The main characteristics of these sites are summarized in Table 3.1.
The three sites are also characterized by different experimental designs : deep
soil profiles, lysimeter and micro-lysimeter. Bromide was used on the three sites as
a conservative tracer. The Vredepeel experimental farm is located in the southeast
of the Netherlands. 55 The field consists of 16 plots, each divided into 27 subplots
3.3 m long and 3 m wide. Each subplot was sampled only once during the duration
of the experiment. One specificity of this site is the presence of a shallow aquifer.
The field was growing winter wheat during the two years of interest, with mustard
as a catch crop in between. bentazon, a selective contact post-emergence herbicide
and bromide were successively applied a few hours after sowing in winter. Soil
profiles were sampled several times over 16 months from one randomly selected
subplot. Soil moisture, bromide and bentazone contents were analyzed on samples
whith sizes ranging from 7.5 cm at the top to 30 cm in the deeper soil. The Kerlavic
experimental farm is located in western France. 56 The experiment was performed
on 1-m-deep lysimeter boxes, with 3 replicates. The lysimeter boxes were filled
with soil, the structureof which was preserved as far as possible. After one year
of stabilization, the three replicates were cropped with winter wheat for 2 years.
The isoproturon herbicide and bromide were successively sprayed a few days after
sowing. Monitoring was performed during autumn and winter, by collecting the
drainage flow at the bottom of the lysimeter box daily. Pesticide and bromide were
analyzed in the accumulated drainage flow every four days. The experimental site
of Thiverval-Grignon 57,58 is located close to Paris, France. The experiment was
performed on a 0.5 hectare field equipped with micro-lysimeters, composed of 20-
cm-diameter cylinder 35 cm deep. The cylinders were first driven into the soil and
the 5 cm at the soil bottom was replaced by a layer of activated carbon and a layer

72