7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
8 Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
Abstract Modelling long-term pesticide transfer to rivers at the catchment scale is
still difficult due to a lack of knowledge of agricultural practices and poorly adapted
field observation. The Orgeval experimental catchment was first investigated to
validate a modelling approach. In addition to pesticide practices investigated over
20 years, directly collected from farmers, monthly integrated river samples were
analysed for 10 years. To explicitly integrate agricultural practices and crop rotation,
the STICS crop model was adapted to simulate pesticide transfer in soil. Annual load
simulations were compared to observed pesticide fluxes in rivers. To simulate the
contamination of groundwater, the STICS-Pest model was coupled to the MODCOU
hydrogeological model. The results are discussed at the subbasin scale in relation to
available data. To upscale the approach at the Seine River basin scale, other
strategies need to be developed.
Keywords eLTER, Modelling, Monitoring, Pesticide, Phytosanitary practices,
PIREN-Seine, STICS-Pest, Zone Atelier Seine
1 Introduction
River contamination by pesticides is currently observed in the Seine basin. This is
mainly due to long-term uses of such substances in agricultural areas. Still today,
atrazine and its metabolite deethylatrazine (DEA) are one of the most frequently
detected pesticides in the Seine basin, even though atrazine has been banned since
2003. In this context, we need to better understand and simulate pesticide fate in
watersheds consistent with long-term uses.
Numerous pesticide fate models are available. The main differences between
them are mostly related to water flow from simple water transfer such as PELMO [1]
and PRZM [2] to more physically oriented models such as PEARL [3], MACRO [4]
and RZWQM [5]. However, modelling pesticides requires understanding the relationship between land management practices and the dynamic of contaminants, and
only a few models are able to take into account specific agricultural practices [6].
Pesticide transfer models need to know when each active ingredient (AI) is applied
and the quantity applied. However, unlike the other data needed for implementing
models (climate, soil characteristics or land use), there is no database available on
phytosanitary practices that can directly feed the models [6–8]. In addition, there are
few attempts to synthesise these practices at the catchment scale. This deficiency led to
the development of new methods of acquiring and processing data [6, 7].
In the PIREN-Seine programme, studies were conducted to better define crop
rotations and landscape diversity in the Seine catchment area. Agricultural practices
were investigated, and the first models simulated nitrate contamination behaviour in
groundwater [9] and surface water [10]. New developments have been carried out to
142
H. Blanchoud et al.
8 Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
Abstract Modelling long-term pesticide transfer to rivers at the catchment scale is
still difficult due to a lack of knowledge of agricultural practices and poorly adapted
field observation. The Orgeval experimental catchment was first investigated to
validate a modelling approach. In addition to pesticide practices investigated over
20 years, directly collected from farmers, monthly integrated river samples were
analysed for 10 years. To explicitly integrate agricultural practices and crop rotation,
the STICS crop model was adapted to simulate pesticide transfer in soil. Annual load
simulations were compared to observed pesticide fluxes in rivers. To simulate the
contamination of groundwater, the STICS-Pest model was coupled to the MODCOU
hydrogeological model. The results are discussed at the subbasin scale in relation to
available data. To upscale the approach at the Seine River basin scale, other
strategies need to be developed.
Keywords eLTER, Modelling, Monitoring, Pesticide, Phytosanitary practices,
PIREN-Seine, STICS-Pest, Zone Atelier Seine
1 Introduction
River contamination by pesticides is currently observed in the Seine basin. This is
mainly due to long-term uses of such substances in agricultural areas. Still today,
atrazine and its metabolite deethylatrazine (DEA) are one of the most frequently
detected pesticides in the Seine basin, even though atrazine has been banned since
2003. In this context, we need to better understand and simulate pesticide fate in
watersheds consistent with long-term uses.
Numerous pesticide fate models are available. The main differences between
them are mostly related to water flow from simple water transfer such as PELMO [1]
and PRZM [2] to more physically oriented models such as PEARL [3], MACRO [4]
and RZWQM [5]. However, modelling pesticides requires understanding the relationship between land management practices and the dynamic of contaminants, and
only a few models are able to take into account specific agricultural practices [6].
Pesticide transfer models need to know when each active ingredient (AI) is applied
and the quantity applied. However, unlike the other data needed for implementing
models (climate, soil characteristics or land use), there is no database available on
phytosanitary practices that can directly feed the models [6–8]. In addition, there are
few attempts to synthesise these practices at the catchment scale. This deficiency led to
the development of new methods of acquiring and processing data [6, 7].
In the PIREN-Seine programme, studies were conducted to better define crop
rotations and landscape diversity in the Seine catchment area. Agricultural practices
were investigated, and the first models simulated nitrate contamination behaviour in
groundwater [9] and surface water [10]. New developments have been carried out to
142
H. Blanchoud et al.