3.2.1 Atmospheric Deposition
To quantify the PAH atmospheric deposition on rural areas, the investigated area
was divided by discriminating between agricultural and forested land. Indeed,
specific deposition processes over vegetated areas have been previously reported
due to the forest filter effect [19, 45, 46], thereby justifying different approaches. To
quantify the fallout flux over agricultural areas, a large database gathering 92 samples
from two studies was available [11, 47]. A significant difference was measured
between the warm and cold periods (Wilcoxon-Mann-Whitney test, p < 0.0001).
This temporal variation was included in the flux estimation, as for F12. Thus, the
atmospheric deposit over agricultural land reached F10 ¼ 5,680 Æ 860 kg year
À1
(Σ15). Because of the representative database, this estimation was awarded a grade
of 3. To estimate the atmospheric deposition over forests, only the data of five
samples from one study were available [47]. The same calculation method as for
other fallout fluxes (F12 and F10) was used, although not enough data were available
to distinguish between the warm and cold period fallouts. Thus, the atmospheric
deposition over forested areas reached 1,280 Æ 240 kg year
À1 (Σ15). Due to the
number of samples, the spatial and temporal representativeness was limited, however. Consequently, the quality rate of this flux estimation was downgraded to 1. The
forest filter effect was quantified based on the following values reported in the
literature: average litter production of 390 Æ 50 g m
À2 year
À1 [45, 48] and average
leaf contaminations of 30 Æ 20 ng g
À1 in remote forests [49, 50]. In addition, data
from a previous study gathering 77 samples collected within the Seine River basin in
forests directly exposed to PAH anthropogenic sources, including roads, were used
[51]. Thus, the annual flux related to defoliation was F11b ¼ 380 Æ 130 kg year
À1
(Σ15). Since the remote forest contamination could not be supported by actual
measurement, the flux quantification was awarded a grade of 1.
3.2.2 Road Runoff Outside of Urban Areas
An additional runoff flux specific to the road network outside of urban areas
was estimated. The surface concerned was estimated from the IGN database based
on the road length and number of lanes provided. The road surface totalled 236 km
2
in the investigated area, and a runoff coefficient of 0.9 was assumed. Since no data
were available for rural road runoff, this volume was matched to the database of
PAH concentrations in urban runoff water (n ¼ 31), and the flux totalled
E14 ¼ 220 Æ 30 kg year
À1 . This flux reached E14 # ¼ 310 Æ 50 kg year
À1 when
considering runoff flux in urban areas ending up in permeable soils. However, this
database was representative of a rural environment, and this quantification was
awarded a grade of 0.
Mass Balance of PAHs at the Scale of the Seine River Basin
175
To quantify the PAH atmospheric deposition on rural areas, the investigated area
was divided by discriminating between agricultural and forested land. Indeed,
specific deposition processes over vegetated areas have been previously reported
due to the forest filter effect [19, 45, 46], thereby justifying different approaches. To
quantify the fallout flux over agricultural areas, a large database gathering 92 samples
from two studies was available [11, 47]. A significant difference was measured
between the warm and cold periods (Wilcoxon-Mann-Whitney test, p < 0.0001).
This temporal variation was included in the flux estimation, as for F12. Thus, the
atmospheric deposit over agricultural land reached F10 ¼ 5,680 Æ 860 kg year
À1
(Σ15). Because of the representative database, this estimation was awarded a grade
of 3. To estimate the atmospheric deposition over forests, only the data of five
samples from one study were available [47]. The same calculation method as for
other fallout fluxes (F12 and F10) was used, although not enough data were available
to distinguish between the warm and cold period fallouts. Thus, the atmospheric
deposition over forested areas reached 1,280 Æ 240 kg year
À1 (Σ15). Due to the
number of samples, the spatial and temporal representativeness was limited, however. Consequently, the quality rate of this flux estimation was downgraded to 1. The
forest filter effect was quantified based on the following values reported in the
literature: average litter production of 390 Æ 50 g m
À2 year
À1 [45, 48] and average
leaf contaminations of 30 Æ 20 ng g
À1 in remote forests [49, 50]. In addition, data
from a previous study gathering 77 samples collected within the Seine River basin in
forests directly exposed to PAH anthropogenic sources, including roads, were used
[51]. Thus, the annual flux related to defoliation was F11b ¼ 380 Æ 130 kg year
À1
(Σ15). Since the remote forest contamination could not be supported by actual
measurement, the flux quantification was awarded a grade of 1.
3.2.2 Road Runoff Outside of Urban Areas
An additional runoff flux specific to the road network outside of urban areas
was estimated. The surface concerned was estimated from the IGN database based
on the road length and number of lanes provided. The road surface totalled 236 km
2
in the investigated area, and a runoff coefficient of 0.9 was assumed. Since no data
were available for rural road runoff, this volume was matched to the database of
PAH concentrations in urban runoff water (n ¼ 31), and the flux totalled
E14 ¼ 220 Æ 30 kg year
À1 . This flux reached E14 # ¼ 310 Æ 50 kg year
À1 when
considering runoff flux in urban areas ending up in permeable soils. However, this
database was representative of a rural environment, and this quantification was
awarded a grade of 0.
Mass Balance of PAHs at the Scale of the Seine River Basin
175