the Riverstrahler model runs on an ensemble of connected objects, either basins,
grouping EBs according to an idealised scheme of confluence of tributaries with
average characteristics by stream order, or branches, with a detailed and exhaustive
representation of the morphology at 1 km spatial resolution. Reservoirs and ponds
are also taken into account with a description of their morphology (depth and surface
area) and hydrology (water inflow and outflow).
Hydrological inputs are provided by the HYDRO database (http://www.hydro.
eaufrance.fr), averaged at a daily time step and separated into surface runoff and
baseflow, using Eckhardt’s recursive filter [45]. Surface runoff and baseflow are
generated using observed discharge time series at, e.g. 50–100 gauging stations for
simulation periods ranging from 1 year [23, 46] to yearly contrasted hydrological
conditions [47, 48] or plurennial time-windows [22, 49, 50].
The Seine basin comprises about 1900 wastewater treatment plants, each being
georeferenced and characterised by the connected population and the type of treatment applied [51]. These data are provided by the Seine–Normandy database and are
transformed into variables compatible with the model (nutrients and carbon under
their specific forms [37, 52, 53]).
Diffuse sources enter the river through surface runoff and the baseflow. Mean
annual nutrients and carbon concentrations are associated with each component of
the water: N (nitrate, ammonium), P (total inorganic phosphorus, TIP), Si (dissolved
and biogenic), suspended solids and organic and inorganic carbon. These concentrations differ according to major land use classes (cropland, permanent grassland,
forests, urbanised areas) and lithological features, which are spatially determined
using the Corine Land Cover database [54] and lithological information [51]. For
nitrate, two approaches are available. The GRAFS approach [55, 56] calculates
leaching fluxes and concentrations [57], for both surface runoff and baseflow, for
each land use class, on the basis of regional agricultural statistics at the département
level. The other approach mobilises complex physically based models: the STICS
agronomic model and the MODCOU hydrogeological model [18, 58]. Inorganic
carbon was recently added to the modelling approach [34, 59], the aquifer’s lithology being a major controlling factor.
The Riverstrahler model calculates seasonal variations of water quality and
ecological functioning for any tributary of the river system given point source and
diffuse source forcings.
3 Long-Term Trends in Water Quality
3.1 Changes in Organic Pollution from Urban Point Sources
From about 11 million inhabitants in 1955, the population of the Seine basin had
increased to 16 million in 2018 [51], with an uneven geographical distribution, with
75% concentrated in the Paris metropolis and along the main Seine branch from
Paris to the estuary.
194
J. Garnier et al.
grouping EBs according to an idealised scheme of confluence of tributaries with
average characteristics by stream order, or branches, with a detailed and exhaustive
representation of the morphology at 1 km spatial resolution. Reservoirs and ponds
are also taken into account with a description of their morphology (depth and surface
area) and hydrology (water inflow and outflow).
Hydrological inputs are provided by the HYDRO database (http://www.hydro.
eaufrance.fr), averaged at a daily time step and separated into surface runoff and
baseflow, using Eckhardt’s recursive filter [45]. Surface runoff and baseflow are
generated using observed discharge time series at, e.g. 50–100 gauging stations for
simulation periods ranging from 1 year [23, 46] to yearly contrasted hydrological
conditions [47, 48] or plurennial time-windows [22, 49, 50].
The Seine basin comprises about 1900 wastewater treatment plants, each being
georeferenced and characterised by the connected population and the type of treatment applied [51]. These data are provided by the Seine–Normandy database and are
transformed into variables compatible with the model (nutrients and carbon under
their specific forms [37, 52, 53]).
Diffuse sources enter the river through surface runoff and the baseflow. Mean
annual nutrients and carbon concentrations are associated with each component of
the water: N (nitrate, ammonium), P (total inorganic phosphorus, TIP), Si (dissolved
and biogenic), suspended solids and organic and inorganic carbon. These concentrations differ according to major land use classes (cropland, permanent grassland,
forests, urbanised areas) and lithological features, which are spatially determined
using the Corine Land Cover database [54] and lithological information [51]. For
nitrate, two approaches are available. The GRAFS approach [55, 56] calculates
leaching fluxes and concentrations [57], for both surface runoff and baseflow, for
each land use class, on the basis of regional agricultural statistics at the département
level. The other approach mobilises complex physically based models: the STICS
agronomic model and the MODCOU hydrogeological model [18, 58]. Inorganic
carbon was recently added to the modelling approach [34, 59], the aquifer’s lithology being a major controlling factor.
The Riverstrahler model calculates seasonal variations of water quality and
ecological functioning for any tributary of the river system given point source and
diffuse source forcings.
3 Long-Term Trends in Water Quality
3.1 Changes in Organic Pollution from Urban Point Sources
From about 11 million inhabitants in 1955, the population of the Seine basin had
increased to 16 million in 2018 [51], with an uneven geographical distribution, with
75% concentrated in the Paris metropolis and along the main Seine branch from
Paris to the estuary.
194
J. Garnier et al.