Cet article s'intéresse à la dynamique du transfert du phosphore dans un bassin rural de la région lémanique, le bassin versant du Foron (51,5 km2). Nous montrons qu'en établissant, pour les périodes de tarissement, la relation concentration moyenne en phosphore - débit moyen à l'exutoire, il est possible d'évaluer globalement:
1) la rétention du P dans le réseau hydrographique lors de ces périodes;
2) la contribution moyenne des sources ponctuelles de P aux exportations du bassin.
La rétention est un phénomène temporaire, mais c'est une étape clé de la dynamique du phosphore. Elle crée d'importants stocks de P dans le réseau hydrographique; ces stocks se situent pour l'essentiel au niveau des sédiments comme l'ont montré diverses études antérieures et sont exportés lors des crues. Les quantités stockées à l'issue de longs étiages dépassent souvent largement les exportations enregistrées pendant les plus fortes crues. La rétention s'accompagne d'une transformation partiellement irréversible de la spéciation du phosphore transféré.
Ces phénomènes sont pris en compte pour établir des bilans annuels de pollution, et notamment la balance diffus - ponctuel.
- bassin versant,
- pollution ponctuelle,
- pollution diffuse
Phosphorus mass balance and the evaluation of the origin of phosphorus outputs in a sub-watershed of lake Léman
The effective management of phosphorus in watersheds requires knowledge of the origin and magnitude of phosphorus inputs. However, for large and complex watersheds, it is impossible to determine the phosphorus mass balances by measuring the flow from each source and the rate of transfer of phosphorus originating from upstream. Because of these difficulties, we have developed a strategy based on the measurement of phosphorus mass balances during base flow periods (with no storm-flow, which means with a constant or decreasing flow). This approach was applied to the Foron River watershed (drainage area 51.5 km2, annual average water discharge 0.5 m3 /s).
The Foron River, a hillside stream with a torrential flow regime, is a tributary of Lac Léman (a large meso-eutrophic lake also known as Lake Geneva). Land use within the Foron watershed is diverse, with 50% of the watershed area being in a naturally forested area (on steep upland slopes), 36% in agricultural land lying on a gentle relief of moraine deposits (70% permanent pasture and hay meadows, 30% cereal grain crops) and 3% in marshes. The population is scattered amongst four small villages, with an overall density of about 100 inhabitants per km2. Stream flow in the river was continuously monitored and automatically sampled at the outlet from 1990 to 1993. Additional grab samples were taken at various locations, springs, agricultural sub-watersheds and brooks in the naturally forested area. Total suspended solids, soluble total phosphorus, orthophosphate and total phosphorus were determined according to standard methods. In addition, sampling was carried out to determine phosphorus stored in river sediments in places that are known to be temporary sinks for sediments.
To validate the efficiency of the approach, we carried out a detailed study of the point sources and the sewage systems in the watershed. Only two-thirds of domestic wastewaters were processed in treatment plants. The minimum load of point source inputs was established to be about 60 kg total P per week and many point sources (that is to say discharged at discrete locations) remained unidentified (i.e., not all households were connected to sewers, farmsteads contributed for unknown quantities of total P).
During base flow periods, phosphorus was predominantly soluble and orthophosphate was found throughout the hydrographic network. Particulate phosphorus and suspended matter concentrations were negligible (respectively < 20 µg/l and <30 mg/l), were not related to flow discharge and originated mainly from point source discharges. The inputs from surface runoff on the watershed were also negligible, as reflected in the low concentrations of total phosphorus in the agricultural or forested sub-watersheds (less than 15 µg/L).
Total inputs of phosphorus into the river included not only this background phosphorus (evaluated to be less than 2 kg total-P per week), but also P mainly originating from point sources. Total phosphorus exported from the outlet varied from 20 to 110 kg total-P per week but was often much lower (up to 80%) than the minimum inputs due to point sources. During these periods phosphorus can accumulate in the hydrographic network, probably in the riverbed sediments and periphyton (macrophytes were not present).
The detailed pattern of phosphorus exports during these low flow periods showed that the amount of total phosphorus exported (90% soluble P) tended to increase with the weekly mean water flow, up to a water discharge rate of 0.8 m3 /s. Above this value, phosphorus export reached a plateau corresponding to a mean constant value (about 80 kg per week). Since weekly average inputs from point sources remained nearly constant, independent of stream flow, this increase in P export indicated an increase in the rate of transfer of total phosphorus from point sources discharged into the river. When the water discharge was > 0.8 m3 /s, a constant P export regime was reached as all the inputs were transported to the outlet. Consequently, output at the outlet was equal to input into the river. This approach permits the calculation of the total point sources. Knowing the inputs from point sources and the total P export at the outlet, we can calculate the diffuse sources.
Significant relationships between weekly mean total-P, soluble-P or orthophosphate concentrations at the outlet and weekly mean discharge were established to validate these observations. The best fit of these measured values didn't follow a standard law of dilution because the dilution effect was attenuated by the antagonist effect due to the increase in P transport inputs from sewage with increasing discharge. ([soluble P]=0.02+ 0.148 Q-1/2 ; n=82; r2 =0.8).
During base flow when Q < 0.8 m3 /s, a portion of phosphorus inputs accumulated in the river biota and sediments through deposition, biological uptake and sorption. This accumulation may explain the higher levels of total-P and the easily exchangeable P in the sediments downstream of point source discharges. The efficiency of these storage processes depends not only on water discharge, but also (secondarily) on some characteristics of the antecedent hydrological conditions. We found that the amount stored increased significantly in the weeks following large storm-flows. We assumed that this indicated the effect of an intense exchange of the P loaded sediments with unsaturated sediments that were generated by erosion of natural upstream areas. As in-stream processes alter phosphorus speciation, a part of soluble-P becomes particulate. During storm-flows following base flow periods with a discharge lower than the limit discharge, the stored phosphorus was released from both particulate and soluble forms and exported. We calculated an annual mass balance of soluble phosphorus to evaluate the amount of particulate phosphorus exported at the outlet due to the transformation of phosphorus from soluble to particulate forms. The result obtained suggested that ¼ of particulate-P (1 t per year) resulted from this speciation change.
To summarize, analyses of phosphorus mass balances during base flow could lead to a new procedure to evaluate the mean total mass of P originating from point sources, without referring to field surveys. In addition, this procedure allowed the quantification of both total-P storage as well as changes in speciation that occur during transfer of phosphorus in the hydrographic network.
- point pollution,
- diffuse pollution