Estimation of nutrient loadings to the Fleet lagoon from diffuse sources

Background and methodology

Summary of results

Further work

Background and methodology

The objective for this investigation carried out by WRc for the Environment Agency was to >estimate nutrient loadings from diffuse sources to the Fleet lagoon, including those arising from the swannery= (Mainstone & Parr 1999).

The topographical catchment of the Fleet is small (28 km2), stretching along the length of the Fleet. Much of the catchment is under pasture, with sheep and dairy farming being important land uses. Arable farming is concentrated around the village of Langton Herring and in scattered areas largely to the south and east. At Abbotsbury Swannery, at the extreme western end of the Fleet, mute swans have been managed for around 600 years. This history of management has led to the development of a large breeding colony (around 150 pairs) in the 50 acre reed bed of the swannery, accompanied by a large flock (approximately 350) of non-breeding birds. In winter, numbers increase to around 1500 as birds from other areas fly in.

Freshwater inputs enter the Fleet from seven small streams, with further inputs from direct runoff along the lagoon shore. Groundwater seepage may also be a substantial though uncertain input given the presence of chalk and greensand aquifers (water bearing rock strata) underlying the whole area.

Any estimate of diffuse loads produced from export coefficients needs to be calibrated in some way against measured loads in the receiving waters. This requires that point source loads are also estimated and the sum of point and non-point source estimates are compared with observed loads entering the lagoon. Export coefficients may then need to be adjusted to provide a better fit with measured loads. Five stages to producing the nutrient budget may therefore be identified (Mainstone and Parr 1999):

  • estimating point source nutrient loads;
  • estimating diffuse nutrient loads;
  • estimating nutrient loads in waters entering the Fleet;
  • comparing nutrient loads estimated by export coefficients with loads estimated from receiving water monitoring;
  • modifying export coefficients if necessary.

Annual nutrient budgets were constructed for both nitrogen and phosphorus. No assessment of the seasonality or bioavailability of the nutrient loads was made during this study (see table below). MAFF agricultural census information on crop areas and livestock numbers for three parishes within the catchment area of the Fleet was used, and data on bird (particularly swan) populations and estimated loads were incorporated. The above estimates for diffuse sources of nutrients (combined with estimates for point sources) were calibrated against measured loads for the Fleet.

Point sources:

Estimated loads for point sources (consisting of two public sewage treatment works at Abbotsbury and Langton Herring, and two private sewage treatment works at Moonfleet Manor Hotel and Royal Engineers Training Camp at Chickerell) were necessarily crude, due to a lack of sufficiently detailed information on flows and nutrient concentrations of sewage effluent from these works. The results indicated that the loads from the two private works were negligible, and that Abbotsbury sewage treatment works contributes a much larger load than that from Langton Herring (1.19-2.90 tonnes N per year and 0.38-1.19 tonnes P per year at Abbotsbury, compared to 0.35 and 0.10 per year respectively for Langton Herring).

Diffuse source:

Loads estimated were from atmospheric deposition, agriculture, bird populations and groundwater. As with the estimations of loads from point sources, there were considerable sources of error for each of the estimates from diffuse sources.

Atmospheric deposition:

Background nutrient loads from atmospheric deposition were estimated at 18.48 tonnes per year for nitrogen, and 0.42 tonnes per year for phosphorus (the latter is based on the whole catchment).

There are two possible methods for calculating the nitrogen loading using the information available, one assumes a fixed proportion of modelled atmospheric deposition is exported to surface waters whilst the other uses the proportion of rainfall which is lost to surface runoff. Partly because of the complex geology of the catchment, the two methods give estimates that are an order of magnitude apart. Therefore, a mean of the two values obtained has been used.


Data on agricultural land is typically only available for aggregated groups of parishes. In the case of the Fleet, the relevant aggregation covered a much larger area than the topographical catchment. Fortunately, three of the parishes constituted 80% of the catchment area of the Fleet, and MAFF was able to separate out data for these three parishes, although data for some crops and livestock was not provided on this basis due to commercial confidence constraints on its use. The catchment is predominantly under pasture (70%), supporting sheep and dairy farming. 24% of farm holdings were devoted to arable, with most of this under cereal crops (predominantly wheat, with significant amounts of winter barley and maize). Data on pig and poultry numbers were withheld. Estimations of nutrient inputs from the various types of agriculture were made using assumptions about fertiliser use for different crops and rates of loss to receiving waters. The calculations produced estimates of 64 tonnes of N and 0.8 tonnes of P per year exported to the Fleet from agricultural fertiliser use and 44 tonnes N and 1.5 tonnes P derived from livestock.

Bird populations:

Colonial breeding in mute swans as occurs at Abbotsbury is traditionally accompanied by high juvenile mortality, as families leave the nest early, and cygnets fail to >imprint= properly on the parents. To assist successful imprinting, swans are fed at the nest with grass clippings (from the swannery lawns) in floating dustbin lids. The non-breeding flock represents a threat to the breeding colony, as if these birds were to descend on the breeding colony, it is likely that the ensuing havoc would result in high juvenile mortalities. The flock is therefore distracted during nesting time by supply of wheat grain away from the reed bed. Wheat is delivered directly to the water and ingested by the birds off the lagoon bed. Feed is not normally supplied to the overwintering flock, except in exceptionally cold weather.

The inputs of nutrients to the Fleet from the swans were estimated by using the nutrient load delivered to the swans in feed (whole wheat, crumbs and pellets), and applying a conversion efficiency to estimate the loads generated in excreta. Figures of 0.25-0.29 tonnes N and 0.05-0.06 tonnes P per year from swan feeding were obtained by this method.

Loads from other bird species in the Fleet (Canada geese, Brent geese, and lesser black-backed, common, black headed and herring gulls) were estimated using monthly counts to obtain bird numbers, multiplied by figures for nutrients in faeces obtained from the literature. Inputs to the Fleet of 0.17 tonnes N and 0.06 tonnes P were obtained by this method for birds other than mute swans.


Figures for nutrient loads from groundwater could not be quantified due to lack of information on both nitrogen and phosphorus concentrations in groundwaters within the catchment. However, this source of nutrients is potentially significant and cannot be ignored, particularly for nitrogen as Environment Agency borehole data indicate that total inorganic nitrogen (TIN) concentrations from groundwaters to the north and west of the catchment are high (around 6 mg/l TIN).


Comparisons were made between the estimates of loadings obtained by the above methods with nutrient loadings measured for the Fleet. Unfortunately, there were no flow data for the seven small streams entering the Fleet, so flows had to be estimated using the Institute of Hydrology=s Micro Low Flow methodology. Nutrient data for the streams were also sparse due to the fragmentation of the riverine load to the Fleet into a number of minor watercourses which would not normally receive much monitoring attention. There was a discrepancy between the calculated loads for phosphorus from the various sources and the calculated loads in the streams, which could be due to a number of reasons, but is probably due to the low accuracy of the methods for estimating both the theoretical inputs and the measured inputs from the streams. Agreement between the two methods for nitrogen was good.

NB: Micro Low Flows is a piece of software developed by the Institute of Hydrology, particularly for catchments subject to artificial influences such as impoundment, discharges and abstraction, to allow the estimation of natural low flows at ungauged sites.

Summary of estimated annual nutrient loads to the Fleet








Point sources (sewage works)


1 - 2.5


12 B 39





37 B 47

Fertiliser application


49 B 50


18 B 23

Background load




10 B 13

Abbotsbury swannery





Other bird species






129.3 B 131.0



3.31 B 4.12



Source: Mainstone & Parr 1999

Summary of results:

Although there were concerns over the reliability of a number of aspects of the data on which the estimates were based, agricultural sources were found to be highly important for both nitrogen and phosphorus. Around 80% of the annual load of nitrogen, and over half, perhaps as much as 70%, of the annual total phosphorus load to the Fleet is estimated to come from agricultural sources. In addition, contributions from pigs and poultry were not included in estimations of inputs from livestock, which might make the agricultural load even more important to the annual budget. The two sewage works at Abbotsbury and Langton Herring may contribute as much as 40% of the annual phosphorus load to the Fleet, but this figure may be as low as 12%. Mute swans and feeding from the swannery and other bird species do not appear to be making a major contribution to nutrient loads entering the lagoon as a whole but they may be important in the Abbotsbury sub-catchment.

Further work:

More detailed modelling was recommended to gain a better understanding of the spatial and seasonal distribution of loads and the effects on water quality within the Fleet. The immediate identification and implementation of practical control measures across the catchment, using catchment walk overs to identify critical practices and run-off pathways, was also recommended. Further modelling will eventually help to focus attention on high risk areas and will help to predict the likely effect of control measures on nutrient status and eutrophication risk. Improved information on the loads entering the Fleet via point sources, feeder streams, direct run-off and groundwater seepage is also recommended. It should be borne in mind, however, that the above studies are approximate calculations of annual nutrient inputs. The timing and availability of loads from different sources may alter the ecological importance ascribed to different human activities.

Next section             References