The fluxes and transformations of suspended particles, carbon and nitrogen in the Humber estuarine system (UK) from 1994 to 1996: results from an integrated observation and modelling study.

Results from the first multi-constituent dynamic modelling study of the Humber estuarine system over seasonal time scales are presented. The model, constructed within the ECoS3 simulation software, has been calibrated using data obtained from 27 surveys of the Yorkshire Ouse and Humber estuaries during 1994-1996. The model is tidally averaged, and is one-dimensional in the vertical and in cross-section. Daily-varying longitudinal concentration profiles of SPM, POC, nitrate, nitrite and ammonium have been simulated, and these have been compared with the observations used for model calibration (model confirmation). The fits are generally good. The model captures the marked seasonal variability in concentrations of the particulate constituents (SPM, POC), and places the turbidity maximum in approximately the correct region of the estuary. There was a high degree of scatter in the SPM and POC concentrations measured due to tidal resuspension, which was not reproduced by the tidally averaged model. Comparisons between measured and simulated nitrate+nitrite are satisfactory, although for a number of surveys the modelled values are systematically too low in the lower Yorkshire Ouse and upper Humber. These discrepancies are not readily explained, but may be linked to inadequately characterised source terms for this nitrogen. The general trends in the concentrations and distributions of ammonium are reproduced by the model, despite the plethora of external inputs of this constituent, and significant modifications to ammonium transport by in situ nitrification and benthic exchange. The simulated concentrations and distributions have also been tested against independent data encapsulating axial transects in the Trent and Humber estuaries, and temporal variations at fixed sites on the Trent and Yorkshire Ouse estuaries. With some exceptions, the fits between the model results and these data are also good. The model has also been used to construct constituent budgets (external/internal sources and sinks) for each estuary for 1994, 1995 and 1996. Riverine inputs of SPM, POC, nitrate, nitrite and ammonium decreased over the 3 years, presumably reflecting changing patterns of terrestrial runoff coupled to a reduction in effluent inputs. This trend was also observed in the point discharges of these constituents direct to the estuary. The simulated estuary was a source of sediments to the North Sea during 1994 and 1996, but a sink during 1995. The difference has been ascribed to the exceptionally dry summer and autumn of 1995, in which up-estuary tidal 'pumping' of SPM was dominant over down-estuary flushing by river water. Fluvial inputs represent an important external source of POC to the model estuary (27-55%), as does the North Sea (26-58%). These inputs are exceeded by POC loss via bacterial remineralisation, suggesting that the estuarine bed is a net source of POC. Riverine inputs account for 76-77% of the external inputs of nitrate. This constituent behaves quasi-conservatively in the estuary, reflecting the dilution of nitrate-rich river waters with low-nitrate coastal water, although nitrification is a significant input (approx. 21%). Denitrification is small, and consequently all nitrate added to the estuary is lost to the North Sea. External loads of nitrite are small and ammonium nitrification accounts for approximately 93% of its annual input. Practically all of the nitrite is oxidised to nitrate, and little escapes to the North Sea. The fluvial and direct waste inputs of ammonium to the estuary are of similar size. Most of the ammonium in the model is nitrified (98-100%) and only a few tons are exported to the North Sea. In 1996 the estuary appears to have been a sink for coastal water ammonium. Constituent budgets are compared with independently evaluated fluxes. There is good agreement in a number of cases, but the comparison highlights the main uncertainties in the simulated fluxes. It is concluded that the model as currently formulated provides an excellent basis for heuristic studies of the Humber estuarine system.