Background

 

River basin management consists by definition of interventions targeted to improve the ecological status of a very dynamic and complex river ecosystem. Predicting and assessing the effects of management measures in the context of highly variable catchment characteristics is a major challenge in the decision process. When confronted with complexity, managers and scientists often seek salvation in mathematical process models, which should picture the reality as accurately as possible and, not negligible either, save a lot of money as compared to more empirical approaches. With increasing computer power, the initial euphoria for catchment models recently ceded to a more critical use of distributed hydrological and attached quality models (Silberstein 2006).

Several EU-funded FP5 projects (HarmoniRiB, Harmoni-CA) addressed quality issues when using models to support river basin management. These range from structural uncertainty originating from the variety model algorithms employed (Refsgaard et al. 2006) to the assessment of input parameter variability, with a special focus on uncertainty of spatial data (Refsgaard et al. 2005a). The afore-mentioned scientific projects developed procedures to deal with and analyze model output uncertainty in view of a more thoughtful application of models in the water management practice (Refsgaard et al. 2005b, 2007).

Integrated modeling of river basins covers both inputs to the rivers (emissions) that are supplied by loading models and processes in the river channels themselves that are treated with water quality models. An analysis of uncertainty sources in river basin quality models identified the transport model and emission data as the main causes of error (De Blois et al. 2003). A wide array of models simulating suspended matter and nutrient emission in catchments is available with different levels of complexity and data requirements (Kalin & Hantush 2003).While being able to provide (upon calibration) reasonable integrated catchment yields at specific gages, these models are, despite of their distributed character, not designed to identify and unravel the sources of the sediments and adsorbed nutrients (Collins & Walling 2004). Furthermore, sediment-loading models are trimmed on theoretical consistence but neglect empirical testing and integration of management practice evaluations (Kalin & Hantush 2003). When compared to empirical estimation techniques, catchment wide loading models often lack accuracy and precision but require a lot more data input, which is often not available without additional monitoring efforts and subsequent calibration (Letcher et al. 2002). Loading models addressing urban stormwater emissions face very similar problems: simple regression models generally perform better than data intensive process-based models if event loads are required (Vaze & Chiew 2003). Finally, both types of models fail if no local calibration data is available.

River monitoring data, both for calibration and validation purposes, suffer from uncertainties that are less due to analytical errors than to sampling design. Especially particulate-phase transported substances have very site-specific temporal and location/scale-related variabilities that need to be representatively monitored (Rode & Suhr 2007). The variety of different weightings and methods to calculate substance flows cannot compensate for poor sampling design structure (De Vries & Klavers 1994). Long-time series of river mass loads are for these reasons often unable to display any significant trend (Littlewood & Marsh 2005). Water managers find themselves caught between surveillance monitoring that should document risks as well as long-term trends and operational monitoring for program of measures assessment, with little guidance on how to design economically viable sampling programs (Greig 2004). Statistical approaches to define optimal sampling frequency (and location) have been reported in the literature though (Coats et al. 2002, Moosmann et al. 2005): They rely on highly resolved measurement data sets that are sub-sampled with Monte-Carlo routines and analyzed for prediction uncertainty with rating curves or other calculation methods. The M3 project takes up recent developments in uncertainty assessment in river basin modeling and monitoring and will demonstrate the practical usefulness of the methods in catchment of different scale, structure and pollution.

Water quality models have traditionally focalized on eutrophication with different degrees of complexity (inclusion of benthic habitats, macrophytes). Their performance in terms of oxygen and nutrient concentration dynamics simulation is satisfactory in most situations encountered in the field (Cox 2003). Models linking eutrophication, pollutant fate and toxicological effect are much scarcer and their application in field studies remains limited due to their high data requirements (Koelmans et al 2001). Within the M3 projects the applied water quality models will therefore be mainly applied to impacts linked to eutrophication (SOBEK-WQ and DWA waterquality model) and pollutant exposure (metals and organic micro-pollutants with DWA Water Quality model).

References

 
DE BLOIS, C.J., WING, H :G ., DE KOK, J.L., KOPPESCHAAR, K. (2003) : Robustness of river basin water quality models. Journal of water resources planning and management 129(3), 189-199
 
COATS, R., LIU, F., GOLDMAN, R.J. (2002): A Monte Carlo test of load calculation methods, Lake Tahoe Basin, California-Nevada. Journal of the American Water Resources Association, 38(3), 719-730
 
COLLINS, A.L., WALLING, D.E. (2004): Documenting catchment suspended sources: problems, approaches and prospects. Progress in Physical Geography 28 (2), 159-196
 
COX, B.A. (2003): A review of currently available in-stream water-quality models and their applicability for simulating dissolved oxygen in lowland rivers. The Science of the Total Environment 314-316, 335-337
 
DE VRIES, A., KLAVERS, H.C. (1994): Riverine fluxes of pollutants: monitoring strategy first, calculation method second. - European Water Pollution Control, 4 (2), 12-17.
 
GREIG, S.M.(2004): Diffuse pollution and the Water Framework Directive: developing a practicable diffuse pollution monitoring strategy. www.sepa.org.uk/pdf/dpi/resources/ dpi_reports/dpi_and_wfd.pdf
 
KALIN, L., HANTUSH, M.M. (2003): Evaluation of sediment transport models and comparative application of two watershed models. EPA report EPA/600/R-03/139
 
KOELMANS, A.A., VAN DER HEIJDE, A., KNIJFF, L.M., AALDERINK, R.H. (2001): Integrated modeling of eutrophication and organic contaminant fate and effects in aquatic ecosystems. A review. Water Research 35 (18), 3517-3536.
 
LETCHER; R.A., JAKEMANN; A.J:, CALFAS, M:, LINFORTH, S:, BAGINSKA, B:, LAWRENCE, I. (2002): A comparison of catchment water quality models and direct estimation techniques. Environmental Modeling & Software 17, 77-85.
 
LITTLEWOOD, I.G., MARSH, T.J: (2005): Annual freshwater river mass loads from Great Britain, 1975-1994: estimation algorithm, database and monitoring network issues. Journal of Hydrology 304, 221-237
 
MOOSMANN; L., MULLER, B., GACHTER, R:, WUEST, A., BUTSCHER, E:, HERZOG, P. (2005): Trend oriented sampling strategy and estimation of soluble reactive phosphorus loads in streams. Water Resources Research, 41, W01020, doi:10.1029/2004/WR003539
 
REFSGAARD, J:C., et al. (2005a) : Harmonised techniques and representative river basin data for assessment and use of uncertainty information in integrated water management (HarmoniRiB). Environmental Science & Policy 8, 267-277.
 
REFSGAARD, JC., HENRIKSEN, H.J., HARRAR, W:G:; SCHOLTEN, H:, KASSAHUN, A. (2005b): Quality assurance in model based water management- review of existing practice and outline of new approaches. Environmental Modelling and Software 20, 1201-1215
 
REFSGAARD, J.C., VAN DER SLUIJS, J.P., BROWN, J:, VAN DER KEUR, P.(2006): A framework for dealing with uncertainty due to model structure error. Advances in Water Resources 29, 1586-1597
 
REFSGAARD, J.C:, VAN DER SLUIJS, J.P., HOJBERG, A.L:, VAN ROLLEGHEM, P.A: (2007): Uncertainty in the environmental modeling process- A framework and guidance. Environmental Modelling & Software 22, 1543-1556.
 
RODE, M., SUHR, U. (2007) : Uncertainties in selected river quality data. Hydrology and Earth System Sciences 11, 863-874
 
SILBERSTEIN, R:P. (2006): Hydrological models are so good, do we still need data? Environmental Modelling & Software 21, 1340-1352.
 
VAZE, J:, CHIEW, F:H.S. (2003): Comparative evaluation of urban storm water quality models. Water Resources Research, 39 (10), 1280, doi: 10.1029/2002WR001788,2003