A Water Quality Model for Shallow River-Lake Systems and its Application in River Basin Management David Kneis A dissertation submitted to the Faculty of Mathematics and Natural Sciences at the University of Potsdam, Germany for the degree of Doctor of Natural Sciences (Dr. rer. nat.) in Geoecology Submitted 23 April 2007 Defended 19 June 2007 Published July 2007 Referees Prof. Axel Bronstert University of Potsdam, Institute of Geoecology PD Dr. Dietrich Borchardt Kassel University, Center for Environmental Systems Research Prof. Jesús Carrera Institute of Earth Sciences ’Jaume Almera’, Barcelona This work is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 2.0 Germany License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-nd/2.0/de/ or send a letter to Creative Commons, 171 Second Street, Suite 300, San Francisco, California, 94105, USA. Elektronisch veröffentlicht auf dem Publikationsserver der Universität Potsdam: http://opus.kobv.de/ubp/volltexte/2007/1464/ urn:nbn:de:kobv:517-opus-14647 [http://nbn-resolving.de/urn:nbn:de:kobv:517-opus-14647] Contents Abstract 9 German abstract 11 1 Introduction 13 1.1 Context of the study .................................... 13 1.2 Focus of this thesis .................................... 14 1.3 Outline of the chapters .................................. 15 2 State of the art and intentions for model development 17 2.1 The scope of water quality modeling ........................... 17 2.2 A short overview on water quality modeling approaches ................. 18 2.2.1 Type of water body ................................ 18 2.2.2 Implementation of turnover processes ...................... 19 2.2.3 Spatial dimensionality .............................. 20 2.2.4 Simulation time scale ............................... 20 2.3 Objectives of model development ............................. 21 2.3.1 Expert recommendations ............................. 21 2.3.2 Specific targets .................................. 21 3 The water quality simulation tool TRAM 25 3.1 Representation of the river network ............................ 25 3.1.1 The reactor concept ................................ 25 3.1.2 Discretization rules ................................ 26 3.1.3 Network topology ................................ 27 3.2 Hydrodynamics ...................................... 27 3.3 Turnover processes .................................... 28 3.3.1 Terminology ................................... 28 3.3.2 Use of the process matrix for model presentation ................ 30 3.3.3 User-defined turnover processes in TRAM ................... 31 3.4 Handling of time series .................................. 35 3.5 Computation algorithms ................................. 35 3.5.1 Objectives of computations ............................ 35 3.5.2 Reactive transport in stirred tank reactors .................... 35 5 6 Contents 3.5.3 Reactive transport in plug-flow reactors ..................... 38 3.5.4 Routing algorithm ................................ 39 3.6 TRAM’s input data .................................... 40 3.6.1 General remarks on data input .......................... 40 3.6.2 Model units .................................... 40 3.6.3 Network description ............................... 41 3.6.4 Preprocessing of hydrodynamic information ................... 41 3.6.5 Definition of the turnover model ......................... 45 3.6.6 Values of constants and initial concentrations .................. 49 3.6.7 Time series data ................................. 49 3.6.8 Simulation control parameters .......................... 50 3.7 TRAM’s output ...................................... 51 3.8 Possible enhancements of the current model version ................... 52 3.9 The source code ...................................... 52 4 Model application to the Lower Havel River 55 4.1 Objectives of modeling .................................. 55 4.2 Introduction to the study site ............................... 56 4.2.1 Hydrology .................................... 56 4.2.2 The eutrophication problem ........................... 57 4.2.3 Water quality overview .............................. 61 4.3 Hydrodynamic submodel ................................. 67 4.3.1 Sources and preprocessing of geometry data ................... 67 4.3.2 Boundary conditions ............................... 68 4.3.3 Model calibration ................................. 71 4.3.4 Exchange of data between the hydrodynamic model and TRAM ........ 71 4.4 Water quality submodel .................................. 71 4.4.1 Simplified outline of the aquatic nutrient cycle ................. 73 4.4.2 Phosphorus storage in sediments and remobilization .............. 76 4.4.3 Representation of nutrient retention and remobilization in TRAM ....... 83 4.4.4 Parameter estimation and model performance .................. 91 4.5 Use of the calibrated model for system analysis ..................... 99 4.5.1 Significance of phosphorus release ........................ 99 4.5.2 Significance of nitrogen retention ........................ 100 4.5.3 Visualization of patterns ............................. 102 4.6 Management scenarios .................................. 102 4.6.1 Integration of catchment models and TRAM .................. 102 4.6.2 Runoff and nutrient emissions from non-modeled subcatchments ........ 105 4.6.3 Description of the scenarios ........................... 106 4.6.4 Simulation results ................................ 108 4.6.5 Uncertainty of predictions ............................ 112 Contents 7 5 Summary and conclusions 123 5.1 Model design ....................................... 123 5.2 Implications for water quality management of the Havel River ............. 124 5.2.1 Significance of internal nutrient turnover .................... 124 5.2.2 Implications of simulated medium-term scenarios ................ 124 5.3 Challenges for future research .............................. 126 5.3.1 Catchment modeling ............................... 126 5.3.2 Classification of the ecological status ...................... 126 5.3.3 Simulation of nutrient turnover in rivers and lakes ................ 126 5.4 Final remarks ....................................... 128 List of figures 130 List of tables 135 Bibliography 136 Acknowledgments 147 8 Contents 9 Abstract port the implementation of the WFD in the lowland catchment of the Havel River located in North-East This work documents the development and appli- Germany. cation of a new model for simulating mass trans- Within the above-mentioned study TRAM was port and turnover in rivers and shallow lakes. applied with two goals in mind. In a first step, The simulation tool called ’TRAM’ is intended the model was used for identifying the magnitude to complement mesoscale eco-hydrological catch- as well as spatial and temporal patterns of nitro- ment models in studies on river basin management. gen retention and sediment phosphorus release in TRAM aims at describing the water quality of in- a 100 km stretch of the highly eutrophic Lower dividual water bodies, using problem- and scale- Havel River. From the system analysis, strongly adequate approaches for representing their hydro- simplified conceptual approaches for modeling N- logical and ecological characteristics. The need for retention and P-remobilization in the studied river- such flexible water quality analysis and prediction lake system were obtained. tools is expected to further increase during the im- In a second step, the impact of reduced external plementation of the European Water Framework nutrient loading on the nitrogen and phosphorus Directive (WFD) as well as in the context of cli- concentrations of the Havel River was simulated mate change research. (scenario analysis) taking into account internal re- The developed simulation tool consists of a tention/release. The boundary conditions for the transport and a reaction module with the latter be- scenario analysis such as runoff and nutrient emis- ing highly flexible with respect to the description sions from river basins were computed by project of turnover processes in the aquatic environment. partners using the catchment models SWIM and Therefore, simulation approaches of different com- ArcEGMO-Urban. Based on the output of TRAM, plexity can easily be tested and model formulations the considered options of emission control could fi- can be chosen in consideration of the problem at nally be evaluated using a site-specific assessment hand, knowledge of process functioning, and data scale which is compatible with the requirements of availability. Consequently, TRAM is suitable for the WFD. Uncertainties in the model predictions both heavily simplified engineering applications as were also examined. well as scientific ecosystem studies involving a According to simulation results, the target of the large number of state variables, interactions, and WFD – with respect to total phosphorus concentra- boundary conditions. tions in the Lower Havel River – could be achieved TRAM can easily be linked to catchment mod- in the medium-term, if the full potential for re- els off-line and it requires the use of external hy- ducing point and non-point emissions was tapped. drodynamic simulation software. Parametrization Furthermore, model results suggest that internal of the model and visualization of simulation re- phosphorus loading will ease off noticeably until sults are facilitated by the use of geographical in- 2015 due to a declining pool of sedimentary mo- formation