David C.-L. Lam

Modeling as a Tool for Nutrient Management in Lake Erie: a Hydrodynamics Study

Abstract Coupled physical-biological numerical models are useful tools for understanding the relevant processes and the influence of biota and human activity on the ecological conditions in the lake, and such a suite of models has been used to assess the impact of zebra mussels on the nutrient cycling in the lake. This paper presents the hydrodynamic part of a Lake Erie modeling exercise using the 3D ELCOM model. Validation runs were performed with 1994, 2001, 2002, and 2003 data where vertical thermistor chain data are compared against model calculations and mean circulation patterns are presented for the different runs. The validated model was then used to understand the flushing of the deep water, the internal wave dynamic and the residual circulation. For example, the presence of two gyres in the west-central basin that entrain nutrient-rich western basin and Sandusky Bay water and are probably a key mechanism for retaining externally supplied nutrients in this region, contributing to variability of primary productivity and its spatial distribution in the central basin. External nutrient loads are transported eastward more quickly than would occur without gyres, and would support less extensive phytoplankton development in the west-central basin. The hydrodynamic results will eventually be used as the drivers for future simulations aimed at studying the fate and transport of nutrients.

Design and implementation of an environmental decision support system

Abstract An environmental decision support system is a specific version of an environmental information system that is designed to help decision makers, managers, and advisors locate relevant information and carry out optimal solutions to problems using special tools and knowledge. The RAISON (Regional Analysis by Intelligent Systems ON microcomputers) for Windows decision support system has been developed at the National Water Research Institute, Environment Canada, over the last 10 years. It integrates data, text, maps, satellite images, pictures, video and other knowledge input. A library of software functions and tools are available for selective extraction of spatial and temporal data that can be analysed using spatial algorithms, models, statistics, expert systems, neural networks, and other information technologies. The system is of a modular design which allows for flexibility in modification of the system to meet the demands of a wide range of applications. System design and practical experiences learned in the development of a decision support system for toxic chemicals in the Great Lakes of North America are discussed.

Multi-model integration in a decision support system: a technical user interface approach for watershed and lake management scenarios

Abstract Computer simulations using mathematical models provide useful tools to investigate different scenarios based on watershed management strategies and environmental conditions. They often require the combined knowledge of meteorological, hydrological, hydrodynamic and biochemical processes in air, soil and water. While existing models for individual processes are available, computational issues (e.g. software compatibility and consistency of model assumptions) on linking and integrating these models are challenging. To resolve these issues, we propose using a technical user interface approach based on expert system technologies that provide intelligent access to databases, models, scenarios and decision support output. As an example, we applied the multi-model integration approach to a watershed management study on Lake Seymour, BC, Canada, where sediment erosion due to precipitation events or forest fires may lead to concerns of high turbidity conditions in a reservoir.

Lake Erie Thermocline Model Results: Comparison with 1967–1982 Data and Relation to Anoxic Occurrences

A one-dimensional thermocline model is presented which is used to estimate daily vertical temperature distributions, thermal layer thicknesses, thermal interface depths, and vertical diffusion. Comparison of computed temperature with observed vertical profile data (1967–1982) for the three basins shows good agreement. The median relative error between observed and computed mean temperature for the central basin hypolimnion is approximately 5 percent. These computed temperature profiles are used to derive statistical distributions which can be useful for designing sampling frequencies in the Great Lakes Surveillance Program. In particular, analyses are presented on the frequency of occurrence of certain thermal structures particularly favorable for the development of anoxia. It is found that, for a hypolimnion thickness of less than 4-m depth and a turbulent diffusivity less than 1 cm2/s, there is a high likelihood of anoxia development in the central basin of Lake Erie.

Towards coupling a 3D hydrodynamic lake model with the Canadian Regional Climate Model: Simulation on Great Slave Lake

Recently, it has been recognized that large lakes exert considerable influence on regional climate systems and vice versa and that the Canadian Regional Climate Model (CRCM), which does not currently have a lake component, requires the development of a coupled lake sub-model. Prior to a full effort for this model development, however, studies are needed to select and assess the suitability of a lake hydrodynamic model in terms of its capability to couple with the CRCM. This paper evaluates the performance of the 3-dimensional hydrodynamic model ELCOM on Great Slave Lake, one of Canadas largest lakes in the northern climatic system. Model simulations showed dominant circulation patterns that can create relatively large spatial and temporal gradients in temperature. Simulated temperatures compared well with cross-lake temperature observations both at the surface and vertically. Sensitivity analysis was applied to determine the critical meteorological variables affecting simulations of temperature and surface heat fluxes. For example, a 10% increase in air temperature and solar radiation was found to result in a 3.1% and 8.3% increase in water surface temperature and 8.5% increase in latent heat flux. Knowledge of the model sensitivity is crucial for future research in which the hydrodynamic model coupled with the atmosphere will be forced from the CRCM output.

Oxygen Depletion in Lake Erie: Modeling the Physical, Chemical, and Biological Interactions, 1972 and 1979

Models of the thermal structure, transport, and oxygen uptake within the water column and sediment are combined to examine the interaction of physical and biochemical processes resulting in oxygen depletion in the central basin hypolimnion of Lake Erie. While the oxygen demand in water and sediment may vary slightly from year to year, the oxygen supply to the hypolimnion due to physical processes is quite variable. Detailed calculations show that in years with relatively calm weather (e.g., 1979), physical processes supply only about 10% of the amount needed to offset demand due to uptake in the water and sediment during stratified periods. As the hypolimnion also tends to be relatively shallow in such years, the available oxygen is often quickly depleted. In contrast, during windy years (e.g., 1972) physical sources can supply about 40% of the oxygen consumed and, since the hypolimnion tends to be thicker, the oxygen concentration seldom reaches the anoxic level. It is concluded that analysis with oxygen depletion as an indicator of Lake Eries recovery due to phosphorus control must consider variability in oxygen sources as a result of physical processes.

A Post-Audit Anaylsis of the NWRI Nine-Box Water Quality Model for Lake Erie

Abstract Computed results from the Lake Erie nine-box water quality model developed at the National Water Research Institute (NWRI) are compared to observed data for three different periods: calibration (1978), verification (1967–1977), and post audit (1979–1982). It is shown that the median mean relative errors for these three stages of model development are essentially the same, demonstrating the model validity. In particular, the post-audit analysis supports the hypothesis that the sediment oxygen demand rate is related to the phosphorus loading. However, this relationship can only be made evident following the accurate simulation of the dynamic changes in the thermal layers and diffusion processes. Since such processes are directly affected by weather episodes, the anoxic occurrences in the central basin hypolimnion are subjected more to short-term meteorological uncertainties than to the long-term effects of the phosphorus removal program. Post-audit results on previously proposed phosphorus control curves, through the use of a completely independent data set, also confirms this phenomenon. It is concluded that a family of control curves or a set of probability indicators incorporating both loading and weather influences are more meaningful than a single control curve that is commonly proposed.

Regional analysis of watershed acidification using the expert systems approach

Abstract This paper describes a workstation environment for the regional analysis by intelligent system on a microcomputer (RAISON) for studying watershed acidification and other environmental problems. Data is processed with a map-based query language and a spreadsheet paradigm. Spatial and temporal data retrieval is made possible for any chosen region and time frame, so that information on water, air and soil is combined in a number of watershed acidification models. A simple prototype expert system uses each of these models and determine which is the most appropriate according to a set of decision rules defined by the user. Based on the data from 53 watersheds in Southern Quebec, Canada, regionalization of the expert model to 91 watersheds is achieved by using a Monte Carlo procedure. A mean relative error of 19.1% is obtained by the Monte Carlo results when compared to observed data.

Hydrological modelling of acidified Canadian watersheds

Abstract Hydrology plays an important role in the acidification of watersheds in Canada by airborne pollutants such as the wet and dry deposition of sulfate and nitrate. The pathways of the anions through the soil layers, the contact times with soil matrix, the snow accumulation and melting sequences and the groundwater flows are all important hydrological factors that influence the acidification processes. A hydrogeochemical model has been developed (TMWAM) and applied to Turkey Lakes Watershed. Results are presented on the calibration and confirmation of the model using the data from these watersheds. Particulate emphasis is placed on the hydrological factors and their relationships to the varying degrees of acidification observed at these locations. Preliminary results of linking the hydrological and hydrogeochemical models are discussed in the context of model applications as part of the national acidification assessment efforts in Canada.

Assessment of the interaction between the Canadian Regional Climate Model and lake thermal-hydrodynamic models

This paper describes our preliminary assessment of the system requirements (data, interface and process) for implementation of terms describing lake effects in the Canadian Regional Climate Model. We demonstrate test results for one-dimensional (1-D) and three-dimensional (3-D) models for lake hydrodynamics that have been tested and prepared for interfacing with atmospheric circulation models. We discuss the use of a physical interface model, Canadian Land Surface Scheme (CLASS), that is under consideration for the air-water interface. Our preliminary assessment indicates that it is technically feasible to apply the combination of small shallow lake (slab) model, 1-dimensional vertical model and 3-dimensional circulation model (for very large lakes), together with the current land-air linkage used in regional climate modelling in Canada. Whether we can discern the necessary conditions for invoking each of the models to the lakes present in individual regional climate grid cells is still to be determined.