Grantee Research Project Results
Final Report: Natural Convection in Stratified Lakes and ReservoirsImpacts on Pollutant Transport, Oxygen Budget, and Nutrient Dynamics
EPA Grant Number: R825428Title: Natural Convection in Stratified Lakes and ReservoirsImpacts on Pollutant Transport, Oxygen Budget, and Nutrient Dynamics
Investigators: Schladow, S. G.
Institution: University of California - Davis
EPA Project Officer: Aja, Hayley
Project Period: November 1, 1996 through October 31, 1999 (Extended to October 31, 2000)
Project Amount: $426,923
RFA: Exploratory Research - Water Engineering (1996) RFA Text | Recipients Lists
Research Category: Water , Land and Waste Management , Safer Chemicals
Objective:
This objective of the study was to investigate, using a combination of field, laboratory, and numerical techniques, the process of natural convection in lakes and reservoirs. Both vertical transport, as would occur when the surface of a lake was cooled, and lateral transport, as would occur, for example, when the margins of a lake heated or cooled differentially, were part of the study. Three specific issues that are controlled by natural convection were noted for particular attention. These were: (1) quantifying the oxygen transfer through the water surface due to natural convection; (2) quantifying the away-shore transport by differential cooling; and (3) quantifying the process of seasonal convection in deep lakes that only occasionally mix to their full depth.Summary/Accomplishments (Outputs/Outcomes):
Laboratory Experiments. The laboratory experiments were mainly directed toward the examination of oxygen transfer under pure natural convection. One key result has been the development of a new technique, fluorescent oxygen visualization (FOV) to quantify the amount of oxygen transferred from the air into water and to visualize the movement of the oxygen enriched water. The primary concept of FOV is based on the quenching of fluorescence intensity by oxygen and the imaging of a two-dimensional plane of the tank. A charge coupled-device (CCD) color video camera was oriented at a right angle to the light sheet. Two-dimensional light intensity images produced on the light sheet were recorded on the CCD camera (640 ? 480 pixels) with a 500 nanosecond exposure time.To simulate the air-water interface in a lake subject to pure natural convection, a rectangular tank was used. The Plexiglas walls and base were double-glazed to provide the necessary thermal insulation and a 0.5 cm thick aluminum plate formed the base of the chamber. Cold water pumped through the lid chamber provided a constant temperature upper boundary condition that simulated the conditions necessary for natural convection. Crushed ice and dry ice packed into the lid chamber also were used to extend the experimental range.
The spatial distribution of oxygen concentration at the air-water interface in the tank was visualized using the fluorescence imaging technique described above to quantify the oxygen transfer driven by only natural convection. Two-dimensional images were continuously acquired with an intensified charge coupled-device (ICCD) camera. A total of 2,160 images were processed for each 3-hour experiment. High-resolution profiles of temperature and electrical conductivity were used to define the temperature structure induced by the surface cooling.
The oxygen rich plumes that were evident in the images have their origin in the thermal boundary layer. The thickness of the thermal boundary layer was in the range of 2 to 4 mm. This layer thickened in localized areas as parcels of cold, oxygen-rich water accumulated before being shed and transferred downwards. This cycle of plume penetration was repeated many times during an experiment and the locations of these thickening regions were constantly changing. The oxygen transfer coefficients and oxygen fluxes produced by the natural convection were measured at four near-constant heat fluxes. For the range of temperature differences 7?C?24?C (the range of heat fluxes 403?901 W/m2), the oxygen transfer coefficients (KL) were in the range of 0.09?0.51 m/day and the total oxygen fluxes (F) were between 752?4,309 mg/m2/day, and increased with the increase of temperature difference between the air and water. For the range of heat fluxes normally experienced, natural convection may dominate oxygen transfer at wind speeds below 1.5 ms-1. The deep penetration of cooling water also was observed, and it was determined to be a significant oxygen source to the deeper parts of a lake.
The two-dimensional visualization of the flow using a fluorescence technique provided a phenomenological description of plumes under a cooling water surface and the mechanism of oxygen transfer under a physically realizable, ubiquitous set of natural conditions in the lakes. The transfer coefficient that has been derived for oxygen transfer under pure convection conditions (i.e., no wind) can readily be incorporated into existing air-water interface models, and more realistic results can be expected for low wind environments.
Field Experiments. The field experimental sites were very different, selected because they each presented different aspects of convective processes. Clear Lake is very shallow (mean depth 8.1 m), with gently sloping sides. By contrast, Lake Tahoe was deep (mean depth 340 m), with relatively steeply sloping sides. It was, therefore, expected that Clear Lake would be dominated by differential heating and cooling (promoting lateral intrusions), whereas Lake Tahoe would only display one dimensional, vertical convection. One of the surprise findings of the study was that differential cooling effects and lateral motions, albeit on a seasonal scale, were far more dominant at Lake Tahoe and were, in fact, the principal determinants of whether the lake mixed totally in a particular year.
Clear Lake. One part of the field work was to identify and quantify near-shore hydrodynamics in the Upper Arm of Clear Lake. In particular, it was focused on natural convection processes due to differential heating and cooling, and their overall significance towards transport of littoral material to the deeper, far-shore areas. The other part of the Clear Lake field work was concerned with indirect convective effects, produced by the baroclinic response to wind forcing, and relates to the overall, basin-scale transport of contaminants by convective processes. This work was centered in the Oaks Arm of Clear Lake.
Convective currents were observed during all seasons studied and were limited to less than 1,500 m from the shore. The potential for convective currents due to differential cooling became greater later in the summer and in the fall, as was seen during late August through October 1997. Convective currents due to differential heating and cooling were observed during 2 consecutive days in the Rancheria embayment where the study was based. The differential heating flow had a velocity of 2?3.5 cm/s, whereas the differential cooling velocity was about 2 cm/s. Taking the duration of the differential cooling event into account, when colder near-shore water slid under warmer water and was replaced by the warmer surface water from farther offshore, one can show that the entire water of the embayment was potentially transported offshore during this time and replaced with offshore surface water. This has important implications for the transport of contaminants out of the littoral zone of a lake.
The dynamics of the Oaks Arm, was found to be characterized by a marked diurnal periodicity dictated by the wind regime. The wind acts during the afternoon and evening hours on this weakly stratified system to generate horizontal temperature gradients both along and across the longitudinal axis of the Oaks Arm. During the night and early morning hours, when the wind forcing is negligible, the baroclinic pressure gradients that result from the horizontal differences in temperature become the dominant forcing mechanism in the system, driving currents of up to 10?15 cm/s westward at the surface and eastward near the bottom. The setup and relaxation processes are modulated by the influence of the earth's rotation and create a residual circulation, referred to as baroclinic pumping.
Our observations reveal that Clear Lake is an active system that changes dramatically in response to the diurnal wind events. Thus, observations gathered at a given moment in time may not be representative of the state of the lake 12 hours later. Flows in the lake also are spatially variable; in general, the conditions at a given sampling station will not be representative of existing conditions only 1,000 meters away. Basing any interpretation of field experiments on partial observations could lead to incorrect answers. The use of a few velocity transects to characterize the lake circulation, for example, may be very misleading, as there exists a notable bias to obtain observations during daylight hours and calm periods. It is during those hours when the lake adjusts from a type of circulation driven mainly by wind to another mainly driven by convective processes.
Lake Tahoe. During the cooling part of the year (fall, winter, and spring), heat and momentum transfer at the lake surface and the temperature-density relation of water combine to destabilize the water column and drive vertical mixing and transport processes. The deepening of the surface layer that this process produces, transfers nutrients from the hypolimnion into the euphotic zone. It also may resuspend sediments that would have settled under stratified conditions, or redistribute particles that may still be in suspension.
Thermistor chains were deployed in Lake Tahoe during the four winters from 1996 through 1999. Vertical destratification and mixing reached maximum depths of 190 m, 260 m, 385 m, and at least 464 m in those years, respectively. Deep mixing at the lake bottom occurred during both the 1998 and 1999 winters, but an additional mechanism such as horizontal convection may have been active (see below). An energetic internal wave spectrum was observed in the lake. Several seiches with excursions over 200 m were observed. The one-dimensional lake model (DLM) was used to simulate mixed layer deepening for 2 years. The model identified convective cooling (penetrative convection) as the dominant mechanism driving vertical mixing and deepening of the surface layer. However, the deepest vertical mixing resulted from intense storm events and dynamic responses within the lake.
Insights into the precise causes of interannual variations in depth of mixing at Lake Tahoe came with the installation of a third thermistor chain on the north side of Lake Tahoe, on a relatively shallow (100 m plateau). Thermistor chain data show that mixing of the deepest water in Lake Tahoe was caused by horizontal exchange processes, rather than vertical mixing processes during the 1998 and 1999 winters. In February 1998 and 1999, mid-lake bottom temperatures dropped suddenly while the overlying water column remained slightly stratified, and the lake was subsequently observed to "fill" with colder water from the bottom up. As a result, nutrient levels in the water column approached homogeneity, but an actual vertical overturn did not occur. Horizontal temperature gradients existed between shallow and deeper waters. Cold fronts, most likely gravity currents, were observed passing from near-shore areas toward the center and were related to heat flux minima. On a volumetric basis, streamflow could not have accounted for the observations.
Therefore, it is proposed that deep mixing was caused by horizontal convection due to differential cooling. The proposed mechanism is similar to horizontal convection caused by diurnal differential heating and cooling, but the time-scale for cooling is now the entire fall and winter period. This process, whereby the hypolimnion is recharged by differential cooling and horizontal convection of surface water on a seasonal basis, opens a number of important issues of concern for the management of Lake Tahoe and other deep lakes that may be subject to similar conditions. These include the interpretation of conventional estimates of sedimentation flux (as upward vertical circulation may be induced away from lake margins), the potential role of gravity currents in resuspending sediments, and the possibly beneficial effect of nutrients and other contaminants from the littoral zone being advected into the hypolimnion below the depth of light penetration. It also suggests, and this is the subject of continuing studies, that interannual variations in deep mixing are tied to the occurrence of conditions that drive the differential cooling process.
Numerical Simulations. A three-dimensional numerical model, for describing water motions in lakes and reservoirs, has been developed and tested at both of the field sites, Clear Lake and Lake Tahoe. The three-dimensional model (SI3D-L) is based on the continuity equation for incompressible fluids, the Reynolds-averaged form of the Navier-Stokes equations for momentum, the transport equation for temperature, and an equation of state relating temperature to fluid density. It incorporates a form of the Mellor-Yamada 2.5 turbulence closure scheme
The hydrodynamics of the Oaks Arm of Clear Lake has been analyzed using SI3D-L. Temperature observations gathered at three thermistor chains were used for model validation. The numerical results, after an initial spin-up period of 2 days, proved to be relatively close to the observations. The numerical simulations show that the factors driving the cyclonic baroclinic pumping circulation of the Oaks Arm are: (1) stratification, (2) periodic and predominantly uniform westerly winds, and (3) Coriolis effects (Earth's rotation). The spatial variability of the wind proved to be especially important in reproducing the anticyclonic circulation observed in the Lower Arm. The Earth's rotation generates a remarkable asymmetry in the evolution of the temperature field across the longitudinal axis of the Oaks Arm, which is critical to an explanation of the existing residual cyclonic circulation and many of the observations gathered in the field. The asymmetric temperature stratification results in differences in the rate at which the vertical circulation switches from being driven by wind to being driven by baroclinic forces, along the north and south shores of the Oaks Arm. The change of vertical circulation starts at the easternmost end of the arm and moves to the west more rapidly along the north shore. Consequently, there are stations on the southwest side of the Oaks Arm where the baroclinic circulation rarely replaces the wind-driven circulation, while across the basin on the north shore the baroclinic circulation is felt more frequently and more intensely. This difference in the amount of time the baroclinicly-driven vertical circulation dominates the wind-driven motions is reflected in the residual profiles. In the north, the residual currents are westward, while in the south, the residual currents are eastward.
The model also has been used to confirm the nature of the dominant basin-scale internal waves of Lake Tahoe under winter convecting conditions. During that time, the lake was weakly stratified with a well-defined metalimnion between 50 and 150 meters below the free water surface. Three modes of oscillation were isolated in the measured internal wave field, and were positively identified through analysis and display of the model results. Two of the modes have sub-inertial frequencies and were identified as vertical mode 1 Kelvin waves traveling counterclockwise around the perimeter of the lake with periodicities of 4?5 days and 54 hours. The longer period Kelvin wave has horizontal mode 1, while the shorter period Kelvin wave has horizontal mode 2. The third internal wave mode is a Poincare wave having a period of about 18 hours, causing transverse oscillations of the isotherms.
The identification and description of the internal wave field was facilitated by the conjunctive use of both detailed field measurements and a three-dimensional modeling approach. The latter provides spatially intensive information that is not possible to gain with a field measurement program. It also enables one to a posteriori select points of particular interest or importance for analysis. Band-filtered, three-dimensional isotherm displacement plots, which can only be produced by the model, constitute a powerful methodology to visualize and interpret the internal wave field. Similarly, the selection of points for the production of rotary spectra was done with the benefit of the hindsight that the model results provided.
Journal Articles on this Report : 7 Displayed | Download in RIS Format
Other project views: | All 20 publications | 7 publications in selected types | All 7 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Lee M, Schladow SG. Visualization of oxygen concentration in water bodies using a fluorescence technique. Water Research 2000;34(10):2842-2845. |
R825428 (Final) |
not available |
|
Rueda FJ, Schladow SG. Simulating the baroclinic response in stratified basins: a quantitative comparison of two numerical models. Journal of Hydrologic Engineering 2001. |
R825428 (Final) |
not available |
|
Rueda FJ, Schladow SG. Quantitative comparison of models for barotropic response of homogeneous basins. Journal of Hydraulic Engineering-Asce 2002;128(2):201-213. |
R825428 (Final) |
not available |
|
Rueda FJ, Schladow SG. Surface seiches in lakes of complex geometry. Application to Clear Lake, California. Limnology and Oceanography 2002;47(3):906-910. |
R825428 (Final) |
not available |
|
Rueda FJ, Schladow SG, Monismith SG, Stacey MT. Dynamics of large polymictic lake. I: Field observations. Journal of Hydraulic Engineering-Asce 2003;129(2):82-91. |
R825428 (Final) |
not available |
|
Rueda FJ, Schladow SG. Dynamics of large polymictic lake. II: Numerical simulations. Journal of Hydraulic Engineering 2003;129(2):92-101. |
R825428 (Final) R825433 (Final) R826282 (Final) |
Exit |
|
Schladow SG, Lee M, Hurzeler BE, Kelly PB. Oxygen transfer across the air-water interface by natural convection in lakes. Limnology and Oceanography 2002;47(5):1394-1404. |
R825428 (Final) |
not available |
Supplemental Keywords:
water, aquatic, physics, engineering, limnology, modeling, monitoring, western, California, CA, EPA Region 9., RFA, Scientific Discipline, Geographic Area, Water, Ecosystem Protection/Environmental Exposure & Risk, Ecology, Ecosystem/Assessment/Indicators, Ecosystem Protection, exploratory research environmental biology, Chemical Mixtures - Environmental Exposure & Risk, Environmental Chemistry, State, Chemistry, Ecological Effects - Environmental Exposure & Risk, Ecological Effects - Human Health, West Coast, Engineering, Chemistry, & Physics, Ecological Indicators, ecological exposure, nutrient dynamics, fate and transport, reservoirs, contaminant transport, stratified lakes and reservoirs, pollutant transport, Clear Lake, laser based diagnostics, ecological impacts, natural convection, lakes, oxygen budget, environmental contaminants, laser induced flouresence studies, transport models, digital particle thermometry, water quality, lake ecosysyems, Lake Tahoe, contaminated aquifers, flourescence assay, California (CA)Relevant Websites:
http://www.engr.ucdavis.edu/~edllab/![Exit EPA icon](https://www.epa.gov/ncer/images/exit.gif)
Progress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.