UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Hypolimnetic withdrawl from a shallow, eutrophic lake Macdonald, Ronald H. 1995

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1996-0068.pdf [ 10.02MB ]
JSON: 831-1.0050365.json
JSON-LD: 831-1.0050365-ld.json
RDF/XML (Pretty): 831-1.0050365-rdf.xml
RDF/JSON: 831-1.0050365-rdf.json
Turtle: 831-1.0050365-turtle.txt
N-Triples: 831-1.0050365-rdf-ntriples.txt
Original Record: 831-1.0050365-source.json
Full Text

Full Text

H Y P O L I M N E T I C W I T H D R A W A L F R O M A S H A L L O W , E U T R O P H I C L A K E by Ronald H . Macdonald B . Sc., The University of Calgary, 1985 A THESIS S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F A P P L I E D S C I E N C E in T H E F A C U L T Y O F G R A D U A T E S T U D I E S Department of C i v i l Engineering We accept this thesis as conforming to the required standard The University of British Columbia December, 1995 © Ronald H . Macdonald, 1995 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia Vancouver, Canada Department DE-6 (2/88) Abstract Hypolimnetic withdrawal is a technique for reducing trophic status in a high productivity, or eutrophic, lake. This thesis presents a case study of a hypolimnetic withdrawal application to a shallow lake in the B.C. interior. The installation demonstrates an appropriate technology for lake restoration suitable for recreational properties where the population base is small, or residents may not be present continuously. The polymictic nature of the lake is evident from daily average water temperatures. Partial destratifications occur within a day and complete de-stratifications occur over a period of a few days. The difference in stratification between the two observed summers highlights the significance of inter-year meteorological variation in determining lake response. Observed responses to wind forcing included surface mixing, isotherm tilting and seiching. A typical summertime seiche period of 10 hours was observed during 1994. Theoretical predictions of the seiche period using two-layer stratification and linear stratification estimates provide approximate upper and lower bounds for the observed response, as determined from spectral analysis. Wind events and lake thermal response can be parameterized by the Wedderburn number. Substantial upwelling and destratification was observed when the Wedderburn number was between 1 and 2. Higher Wedderburn numbers resulted in seiching responses. A small deep area provides a quiescent region in the lake that is thermally stratified from the main water column. This area was low in oxygen through the entire 1994 withdrawal season and had elevated concentrations of phosphorus, iron, and manganese indicating that this area may not mix during spring turnover, or that oxygen depletion rates are sufficient to result in anoxia only four weeks after ice off. The chemistry of the lake agreed with our expectation of an anoxic, internally loaded system. Following the depletion of oxygen, concentrations of phosphorus, iron, and manganese increased in the lower levels of the lake. This build-up was sustained for two i i weeks without mixing to the surface due to thermal stratification. Subsequent surface cooling resulted in partial mixing. The resulting circulation of nutrients initiated an algae bloom. During the first year of operation, a total of 0.63 m of lake (10% of volume) was removed through the withdrawal. This resulted in the export of about 30 kg of total phosphorus. Comparison of the maximum phosphorus observed in the lake for several years indicates that relative to other years, the phosphorus in Chain Lake was about the same or slightly lower than in other years. Rough calculations indicate that the withdrawal may achieve a net phosphorus export of 5-20 kg/yr, depleting the sediment phosphorus pool in times scales of centuries. A correlation between the daily average air temperature in Princeton, and the lake stability shows that lake stratifications are coincident with long periods of above average temperatures in Princeton. The correlation permits use of the long term meteorological record to interpret the frequency distribution of lake stratifications. The results indicate that the stratification observed in 1994 was a once in 7 seven year event in terms of the length of stratification. Monitoring evaluated the environmental impacts on Chain Lake and Hayes Creek. The withdrawal is expected to have a negligible impact on the thermal structure of the lake. Dissolved oxygen deficits were observed throughout the summer at the withdrawal and an attached fountain was able to increase the dissolved oxygen by about 1.5 mg/L. High levels of ammonia, iron, and manganese were observed during the July stratification. The ammonia was quickly consumed, or lost in Hayes Creek, while iron and manganese exhibited an attenuation in which the peak levels were reduced, and the duration increased. After 5 weeks of stratification, sulfide was detected on August 8. Overall, the environmental impacts of the withdrawal are localized within the area of the Hayes Creek between the Fountain and the Jellicoe Road station (about 500 m). i i i Table of Contents Abstract ii Table of Contents iv List of Tables vii List of Figures viii Acknowledgments xi 1. INTRODUCTION 1.1 Eutrophication and Lake Restoration 1 1.2 Objectives and Scope of Research 1 1.3 Study Site 2 1.3.1 Chain Lake : 2 1.3.2 Previous Restoration Efforts -.. 3 2. LITERATURE REVIEW 2.1 B ackground , 5 2.1.1 Stratification 5 2.1.2 Oxygen Depletion 7 2.1.3 Internal Loading 7 2.1.4 Nutrient Loading and Water Column Concentration 9 2.1.5 Meteorological Forcing and Lake Response 11 2.1.6 Vertical Transport of Nutrients 16 2.2 Hypolimnetic Withdrawal for Lake Restoration 17 2.2.1 Nutrient Export 17 2.2.2 Reduced Anoxia 17 2.2.3 Other Hypolimnetic Withdrawal Implementations 18 2.3 Chain Lake as Withdrawal Candidate 20 2.3.1 Trophic Status 21 2.3.2 Sediments 22 3. METHODS, MATERIALS, and DATA PROCESSING 3.1 Withdrawal System 23 3.2 Physical Samples 24 3.2.1 Sample Locations 24 3.2.2 Handling, Storage, and Preservation 25 3.2.3 Sample Analysis 26 3.3 Dissolved Oxygen 26 3.4 Electronic Data 26 3.4.1 Meteorological Data 26 3.4.2 Thermistor Data ....27 3.5 Hydrologic Data 27 iv 3.6 Data Processing 28 3.6.1 Time Series Data 28 3.6.2 Areal Data Averaging : 29 3.7 Quality Assurance / Quality Control (QA/QC) 30 3.7.1 Iron and Manganese 30 3.7.2 Digested Phosphorus ..32 4. OBSERVATIONS and RESULTS 4.1 Meteorological Data •.. 35 4.1.1 Regional Weather 35 4.1.2 Local Meteorological Forcing 36 4.2 Water Temperatures.... ; 37 4.3 Hypolimnetic Withdrawal Characteristics 38 4.4 Lake Physics 41 4.4.1 Diversion Inflow / Withdrawal Outflow Water Temperatures 41 4.4.2 Surface Mixing 44 4.4.3 Wind Forcing - Seiching 44 4.4.4 Wind Set-Up 49 4.4.5 Stratification, Stability, and Wind Work 51 4.4.6 Dredged Hole 51 4.4.7 Expected Stratification 52 4.5 Thermal and Water Budgets 53 4.5.1 Heat Content 53 4.5.2 Water Budget 54 4.6 Chemical / Biological Measurements 54 4.6.1 Dissolved Oxygen 54 4.6.2 Secchi Depth / Chlorophyll-a / Total Phosphorus 55 4.6.3 Phosphorus Components 57 4.6.4 Iron and Manganese 58 4.6.6 Destratification and Elimination of Anoxia 61 4.7 Summary 61 5. DISCUSSION 5.1 Defining a Successful Withdrawal 63 5.2 The Technological Applicability of Withdrawal 64 5.3 Limnological Perspective of Observations 65 5.3.1 Growth Limitation 65 5.3.2 Reduction of Lake Phosphorus Content 67 5.3.3 Changes in Secchi Transparency 68 5.3.4 Sedimentation and Release Rates 69 5.4. Effectiveness of the Withdrawal for Nutrient Export 71 5.4.1 1994 Nutrient Export 71 5.4.2 Phosphorus Removal Rate 71 5.4.3 Net Export and Long Term Potential 72 5.4.4 Phosphorus Pool 73 v 5.5 Seasonal Variability of Anoxia 75 5.5.1 Inter-year Variability 75 5.5.2 Meteorological and Simplified Prediction ....76 5.5.3 Correlation of Stability to A i r Temperature 77 5.5.5 Distribution of Warm Periods 79 5.6 Summary 79 6. ENVIRONMENTAL IMPACTS and MITIGATION 6.1 Potential Impacts 81 6.2 Environmental Monitoring Program 82 6.2.1 Program Design ". 82 6.2.2 Previous Data Collection 83 6.2.3 1993 'Before withdrawal' Monitoring 85 6.2.4 1994 'During Withdrawal' Monitoring 85 6.3 Environmental Concerns 86 6.3.1 Lake Heat Content 86 6.3.2 Destratification 87 6.3.3 Water Budget / Lake Level .' 87 6.3.4 Dissolved Oxygen - Aerator-Fountain 87 6.3.5 Dissolved Oxygen - D O Sag 88 6.3.6 Nutrient Loads 89 6.3.7 Ammonia 90 6.3.8 Iron and Manganese ....91 6.3.9 Hydrogen Sulfide 93 6.3.10 Turbidity ...94 6.4 Summary 95 7. CONCLUSIONS and RECOMMENDATIONS 7.1 Summary and Conclusions 96 7.2 Recommendations 100 7.3 Further Research 101 8. REFERENCES .....102 9. FIGURES 109 10. APPENDICES 163 APPENDIX A - Quality Assurance and Quality Control 164 APPENDIX B - Summary of Lake Sample Values 174 APPENDIX C - Calculation of Heat and Water Budgets 177 APPENDIX D - Stream Monitoring Program - 1994 181 vi List of Tables Table 1.1 Limnological Parameters for Chain Lake 3 Table 2.1: Summary of Physical Parameters 16 Table 3.1: Summary of Iron QA/QC 31 Table 3.2: Summary of Manganese QA/QC 31 Table 3.3: Summary of Phosphorus Runs - 1994 33 Table 3.4: Summary of Digested Phosphorus (Soluble and Total) QA/QC - 1994 33 Table 4.1: Expected and Observed Seiching Period July 24-29,1994 46 Table 4.2: CTD Profiles: Isotherm Slope Summary , 50 Table 5.1 Results of Chain Lake Core Incubations 70 Table 5.2 Estimation of Sediment Bioavailable Phosphorus Pool 75 Table 6.1: Summary of Hayes Creek Water Quality 84 Table 6.2: Summary of Shinish Creek Water Quality 84 Table A - l Method Detection Limit Runs for Fe and Mn 166 Table A-2 Fe and Mn Replicate Analysis (of Same Sample) 1994 167 Table A-3 Fe and Mn Triplicate Analysis (of Same Sample) 1994 167 Table A-4 Fe and Mn Replicate Sampling and Analysis - 1994 167 Table A-5 Fe and Mn Triplicate Sampling and Analysis - 1994 167 Table A-6 Fe and Mn Spikes and Standard Additions - 1994 '. 168 Table A-7 Comparison of HN03 vs. H2S04 Preservation for Fe and Mn-1994 169 Table A-8 Method Detection Limit Runs for Phosphorus 170 Table A-9 Phosphorus Replicate Analysis (of Same Sample) 1994 171 Table A-10 Phosphorus Triplicate Analysis (of Same Sample) 1994 171 Table A - l 1 Phosphorus Replicate Sampling and Analysis 1994 172 Table A-12 Phosphorus Spikes and Standard Additions - 1994 172 Table A-13 A N O V A Table for Phosphorus Replicate Test - Sampled Oct 5, 1994 172 vii Table D - l - Fountain Station 182 Table D-2 - Spillway Station 182 Table D-3 - Beaver Dam Station 183 Table D-4 - Jellicoe Road Station 183 Table D-5 - Hayes Creek above Hayes-Shinish Junction 184 Table D-6 - Shinish Creek Station 184 Table D-7 - Shinish Creek above Hayes-Shinish Junction 184 Table D-8 - Hayes Creek Station 184 viii List of Figures Figure 1.1 Morphometric Map of Chain Lake (after Murphy et al, 1990) I l l Figure 1.2 Aerial Photo of Chain Lake 112 Figure 2.1 Comparison of Two-layer and linear Stratification 113 Figure 2.2 Typical Sediment Structure at Chain Lake :... 114 Figure 2.3 Sediment Composition: Iron, Phosphorus, Water Content, and Age vs. depth (after Murphy et al, 1990) 115 Figure 3.1 Schematic of Withdrawal Chain Lake Withdrawal 116 Figure 3.2(a) Photograph of Withdrawal Box 117 Figure 3.2(b) Photograph of Aerator Fountain 118 Figure 3.3 Stream Sampling Locations 119 Figure 3.4 Extract of Times series and Daily Average Times Series 120 Figure 3.5 Chain Lake Sample Schematic 94-07-23 121 Figure 4.1 Meterolrologic Conditions at Princeton 122 Figure 4.2 Meteorlogical Measurements at Chain Lake -1994 123 Figure 4.3 Daily Average Water Temperatures 124 Figure 4.4 Hypolimnetic Withdrawal Composition - 1994 125 Figure 4.4 (continued) Hypolimnetic Withdrawal Composition - 1994 126 Figure 4.5 Temperatures of Shinish Creek and hypolimnetic withdrawal 127 Figure 4.6 CTD transect 95-07-07 128 Figure 4.7 Thermistor Data from July 24 to July 27, 1994 129 Figure 4.8 Expected Chain Lake Seiche Period and Observed FFT Components -1994 130 Figure 4.9 Wind Forcing and Isotherm Response July 15 to July 26,1994 131 Figure 4.10 CTD Transect Profile Summary 132 Figure 4.11 Lake Stability, Wind Work and Daily Precipitation - 1994 133 Figure 4.12 Temperature and Dissolved Oxygen Profiles in the Dredged Hole 134 ix Figure 4.13 Heat Content of Chain Lake 1993 and 1994 135 Figure 4.14 Water Budget for Chain Lake - 1994 136 Figure 4.15 Dissolved Oxygen Profiles - 1993 137 Figure 4.16 Dissolved Oxygen Profiles - 1994 , 138 Figure 4.17 Chain Lake 1993: Secchi Depth, Chla, and Total Phosphorus 139 Figure 4.18 Chain Lake 1994: Secchi Depth, Chla, and Total Phosphorus 140 Figure 4.19 Chain Lake 1994 Phosphorus Concentrations 141 Figure 4.20 Chain Lake 1994 Iron Concentrations 142 Figure 4.21 Chain Lake 1994 Manganese Concentrations 143 Figure 4.22 Ratio of Iron to Manganese in the withdrawal flow - 1994 144 Figure 4.23 Hypolimnetic Withdrawal Composition August 4 to Sept 1, 1994 145 Figure 5.1 Flow Test of Withdrawal System 95-05-20 146 Figure 5.2 Chi a vs. Total Phosphorus 147 Figure 5.3 Maximum Measured Phosphorus Content in Chain Lake 148 Figure 5.4 Chain Lake Secchi Measurement - Historical Perspective 149 Figure 5.5 Apparent Net Phosphorus Concentration Change - 1994 150 Figure 5.6 Chain Lake Phosphorus Export - 1994 151 Figure 5.7 Chain Lake Phosphorus Renewal Time - 1994 152 Figure 5.8: Lake Stability Correlation to Princeton Air Temperature 153 Figure 5.9: Princeton Mean Air Temperature during summer - 1937 to 1994 154 Figure 5.10: Results of 'Worm Count' of Warm Periods 155 Figure 6.1 Hayes and Shinish Creek Total Phosphorus, 1993-08-19 156 Figure 6.2 Chain Lake Withdrawal Fountain Dissolved Oxygen Levels -1994 157 Figure 6.3 Hayes Creek Dissovled Oxygen Levels -1994 158 Figure 6.4 Withdrawal Ammonia Levels - 1994 159 Figure 6.5 Hayes and Shinish Creek Ammonia Levels - 1994 160 Figure 6.6 Hayes and Shinish Creek Iron Levels - 1994 161 x Figure 6.7 Hayes and Shinish Creek Manganese Levels - 1994 xi Acknowledgments Thanks to the residents of Chain Lake, who worked hard to put the project together. In particular, I am indebted to Bob and Ann Kidd, Jim Dawson, and Rolf and Toni Mauer. Tom Murphy initiated this project and stayed with it to completion - providing his time, some $, and lots of tools to work with. Thanks also to Mike Mawhinney for guiding the installation. Thanks to my supervisor Greg Lawrence - he did all the usual stuff. Plus, he met with me on weekends, funded me to do extra monitoring 'just because', read my drafts, and found a little more funding when I needed it. Ken Hall, my second reader, lent me his limnology equipment, and turned my drafts around quickly. Thanks to Susan Harper, Paula Parkinson, and especially Zufang Zhou, all in the environmental lab. Their assistance was invaluable arid always forthcoming - even when I walked in with a question at 2 minutes to 5:00. Jason Vine built the darn thing and deserves recognition. Plus he stuck around when things weren't as exciting - going to public meetings, sizing culvert pipe, etc. He also turned me into a big Mac-head! His unbounded computing expertise and endless generosity with it saved me at least a zillion hours. Craig Stevens mentored me in the physics of lakes - a dubious honor and provided useful feedback on my results - even on my most esoteric tangents! Thanks to Al and Don who were always willing to talk shop back when we were all still curious and an M D L was important. Thanks to my many friends who have offered me support through out the past couple of years. Special thanks to my dear Loralee, who demonstrated the true depth of her love, patience and understanding while I missed camping trips to do field trips, Saturdays to do lab-days, and crunched late into the night too, too many times. You are wonderful!! xii 1. INTRODUCTION: 1.1 Eutrophication and Lake Restoration Eutrophication is the enrichment of natural waters with nutrients, usually phosphorus and nitrogen, leading to high levels of algae or aquatic plants. This often makes the water undesirable for human purposes. Eutrophication is a natural process in the life cycle of a lake, occurring as the lake fills with organic matter over time scales of thousands of years (Stumm and Stumm-Zolinger, 1972). It is often accelerated by anthropogenic activities, to the point where eutrophication can occur on time scales of decades. Eutrophication may have detrimental effects on a lake. These include severe growth of algae or aquatic plants, oxygen depletion during the summer or winter due to the oxygen demand of decaying organic matter, and a reduction in the numbers of species in a water body (diversity), though a proliferation of the species that do remain (Moss, 1988). If the water is a drinking water source, the high productivity can result in taste and odor problems and high levels of organic compounds and possibly result in undesirable disinfection by-products. Since eutrophic waters are often unsuitable for human purposes, significant resources are spent on lake restoration activities. Lake restoration may be loosely defined as intervention to "return a [lake] system to its condition prior to disturbance" (Cooke et al, 1993). The range of restoration activities is broad. It can include control of nutrient inputs, nutrient inactivation by chemical precipitation or oxidation, sediment dredging, algaecide treatment, water level control, sediment capping or shading, artificial destratification, hypolimnetic aeration, biomanipulation, and plant harvesting (Cooke et al, 1993). The restoration technique examined here is hypolimnetic withdrawal. 1.2 Objectives and Scope of Research This thesis describes a case study of a hypolimnetic withdrawal application. The lake studied is representative of many lakes in B C that are highly productive. A s such, it provides 1 Chapter 1: Introduction an opportunity to 'pilot' test hypolimnetic withdrawal as a lake restoration technique in the province. The objectives of this research are to: • demonstrate that a withdrawal system can be technologically appropriate for application in interior BC lakes. • monitor the lake conditions and evaluate the short term (i.e. first year) effects of the withdrawal system. • observe the physical nature of a shallow eutrophic lake and the effect of different environmental forces on the lake's behavior, with particular emphasis on those events which will have an effect on a withdrawal system. • monitor and evaluate the environmental impacts of a withdrawal system on the lake, and the receiving stream. 1.3 Study Site This project was initiated by the residents of Chain Lake. Funding was obtained from the Habitat Conservation Fund of the BC Ministry of Environment in light of the popular recreational trout fishery at the lake which has been valued at 9000 angler days per year (Steve Mathews, pers comm, 1992). 1.3.1 Chain Lake Chain Lake is situated 45 km NE of Princeton, British Columbia. It is rectangular in shape and relatively flat bottomed. A map of the lake morphology is shown in Figure 1.1. An aerial photo of Chain Lake is shown in Figure 1.2. The greatest 'natural' lake depth is about 6.5 m. In 1987/88 a 70 m x 80 m x 3m hole was dredged in the bottom of the south end of the lake and as a result, the maximum depth in the lake is now 9 m (Weigand, 1989a). The volume of the dredged hole (16,800 m )^ represents a 0.6% increase in lake volume. A summary of the morphometric and limnological parameters is given in Table 1.1. They indicate that Chain Lake is a small, shallow lake with high algal productivity. 2 Chapter 1: Introduction Table 1.1 Limnological Parameters for Chain Lake Physical Parameters Chemical Parameters Ave Depth (m) Max. Depth (m) Length (km) Width (km) Area (ha) Volume (106m3) Residence Time(yr) 6 9 1.6 0.3 46 2.76 0.5 - 3.0 Alkalinity (mg/L CaC0 3 ) Hardness (mg/L CaCC>3) Conductivity (|iS/cm) Total Phosphorus (|ig/L) Maximum chl a (ug/L^1) Minimum Secchi Depth(*) 80 65 135 13-282 > 100 < 1.0 Sources: Murphy (1987), Water Investigations Branch (1977), this study Notes: (1) During reported algae blooms 1.3.2 Previous Restoration Efforts Local residents have been actively involved in the management of Chain Lake since the 1960's. Several restoration projects have been implemented including copper sulfate treatment, small scale aeration, water diversion, and sediment dredging. These met with varying degrees of success. For example, the copper sulfate (CUSO4) treatment was extremely effective in removing algae from the lake. However, the treatment was short lived and within weeks an algae bloom had reappeared.1 This is not surprising as copper treatment is a temporary measure (Cooke et al, 1993). A diversion was built in the 1960's to bring water from the Shinish Creek into Chain Lake. It attempted to increase the hydraulic flushing rate of the lake. Shinish Creek is suitable for diversion because it has lower nutrient levels than Hayes Creek. The benefits of the diversion have not been clearly defined. Ennis (1972) observed no significant change in the algal composition during the first years of operation and the Water Investigations Branch of the Ministry of Environment (1977) found the flow rate through the diversion too low to 1 It has been speculated that the bloom following the CuS04 treatment was accelerated by the decomposition of the algae that the CuS04 had killed. This is supported only by anecdotal evidence that the bloom came back 'worse than before' (Dawson, pers comm, 1993). 3 Chapter 1: Introduction alter the flushing rate of the lake. As well, the influence of internal loading has reduced the flushing benefit expected from the diversion (Murphy, 1987). Conversely, there is anecdotal evidence that fish kills have become less frequent since the installation of the withdrawal (Murphy, 1987). In the fall of 1987 and spring of 1988 an experimental dredging program took place at Chain Lake. An area at the south end of the lake was dredged 3 m deep (70m x 80m area). The dredging was intended to create a sediment focus, in which suspended sediments from the entire lake would be preferentially deposited, and to increase the stability of thermal stratification in the lake by creating a deeper hypolimnion. Increased stability can be achieved by dredging (e.g. Stefan and Hanson, 1980) though it is uncertain whether the small dredged hole would have an effect of lake wide stratification. From 1989 to 1992, the residents conducted their own maintenance dredging program which removed accumulated sediment from the dredged area. This practice was successful in maintaining the maximum depth though the effects on the lake status is uncertain, primarily due to insufficient long term monitoring. Statistically significant evidence for the success of these restoration projects is rarely available. Ennis (1972) found no significant change in algal composition or chemical composition of the lake during the early years of the diversion operation. The monitoring program following the dredging program was only possible to the end of 1988, and so a quantification of the long term effects has not been made. Anecdotal evidence does not verify whether the various restoration measures have had some positive effect. Some local residents at the lake make reference to how 'bad things used to be' implying that the water quality is improving. Contrary to this though is the general impression of poor (and declining) water quality in the lake during the 1980's which was a driving force behind the dredging project. The difficulty of evaluating lake restoration effectiveness is not unique to Chain Lake (Cooke et al, 1993). 4 2. L I T E R A T U R E R E V I E W : 2.1 Background 2.1.1 Stratification Temperate lakes are approximately isothermal at the time of ice-off and under isothermal conditions vertical mixing is usually complete. Physical properties such as temperature and dissolved oxygen are uniform through the water column. Following ice-off, solar radiation transfers heat into lakes through the surface, and most of the energy is captured in the upper layers of the lake. The surface of the lake warms and the wind mixes the upper region into a homogeneous warm layer. This is called the epilimnion or mixed layer. Beneath the epilimnion is a cooler layer called the hypolimnion. The epi- and hypolimnion are separated by a region of high thermal gradient (and therefore high density gradient) called the metalimnion or thermocline. Metalimnetic temperature gradients are typically 17m or greater (Wetzel, 1975, Hutchinson, 1957). The relative thickness of the metalimnion may be on the order of several meters which, for a deep lake, is a small fraction of the total depth. Further input of heat during spring and summer results in continued warming of the epilimnion, which strengthens the density difference separating the layers. In many lakes the thermal structure is stable throughout the summer season. In the fall, heat losses through the surface cool the mixed layer. As it cools, the epilimnion mixes with water from the metalimnion creating a thicker mixed layer. The mixed layer continues to cool and increase in thickness until isothermal conditions are reached - i.e. fall overturn. Shallow lakes often do not behave in a simple two-layer fashion. The stratifications may only last for short time periods ranging from diurnal surface stratification up to quasi-seasonal stratifications of several weeks. Distinctions between an epi- and hypolimnion may not be evident. Rather a more continuous stratification may be observed (Hutchinson, 1957). The existence of stratification at all may be temporary and dependent on the morphology of the lake, the areal cover around the lake which affects the wind field, or the weather 5 Chapter 2: Literature Review conditions at that time. (Gorham and Boyce, 1989, Hutchinson, 1957). An important feature of shallow lakes is that breakdowns in stratification may be frequently driven by storms or strong wind mixing events and not solely by fall cooling. For the discussions here the term 'hypolimnetic' will often be used although the presence of a 'text book' hypolimnion may not be appropriate. In this context hypolimnion will refer to the region at the bottom of the lake, which is somewhat cooler than the surface water, and may be isolated from contact with the atmosphere due to these thermal density differences, whether step-wise or continuous. The strength of thermal stratification may be described by the stability (S). Stability is the work required to convert a given thermal stratification into a homogeneous temperature stratification. It represents the sum of the work required to raise or lower individual parcels of water to the level of the midpoint of stratification. Idso (1973) defined the stability S as the work (in Joules) required to mix the water column, normalized by the lake surface area (m2): where: S = stability (Jm"2) z = depth (m) zm = depth of mean density pm (m) p = density at depth z (kg m"3) pm = volumetric mean density (kg m"3) H = maximum lake depth (m) Aj = lake surface area (m2) g = gravitational acceleration (9.81 m s"2) Stability is a useful parameter because it incorporates the effects of the non-linear nature of water density with temperature (through the p terms), the depth of thermocline 6 Chapter 2: Literature Review (through the z m term), and lake morphometry (through the volumetric average density p m term). As such it is a more useful parameter than simply using temperature profiles, particularly in lakes with variable morphometry. 2.1.2 Oxygen Depletion The hypolimnia of lakes are isolated from the surface and thus from the natural aeration by wind and contact with the atmosphere. Metabolic respiration continues as bacteria degrade deposited organic matter and other species consume oxygen. The result is that hypolimnetic oxygen is consumed. If the oxygen is not replaced by photosynthesis, or if the consumption rate is fast enough, then the hypolimnion will become anoxic. (Wetzel, 1975, Henderson-Sellers, 1984). The rate of oxygen depletion is primarily dependent on the volume of the hypolimnion (i.e. the size of the oxygen reservoir at the onset of stratification), the temperature, and the amount of organic matter present for metabolism. Cornett and Rigler (1984) suggest that the rate of hypolimnetic oxygen depletion often occurs at a constant rate over time and thus is not a function of the oxygen concentration (i.e. a zero order process). In deep, or low productivity lakes the process of oxygen depletion occurs over seasonal time scales and complete oxygen depletion may not occur before fall overturn. On the other hand, shallow lakes with relatively small hypolimnetic volumes, and high loads of organic matter to the hypolimnion may become depleted very quickly. For example, Stefan and Hanson (1981) evaluated five shallow eutrophic lakes and found oxygen depletion rates from 0.67 mg/L/day to 1.65 mg/L/day. At these rates, a fully aerated hypolimnion with dissolved oxygen of =10 mg/L would become anoxic within 6 to 15 days. 2.1.3 Internal Loading Internal loading is the movement of nutrients, usually phosphorus, into the water column from within the lake. Internal loads are often from the sediments but may also originate from aquatic plants, or decay of particulate matter within the water column. In 7 Chapter 2: Literature Review contrast to this, external loads originate outside of the lake body and include watershed runoff, stream inflows, airborne precipitation. The sediment pool represents a large store of nutrients which, when liberated, may promote productivity. The nutrients released by internal mechanisms are often of the same magnitude or greater than the external loads. Internal loading may occur under aerobic conditions. This is general attributed to mineralization of organic matter (i.e. biological and chemical decay). While not always considered important, the aerobic cycling of nutrients has been reported as significant (e.g. Lee, 1970, Golterman, 1995). Mineralization may occur anaerobically as well, although anaerobic mineralization processes are typically much slower than aerobic ones. Internal loading is most often an anaerobic phenomenon. The classic internal loading mechanism is a chemical liberation associated with iron precipitation and release. Under aerobic conditions iron is in the F e 3 + state, which forms insoluble Fe(OH)3 (or FeOOH). Phosphate ions adsorb to the surface of precipitated ferric hydroxide. In this state, the phosphate ions are essentially immobilized from either the water column, or the pore water within the sediments. As a result, aerobic waters containing iron are generally regarded as nutrient sinks, removing phosphorus from the water column. However, under anaerobic conditions, the oxidation-reduction potential (ORP) of the water drops below 200 mV and F e 3 + is reduced to F e 2 + , which is highly soluble (Henderson-Sellers, 1984). Reduced iron moves from the particulate phase of the sediments, to the aqueous phase of the sediment pore water. This solubilization is accompanied by the release of the adsorbed phosphorus into solution. This model of iron bound phosphorus was shown very clearly in a classic experiment on lake cores by Mortimer (1941, 1942, 1971). This ORP (i.e. aerobic-anaerobic) regulated release of phosphorus may dominate the entire lake system. Lee et al (1976) noted that anaerobic phosphorus release rates are typically 10 times that of aerobic release rates, demonstrating the significance of this mechanism. From the interstitial pore water iron and phosphorus can move upward in the sediment to the water column. Many factors interact to control the relative rates of 8 Chapter 2: Literature Review phosphorus release and movement into the water column. These may be physical (e.g. diffusion rates, advection, sediment re-suspension, sediment compaction), chemical (pH, dissolved oxygen, redox potential, solubility, surface adsorption), or biological (microbial decomposition, zooplankton grazing) (Lee, 1970). Release rates are expressed as the mass of phosphorus released per unit time. Usually this value is normalized by the lake sediment area and expressed as mass of phosphorus released per unit area per unit time (e.g. mg/m2/day). Niirnberg (1988) surveyed published data on lakes which released phosphorus during periods of anoxia. Typical rates of sediment phosphorus release in eutrophic lakes are from 10 to 30 mg/m2/day during periods of anoxia. For a lake the size of Chain Lake, an internal load of 20 mg/m2/day, if distributed throughout the entire lake, will increase the phosphorus concentration by 3.3 Jig/L per day. Since a typical phosphorus concentration water quality objective for drinking, recreation or aquatic life is in the range of 15 fig/L (Nagpal, 1994), this shows how quickly a sediment release could increase the phosphorus concentration beyond a desirable level. 2.1.4 Nutrient Loading and Water Column Concentration The simplest models of lake nutrient loading are derived from Completely Stirred Tank Reactor (CSTR) models (Rechow and Chapra, 1983). These are based on simple steady state models, and are most often used to predict a response to changing loads over long time periods. Common to these models is an assumption that lakes tend to be phosphorus sinks and that some of the incoming nutrient is deposited in the sediment sink, and unavailable for growth. This loss has been represented alternatively as a sedimentary loss coefficient (fs), an apparent settling velocity of phosphorus (vs), and as a lake retention coefficient (Rp) representing the fraction of incoming phosphorus that is retained in the lake (Rechow and Chapra, 1983). These different derivations of loss coefficients are simplifications of the true nature of phosphorus movement to and from sediments. 9 Chapter 2: Literature Review The CSTR models have difficulty incorporating short time effects of internal loading. Some attempts have been made to try to account for internal loading in simple tank reactor models. Niirnberg (1984) included internal loading in a phosphorus retention type model. This model accounted for the nutrients from an internal load, but was based again around a steady-state (i.e. year-long) approach. As such it used seasonal estimates of the internal loading to estimate the long term average phosphorus concentration. Yoshida (1982) included the temporal variability of internal loading rates (as a function of sediment temperature over the course of a season). This analysis, however, did not include the effects of sedimentation over the season which may be a significant sink term from the water column in eutrophic lakes. The use of personal computers and spreadsheet programs allows many applications of the simple loading models to be continuously integrated over time. Thus a steady state relation can be applied for each time step (a day for example), and for each time step the functions that may be time varying (internal loading for example) can be included. A problem that has been noted for example is that sedimentation processes (or phosphorus retention) vary seasonally and require a certain level of study to be included in a model (Carlson, 1995 pers comm). Finally there are several available reservoir simulators which are capable of modeling multiple reactions occurring in the water column and at the sediment bed. These are limited by the large data sets required to initiate simulation runs, and the numbers of parameters which are frequently 'guess'-timated in order to run the model. In summary, there are no simple methods to accurately determine effects of internal loading coincident with variable sedimentation in eutrophic systems. Extensive field and/or laboratory measurements are often required to quantify these fluxes, and general applicability to different lakes is not assured. This is an area of limnology that still requires research and is hampered by the difficulties in defining standardized definitions and techniques (Niirnberg, 1995, pers comm). 10 Chapter 2: Literature Review 2.1.5 Meteorological Forcing and Lake Response Wind stress promotes destabilization (and hence vertical transport) by a number of mechanisms: (i) it may impart kinetic energy to the water column - the turbulence of the mixed layer, (ii) it may drag the surface layer downwind - wind 'set-up', which may result in seiching, and (iii) it up may result in upwelling - hypolimnetic waters appearing at the surface at the upwind end of the lake. Wind Stress The wind stress, is a drag force on the water surface due to the friction between the air and the water. This shear stress (t) is defined as: 2 . 2 ' U = C d P a ULd where : T = surface shear stress (N m~2) Cd = drag coefficient ( » 1 . 0 * 1 ( T 3 ) pa = air density (= 1.25 kg m"3) uwind = W U 1 Q speed (m s~l) Values of Cd have been determined to be around 1 .0*10" 3 to 1 . 3 * 1 0 - 3 , with some dependence on the wind velocity. The dependence on wind speed is reduced as the lake depth decreases. Without direct information to the contrary a constant value of 1.0 * 1 0 ~ 3 is acceptable (Fischer et al, 1 9 7 9 ) . The shear stress results in a downwind movement of the surface water characterized by a shear velocity, M = k r T w* = J — 2 .3 11 Chapter 2: Literature Review Where po is a representative density of water (=1000 kg/m3). The shear velocity is not the actual velocity of the surface water, but rather a velocity scale which is a measure of the shear stress. Wind Work Wind energy acts to disrupt the thermal stability described by eqn 1.1. The rate of work into the lake, or power applied, Pwind ( m J is given'as: Pwind = T Udrift 2 4 Since udrift has rarely been measured, provides an order of magnitude estimate for the wind work applied. Actual drift velocities are typically in the range of 4 to 20 times u% • Hutchinson (1957) reported the work of Ekman that udrift =1.8 % of wind speed at 30 ° latitude, and 1.4 % of wind speed at 60 ° latitude. For typical po and p a of 1000 and 1.2 kg/m3, respectively, and equations 2.2 and 2.3 this implies that udrift is about 12 to 16 times u* ' Wind Set-up In response to the wind stress, surface water 'piles-up' at the downwind end of the lake. To maintain equilibrium, the thermocline tilts in response. Thus the isotherms (representing iso-pycnals) will slope downward in the leeward direction. In response, the hypolimnetic water moves to the upwind end of the lake. In a two-layer system this is visualized best as a wedge of epilimnion ('pointing' upwind) floating on a wedge of hypolimnion (pointing downwind). The two wedges have approximately the same angle so that the free surface remains essentially horizontal. In reality, the surface slopes upward in the downwind direction but the vertical displacement is orders of magnitude smaller than the displacement of the density interface. In a shallow, linearly stratified lake, the wind stress moves surface water downwind, tilting the entire thermal structure. Turbulence, and mixing 12 Chapter 2: Literature Review with upwelled, slightly heavier fluid quickly produces a mixed surface layer. This was observed at the lab scale by Monismith (1986). This mixed layer formation occurs much faster than the establishment of isotherm tilt. Seiching Following wind set-up, if the wind shifts or relaxes, the thermal structure may seiche with a natural frequency Ti given by: 2.5 where: L = length of lake (m) c = phase speed (m s~l) For a two-layer system the phase speed, c, is: f c = 2.6 V J where: surface layer depth (m) bottom layer depth (m) surface layer density {kg m - 3) bottom layer density (kg m~3) P i Pi and g' is the reduced gravitational acceleration: Pi 8 2.7 For a linearly stratified system the first mode phase speed is NH c = 2.8 13 Chapter 2: Literature Review where: 2.9 When continuous profile data is not available the Brunt - Vaisalla frequency may be estimated from interval data (e.g. top to bottom temperature difference) as: where Ap is the density difference (kgrrr3) over a vertical distance Az (m). In a linearly, or continuously stratified system, there is no clearly defined epilimnion. As such an upper and lower water density are not clearly defined. For this situation Monismith (1986) defined the reduced gravitational acceleration from the Brunt-Vaisalla frequency as: The formulation of equation 2.11 is dimensionally correct and substitution of 2.10 into 2.11 yields a relation analogous to 2.7, with the exception of the V2 factor in eqn 2.11. Monismith did not explain his choice of definition for g' other than it provides a convenient basis for comparing linear stratifications to two layer stratifications1. A pictorial definition of these two terms is shown in Figure 2.1. Table 2.1 summarizes the different calculations used for two-layer and continuous stratifications. 2.10 2.11 'Later, when defining the Richardson number, Monosmith used H/2 as the representative depth for defining the behavior. The factor of 1/2 in eqn 2.11 is unexplained. 14 Chapter 2: Literature Review Richardson Number The non-dimensional parameter defining the ratio of stabilizing buoyancy forces to destabilising wind forcing is the Richardson number (Ri). Kl~~J 2.12 where h is a representative depth. For a two-layer system h is the depth of the mixed layer. For a linear stratification Monismith (1986) defined h as H/2 with the intention of enabling comparison of the stress response between the two-layer and linear stratifications. In a two layer system, the Richardson number is a useful indicator of the degree of wind set up. Spigel and Imberger (1980) and Harleman (1982) show the slope of the thermocline under steady state conditions is proportional to 1/Ri. In a continuously stratified system no analogous relationship has been reported. Wedderburn Number The Wedderburn number (W) describes the buoyancy to wind ratio, including the effect of the lake depth to fetch ratio. W = Ril 2.13 Another interpretation of the Wedderburn number is that it represents the ratio of surface layer depth (h) to the maximum isotherm displacement (L/Ri). A large Wedderburn number indicates that the displacements are small (weak winds or strong stratifications). The opposite is true with small W's indicating large displacements. The critical Wedderburn number for a two layer system is about 1. This is the point at which upwelling occurs. The point of upwelling is also the point at which most of the assumptions required for the derivation of two-layer theory breakdown. The Wedderburn number is a useful tool for categorizing lake response to wind forcing as it includes the effect of density stabilization, wind de-stabilization, and the fetch of the lake (Spigel and Imberger, 1980). 15 Chapter 2: Literature Review Table 2.1 presents a summary of the parameters defined and used for calculation. Table 2.1: Summary of Physical Parameters Two -Layer Linear or Continuous Parameter Stratification Stratification Pi epilimnion surface P2 hypolimnion bottom of water column Ap P 2 - Pi P 2 - P 1 N 2 — Po Az g' g (Ap/p2) y2N*H Ri 4 ( f ) w* 2 W L —(—> L 2 2.1.6 Vertical Transport of Nutrients The position of nutrients in the water column may affect whether an algae bloom occurs. In a eutrophic lake where Secchi depths might be on the order of 1 or 2 m, there is typically little or no photosynthesis occurring at the bottom of the lake2. The nutrients released from the sediments are not taken up immediately by plants or algae. Their presence becomes significant when the hypolimnetic waters are mixed to the epilimnion. In deeper lakes, nutrient mixing is associated with fall overturn and results in a fall bloom. At this time solar radiation has declined from it's summer peak, and the lake temperature is cooler than 2 A rule of thumb indicates that the rate of production = rate of respiration at a depth of 1.2 to 2.7 times the Secchi depth (mean of 1.7 x SD) (Moss, 1988 pg 164). Below this depth the rate of oxygen consumption (via respiration) will exceed that of production (photosynethesis). 16 Chapter 2: Literature Review mid-summer. As a result, a fall bloom may grow relatively slowly. In a polymictic lake vertical mixing may occur at any time of the summer and could promote growth at a fast rate. It is important to note that hypolimnetic phosphorus is generally bioavailable when mixed with surface water. For example Niirnberg (1985) determined that more than 60% of the phosphorus transported to the epilimnion from anoxic hypolimnion is available for growth. 2.2 Hypolimnetic Withdrawal for Lake Restoration Hypolimnetic withdrawal is the removal of water from deeper regions of the lake, in preference to the surface water. The objectives of this technique are: • to remove limiting nutrient (usually phosphorus) from the water column before it becomes available to algae or plants. .• to reduce the severity of anoxic conditions at the lower levels of the lake, thus reducing the level of internal loading. 2.2.1 Nutrient Export Hypolimnetic withdrawal attempts to break the cycle of anoxia-internal loading-eutrophication-sedimentation-decay-anoxia by exporting nutrients from the lake-sediment system. This implies that hypolimnetic withdrawal is appropriate to internally loaded lakes. The nature of many eutrophic systems is that hydraulic residence times are not sufficient to flush nutrients from the lake (Cooke et al, 1993). The withdrawal maximizes the phosphorus content of the water leaving the lake, so that the phosphorus residence time can be reduced without requiring a reduced hydraulic residence time. 2.2.2 Reduced Anoxia The withdrawal aims to maintain aerobic conditions at the deeper levels of the lake. The low oxygen water that is withdrawn is replaced by the lowering of the thermocline, by entrainment of the thermocline down into the hypolimnion, or by direct replacement if a 17 Chapter 2: Literature Review surface inflow is available that can be directed to the bottom of the lake (e.g. Lake Ballinger, Cooke et al, 1993). The success of such an effort to reduce anoxia depends on the relative rates of hypolimnetic oxygen depletion and supply, and the hypolimnetic residence time of water via a withdrawal. Since oxygen depletion can occur quickly in eutrophic lakes then efforts to flush out the anoxic water faster than the oxygen depletes may be difficult. 2.2.3 Other Hypolimnetic Withdrawal Implementations The first bottom withdrawal intended for the purpose of lake restoration was on Lake Kortowo, Poland (Olszewski, 1961, 1973). The objective was to remove the anoxic hypolimnion and increase the amount of the water column available for aquatic life. The withdrawal was rather 'aggressive' (flow rates not documented) and succeeded in lowering the thermocline from a depth of 6 m to 16 m, and accelerated the fall turnover by as much as 2 months, but usually by 1 month. This definitely increased the aerated portion of the water column, but also raised sediment temperatures from the 2 - 8 °C range as high to as 16 °C. The higher temperatures accelerated oxygen consuming metabolism and prevented the elimination of anoxia. As such the withdrawal 'fought itself by increasing the oxygen demand as it increased the oxygen supply. Niirnberg (1987a) reviewed the known lakes employing hypolimnetic withdrawal. She found that epilimnetic phosphorus concentrations decreased with the amount of phosphorus exported and the duration of withdrawal. While declines in epilimnetic phosphorus seem to occur, the data are scattered, and several years operation was required before noticeable changes occur. Niirnberg's review also examined the effects of withdrawal on anoxia and found that the depth at which anoxia occurs (the depth of the oxycline) may move deeper in the water column due to withdrawal operation. This is reasonable since the hypolimnion is being removed during withdrawal, and being replaced from above. However, the time period of anoxia was not significantly reduced in the larger lakes studied (i.e. volume > 2.5 x 106 m3), indicating that the depletion of oxygen occurred much faster than 18 Chapter 2: Literature Review anoxic water could be removed and replaced. This leaves the objective of 'reduction of anoxia' open to definition. The withdrawals seemed able to lower the oxycline, thus reducing the volume of anoxic water, but didn't reduce the time period of anoxia. A withdrawal was installed in Lake Ballinger (near Seattle) in 1982. It was accompanied by a water diversion directing the inflow stream into the hypolimnion. It eliminated anoxia in the hypolimnion in 3 of the first 4 years of operation (KCM, 1986). This may be a result of inlet modifications which diverted the inflow directly to the hypolimnion, thus ensuring a constant supply of oxygen to the lake bottom. Recently though the effectiveness has been reduced by severe increases in external nutrient loads resulting from the development of the surrounding watershed which have 'swamped' the restoration benefits of the withdrawal (Gibbons, 1993, pers comm). Livingstone and Schanz (1994) present the limnological data collected for 5 years before, and 11 years after the installation of a hypolimnetic withdrawal at Liitzelsee, Switzerland. They observed: (i) a reduction in the total phosphorus content of the lake mostly as a result of a reduction in hypolimnetic phosphorus concentration, (ii) an increase of Secchi depths in the summer and fall, (iii) an increase in the oxygen concentrations in the late summer, (iv) an increase in lake thermal stability in the spring due to retention of the epilimnion during runoff3, and (v) a decrease of stability in the fall due to the slow summer long removal of the hypolimnion. No comment is made if the enhanced stratification in the spring accelerates the onset of anoxia. As well, the decrease of stability in the fall accelerated turnover and is likely responsible for the increased oxygen concentrations in the late summer. Livingstone and Schanz data also highlights the significance of inter-year variability. They present most of their data as seasonal data (i.e. different years overlain together with the day of year as the x-axis.) There are points that diverge from the quoted trends (indicating 3The newly forming epilimnion is usually flushed out during freshet at 11 cm water depth per day. The withdrawal removes 3.5 cm of hypolimnion during this time and 'held' back this amount of epilimnion. Thus, the establishment of stratification, (and stability) was enhanced by the withdrawal. 19 Chapter 2: Literature Review single anomalous years, or seasons). It is only within the context of 5 years of 'before' data and 11 years of 'after' data that the trends become clear. Drawbacks to hypolimnetic withdrawal have focused on the effects of anoxic water on downstream creeks. Anoxic water may contain high levels of iron, manganese, ammonia, and sulfide. Several tactics have been employed at other withdrawal sites to reduce environmental impacts. Niirnberg (1987a) reported one withdrawal installation that used fountains to aerate the withdrawn water, one that discharged the withdrawn water into a treatment plant, and one that mixed epilimnetic water with withdrawal water. These efforts were to reduce iron levels and contain H2S odors. An installation reported by Kortmann et al (1982) experienced iron levels above 1 mg/L, which exceeded a permitted discharge level. As a result the withdrawal inlet was raised to a point higher in the water column where it withdrew water with lower iron concentrations. The anoxic water removed from the Lake Ballinger withdrawal (near Seattle, Wa) resulted in reduced benthic activity, precipitation of hydroxides in the receiving stream (Gibbons, 1993, pers coram), and enhanced periphyton growth (Cooke et al, 1993). As well, the operation resulted in odor complaints from guests at a nearby golf club. The solution was to stop the withdrawal operation during events at the clubhouse, thus reducing the nasal inconvenience (KCM, 1986). 2.3 Chain Lake as Withdrawal Candidate Chain Lake has been the subject of several studies primarily as a result of the different restoration measures that have been attempted. The studies have resulted in relatively intensive data during single seasons, around the time of a management strategy change (Northcote, 1967, Taylor, 1971, Ennis, 1972, Water Investigations Branch, 1977, Murphy and Urciouli, 1984, Murphy, 1985, Weigand, 1989). Some more general studies have been done to understand the nature of water circulation (Mathews, 1982), external nutrient loading (Lacelle, 1986), internal loading (Murphy, 1987), and historical trophic 20 Chapter 2: Literature Review status (Murphy et al, 1990). This historical data set as well as the history of restoration activities make Chain Lake a suitable candidate withdrawal. The hypothesis of anoxic release of internally loaded (particularly iron-bound) phosphorus is well supported in Chain Lake. This makes it a potential candidate. As well, the lake is small enough that an affordable system could be able to export a sufficient volume of lake water to have a positive effect. The earlier attempt to increase the flushing rate of the lake by building the diversion was limited due to engineering and hydrology constraints. Thus it seems reasonable that the most effective use of diversion water would be to flush high nutrient water from the bottom of the lake. 2.3.1 Trophic Status Values of total phosphorus in the late summer can be as high as 300 |ig/L and result in chla levels up to 100 |ig/L. Secchi disk measurements of transparency are often below 1 m in the late summer. Trophic status is somewhat of a subjective classification since all lakes are different, and different methods and measuring scales exist for trophic status (e.g. Carlson, 1977), however, during a summer algae bloom, the lake would certainly be described as eutrophic. Chain Lake has not always been eutrophic. Murphy et al (1990) examined the sediment core diatom record and found that several periods of high and low productivity have occurred, and are associated with changes in the water depth of the lake. The current shift to eutrophic status of the lake appears to have been exacerbated by an increase in lake level in 1957. At this time the water level was raised by about 4 ft. Prior to this date the diatoms present in the lake are those indicative of a less eutrophic environment. The hypothesis is that the deepening of the lake permitted some degree of thermal stratification to occur, which resulted in hypolimnetic oxygen depletion and anoxic release of nutrients from the sediments. High levels of productivity have been associated with fish kills in lakes. The 'usual' cause is that the decay of large amounts of organic matter deplete the oxygen in the lake, 21 Chapter 2: Literature Review increase arnmonia levels and stress the fish to the point of mortality. This could occur in either summer or winter. There are anecdotal records of fish kills at Chain Lake, though they apparently are less common since the installation of the Shinish diversion. The Fisheries Branch files in Penticton contain one documented reference to a summertime fish kill in 1975 (Bull, 1975). The report on the 1975 fish kill is not conclusive as to the cause, as high oxygen levels were found in the lake. Presumably a combination of factors resulted in that event including low oxygen at the bottom of the lake and high temperatures at the surface. 2.3.2 Sediments The bottom sediments of the lake are highly unconsolidated and composed of mostly decaying organic matter. The region between the first sediment-water interface and 'solid' sediments can be over 1 m. This is demonstrated in Figure 2.2 which shows the bottom depth recorded using different weight tools. The region between 'top' and 'firm' sediments is often anoxic, and may be persistently so in spite of aerobic or anaerobic conditions in the water column. This is not uncommon in lakes with high organic contents (e.g. Lee, 1970) The chemical composition of the sediments has been studied with the objective of identifying significant chemical factors in sediment nutrient release (Murphy, 1987). Profiles of iron, phosphorus, water content, and sediment age with depth are shown in Figure 2.3 (after Murphy, 1987). The profiles show significant iron content throughout the sediment depth. The phosphorus profiles indicate that as sediment is buried, it's phosphorus content approaches a constant value. 22 3. M E T H O D S , M A T E R I A L S , and D A T A PROCESSING: 3.1 Withdrawal System The withdrawal system was installed in August, 1993 at the south end of Chain Lake (see Figure 1.1). The withdrawal pipe extends along the lake bottom from the outlet dam to a point 170 m from shore. The pipe is made of high density polyethylene (HDPE) pipe, sold under the brand name PLEXCO by Chevron Chemicals. This is more flexible than PVC pipe and is reputed to withstand internal freezing. The inlet is located at 6.2 m depth within the dredged hole created in 1987/88. The design makes use of a pre-existing drop valve and culvert at the dam site. A schematic of the withdrawal design is shown in Figure 3.1. The withdrawal is designed to flow using gravity and thus requires no mechanical pumps. Water drawn from depth in the lake flows through the pipe and empties into a box/coffer dam structure which surrounds the drop valve. This coffer dam provides an hydraulic seal between the lake and the existing culvert. The driving force through the withdrawal pipe is the difference in elevation between the lake, and the water level inside the box. The head losses within the pipe are small and flows of 80 L/s are possible with differences in water level of 7.5 cm. The water level inside the box is maintained lower than lake level by the continuous draining of the box through the culvert pipe. The system is able to flow by gravity because the pre-existing culvert is located below lake level. From the box, the withdrawn water flows into the drop valve, and through the dam via the existing culvert pipe. A fountain/aerator was added to the end of the culvert pipe. The water flow out of the box is driven by the head difference between the inside of the box, and the water level in the outlet aerator/fountain. The most significant head losses occur through the fountain aerator. This component has decreased the maximum flow to 100 L/s from values approaching 200 L/s without the fountain. Figure 3.2 shows a picture of the box installation, and the aerator fountain as installed. The system was tested during August and September, 1993. Flow tests indicated several leaks between the box and the lake bottom. These were sealed using concrete filled 23 Chapter 3: Methods and Materials burlap bags. A final hydraulic test was made in June 1994 to determine the quality of the seal. In this test, the pipe inlet was temporarily sealed, and the drop valve opened. Water was drained from the box up to a maximum head difference of about 0.5 m. Any leaks detected up to this point were sealed. The withdrawal system was operated for the first year from June to October, 1994. Typical rates were on the order of 35-50 L/s through the summer, but rates as low as 15 L/s were established and maintained. 3.2 Physical Samples Lake water samples were collected with a three litre plastic Van Dorn bottle. The bottle was held vertical in the water and represents approximately a 45 cm sample of the water column. Stream water grab samples were collected from the shore using a 4 litre plastic jug connected to a two meter pole. 3.2.1 Sample Locations The lake sampling and stream sampling stations are shown in Figure 3.1. The lake sampling stations are: • the dredged hole station is the dredged hole created in 1987/88. • the met station is the location of the EFM-UBC meteorological station and thermistor chain located at the south end of the lake. The location is approximately 50 m north of the dredged hole. • the Mid lake station is a sampling point at approximately the mid point of the lake. • the north end station is a sampling point at the north end of the lake. The stream sample locations are defined as: • the Fountain is the fountain outlet • the Spillway is the spillway weir outlet. • the Beaver Dam is a station at the south end of Chain Estates 24 Chapter 3: Methods and Materials • the Jellicoe Road station is the point where Hayes Creek flows through a pair of culverts under Jellicoe Road. Samples were taken from the downstream end of the culvert. • the Diversion is the Shinish Creek at the point where the diversion draws off water. • the Fish Barrier is the fish barrier on Hayes Creek approximately 5 km downstream of Chain Lake. 3.2.2 Handling, Storage, and Preservation From the sample collector, samples were transferred to acid washed, distilled water rinsed, in verted-air-dried plastic bottles for storage and transport. Samples required for determination of soluble components (phosphorus and metals) were filtered through 0.45 |im cellulose filters using a plastic filtration unit by either a hand vacuum pump, a battery operated pump or an A C pump. Until July '94 samples were filtered at the moment of collection. From August '94 samples were collected in a sequence and all filtered after collection. In the former case, filtering was performed within 20 minutes of collection, and in the latter case within two hours of collection. Samples to be analyzed for total phosphorus, soluble phosphorus, nitrate, and ammonia were preserved to pH less than 2 with ~2 mL per L of 25% H 2 S O 4 (Standard Methods, 1992). A sub sample of the filtered sample (without acid preservation) was frozen in a 30 mL or 60 mL Nalgene bottle for analysis for soluble reactive phosphorus (SRP). Samples for iron and manganese analysis were preserved to pH less than 2 with concentrated nitric or 25% sulphuric acid. The standard method implies nitric acid, however extensive duplicate comparisons indicate that for the Chain Lake samples the different preservations do not affect the determined value. (See Section 3.7 and Appendix A for sample preservation quality control). Samples for chl a analysis were filtered through a GFC filter (pore size = 1.2 [im retention size), and the filters frozen. Two different volumes (e.g. 50 mL and 100 mL) of the same sample grab were filtered to act as extraction and analysis duplicates. 25 Chapter 3: Methods and Materials 3.2.3 Sample Analysis Total phosphorus (TP) and total soluble phosphorus (TSP) were determined by digestion of the unfiltered and filtered samples respectively using a 5:1 nitric:sulphuric acid solution, and spectrophotometric analysis using stannous chloride/ammonium molybdate reagents at 690 nm (Standard Methods, 1992). Soluble reactive phosphorus analysis was performed in the Oceanography Department at U B C by ammonium molybdate-ascorbic acid method using a Technicon auto-analyser II. Ammonia was determined using a Lachat QuikChem A E auto-analyzer in the Environmental Engineering Lab of the C i v i l Engineering Department at U B C and the method of salicylate/hypochlorite reaction (QuikChem method 10-107-06-1-Z). Prior to analysis all samples were p H adjusted to three. Nitrate + nitrite was determined using the same Lachat QuikChem A E auto-analyzer by the method of cadmium reduction (QuikChem method 10-107-04-1-Z). Samples for metals analysis, both unfiltered (total metals) and filtered (soluble metals) were digested with nitric acid, concentrated volumetrically 5 to 1, and then analyzed simultaneously for iron and manganese using a dual beam Thermo Jarrell Ash Flame Atomic Absorption Analyzer. 3.3 Dissolved Oxygen Dissolved oxygen profiles were taken regularly through the two years using a Mode l 57 Y S I dissolved oxygen probe. On an occasional basis in .1993 a Hydrolab® multilogger was available to take profiles. 3.4 Electronic Data 3.4.1 Meteorological Data Meteorological data was collected using a floating raft station owned by the Environmental F lu id Mechanics ( E F M ) group at U B C . The E F M met station was positioned at the south end of the lake just north of the dredged hole area. At regular scan times the station recorded instantaneous readings of air temperature, humidity, incoming and reflected 26 Chapter 3: Methods and Materials radiation, and wind direction. Wind speed was averaged over the entire scan interval and recorded at the scan time. Data was collected from May 6 to November 5, 1993 and from A p r i l 22 to September 9, 1994. The record is continuous except for the period from June 12 to July 3, 1993. Most of the data were collected at 30 minute intervals, however short tests were done at 10 and 5 minute intervals for comparison with other logging equipment. A second weather station, on loan from the National Water Research Institute was deployed in August 1994 until September 15. This station provided a verification of the data collected by the E F M station and was used in a simultaneous thermistor monitoring experiment to observe the physical seiching behavior of the lake. 3.4.2 Thermistor Data The E F M met station also includes a chain of six thermistors which hang from the float and record instantaneous water temperature at each scan interval. The thermistors recorded water temperature at depths of 0.5, 1.5, 2.5, 3.5, 5.0, and 6.5 m. The last thermistor is resting in the unconsolidated sediments at the bottom of the lake. The N W R I met station included 6 thermistors which were placed at the same depths. Finally a set of four stand alone WaDaR thermistors were deployed from mid August to mid September at the mid lake station. The four were deployed at 1.5, 2.5, 3.5, and 5.0 m water depths. Two others were used to monitor the Shinish Creek diversion inflow and withdrawal water temperature during this period. 3.5 Hydrologic Data Flow rate measurements were made on the Hayes Creek inflow, the Shinish Creek diversion, and on the hypolimnetic withdrawal flow. A l l hydrology measurements were made at discrete time points - i.e. no continuous recorders were available. The Hayes creek flow was measured using a propeller type velocity gauge (manufactured by A . Ott) at the site of the culverts at the north end of the lake. A t this location, no flow was measured during 27 Chapter 3: Methods and Materials 1994. The hypolimnetic flow rate was measured using the same type velocity meter in the dam culvert just at the point of junction with the fountain/elbow. Measurements were made at several places across the cross section of the flowing pipe and averaged to produce an average velocity. This velocity enabled computation of the volumetric flow rate. The diversion flow was measured using a 90 degree V-notch weir and water depth marker positioned in the flume box located along the ditch just before the point of entry to the lake. 3.6 Data Processing 3.6.1 Time Series Data The E F M meteorological station collected data from the south end of the lake, near the inlet to the withdrawal pipe. As a result, temperature observations are not always representative of the entire lake. When the lake is oscillating through a seiche period the temperature at the thermistors will rise and fall as the thermistors encounter alternatively surface and bottom water. As the thermocline rises at one end of the lake, then more of the thermistors in the chain are 'submerged' in cooler water. This may occur without a change in total heat content of the lake. To account for this, two time scales of analysis are employed with reference to lake temperatures. Daily average temperatures are used for examining data over time scales of days or longer. This is useful for examining general lake stability. To study the shorter term effects such as seiching and diurnal heating and cooling, either the raw data or hourly averaged data are used. Daily averaging filters the effects of seiching observed at a thermistor. This averaging also filters out the diurnal effects of daytime heating and nighttime (or weather induced) cooling. These effects may be considerable as the surface waters were observed to vary by several degrees during a single day. The seiching features of the lake are often established on times scales of T/4, and as a result the wind forcing is often averaged on time scales approximating this period. 28 Chapter 3: Methods and Materials To demonstrate the effects of averaging, a sample extract of the thermistor record from the warmest period of 1994 is shown in Figure 3.4. Also shown are the average daily temperatures. The surface of the water can be seen to vary by several degrees. The sediment surface shows virtually no daily variation in temperature. The daily fluctuation in temperature at the surface (i.e. 0.5 m thermistor) is greater than 2 °C as a result of diurnal heating and cooling. The daily variation within the water column, however, may be up to 4 °C per day due to seiching (e.g. the 2.5 m thermistor). This can be as almost as great as the total stratification represented by daily averaging which from 0.5 to 5.0 m is only 6 °C. Daily fluctuations are important to bear in mind when examining daily average plots. Over the course on a day, the surface temperature varies more than the bottom water temperature. Thus the 'daily average' temperature tends to underestimate the difference between top and bottom (i.e. strength of stratification) during the afternoon, when we expect to experience stronger wind events. 3.6.2 Areal Data Averaging Lake samples were typically taken at three stations along the length of the lake, at several different depths. This repetition allowed any areal variation to be detected, and reduced uncertainty of analysis by increasing the number of samples. In order to produce seasonal contours showing concentrations varying with depth and time, the horizontal dimension was removed. That is, horizontal variations were not considered significant in seasonal analysis. Samples were averaged along the horizontal dimension, for each depth, (i.e. the 1 m samples averaged, the 3 m samples, etc.). This is acceptable since in many cases significant differences were observed vertically, while significant horizontal gradients were not expected or observed. An example is shown in Figure 3.5, which shows analysis values for the samples taken on July 23, 1994. The samples collected this day show increasing concentrations of sediment related compounds (phosphorus, ammonia, iron and manganese) with increasing depth. Particularly dramatic are the elevated concentrations shown in the 29 Chapter 3: Methods and Materials dredged hole area. There is a statistically significant difference between the averages at different depths, but not the averages at different stations (based on a t test of the two means, Walpole and Meyers, 1978). Maintaining vertical definition during averaging allows for stratification effects to be retained when they do occur, and does not reduce accuracy when the lake is destratified. 3.7 Quality Assurance / Quality Control (QA/QC) Extensive Quality Assurance (QA) and Quality Control (QC) procedures were maintained during the laboratory analysis procedures. This was required because several of the chemical species analyzed were expected to occur in trace amounts (i.e. ppb range) and because the lab analysis required extensive digestion and sample handling which represent a possible source of contamination. Proper QA/QC provides some estimation of the confidence limits of the analyses. The discussion here is focused on the analyses for Fe, Mn, and Phosphorus (TP and TSP). Appendix A details the QA/QC program, defines the terms used, and provides full data tables. 3.7.1 Iron and Manganese Manufacturers specifications of machine sensitivity for the Atomic Absorption analyses are reported as 0.040 mg/L for iron and 0.020 mg/L for manganese (Thermo Jarrel Ash, 1986). The samples were concentrated 5:1 so the sensitivity to original sample concentrations is -0.008 mg/L and =0.004 mg/L for iron and manganese, respectively. The M D L analysis (Table A - l of Appendix A) indicates an average M D L of 0.034 mg/L for Fe and 0.016 mg/L for Mn (based on three M D L evaluations). These values are similar to other reported values (e.g. Stauffer, 1987). Also important in trace analysis is the method of preservation. Standard Methods (1992) prescribes nitric acid preservation for metals analysis, though sulphuric preservation is not uncommon. A series of replicate samples were taken, one preserved with nitric, and one 30 Chapter 3: Methods and Materials preserved with sulfuric acid. The results are shown in Table A-7 (of Appendix A) and indicate that the preservations yield similar results. Standard Methods (1992) defines replicate analysis quality as 'acceptable' when determined values are ±15 %, and as 'excellent' when values are ±5 %. A summary of the QA/QC program for iron is shown in Table 3.1 and for manganese in Table 3.2. These tables summarize the number of acceptable and excellent replicates. The replicates indicate no systematic errors and close agreement with the original samples (most replicates are acceptable or better). The manganese results appear somewhat better than the iron results. This is expected as the A A response for manganese is stronger than for iron (Thermo Jarrel Ash, 1986) Table 3.1: Summary of Iron QA/QC Procedure Total Number # of # of ref table in Performed acceptable excellent App. A Replicate Analysis Triplicate Analysis Replicate Sampling Triplicate Sampling Spikes Nitric vs. Sulphuric 7 6 ' 4 A-2 2 2 1 A-3 6 5 1 A-4 2 2 1 A-5 10 10 1 A-6 36 33 17 A-7 Table 3.2: Summary of Manganese QA/QC Parameter Total Number # of # of ref table in Performed acceptable excellent App. A Replicate Analysis Triplicate Analysis Replicate Sampling Triplicate Sampling Spikes Nitric vs. Sulphuric 7 6 5 A-2 2 2 2 A-3 6 4 2 A-4 2 1 1 A-5 10 10 7 A-6 36 28 19 A-7 31 Chapter 3: Methods and Materials 3.7.2 Digested Phosphorus An important component of this work is the evaluation of phosphorus. A manual digestion and spectrophotometric technique was used. This presents the possibility of sample contamination. As a result, extensive check standards, replicates, spikes and standard additions were used to ensure quality control. Samples were analyzed in batches of 24 flasks, some devoted to standards, replicates, and spikes. In total, over 40% of the 'samples' run were standards, replicates, and spikes. Method Detection Limit (MDL) evaluations were run using the EPA method (1988) (see Appendix A). In it seven samples or standards near the detection limit are analyzed by the analytical procedure and the variance in the results used to estimate the M D L 1 Based on the EPA method an M D L value for digestion and evaluation of phosphorus varies from 8 to 40 ug/L phosphorus depending on the run (eight different M D L series were run). A point to consider here is that the M D L procedure is based on a normal distribution of replicate values. In the M D L runs, the distributions seemed skewed to the positive side. The technique is spectrophotometric and any contamination is likely to produce a higher absorbance (ex: due to scattering). This is supported by the low level M D L runs which indicate that the method is slightly biased high at near M D L levels (=6-12 |i.g/L). This skewed distribution means that samples at the estimated three times standard deviation M D L level actually are greater than the 99% confidence level. In summary, the calculated M D L (of approximately 25 |ig/L) does represent a confidence level for quantification, whereas a level much lower level is probably accurate as a 99% confidence detection level (say 10 itg/L). That is, we can say that detected levels between 10 and 25 are positively detected, though not accurately quantified. The phosphorus analysis was performed in batches of 24-28 flasks for digestion, neutralization, and reaction. A summary of the runs made in 1994 is shown in Table 3.3. 1 The EPA method defines the MDL as the standard deviation of the n replicates times the t statistic for n-1 degrees of freedom. For seven replicates, the t statistic is 3.174 which is gives rise to the "three times standard deviation" rule for MDL. 32 • Chapter 3: Methods and Materials Table 3.3: Summary of Phosphorus Runs - 1994 Total Number of Runs 25 Number of Samples 244 Number of Replicates 18 Number of Spikes / Std Additions 9 Number of Calibration Standards 75 Number of Check Standards 74 A regular program of spikes and replicates was run for the phosphorus analysis, similar to the iron and manganese analysis. The results are summarized in Table 2.4. They show good recovery of spikes, but less accurate results for replicates. The low number of 'excellent' results indicates a relatively large variability in the analysis procedure. This is unfortunate, but unavoidable with this method. Table 3.4: Summary of Digested Phosphorus (Soluble and Total) QA/QC - 1994 Parameter Total Number #of #of ref table in Performed acceptable' excellent' App. A Replicate Analysis 20 3 3 A-9 Triplicate Analysis 12 6 2 A-10 Replicate Sampling 5 1 0 A - l l Spikes 10 9 3 A-12 The high variability associated with the phosphorus analysis was a major factor in designing the sampling program to cover three lake stations. To better understand this variability of sampling and handling, a final evaluation of variability was run at the end of the 1994 season. It was a sampling, storage and analysis evaluation to determine the sources of variation. Four samples were taken sequentially from the same depth in the lake, each sample was split into four identically prepared containers for preservation, transport and storage. Each container was sub-sampled and analyzed four times. This 4 x 4 x 4 matrix of results allows an analysis of variance (ANOVA) be used to determine sources of variation. 33 Chapter 3: Methods and Materials The results of the A N O V A analysis are presented in Table A13 of Appendix A. This analysis indicates that different grabs, and different analysis runs can be significant sources of variation (at the 5% level). This should be taken with caution since the different grabs were taken one after another. The sampling itself may disrupt the water column (ex: the weight on the bottom of the van Dorn is 1 m below the sample point and may have disturbed the sediments). The total of 64 analyses, each representing the concentration in the water column at that time, had an average value of 88 |ig/L, and a standard deviation of 13.3 \ig/L (coefficient of variation =15 %). This variability is not unusual for this type of analysis. Standard Methods (1992) documents a round robin test of 20 labs analyzing a check sample with a total phosphorus content of 210 fig/L. They reported a coefficient of variation of 20.8% (standard deviation = 44 |lg/L). The variability in this study is less than that reported by Standard Methods, and at lower concentrations. 34 4. OBSERVATIONS and RESULTS: 4.1 Meteorological Data 4.1.1 Regional Weather A weather record has been maintained at Princeton (35 km to the SW of Chain Lake) for 58 years (AES, 1995). This location is comparable to Chain Lake because of its close proximity, and the absence of major geographical features such as mountains between the two locations. A plot of the 58-year mean of daily average air temperatures at Princeton through the summer is shown in Figure 4.1. Daily average temperature is the average of a daily maximum and a daily minimum recorded on the day. The 58-year mean daily temperature is the mean of the 58 daily average temperature values recorded on that calendar day. The 58-year mean temperature increases through the summer until the end of July and then begins to decline in early August. The early onset of cooling is likely because Chain Lake is at reasonably high elevation (1000 m) and in an interior location. The plot of 58 year mean daily average temperatures is overlain with the daily average temperature recorded during the study years 1993 and 1994. This provides insight into the degree of anomaly of each year. The first year, 1993, was characterized by a very warm spring with temperatures as much as 10 °C above average during May, variable but approximately average temperatures in June, and below average temperatures in July. In 1994, there was again a warm period in May, though much shorter in duration than the previous year, and approximately average temperatures in June. During July 1994 high temperatures were recorded of up to 10 °C above average. In fact, during the third week of July, daily maximum temperatures were between 35 and 40 °C (not shown), establishing several new records at the Princeton Station1. The period of July 1994 was extremely dry in much of southern B.C. and was characterized by large forest fires in the Penticton, Hedley and Tulameen areas, among others. 35 Chapter 4: Observations and Results Also shown in Figure 4.1 is the monthly precipitation recorded at Princeton (again the 58 year mean overlain with the 1993 and 1994 data). The times of greatest precipitation are November to January, and June (though June precipitation is only marginally greater than May, July, or August). The 1993 and 1994 plots show how variable the precipitation may be in individual years. In 1993 Jan. to April precipitation was below average and the summer precipitation was well above average (e.g. over 80 mm in July vs. 25 mm average). In 1994 the precipitation did not follow a trend relative to the average - it varied above and below from April to October. Of particular interest is the July precipitation which was less than 10 mm, far below the 25 mm average. The temperature and precipitation plots are not independent of each other. It seems reasonable that cloudy or rainy days would be cooler than clear ones, or that record hot temperatures would not occur on overcast or rainy days. Thus, July 1993 was a rainy month and also a cold month while July 1994 was a hot, dry month. Since air temperatures begin to decline in early August, July appears to be a defining month in determining the 'nature' of the summer season. The two field seasons at Chain Lake may be qualitatively defined as a cold wet year, though with a warm spring (1993), and a hot, dry year (1994). Chain Lake is in the uplands areas of the watershed, and receives most of it's inflow from surface run-off and springs, and not through flow from Hayes Creek. Compounding the dry, wet year of 1994 was a low snow pack at the start of the season (Mathews, pers coram, 1994), which resulted in smaller than normal snow melt flows during May and June. 4.1.2 Local Meteorological Forcing The data collected by the E F M met station provide a record of the local meteorological conditions at the lake. A summary plot of solar radiation, air temperature and wind speed recorded at the lake is shown for 1994 in Figure 4.2. Also shown is the daily 36 Chapter 4: Observations and Results precipitation record from Princeton2. This plot highlights the important events during the 1994 summer of withdrawal operation. The warm period in the beginning of May is evident on the temperature plot. The wind is variable with several strong events in May and June. During July, there is a noticeable increase in air temperatures and dramatic decrease in wind speed. July is also a month with virtually no precipitation. These data together indicate that the hot, dry weather, with no cloud cover that was measured in Princeton (Figure 4.1) also occurred at the field site. These forcing functions form the basis for the interpretation of lake response during the summer. 4.2 Water Temperatures The daily average water temperatures recorded by the EFM-met station thermistors are shown in Fig 4.3 for the two years of data collection. These plots show the thermal stratification. That is, when the temperature traces diverge, there is stratification in the water column. In neither year did a strong season-long stratification develop. There are several periods of stratification during each summer, with destratification events occurring between them. Common to both years are stratification periods in May. This effect was especially pronounced in 1993 and coincides with the above average air temperatures observed in Princeton and at the lake. In spite of the strong stratification in May, cooler weather and/or rain in June resulted in destratification. This occurred in both years. During July 1993 little stratification occurred - a cold, wet month. Short periods of weak stratification are evident in August and September. In 1994, several stratification events were observed. The longest commenced on July 5 t h, increased until the 24 t h, then weakened slightly until August the 7 t h. The strength of stratification increased again slightly, then waned steadily and full mixing occurred by the 2 No daily or continuous rain guages were established at the lake site, so the precipitation record at Princeton is taken as applicable. 37 Chapter 4: Observations and Results middle of August. This is the longest period of stratification observed, and directly correlates to the high temperatures and low wind speeds observed at the lake (shown in Figure 4.2). The daily temperature plots (Figure 4.3) indicate the polymictic nature of Chain Lake. Stratifications develop within several days, and disappear as quickly. Comparison of the two years of field data hints at the wide variability between years. That is, the shapes of the two plots are different with 1993 warm early in the year and maintaining constant temperature, forming a 'flat topped' plot, and 1994 showing a constant increase in temperature to a peak, followed by a continuous decline, forming a peaked plot. The two plots indicate the inter-year variation which makes generalizations about the time of stratification difficult. Also shown from the water temperature plots is a decline in temperature beginning in early August, coincident with the decline in air temperatures. The maximum air temperature actually occurs at the end of July, slightly before the decline in water temperature. This reflects the thermal inertia of the lake. 4.3 Hypolimnetic Withdrawal Characteristics Composition of the withdrawal flow during 1994, measured at the aerator-fountain is shown in Figure 4.4. Withdrawal flow was initiated on June 20. Flow was maintained continuously until Oct. 5, 1994, though the rates were adjusted a couple of times. The normal flow rate was about 35 L/s, though during July was reduced to 15 L/s for a.short period. The temperature of the fountain release was 14.5 °C at start-up on June 20 and rose to 18.5 °C during the mid-summer. From mid August until closure temperatures fell gradually to 13 °C. Some continuous monitoring was done which indicated that the withdrawal temperature did not vary diurnally. In comparison, the lake surface water temperature often varied by several degrees each day. Dissolved oxygen concentrations are low from the start-up until the end of August. The observed concentrations do not reach zero due to some leaks between the box and the 38 Chapter 4: Observations and Results lake. That is, some surface water is mixed with the withdrawal water in the box before release through the aerator fountain. The oxygen data indicate that the bottom of the lake either around 6 m depth, or at least in the dredged hole is a persistently anoxic region throughout the summer. The short term stratifications and destratifications indicated by the daily temperature plots in Figure 4.3 may not have an effect on the oxygen concentration at the withdrawal depth (i.e. at 6.2 m). The fall turnover is complete, however, as the oxygen level has risen to near saturation value by September. The pH record is crude and not complete. Operational problems prevented accurate recording of pH on all site visits. The values collected do show a range of pH values from near 7 at the start of July to 9 in October. This coincides with our expectations, as anoxic conditions are accompanied by a decrease in pH, and aerobic ones by an increase (Wetzel, 1975). Phosphorus concentration in the withdrawal flow increases during the July stratification. Total phosphorus (TP) at the beginning of operation was less than 50 |J.g/L, and increased during July stratification to greater than 300 |ig/L. Matching this were increases in total soluble phosphorus (TSP), and the soluble reactive phosphorus (SRP). Based on the simple iron bound model, it is expected that the increase in total phosphorus would be composed of mostly SRP (i.e., the PO4 ion) which is typically adsorbed to the iron-hydroxide precipitate. The data show that much of the phosphorus is not in SRP form. This could be due to several factors. First an active algae population deep in the lake may be incorporating the phosphorus as quickly as it is released. If the system were phosphorus limited, then the algae would accumulate near the point of phosphorus release - the sediments. Another reason may be chemical adsorption either occurring in the box, or more likely in the water column before withdrawal. This is not predicted under anoxic conditions by the standard oxic-anoxic model, though caveats are often made to this model that iron complexation with organic matter and bicarbonate may complicate the processes. Thirdly, the possibility of analytical error cannot be neglected as the samples were frozen and thawed 2-4 weeks later 39 Chapter 4: Observations and Results for analysis. Coagulation and flocculation reactions may have occurred during freezing that could remove the SRP from reactive form. Nitrogen was measured as total ammonia, and N O x , which is nitrate plus nitrite (NO3 + NO2). N O x values were low throughout the year however, ammonia levels rose twice during the year. The first increase in ammonia is associated with the anoxic conditions of the July stratification with values rising to nearly 0.8 mg/L. The reducing conditions in the dredged hole result in the formation of ammonia. The second increase in ammonia occurs when the withdrawal flow is well aerated. Thus, it is not likely that it is from an 'anoxic source' - meaning as a chemical reduction of nitrate. More likely is that during the fall, the watershed is loaded with a lot of organic matter (e.g. leaves, settling algae, animal feces) which breaks down quickly to release proteins and amino acids as organic nitrogen. This organic nitrogen breaks down to form ammonia. Iron and manganese concentrations both increased dramatically during the summer stratification of 1994. In June and September iron levels are less than 0.30 mg/L and manganese levels around 0.05 mg/L. During the summer stratification total iron levels increased to over 1 mg/L with a single value of 2.7 mg/L on August 8. Manganese values increased to over 0.3 mg/L with a maximum value of 0.53 mg/L on August 8. The increase from the beginning of operation to the maximum measured value represents a 9 fold increase in iron and an 11 fold increase in manganese concentrations. A large fraction of the iron is in a non soluble form at the withdrawal fountain even when the total iron content is > 1 mg/L. The high total iron level indicates anoxic release and this iron should be in reduced F e 2 + (i.e. soluble) form. The observation of a large non soluble component could be the result of fast oxidation in the box, the fountain, or the sample container prior to filtering. The reaction time for the oxidation of F e 2 + to F e 3 + is on the order of 40 minutes to 1 hour (Imboden and Schwarzenbach, 1985). Although this occurrence may seem unlikely as the residence time in the 'box' is only about ten minutes, the residence time is enough to increase the dissolved oxygen content from zero to about one mg/L. Another 40 Chapter 4: Observations and Results cause for non soluble iron could be the presence or introduction of some aerobic water to the dredged hole region (movement via seiching perhaps, or the Shinish diversion). It has been observed that because iron is so quickly oxidized, that it is rarely observed in soluble form in an aerobic water column. That is, the soluble F e 2 + may migrate away from the sediment region, but once it contacts aerated water, it quickly forms precipitate and is detected as particulate. The supposition then, is that the water in the region of the withdrawal inlet may mix with enough aerated water to partially oxidize the iron, but not enough oxygen is mixed in to leave a dissolved oxygen 'residual'. Conversely, manganese oxidation reactions are slow, on the order of days or weeks (Stokes et al, 1988). Thus any liberated manganese will stay in soluble form, at least until withdrawn. This is indicated by the fact that virtually all of the withdrawn manganese is soluble. All soluble manganese is generally regarded as being M n 2 + (Delfino and Lee, 1968, 1971, Gordon etal, 1983). 4.4 Lake Physics This section describes the physical and hydrodynamic observations made during the field program. These observations are discussed here because the physical features of the system have an effect on, and are reflected in, the chemical and biological behavior of the lake. Though this section presents all the observations made emphasis will be placed where the behavior has implications for the hypolimnetic withdrawal. 4.4.1 Diversion Inflow / Withdrawal Outflow Water Temperatures Water from the Shinish Creek enters the lake along the east side approximately 200 m from the dam at the south end. Whether the Shinish water circulates widely through the lake, or short circuits and flows directly out over the spillway has been debated in the past (Mathews, 1982). The hypolimnetic withdrawal has the potential to exacerbate any short circuiting by removing the Shinish diversion water directly from the bottom of the lake. This 41 Chapter 4: Observations and Results project was unable to quantify the potential for short circuiting, however, previous measurements permit the issue to be addressed. Mathews (1982) measured the areal and vertical distribution of temperature and conductivity in the lake. The temperature data showed that Shinish diversion water was much cooler than the lake (diversion at 8.2 °C, lake at about 15 °C on test day), yet this temperature difference was undetectable as a tracer of the Shinish water even a few tens of meters offshore. Similarly the difference in conductivity between the diversion (24 uS/cm) and the lake (130 |0,S/cm) was not suitable as a tracer either. Mathews also conducted two fluorescein dye tests. These tests were able to track the Shinish water and showed that Shinish water from the diversion did circulate over almost half the lake area within 24 hours, and mostly below 2 m water depth. During the 1994 field season temperatures of the Shinish Creek at the point of the diversion and the hypolimnetic withdrawal outflow were measured. These are shown in Figure 4.5. A plot of single point measurements taken during the creek sampling runs demonstrates that the diversion is always cooler than the withdrawal (Figure 4.5 a). Thus we would expect the diversion water to sink to the bottom of the lake, and even to the bottom of the dredged hole underneath the withdrawal inlet. However, it is reasonable to expect mixing to occur upon lake entry because the entry point is delta shaped which 'spreads' the inflow. As well, the temperature of the diversion may warm during the 2 mile transit to the lake. A temperature reading taken on July 23, 1994 shows that the flow had warmed by 3 °C (not shown in Figure 4.5a). This likely represents a maximum value for temperature increase as July was the hottest month of the year. It is reasonable then that a portion of the diversion water inflow would sink to the bottom of the lake, and spread over the bottom of the south end of the lake. Some of this water must travel to the dredged hole area. It is plausible that this water could cause the partial oxidation which resulted in the low soluble iron values observed at the fountain, despite the anoxic conditions. 42 Chapter 4: Observations and Results Part of the record is a detailed times series obtained using WaDaR® thermistors from August 17 to Sept. 12, measuring the actual diversion inflow temperature (Figure 4.5b). This period is a time of turnover and mixing in the lake with nighttime temperatures often at or below dew point. The WaDaR® record highlights the complexity of the inflow dynamics. The diversion flow varies by over 5 °C daily, which would alter the depth it would fall to in the lake. Note that the diversion entry to the lake is consistently 5 °C or more cooler than the withdrawal water (not just the surface water!), indicating that it should plunge to below the withdrawal depth. The Shinish inflow then brings cool water saturated with dissolved oxygen into the lake. If that water did short circuit to the withdrawal, it was not detected during the strong stratification of 1994. The reason for this is that the O2 supply would have been small compared to the sediment and hypolimnetic oxygen demands (SOD and HOD, respectively). For example, a typical inflow of 50 L/s, with a DO of 10 mg/L, represents an oxygen influx of 43.2 kg/day. If 1/2 of this oxygen rich water sank to the bottom of the water column, and covered the south 1/3 of the lake bottom this would be a supply of 0.15 g O2/ m2/day to the sediment surface. In contrast, Cole and Buchak (1993) compiled a range of SOD values for water quality simulation. They found SOD rates to be in the range of 0.1 to 1.0 g O2/ m2/day but that values greater than 5.0 have been documented. Murphy's Chain Lake core incubations indicated SOD of about 0.3 g O2/ m2/day, more than double our estimated supply rate. This does not even account for the hypolimnetic oxygen demand (HOD) which may be substantial and would consume oxygen from diversion water that mixes and doesn't sink completely to the sediment surface. These order of magnitude numbers indicate that the Shinish diversion alone is not able to maintain oxygen levels at the lake bottom during a period of strong stratification and warm temperature. 43 Chapter 4: Observations and Results 4.4.2 Surface Mixing On several occasions, CTD transects were conducted along the length of the lake. These provide insight into the response of the thermal stratification to wind forcing. An example taken on July 7, 1994 is shown in Figure 4.6. Winds are weak from midnight until noon. During this time the lake surface warms uniformly due to solar radiation. At 1:00 PM the wind starts blowing at about 2 m/s from the south. A series of five profiles were taken along the length of the lake between 1:33 PM and 2:17 PM. These show that the surface water has been driven downwind, and stirred into an isothermal region of about 18.25 °C. The north end (downwind end) profile indicates a mixed layer to a depth of 4 m, while the south end profile (upwind end) shows a continuous stratification at the surface and a relatively constant temperature 'hypolimnion' from about 3 m down, at a temperature around 16 °C. This hypolimnion is clearly evident in the profiles (Figure 4.6 b). The southern most profile, taken in the dredged hole, shows a thermocline at about 7 m where the temperature drops from 16 °C to less than 14 °C. A contour plot of the profile information (Figure 4.6 c) demonstrates the wind set-up effect. The wind stress has resulted in a tilted thermal structure, overlain by a wedge of 'epilimnion' on the top of the lake. The volume of the wedge is about a third of the lake which upon relaxation of the wind would cover the top 2 m of the lake. This tilted thermal structure, and mixed layer was observed approximately 1-2 hours after the commencement of the wind demonstrating that the stratifications of the surface 1-2 m can be mixed up quickly. 4.4.3 Wind Forcing - Seiching The thermistor data set provides insight into the seiching behavior of the lake. The thermistor chain is located at the south end of the lake, and so the rising and falling of the thermal structure is detected by the thermistors which are exposed alternatively to cooler and warmer water. An extract from the thermistor record over the period July 24 to July 27 is shown in Figure 4.7. Also shown is the FFT spectral density of a 5.3 day (i.e. 256 point) 44 Chapter 4: Observations and Results times series extract beginning of July 24. The thermistor time series data show the surface of the lake heating and cooling in response to diurnal forcing. Water temperatures at the 0.5 m thermistor are a minimum at about 7:00 or 8:00 A M and a maximum at about 5:00 to 6:00 PM. Spectral density of the 0.5 m thermistor trace show a dominant period of 24 hours (shown in Figure 4.7 b). The deeper thermistors show a different response than the surface thermistor. Firstly, the variations in temperature are greater than those observed at the surface. For example, the temperature at 3.5 m varies 4 °C while the surface temperature varies by ~2 °C. Secondly, the spectral density for the 3.5 m thermistor shows a dominant period of 10 hours, and a secondary period of 24 hours. If all temperature variations were the result of a daily cycle of heating and cooling through the surface then the dominant period at all thermistors would be 24 hours, and the magnitude of variation would decrease with depth. This is not the case and is explained easily as seiching of the water column. In Figure 4.7 the spectral densities are plotted on a log scale and the magnitude of the 10 hour peak represents three times the energy of the 24 hour peak. An important feature of this observation is that the dominant frequency is so clearly at 10 hours indicating a seiching (or at least non-diurnal) response, and not 12 hours which might suggest some variation on a daily cycle. To compare theoretical and observed seiche periods requires an estimate of the thickness of the mixed layer. The previous discussion of surface layer mixing highlighted that the top 2-3 m of the lake will mix very quickly under wind stress. The time series trace also indicates the thickness of the mixed layer. When the thermocline is depressed at the met station, all the thermistors are at approximately the same temperature. That is, when the temperature of the water column rises (see Figure 4.7 a), the top thermistors are at the same temperature, indicating a homogeneous surface layer. For example, at the beginning of day 207 the top 4 thermistors (to a depth of 3.5 m) are all at 22 °C, indicating a well mixed zone to at least 3.5 m. The temperature at 5 m rises slightly at these times indicating that the depth 45 Chapter 4: Observations and Results of the thermal tilt may be as deep as 5 m. Alternatively, when the thermocline rises at the met station (i.e. falling temperatures), the lower depth temperatures diverge indicating stratification in the lower layer. At this point only the 0.5 and sometimes the 1.5 m thermistor are in surface water. If we assume a linear thermocline during the seiching, and that the depth varies from about 1 to about 4 m, the average depth of the surface layer is 2.5 m. This is consistent with the surface mixing response observed during in the CTD transects. A comparison of the theoretical and observed seiching period for the extracted period in Figure 4.7 is shown in Table 4.1. The observed seiche period is 10 hours whereas the theoretical two layer and linear stratified periods are 7.5 and 11.5 hours, respectively. It appears that the natural system is somewhere between the two. This is reasonable as the actual system is neither two layered nor continuously stratified but rather composed of a well mixed surface layer over a continuously stratified lower layer. The period of a two layer system is increased by the presence of a finite thermocline thickness between the layers (Thorpe, 1968). It follows that the stratification below the surface layer would also increase the seiche period. Table 4.1: Expected and Observed Seiching Period July 24-29,1994 Observed Theoretical Two-Layer Theoretical Continuous Surface Temperature (°C) 22 22 Bottom Temperature (°C) 17 17 Ap (kg/m3) 1.005 1.005 g' (m/s2) 10 x IO"3 h(m) =2.5 H(m) 6 6 L(m) 1600 1600 Az (m) 6 N 2 (rad/s) 40 x IO"3 Seiche Period (hr) 10.0 7.4 11.5 Notes: (1) Temperature at 5 m depth taken as bottom temperature (2) Two layer period from equations 2.5 and 2.6 (3) linear stratification period from equations 2.5, 2.8, and 2.10 46 Chapter 4: Observations and Results The observations can be broadened to include the range of stratifications observed. A suitable mixed layer depth need not be estimated for each excerpt. We can expect the top 2-3 m to be mixed, and from equation 2.6, the two-layer seiche period is relatively insensitive to mixed layer depth when h=H/2. For mixed layer depth from 1.75 to 4.25 m, the theoretical seiche period is between 7.3 and 8 hours. Only when one of the layers becomes small does the seiche period change rapidly with depth. If we use half the lake depth as the mixed layer depth, we will calculate the minimum two layer period for a given stratification. The calculation is made using daily average temperatures, and assuming that the 1.5 m thermistors represents the average mixed layer temperature during seiching, and the 5.0 m thermistor represents the average lower layer temperature. The expected linear stratification period is estimated using the temperature difference between the 0.5 m and 5.0 m thermistors. Together it is expected that these two estimates would bound the actual behavior. These periods are shown for the entire 1994 season in Figure 4.8. The seiche period is not a smooth seasonal function. It changes rapidly. This is not surprising since the seiche period is a function of the lake stratification, and we have already observed that stratifications on Chain Lake are formed and removed over time scales of a few days. For comparison several 128 point (2.6 day) extracts from the time series data were analyzed by FFT to determine the dominant response periods. The extract periods were chosen when the expected seiche was short enough to be discernible from the seasonal warming and the diurnal fluctuations of the lake. Shown plotted for each of these time extracts are the FFT components at the appropriate period, with the plotting symbol scaled in size by the magnitude of the FFT coefficient. The FFT components do reveal some interesting features. This first is a strong diurnal cycle - as was observed in the 5 day extract (Figure 4.7). In the seasonal plot of Figure 4.8, the 24 hour energy is shared between the 32 and 21.3 hour bins. Apart from the diurnal features, the observed FFT components agree well with the theoretical seiche periods, often falling between the two estimates curves. 47 Chapter 4: Observations and Results This representation also highlights a difficulty of using spectral analysis on this type of data. The short time periods sampled result in broad FFT bins at low frequencies. For example, the third to eighth bins represent periods of 21.3, 16.0, 12.8, 10.7, 9.1, and 8.0 hours respectively. This makes it difficult to discern between 7, 10 and 12 hours movements. The only remedy for this situation is to sample for a longer time period during which the stratification may change, or the wind forcing change, which would alter the response. The isotherm response is useful for interpretation of wind events. Figure 4.9 shows the squared wind speed, temperature isotherms, and Wedderburn number for the period from July 15 to July 26, 1994. The Wedderburn number (W) indicates the severity of the wind forcing relative to the thermal stratification. For a two-layer system, upwelling occurs at a theoretical W ~ l . Above this value, the thermocline tilts and may seiche. Below this value upwelling occurs with T/4 and destratification may occur. For a continuously stratified system, upwelling may occur at Wedderburn numbers much greater than 1. Monosmith (1986) observed upwelling at Wedderburn numbers "over an order of magnitude greater" than 1. Several features are apparent from the isotherm plot. One, is that the deeper isotherms do not oscillate as much as the surface isotherms. This indicates less water movement at the deeper thermistors. Secondly, the continuous nature of the stratification is apparent in this representation, as there is generally no consistent isothermal region within the contoured area. The diurnal heating and cooling during this period does not reach below 2.5 m depth. Above about 2.5 m, isotherms 'enter' the water column during heating, and 'exit' during cooling. Also evident is a slow deepening of the isotherms (e.g. the 20° isotherm appears on July 18, and reaches 3-4 m by the end of July 21 - a decline of = 0.5 m/day). Recall that this was an unusually calm, warm period at the lake. Several wind events are apparent in the extract shown in Figure 4.9. Firstly, on July 19 (day 199), strong winds blow for most of the day, and the Wedderburn number is between 48 Chapter 4: Observations and Results 1 and 2 for 16 hours. The low Wedderburn number implies strong upwelling and mixing -observed are the top 4 m mixed to 18 °C. Secondly, on July 22 and 23, (days 203 and 204) the magnitude of wind forcing is virtually the same yet the isotherm response is different. On the 22 n d oscillations are about 1 m and on the 23 r d are about 2 m. (as shown by the 20° isotherm). This could be due to the wind event being in or out of synchronization with the internal seiche. A wind event in phase with the seiche would amplify the isotherm oscillations. Finally, very high oscillations occur on the 25 t h in response to wind forcing on that day. This is a special situation as the wind set-up from the north, changed direction 180° from the south in the same time scale as the seiche period. 4.4.4 WindSet-Up The CTD transect already mentioned demonstrates the thermocline tilt that occurs under wind set-up. For a two layer system under wind stress, the thermocline is expected to tilt with a slope inversely proportional to the Richardson number. The four transects observed, including the one already discussed, are shown in Figure 4.10, and their parameters are summarized in Table 4.1. In this table, the Wedderburn numbers have been estimated based on daily average water temperature assuming a linear temperature gradient (i.e. an approximately linear density gradient), and 2.4 hour averaged wind speeds. A slope of 1/Ri, based on two-layer theory is shown for comparison. These data seems to indicate that it is a reasonable assumption in this case. This estimation has assumed that the observed set-up is approximately stable, or at least is not in seiching motion at the time of profiling. It was mentioned that the initial wind stirs up a 'wedge' of epilimnion at the downwind end of the lake. Monismith (1986) observed that a constant stress would result in a quasi-steady-state mixed layer at the leeward end. This mixed layer forms by upwelling until the wedge is 'quasi-stable' after which it increases much more slowly by deepening at the density interface. This upwelling and mixing was observed to continue until the average 49 Chapter 4: Observations and Results density jump and average mixed layer depth were equivalent to a two layer system with a Wedderburn number of about 1. To see if this effect is observed in the Chain Lake transects, an estimate of the mixed layer is made by estimating the volume of the mixed layer wedge. The two-layer density jump is interpreted from the CTD profiles. The four transects are also summarized in Table 4.2. The mixed layer response does appear to approach a value of W near 1. The assumption inherent is that the transects were made at a period of 'quasi-steady' state i.e. after a period of Ti/4. From Monismith's observations, the quasi-stable wedge forms much faster than the overall wind stress (i.e. tilting response). Thus our CTD transect snapshots, which indicate overall isotherm tilting, should be expected to be at this quasi-steady state point after the 1-2 hours of observed wind stress. Table 4.2: CTD Profiles: Isotherm Slope Summary Mav9 Mav23 June 20 July 7 Surface Temperature-Ti (°C) 15.0 14.0 15.5 18.5 Bottom Temperature-T2 (°C) 11.0 12.5 14.5 16.0 a t i (kg/m3) -0.875 -0.730 -0.952 -1.474 rj t2 (kg/m3) -0.369 -0.535 -0.802 -1.032 g' (10-3m/s2) 5.0 1.91 1.47 4.34 Wind Velocity (m/s) 3 3.25 0.5 3 Shear Velocity u* (mm/s) 3.4 3.6 0.6 3.4 h(m) 1.5 2.7 1.7 2.75 Ri 662 391 8011 1061 W 0.62 0.66 8.51 1.82 1/Ri (m/km) 1.51 2.55 0.12 0.94 Observed slope (m/km) 1.13 3.4 -0 2.45 50 Chapter 4: Observations and Results 4.4.5 Stratification, Stability, and Wind Work Comparison of the thermal stability (eqn 1.1) and the applied wind work (eqn 1.4) indicate the wind driven de-stratification events. These parameters are shown in Figure 4.11 for 1994. The wind work is calibrated by assuming uw = 15 u*. This is arbitrarily chosen so that the wind work plots on the same scale as the stability, but is within the documented range (see section 2.1). Several strong wind events are evident destabilizing the water column. Wind driven partial destratification is also indicated by a decrease in stability. The level of wind is not the only factor in determining stratification. For example, low winds during July 1993 did not result in significant stratification (not shown). The wind work during July of 1993 and 1994 were both on the same order of magnitude, yet July 1993 was a period of low stratification, and July 1994 was a period of significant stratification. Similarly, destratification is not always a wind driven event. During late July and early August 1994, the lake stability declines without significant wind input. This was a period with some rain (August 7 & 8). As well, the lake is at 1000 m elevation and beginning in mid-August is subject to cool, dew point temperature, evenings. Both the rain and the cool weather are driving forces for cooling, and destratification. 4.4.6 Dredged Hole The dredged hole provides a quiescent region in the lake. Weigand (1989a) found water velocities here to be lower and sedimentation rates higher than in the main body of the lake. Temperatures during 1993, and 1994 are consistently lower in the dredged hole. On many occasions there is a definite thermocline of 2-3 °C in this region. This was evident in the CTD profiles in Figure 4.10. Aside from maintaining a stable thermocline, the dredged area may not be completely mixed with the water column during spring turnover. Temperature and oxygen profiles from February to May 1994 are shown in Figure 4.12. They indicate that in spite of the mixis of the main water body the dredged area may be persistently cooler and also persistently anoxic. The dredged hole then represents a stable 51 Chapter 4: Observations and Results seasonal hypolimnion. As well it may provide a reservoir of anoxia that does not turn over during the spring. Some of the morphometric maps have indicated a slight sill rising to 5.5 m in the middle of the lake. If so, then the dredged area could act as a sink for bottom water from the entire south end of the lake. The dredged area represents less than 1% of the entire lake volume, yet it may act as a thermal stabilizer to maintain the stratification discussed, and a collection point to withdraw anoxic water from close to the sediment for the south end of the lake. 4.4.7 Expected Stratification It is appropriate in light of the previous observations of stratification, destratification, to wonder what would be the expected stratification. Chain Lake is relatively shallow (mean depth approximately 6.0 m). It has been observed to stratify intermittently over a summer (Northcote, 1967, Weigand, 1989). This agrees with the regional study of Gorham and Boyce (1989) which indicated that based on its depth and area, Chain Lake is borderline between lakes that stratify for the duration of a season, and those that do not stratify or frequently de-stratify. Stefan & Hanson (1980), modeled a group of temperate lakes in Minnesota with areas from 34 to 225 ha, and mean depths from 2.1 to 3.7 m. Their objective was to determine the depth required of a lake in order to maintain a seasonal stratification, i.e., become dimictic. They found low numbers of stratifications for shallow (i.e. completely mixed) and deep (dimictic) lakes over a season. In between, at about 4 to 6 m depth, lakes of this size would experience a maximum number of stratifications in a season. Chain Lake (46 ha area and 6 m depth) would be expected to be stratified for a large portion of the year but would experience several mixing events per year. 52 Chapter 4: Observations and Results 4.5 Thermal and Water Budgets 4.5.1 Heat Content The heat content of the lake, measured relative to zero degrees C, during 1993 and 1994 is shown in Figure 4.13. The heat content plot is similar to the temperature plots already shown. That is, 1993 experienced a quick rise in temperature early in the year, and a flat-top temperature plot through the summer, while 1994 showed a continuous increase to a peak and then decrease. The maximum heat content of the lake occurred around August 8 in 1993 and July 29 in 1994. From the comparative meteorology of the two summers, it seems obvious that the maximum heat content in the lake is lower in 1993 than in 1994. Somewhat surprising is that the difference is not large (480 MJ/m^ in 1993 and 500 MJ/m^ in 1994), and that the maximums occur at approximately the same time. This may be partially explained by consideration of the heat budget components. The predominant heat sources/sinks to the lake are: (1) solar radiation - the direct beam radiation from the sun, (2) longwave radiation - the black body radiation (infra-red) radiation emitted from the water to the air, and from water vapor in the air to the lake, (3) sensible heat exchange - the heat losses from conduction and convection carried away from the lake by wind, and (4) latent heat exchange - the heat removed from the lake to convert liquid water to water vapor. Descriptions of how each component is estimated are given in Appendix C. The different fluxes change in response to the ambient conditions. The removal by evaporation is a function of the lake surface temperature and air temperature. When the air temperature increases, i.e. during a sunny day with high incident radiation, then evaporation increases. Black body radiation and convection increase as the lake surface temperature increases. Thus, when the lake is very warm at the surface, it looses heat at a faster rate. If the lake does heat up more than usual, then more heat will convect away and resist excessive heating. In summary, inter-year variation in water temperatures are certainly not as great as the inter-year variation in air temperatures. The field data collected indicates that differences 53 Chapter 4: Observations and Results in air temperature between 1993 and 1994 were on the order of 10 °C, and differences in water temperatures were on the order of 2 °C. 4.5.2 Water Budget The primary sources of water input are direct inflow, the Shinish diversion, ground water, precipitation, and surface runoff. Flows of water out of the lake are evaporation, the surface outflow, and the withdrawal pipe. Water flow rates estimated over the 1994 summer are shown in Figure 4.14. For the majority of the season the system flowed at a set point of 35 L/s. This set point was an attempt to match the withdrawal flow to be approximately equal to the diversion inflow. A total of 288 x 10^  m-^  of water was removed via the withdrawal pipe. This is equivalent to about 63 cm of lake level, or =10% of the lake volume. This is lower than anticipated but is mainly due to the extremely dry year and the late start date of the Shinish diversion. The water balance indicates that the diversion, the withdrawal and ground water flows are the most significant water fluxes to / from the lake, and are all of the same order of magnitude. 4.6 Chemical / Biological Measurements This section below summarize the chemical and biological data, i.e. visibility, chla, and chemical analysis. Tabulated values of the samples collected are provided in Appendix D. 4.6.1 Dissolved Oxygen The dissolved oxygen data are shown as profiles in Figures 4.15 and 4.16 for 1993 and 1994, respectively. Several important features are indicated. First, some of the profiles at different locations in the lake are quite different on the same day (e.g. July 20, 1993 in Figure 4.16). This is likely an indication of the seiching response of the lake. Second, some profiles indicate high oxygen levels in the surface water, and low oxygen at the lake bottom 54 . Chapter 4: Observations and Results (e.g. July 23, 1994) This is common in high productivity systems. The high levels are the result of photosynthesis, and the low levels the result of organic decay. Finally, all the profiles terminate in the sediments at zero mg/L D.O. This occurs regardless of whether the water column is aerobic or not, or if lake is stratified or not. This is not surprising given the substantial sediment oxygen demand (SOD). 4.6.2 Secchi Depth / Chlorophyll-a / Total Phosphorus The Secchi disk transparency, chlorophyll-a concentration, and total phosphorus concentration for the field seasons 1993 and 1994 are shown in Figures 4.17 and 4.18 respectively. All three measures are used within limnology to indicate lake productivity and allow for useful interpretation of the lake behavior. The Secchi record for 1993 is sparse, but indicates a slow build up of algae in the lake - indicated by the gradual decrease in transparency. The lowest transparency was recorded in late August. During 1994 an extensive Secchi record was maintained. In the spring transparency is greater than 4 m. There is a gradual decrease through June to about 2.5 m. Following the initiation of the withdrawal, and coincident with the long stratification in July, visibility increases again to over 4 m. During the last week in July the Secchi depth decreases rapidly to about 1 m by the 15 t h of August. This is the result of an algae bloom which remained until the end of the year. Chlorophyll data show a build up chl a at the surface that coincides with the reduced water transparency. Maximum values observed were 107 iig/L in 1993 (measured August 26) and 76 iig/L in 1994 (measured Sept. 15). The phosphorus data show a build up of phosphorus in the bottom of the lake during both years. Later in the season, there are high levels of total phosphorus in the surface waters. The Secchi transparencies decrease as the amount of chl a in the surface water increases. Secchi depth may be thought of as an integrating measure of particulate and suspended solids content in the water - one that always integrates from the surface. Hence, 55 Chapter 4: Observations and Results low surface levels of chl a correspond to high transparency levels. The 1994 record indicates the value of using Secchi depth to indicate water quality, and provides an excellent review of the course of events during that summer. The initial quality is good in May when the flushing rate is high and biological production is slower due to cooler temperatures (SD > 4 m). The transparency decreases during June as biomass increases. The phosphorus concentration increases at the bottom of the lake several weeks before the maximum surface chl a levels were observed. This build up remains stable at the bottom for two or more weeks before it is mixed through the lake - even during relatively isothermal conditions in 1993. During July, 1994 the phosphorus concentration increase at the bottom of the lake is accompanied by an increase in algal biomass below 4 m with chl a levels above 20 |ig/L. This build up of algae is not reflected in the Secchi record because it occurs below the Secchi depth. In 1993, a build up of chl a at the bottom of the lake accompanying the increase of phosphorus is not indicated by the contours. A reason could be that the above average rainfall increased the water inflow during the summer. The increased run off to the lake is likely to contain substantial suspended solids which would block some light, forcing algae production to higher levels. This is supported by comparing the Secchi record to the chl a data. Notice that surface chl a levels were below 10 [ig/L in July 1994 and the Secchi transparency was up to 4.5 m. In July 1993, however, chl a was about the same, yet the Secchi depth was only 3.0 to 3.5 m. This example highlights the difficulty in using Secchi depth to indicate biomass. A large 'reservoir' of phosphorus is apparent in 1994 inside the dredged hole at the beginning of May. The water volume here is small (=0.6% of the lake volume) though the concentration is 10 times higher than the remainder of the water column. This observation implies either (a) that lake mixing is incomplete in the spring, or (b) that the metabolism in the sediments is already depleting oxygen faster than it can be supplied. In the case of incomplete mixing, relatively large amounts of nutrients from the winter stratification are still 56 Chapter 4: Observations and Results available in late May. If so, then the withdrawal could reduce the spring bloom. The sample in May was taken a month after ice-off. The withdrawal can remove the volume of the dredged hole in about 6 days so several residence volumes of the dredged hole could be removed before those nutrients circulate upward. In the case of fast oxygen depletion in the spring, the implication is that the build-up of phosphorus, which is generally regarded as a summer phenomenon is occurring as early as May, and could occur over the lower 0.25 to 0.5 m of the entire lake. Occurrence of anoxia within weeks of ice off is not widely reported for temperate lakes. Whether due to incomplete turnover, or early anoxia, it indicates that the dredged hole contains reduced compounds that should be removed as quickly as possible. Operation of the withdrawal early could certainly remove some of this water. Finally, it is apparent that the algae bloom in August starts before the lake is fully mixed. Two possibilities exist, one is that some entrainment has occurred during the partial destratification at the beginning of August. Another is that the observed seiching in the lake was accelerating vertical transport. This is not unusual as seiching results in significant movement of water in the bottom of the lake. The increased turbulence of moving water could mix up sediments or high concentration bottom water. 4.6.3 Phosphorus Components During 1994 the phosphorus measurements included the breakdown of the total phosphorus (TP) into total soluble phosphorus (TSP) and soluble reactive phosphorus (SRP). Total phosphorus measures the entire phosphorus pool, including phosphorus available to algae, and the phosphorus tied up in the algae themselves. In a low productivity system TP is often used to indicate system productivity because phosphorus is likely the limiting nutrient. Since phosphorus is so scarce in these systems, it is a reasonable 'tracer' of productivity. In higher productivity systems the use of TP to indicate growth becomes less certain. There may be large amounts of recalcitrant phosphorus in suspended or dissolved organic material 57 Chapter 4: Observations and Results that is not bioavailable. The actual pool of phosphorus that cycles through the biota is not represented by the TP pool. Total soluble phosphorus represents the soluble phosphorus available and measures the inorganic phosphorus pool as well as polyphosphates and complex organic phosphorus. Finally, SRP represents the most easily available form of phosphorus - the PO4 ion. Different researchers use different compartments of the phosphorus pool depending on their application. Total phosphorus is the most common measure because it is the simplest to gather. The sample can be easily preserved in the field, and is stable. On the other hand, SRP has an extremely short residence time under phosphorus limited conditions, and samples must be filtered and analyzed promptly to ensure accurate values, particularly when a sample from a dark, or anoxic environment is brought to the surface where growth or aeration could transform the sample. Contours of TP, TSP and SRP against lake depth and time are shown in Figure 4.19 for 1994. The build up of TP in the dredged hole during July, 1994 (already discussed) is also evident in the TSP and SRP contours. The contours highlight the bottom to top concentration gradient, which is indicative of sediment release. As well, the increase of phosphorus during July is observed in all three measured forms. 4.6.4 Iron and Manganese Depth vs. time contours of iron for 1994 are shown in Figure 4.20. There is a build-up of iron during the summer stratification. Total iron levels reach over 2.0 mg/L. Soluble iron levels are low outside of the dredged hole area (all but two measurements < 0.1 mg/L). Inside the dredged hole, soluble iron levels of over 0.5 mg/L were observed. A second region of high iron can be seen in May. This corresponds to the high phosphorus levels already mentioned and supports the hypothesis that the dredged area is either persistently anoxic or has a slow turnover in the spring 58 Chapter 4: Observations and Results Depth vs. time contours of manganese for 1994 are shown in Figure 4.21. Maximum values of total and soluble Mn occur in August and reach 0.53 mg/L and 0.50 mg/L, respectively. The manganese contours show the same trends as the iron contours, namely high values in the dredged area in May, and again during the summer. As well, the soluble component appears to comprise a large portion of the total component. The iron and manganese data agree with our expectations of anoxic, ORP controlled release. Since soluble Fe^+ oxidizes rapidly to F e 3 + ' when iron migrates away from an anoxic area, it quickly precipitates and becomes insoluble. This is evident in the contours in Figure 4.20 where total iron concentrations are over 1.0 mg/L outside the dredged hole, yet the soluble component is less about 0.3 mg/L. Since the oxidation of manganese is much slower it is often present in soluble form in aerobic regions, away from the source. This can be seen in the contours of Figure 4.21 where large fractions of the Mn is soluble. This has been reported in other lakes (e.g. Stuaffer, 1987, Imboden and Lerman, 1978, Hakanson and Jansson, 1983). The oxidation-reduction model indicates that manganese should be reduced (and therefore released from the sediments) prior to iron (Henderson-Sellers, 1984). It is expected that the manganese concentration will increase prior to the iron. Further decline in the ORP results in iron and then sulfate reduction. The presence of sulfide (S 2 -) and reduced iron (Fe 2 +) results in the formation of insoluble FeS. This precipitation may act as a controlling mechanism for the soluble iron concentration. Stauffer (1987) surveyed 28 calcareous temperate lakes and found soluble iron concentrations below 0.18 mg/L whenever sulfide was detected. At Chain Lake, when the withdrawal flow contained sulfide, then the soluble iron concentration was probably low, even though the total iron concentration was 2.7 mg/L. A filtered sample was not available on that date and this is shown in Figure 4.4 with a question mark to indicate the unknown value. It also highlights the uncertainty of connecting the points of soluble iron as a continuous function when the value at this time is unknown. With the exception of this one occurrence of sulfide, the release of iron into the hypolimnion 59 Chapter 4: Observations and Results is not limited by precipitation reactions. This may explain why the level of soluble iron was generally high compared to Stauffer's survey with concentrations of 0.3 mg/L in the withdrawal flow. The greater mobility of manganese under 'slightly' anoxic conditions means that it should accumulate more readily than iron in the hypolimnion of the lake. For example, Division et al (1982) studied the sedimentology of iron and manganese in a seasonally anoxic lake. They found that during aerobic periods the external inputs of iron and manganese were deposited on the sediment. While both iron and manganese concentrations increased in the hypolimnion during anoxia, the released iron is a small fraction of the total input, and most of the iron is retained in the sediments. The manganese is more easily reduced and a much smaller fraction is retained and buried in the sediments. The top-most sediments and the water column become 'manganese accumulators' in a sense cycling the small manganese input. This may explain why manganese is present in relatively high concentrations in the withdrawal water and only below method detection levels in the watershed in general (e.g. Shinish Creek to be discussed in Chapter 6). In fact, the manganese may be washed out on an annual basis and the accumulation effect may only last through the course of a year (Davison et al, 1982). Examination of the withdrawal flow has already shown that both Mn and Fe increase by an order of magnitude during anoxia from aerobic baseline concentrations. A plot of the withdrawal Fe/Mn ratio is shown in Figure 4.22. When the withdrawal is aerated the ratio of Fe/Mn is on the order of 4 - 7. However, during the period of anoxia the Fe/Mn ratio falls, indicating the release of manganese prior to iron. There is one point, on Aug. 7, which does not follow the trend. The Fe/Mn ratio suddenly jumps from 3 to 5. This change occurs while the manganese concentration is also increasing, indicating that the iron release rate is suddenly very high. It is likely that the ORP is reduced to the point where iron release is uninhibited by reduction potential and large amounts of iron are released from the sediments. 60 Chapter 4: Observations and Results Mortimer's original work (1941,1942) shows that release from the sediment of both manganese and iron begins suddenly with a drop of ORP. 4.6.6 Destratification and Elimination of Anoxia The discussion of lake physics has indicated that turnover events could occur within a few days. Some additional samples were taken during August which indicate the speed of chemical oxidation during a physical turnover. These are extracted from the fountain composition and are shown in Figure 4.23. The sharp decline in TP, Fe, and Mn levels in the withdrawal fountain from August 15 to 20, indicate that the turnover of the lake can eliminate reduced compounds within a couple of days. This can make sampling for these events difficult. In summary, the situation is one in which severe anoxia takes several weeks to occur (e.g. sufide detected only after 4 weeks of stratification), yet may be removed by a strong mixing event in several days (e.g. sulfide disappeared between Aug 8 and Aug 15 site visits). 4.7 Summary The meteorological observations allow the characterization of the two field seasons. The first, 1993, was a cold and wet year, which implies slower lake metabolism and greater hydraulic flushing rates. A notable thermal stratification occurred in the spring, but was removed by June. For most of July there was only minimal thermal stratification. The second season, 1994, was a dry, hot summer with virtually no precipitation, and record high air temperatures. The lake was very strongly stratified for all of July and the beginning of August. The polymictic nature of the lake is evident from daily average water temperatures. Partial destratifications occur within a day and complete de-stratifications occur over a period of a few days. The difference between these two summers highlights the significance of inter-year meteorological variation in determining lake response. 61 Chapter 4: Observations and Results Response to wind set-up was observed using CTD transects. Brief wind events were observed to mix the top 2.5 m to a homogeneous layer in 1-2 hours (faster than Ti/4). Observed isotherm tilt during wind set-up indicate that the 1/Ri estimate of isotherm slope is appropriate for a continuous stratification. Seiching was observed on several occasions with a typical summertime first mode period of 10 hours. Comparisons to two-layer stratification and linear stratification estimates suggest the seiche response is bounded by these two estimates, the former providing the low end and the latter the high end estimate for first mode internal seiche period. Wind events and lake thermal response were parameterized by the Wedderburn number. Substantial upwelling and destratification was observed when the Wedderburn number was on the order of 1 to 2. Higher Wedderburn numbers were associated with seiching responses. The dredged area provides a quiescent region in the lake and is thermally stratified by about 2 °C to the overlying water column. This area was low in oxygen through the entire 1994 withdrawal season, regardless of the condition of the main water column. High concentrations of phosphorus, iron, and manganese indicate that this area may not mix during spring turnover, or that oxygen depletion rates are fast enough to result in anoxia only four weeks after ice off. The chemistry of the lake agreed with our expectations of an anoxic, internally loaded system. Phosphorus, iron, and manganese concentrations increased in the bottom of the lake during stratification. This situation was stable in 1994 for over two weeks before being mixed to the surface (even during a period of substantial seiching). An algae bloom was initiated at about the time of a decrease in lake stability. This implies some mixing of the hypolimnion into the surface layer during the partial destratification, or mixing due to seiching. In either case, a complete turnover was not required in order for an algae bloom to be stimulated. 62 5. DISCUSSION: This chapter examines the collected data set from Chain Lake in an attempt to evaluate the prospects for success of the restoration effort of bottom withdrawal. An attempt is made to address the issue of inter-year variability of the weather, and places the observed stratification and anoxic period from 1994 into the context of an anomalous year of extreme stratification. A field site is not a laboratory in which all conditions can be controlled and the effects evaluated and for a small lake such as Chain, the variability between years may have dramatic effects on the lake conditions which complicates the use of baseline data to evaluate the response of a change in watershed management. 5.1 Defining a Successful Withdrawal Defining the success of the project is not an easy task. From the objective of reducing internal loading, success would be considered a noticeable reduction in lake phosphorus loads. To the residents of the lake success will be judged by the 'look' of the lake which could loosely be translated into Secchi transparency. The Environment Ministry (HCF) who provided partial funding for the project are interested in the enhancement of the fishery. Evaluating these different measures is complicated. For example, it is known that Secchi transparency decreases asymptotically with the chl a level. That is, above a certain level of biomass, the Secchi transparency is reduced only very slightly when the biomass increases further. Similarly a significant decrease in biomass might only provide a small improvement in transparency. This was observed in the field data. Maximum Chl-a level was 107 iig/L in 1993 and only 76 ixg/L in 1994, over a 25% reduction. However, the lowest recorded Secchi depth was about the same at 1 m. In fact, the minimum Secchi reading for 1994 was recorded when the chl a was only 35 fig/L. A similar effect is observed for phosphorus. The maximum measured total phosphorus in the 63 Chapter 5: Discussion main lake body was over 200 \xg/L in 1993 and only 120 |!g/L in 1994. It could be argued that there was a 40% reduction in peak measured phosphorus during the first year of withdrawal operation. Many residents considered the withdrawal a success as the lake was clear and warm through July. Whether this effect was the result of the hot dry weather, or the action of the withdrawal, or a combination of the two cannot be answered with certainty. A method of removing the seasonal variability is not available after such a short operational time. This is not surprising as it was anticipated that several years of withdrawal might be required before results could be quantified. 5.2 The Technological Applicability of Withdrawal The operation of the withdrawal during 1994 verified that the system was an appropriate technology for this application. The system was operated with few flow adjustments, primarily to reduce the flow during the driest period and to turn it up later. As well, once the system was opened up for the season, there were no requirements to stop for maintenance. A flow rate test was conducted on May 20, 1995 during which the flow rate was calibrated to the height of water in the fountain, thus providing the residents with a simple and accurate method of gauging the flow. The results of the test are shown in Figure 5.1. The system is capable of flow rates up to about 80 L/s. In normal operation it is expected that the flow rate will be about 50 L/s. These factors are important for B.C. applications as many of the lakes that are potential withdrawal candidates are recreational areas. These situations would often be outside the jurisdiction of municipalities, and system operation would likely be the responsibility of volunteer operators as at Chain Lake. In these cases residents may not be on site continuously to monitor the operation. Systems like this should be designed for intermittent layperson operation if they are to be employed successfully. 64 Chapter 5: Discussion 5.3 Limnological Perspective of Observations 5.3.1 Growth Limitation Up to this point it has been assumed that the limiting factor for growth in Chain Lake is phosphorus, and that removing phosphorus will reduce lake productivity. To completely validate this assumption would require a series of nutrient enrichment experiments to determine which nutrients might stimulate growth. The data available do allow some estimation of growth limitation. Growth limitation as expressed by the 'law of the minimum' is the concept that living organisms need nutrients in certain ratios. Organisms will be limited by the nutrient that is in scarcest supply relative to the ratio required for growth. An approximate model for the biomass of algae is the chemical formula C106H260O110N16P (Schindler, 1985)1 Based on this chemical structure the need for carbon, nitrogen and phosphorus is 106:16:1 on a molar basis, or 40:7:1 on a mass basis. Theoretically when the N to P mass ratio is above 7, there is excess nitrogen for the amount of phosphorus available. If the N to P mass ratio is below 7 then nitrogen is limiting. Blue-green algae are able to fix nitrogen from the atmosphere. They do not have to rely on the aquatic media to supply nitrogen and thus have an advantage over other algae under nitrogen limited conditions. Few generalizations can be made, but it is not uncommon to observe green algae dominating under phosphorus limited conditions (due to their faster growth capabilities), and blue-green algae dominating under nitrogen limited conditions (due to their ability to fix nitrogen) (Wetzel, 1975). In reality, different species require different ratios, and competition between algae populations is more complex than this simple model. For example, N:P values of about 30:1 have been observed to separate blue-green dominance from green dominance (e.g. Smith, This molecule is also expressed as C106H181O45N16P. The ratios of C:N:P are the same but the molecule has different oxygen and hydrogen components. For this discussion the difference is immaterial. 65 Chapter 5: Discussion 1983, Rhee, 1978). The 30:1 cross-over ratio seems to indicate the point at which green algae become nitrogen limited. Note that the presence of cyanobacteria (blue-green algae) is not a sufficient test for nitrogen limitation. The past data collected indicates that Chain Lake may be both nitrogen and phosphorus limited on different occasions. For example, Water Investigations Branch (1977) found that the N:P ratio averaged around 10, though on some measuring days was 6.6, indicating slight N limitation. Much of the data of Weigand (1989b) indicates a similar range. Phosphorus limitation would not be surprising at spring turnover, and other aerobic periods. The external nutrient loads are small, and aerobic conditions would remove much of the available phosphorus from the water column. Evidence of phosphorus limitation can be seen in the phosphorus contours for 1994 shown in Figure 4.19. The figure shows an increase of total phosphorus (TP) and total soluble phosphorus (TSP) during July, yet no significant build up of SRP. This may imply that biomass is increasing (increase in TP), and implies that more nutrients are being cycled in the water column (increase in TSP). However there is still a 'need' for easily available phosphorus which explains the lack of excess SRP. Following any substantial phosphorus release the system might be expected to be nitrogen limited. For example in August, at the height of the algae bloom there are higher levels of SRP (nearly 40 |ig/L by August 15th). At this time the phosphorus is likely in excess, and nitrogen the limiting nutrient in the bottom 3 m of the lake. Murphy (1987) recorded surface SRP levels of 100 [ig/L immediately following an overturn, which again indicates that at these times there is likely excess phosphorus. One approach to verify the nutrient limitation is to examine the ratio of chl a to total phosphorus (Stefan and Hanson, 1981). The ratio of chl a to total phosphorus varies 66 Chapter 5: Discussion somewhat but has been observed to be approximately chl a:TP =1:2 (Ahlgren et al,1988)2. Of particular interest, Ahlgren found that chl a vs. TP ratios were about 0.5 for many eutrophic lakes. If phosphorus is limiting, then algae will grow and produce chl a until the phosphorus limit is reached. If phosphorus does not limit growth then the production of chl a will be limited by another factor, phosphorus will be in excess, and the ratio of chl a:TP will be less than 1:2. Figure 5.2 shows the plot for the data of 1993 and 1994. A point on the 1:2 line indicates that chl a is being produced in an amount defined by the amount of phosphorus. A point to the right of the line implies an excess of phosphorus: Most of the data points are to the right of the line thus indicating other limiting factors, however it seems that on any given sampling day, one or two points in the water column profile are near the line. This means that somewhere in the water column phosphorus is limiting growth. As a result, we should be aware that phosphorus may not always the growth limiting factor in Chain Lake. An objective of the withdrawal however, can be considered to export sufficient phosphorus to make it the limiting nutrient, or export nitrogen with the phosphorus and reduce the amount of limiting nitrogen3. 5.3.2 Reduction of Lake Phosphorus Content The comparison of lake response in 1994 to other years is difficult due to the lack of continuous, consistent data. Since the objective of withdrawal is to reduce phosphorus levels then a comparison of maximum lake phosphorus content should indicate if dramatic phosphorus reductions are being achieved. From the data available a plot of the maximum recorded phosphorus content of the lake is shown in Figure 5.3 This type of comparison is 2Algae are typically about 1% phosphorus by mass and anywhere from 0.5%-2% chl a by mass (Ahlgren et al,1988). 3 Reducing the nitrogen level could have a detrimental effect because species composition might shift further to the nitrogen fixers. Evaluation of this effect is beyond this work, and might take several years to be noticed. If it occurs, and if the withdrawal is effective in reducing biomass, then we might expect that the lake produces less total biomass of algae, but with a higher proportion being blue-greens. Northcote (1967) reported that in August typically as much as 86% of the algae were blue-greens. 67 Chapter 5: Discussion crude because of different sampling intervals and sampling stations used during the different study years. No assurance is made here that the past or current measurements capture the moment of maximum phosphorus content. On the other hand, the algae blooms do not fade away overnight and as most of the studies sampled bi-weekly or monthly they should provide reasonable estimates. A high value of 650 kg was recorded in 1988 and may have been a response to the dredging operations in that year which disturbed the sediments in the south end of the lake. The lowest value of under 100 kg in 1971 does not have an explanation but could be due to sampling program. The maximum measured amount of phosphorus in 1993 and 1994 is comparable with other years at around 300 kg in the late summer. As such, no observable reduction in the maximum recorded phosphorus content of the lake was observed. This representation may bias against observing a restorative effect. For example, if the withdrawal results in a delayed algae bloom of the same 'magnitude', it would be having some positive effect. Examining only the maximum phosphorus level might not indicate this. 5.3.3 Changes in Secchi Transparency Comparing Secchi visibility observed during the first year with the historical data set could indicate trends in lake water quality. The compiled data is shown in Figure 5.4. The data show a lot of variation,. For example, in mid August the Secchi depth was less than 1 m in 1962, and yet 3.5 m in 1971. The former value would be considered poor water quality (or normal algae bloom conditions), and the latter would be considered very good for the mid summer. The overlay of 1994 data shows how visibility may vary during a single year and raises suspicion about the validity of connecting sparse data points from other years. It could be argued using Figure 5.4 that a rough trend of expected Secchi depths would be 2-3 m during June and July decreasing to 1-2 m in August. In 1994 there is definitely a delay of the algae bloom into August compared with 1993. This is not sufficient to validate the 68 Chapter 5: Discussion success of the withdrawal however and several years of data collection may be required before any trends become apparent. As already mentioned, there were reductions in the amount of chl a observed in 1994 compared to 1993 and this coupled with the weather conditions could explain the delay of the algae bloom into August. 5.3.4 Sedimentation and Release Rates Sedimentation can be significant in high productivity systems. Weigand and Munteau (1989) and Weigand (1989b) discuss the sedimentological data collected in 1988. They found sedimentation rates around 5.0 g sed/m2/d over the lake during an algae bloom and 15 g sed/m2/d in the dredged hole following a sediment resuspension event4. At the same time these sediments were high in bioavailable phosphorus (BAP) resulting in phosphorus deposition rates typically 10 - 40 mg BAP/m 2/d and as high as 50 mg BAP/m 2 /d. Phosphorus release rates have been measured from core incubations and from lake mass balances. The latter is much more difficult because of the need to complete a full phosphorus budget. Niirnberg (1987b) compared the two methods and found that laboratory core incubations provided a reasonable prediction of in-lake release rates. Murphy (1985) did a simple core incubation of Chain Lake cores and found release rates of 19 and 40 mg/m2/d. His results are shown in Table 5.1. The release rates are in the same range as typical anoxic release rates compiled by Niirnberg which may be as high as 50 mg P/m 2/d. Taken together this sedimentation rate and anoxic release rate information implies that during a summer bloom there are large fluxes of phosphorus exchanged between the water column and the sediment surface. At these rates the phosphorus content of the lake (about 300 kg) is being turned over in about 15 days by each process - in a different direction! At other times the sedimentation rate may drop by an order of magnitude 4For the sake local comparison, McCallum (1995) reported sediment accumulations in Burnaby Lake, a eutrophic urban lake, that reached 1 g/cm2/yr (27 g/m2/day) in the 1960's. This represents the buried fraction of sedimentation, which is less than 100% of the deposited sediment measured by sediment traps. 69 Chapter 5: Discussion (Weigand, 1989b). At these times the sedimentation turnover time is in the range of 150 days. Table 5.1 Results of Chain Lake Core Incubations Time At SRP A SRP Release Rate (days) (days) (mg/L) (mg/L) (mg/m2/d) 0 7 7 =0 20 13 1.8 1.8 40 40 20 3.1 1.3 19 Sources: Murphy (1987) Notes: (l)Depletion of dissolved oxygen observed on day 7 (2)Cores incubated with 27 cm sediment, 29 cm water depth. The data set collected in 1994 allows an estimation only of the net phosphorus release rate, which is simply the observed change in phosphorus content of the lake normalized to the lake sediment area. This is shown in Figure 5.5 The net release rates are on the order of a few mg P/m2/d. These rates seem small compared to the individual components of sedimentation and release rate. The implication is that both components must be evaluated to do a proper nutrient accounting in a eutrophic system. When estimating release rates the sedimentation rates have sometimes been neglected (e.g. Yoshida, 1982) or their influence minimized (e.g. Riley and Prepas, 1984). To complete a material balance for phosphorus, only one flux term can be un-measured (with closure from the observed response). Without both terms, this work is only able to say that the sedimentation and release rates are on the same order of magnitude and the difference between them is probably small compared to their magnitude. No confirmation or quantification of reduced internal loading due to the withdrawal operation is possible from the data set. 70 Chapter 5: Discussion 5.4. Effectiveness of the Withdrawal for Nutrient Export 5.4.1 1994 Nutrient Export The phosphorus mass exported via the withdrawal is the product of the withdrawal flow rate and the withdrawal phosphorus concentration. This is shown in Figure 5.6. Over the entire withdrawal period in 1994 approximately 30 kg of total phosphorus were exported through the system. The withdrawal is currently limited by hydrology. That is, the amount of water flowing into the lake in the summer is so low that the withdrawal must be operated at a low flow rate in order to maintain the lake level. The hydrologic limits to the withdrawal may be temporary. The diversion has been upgraded and is capable of delivering over 50 L/s throughout the summer. The withdrawal flow could be increased to this amount. As well, there is the option to optimize withdrawal flows by increasing the rate during anoxic times and decreasing it following turnover. The withdrawal could conceivably be run at =75 L/s for one or two weeks in the summer. At these flow rates, and with phosphorus concentrations similar to 1994, the maximum possible phosphorus export is about 60 kg/summer though a more realistic value would be 50 kg/summer. This value (50 kg/summer) will be used for estimating the effects of the withdrawal. 5.4.2 Phosphorus Removal Rate The short term effectiveness is premised on a 'race' to remove phosphorus as fast as it is liberated from the sediment to prevent it becoming available to promote algae growth. The level of phosphorus export can be expressed as a phosphorus turnover rate. That is, the mass of phosphorus in the lake divided by the instantaneous export rate is the instantaneous renewal time. Since the withdrawal is located in the region of highest phosphorus concentration, the phosphorus renewal through the pipe will be less than the water renewal time. This is the objective of the withdrawal after all, and it is particularly important in 71 Chapter 5: Discussion Chain Lake where there is often an hydrology limit to flow rates, and where enhanced flushing alone (through the diversion) has been insufficient to reduce phosphorus concentrations. The instantaneous turnover time5 is shown in Figure 5.7. Plotted on the figure are the export residence times for the system in 1994, and the residence times if the system is operated continuously at 50 L/s. Also shown on the plot is the hydraulic renewal time through the pipe. It shows that while the pipe is removing the water from the lake about every 1000 days, it is flushing the phosphorus faster, with renewal times as low as 200 days. The expected increase in flow rates and optimization of the pattern reduces the quickest phosphorus flushing time to around 150 days. Recall that the sedimentation turnover time ranged from 15 to 150 days and the release rate turnover time was as short as 15 days. It appears that this removal rate would be on the same order of the sedimentation rate, though slower than the possible internal loading rate. Thus, complete removal of internally loaded phosphorus, as quickly as it is released seems unlikely. 5.4.3 Net Export and Long Term Potential If the system is able to export more phosphorus than enters the lake, then over long periods it will result in a net export of phosphorus from the lake - and particularly from the sediments! The most dramatic case of this is Lake Maun reported by Cooke et al (1993) in which the phosphorus export exceeded the external loading and is reducing the supply of available phosphorus from the sediments. In Chain Lake, external loads are not significant nutrient sources at this time. Septic system loads were estimated as 21.8 kg-P/yr (Lacelle, 1986) and 29.3 kg-P/yr (Murphy and Urciuoli, 1984). Groundwater and Hayes Creek inflows for 1983 contained about 30 \ig/L 5 The terms turnover time, residence time, and renewal time are used interchangeably and may refer to asny constituent of the lake (e.g. water, phosphorus, oxygen, etc). The turnover time is the amount of the constituent - the pool (a mass) divied by the rate of movement through the pool - outflow, sediment release, metabolic consumption, etc - (as a mass per unit time). 72 Chapter 5: Discussion total phosphorus (Murphy, 1987). For the estimated ground water flow rates in 1994 (about 20 L/s) this is equivalent to an inflow of 0.05 kg-P/day or a maximum of 10 kg over the summer. Total external inputs would then be about 40 kg/yr. The potential withdrawal rate of about 45 -60 kg for the year would be just greater than this giving a net export of 5 to 20 kg/year. Thus we might expect that over long periods there will be a gradual reduction in the phosphorus content of the sediments as each year more is exported than enters the lake. This process would be slow, as phosphorus pools in sediments are. typically large. 5.4.4 Phosphorus Pool The sediments are a large reservoir of phosphorus that can be recycled back to the water column. Murphy (1987), shown in Figure 2.3, found that the top sediments are rich in bioavailable phosphorus (BAP) with a content of 500 |ig/g dry weight. This is not an unusual amount as Holtman et al (1988) report a range for sediment total phosphorus from 10 ixg/g in sandy sediments to 10,000 itg/g in iron rich sediments. Note that Murphy's values are extractable, 'bioavailable' phosphorus.6 The total phosphorus content of the sediments is greater. The sediment phosphorus profile indicates the amount of active sediment. Deeper in the sediment the bioavailable phosphorus reaches a constant value. This implies that the deeper sediments are inactive, and although defined as bioavailable, are actually not physically available as they are being buried. This approach has been used to model the phosphorus pool in Lake Onondaga (van Orman, 1993). The top 65 cm of the P b 2 1 0 dated cores all appeared to be about 20 years old (see Figure 2.3) and without a noticeable trend of age vs. depth, indicating significant turbation of this region (Murphy, 1987). This is approximately the depth at which the highly variable water content profile changes to a more 6 Bio-available phosphorus is the fraction extractable by 0.5 N NaOH. This is an operational definition but is intended to approximate the fraction bound by Fe and AL (Bostrbm, 1988). 73 Chapter 5: Discussion consistent, steadily declining water content, indicating consistent compaction. Thus it is reasonable that there is an active phosphorus cycling region of about the top 1 m of sediment. Weigand (1989b) reported settling sediment ranged from about 0.1% BAP (1000 |J.g/g) when the sedimentation rate is low (probably inorganic) to 1% BAP (10,000 |J.g/g) when the lake is undergoing an algae bloom and sedimentation rate is high (high organic content). The lower value, and Murphy's surface sediment content of 500 |ig/g means that about 50% of the deposited sediment phosphorus remains in the sediment and 50% released. At the higher value (i.e. high organic content of 10,000 |J.g/g), 5% of the deposited sediment phosphorus remains on the sediment and 95 % released. This compares with other studies. For example Fallon and Brock (1980) found that only about 25% to 50% of the sedimented organic matter that reached the sediment surface was buried. Table 5.2 presents an estimation of the phosphorus pool of bioavailable phosphorus based on Murphy's (1987) core analysis. Given a i m deep 'active' zone we can integrate the BAP profile from 0 m to 1 m, for an average of 235 |ig/g. The sediments are about 95% water (i.e. 5 kg dry sediment per m 3) This 95 % of the sediments that are water typically contain upwards of 5000 |ig/L phosphorus.7 The result is an order of magnitude estimate of the active sediment phosphorus pool of just over 7500 kg. Comparing this pool to the 300 kg in the water column in 1994, this is a sediment pool to water column pool ratio of just about 25. This seems low, as typically the sediment pool contains orders of magnitude more phosphorus than the water column (e.g. 40,000:1 ratio, McCarty, 1970). The reason is partially due to the use of bioavailable phosphorus values determined by Murphy. As well, internally loaded systems return a large fraction of their nutrients back to the water column, thus reducing the pool-to-pool ratio. Samples of pore water, extracted from sediment samples returned to the lab, were allowed to sit sealed at 4°C for several weeks until completely anoxic. Pore waters analysed contained 6 to 13 mg/L SRP. 74 Chapter 5: Discussion From this content of bioavailable phosphorus, and a net export rate from 5 to 20 kg/year, this represents a depletion rate of 380 - 1500 years for the sediment phosphorus pool! It would seem that the system is unlikely to reduce the phosphorus content of the sediments in the near future. Table 5.2 Estimation of Sediment Bioavailable Phosphorus Pool Sediment Pore Total Spaces Active mass under 1 m 2 (@95 % water) (kg) 50 950 P content of sediment (fig/g) 235 — P content of pore water (fig/L) — 5000 Phosphorus Pool (g/m2) 11.75 4.75 16.5 Total Lake Phosphorus Pool (kg) 5405 2185 7590 Sources: Murphy (1987) Notes: (l)Each m 2 is underlain by active sediments to a depth of lm (2)Sediment area= 46 ha=460,000 m 2 Seasonal effects may play a role. Murphy (1985) observed that Chain Lake seemed to be a sink for phosphorus in the spring, yet a net exporter during the summer. If so, then the withdrawal could extend the period over which the lake exports phosphorus and increase the net export rate. 5.5 Seasonal Variability of Anoxia 5.5.1 Inter-year Variability The first study year has already been described as a 'cold and wet' year, while 1994 was a 'hot and dry' year at Chain Lake. These terms were established earlier by examining the weather record. Accounting for these factors is more difficult. There is anecdotal evidence that the water quality in Chain Lake is related to the weather. For example, it is widely regarded among lake residents that the algae blooms are more severe if the summer is 75 Chapter 5: Discussion hot, and less severe if it is cool. Sulfide was detected in the withdrawal flow during 1994. This unexpected event in 1994 (Murphy, 1994, pers comm) was likely the result of high temperatures which accelerate decomposition, and a long stratification period of over 40 days. If the July stratification had lasted only three weeks, for example, then it is possible that sulfide would not have been formed. A turnover would have increased the dissolved oxygen and ORP and stopped the formation of sulfide. It seems that withdrawal applications in shallow lakes could be better understood if the nature of the thermal stratifications is understood, i.e. the frequency and duration of stratifications. This section details an examination of how the lake behavior is related to the local weather, and how the weather record may be used to estimate how 'normal' the stratification of 1994 was or wasn't. 5.5.2 Meteorological and Simplified Prediction Some effort in limnology has been devoted to the correlation of simple meteorological or limnological parameters to lake behavior. These attempts have been based on regional studies that were intended to provide working tools to characterize lakes when little or no data is available. Kalff (1991) attempted to correlate latitude to variation in lake behavior (phosphorus concentration, primary productivity and fish yield). Niirnberg (1988) correlated the date of fall turnover in temperate lakes to the mean summer hypolimnion temperature. Demers and Kalff (1993) found that the onset of spring stratification could be correlated to the mean annual air temperature. Livingstone and Schanz (1994) noted that for their shallow lake, average air temperature and water temperature track each other very closely implying a close correlation of 'weather to water' in small lakes. These previous studies indicate that weather, and simple parameters, may be useful predictive tools in limnology. 76 Chapter 5: Discussion 5.5.3 Correlation of Stability to Air Temperature A correlation is made between lake stability and average air temperature at Princeton. Lake stability is a useful, all-encompassing measure of stratification. The air temperature is useful because it is the simplest measure available and would be the most available parameter from weather records. There is some mechanistic background for the connection. That is we have assumed that: (i) the short term weather is a primary driving force behind the establishment and breakdown of stratification, (ii) the stability varies on time scales similar to the variation in the weather and (iii) the average air temperature at Princeton represents the local weather. The set-up of stratification in Chain Lake occurs quickly, in a matter of 1 to 5 days. As well, stratifications were observed to break down in periods of a day or two. This seems to verify the first two assumptions above. As well, there is intuitive evidence that average air temperature could be considered an integrating or all encompassing measure of the local weather. In simple terms warmer days have warmer average temperatures and allow more stratification than do cooler days. Whether the average temperature includes the effect of wind (i.e. are windy days cooler than calm days?), or the night cooling that drives convective mixing is arguable, though less certain. For the time being, it will be assumed that the average temperature, does reflect aggregate local weather. The lake thermal stability and differenced air temperature (as the daily average temperature minus the 58-yr mean temperature) for 1993 and 1994 are shown in Figure 5.8. There is a strong visual correlation between the two parameters. Most of the changes in the Princeton air temperature record are paralleled by the lake stability. That is, that long periods of above average temperature a matched by substantial lake stability. Recall that our objective is not to define an exact level of stratification, but to understand the frequency and duration of stratification events. 77 Chapter 5: Discussion The 58 year temperature record is now examined for periods where the temperature is above the long term average. This is shown in Figure 5.9. Each year of record is shown as a horizontal line segment. When the temperature is above average, a plot symbol (+) is plotted. The plot symbol is scaled according to its magnitude above the 58-yr mean temperature for that day (i.e. higher temperatures result in larger '+' symbols). Placed side by side, these plot symbols form 'worms' that indicate above average temperature, which we believe coincide with lake stratification. A cutoff of 1 °C above the long term mean is used, i.e. only points 1 °C or greater above the 58-year mean are plotted. The 'worms' show the trends we have already mentioned, i.e. warm May during 1993 and cool July, and a very hot summer in 1994. As well, Murphy (1985) noted that 1983 was a cold year with above average rainfall. This is supported by the 'worm' for 1983 which shows almost no above average temperatures. The fact that the lake is stratified does not ensure that anoxia will occur. Such an event is dependent on other variables, one of which is the temperature. The metabolic reactions which consume oxygen are temperature sensitive, typically doubling in speed for every 10 degree C increase in temperature. A ten day stratification period in May, is much less likely to result in anoxia than a ten day stratification in July. To avoid including 'time of year' factors that would account for hypolimnetic temperature, and recalling that the objective is to characterize the summer stratification period^ a 'summer' interval was chosen from June 1 to August 15. Counts were made of all above average temperature periods commencing between these dates (i.e. counting the number and length of the 'worms' in Fig 5.9). The end day of August 15 was chosen because Chain Lake is at high altitude, and evening mists and dew point temperatures begin to occur regularly in mid-August. Note that this count will include events starting on or before August 15. 78 Chapter 5: Discussion 5.5.5 Distribution of Warm Periods From the count, a frequency distribution was derived for the number of warm weather events per year. This is shown in Figure 5.10a. There are an average of 4.08 warm weather periods per summer season. We can thus expect about four stratification events per summer. A histogram of the warm weather periods indicates that they are approximately log normally distributed. The distribution is shown in Figure 5.10b. The distribution has a log domain mean of 1.63 and a log domain standard deviation of 0.89. The utility of this plot is demonstrated by recalling the situation in 1994, where a stratification of 26 days was observed (40 days of stratification but 26 continuous days with temperatures more than 1 degree above the mean). In the log domain the standard normal variable z=Hn(26)-1.63]/Q ^ = 1.83. From standard normal distribution tables, the probability of an event equal to or greater than 26 days is 0.034, or 3.4 %. Given that there are an average of four events per year the annual probability of the stratification observed during 1994 is 4 * 0.034 = 0.138 or 13.8 %. This is equivalent to a return period of 7.2 years. Recall that this return period estimate is derived from the fitted lognormal parameters. If we look at the 'worm count' itself we find 241 'worms' and only 2 with lengths equal to or greater than 26 days. From this we could estimate the frequency as 2 out of 241 events * 4 events per year which gives a probability of 0.033 (a return period of 30 years!). 5.6 Summary The operation of the withdrawal demonstrates the applicability of hypolimnetic withdrawal to Chain Lake. The installation represents an appropriate technology for sparsely populated areas where the tax base may be insufficient to maintain a high operational cost project. As well, the low operating demands are suitable for recreational properties where residents may not be present at all times. 79 Chapter 5: Discussion From the first year of operation, and the historical data set it is difficult to conclude with certainty the effectiveness of the withdrawal. A total of 0.60 cm of lake (10% of volume) was removed through the withdrawal. This resulted in the export of about 30 kg of total phosphorus. This is in the context of a maximum phosphorus content over the year of 300 kg. Comparison of the maximum phosphorus observed in the lake for several years indicates that relative to other years, the phosphorus content in Chain Lake during 1994 was about the same or slightly lower than in other years. Rough calculations indicate that the withdrawn phosphorus amount is slightly more than the net external loading implying a very slow reduction in the bioavailable sediment phosphorus pool. The net export of 5-20 kg/yr is equivalent to a sediment depletion time of 380 to 1500 years! A correlation between the daily average air temperature in Princeton, and the lake stability shows that long periods of above average temperatures in Princeton are accompanied by lake stratifications. The correlation permits use of the long term meteorological record to interpret the frequency distribution of lake stratification. The results indicate a conservative estimate that the stratification observed in 1994 was a once in 7 seven year event in terms of the length of stratification. 80 6. E N V I R O N M E N T A L IMPACTS and MITIGATION: 6.1 Potential Impacts The environmental effects of hypolimnetic withdrawal are associated with the release of anoxic water into streams and creeks leaving the lake. These have already been discussed in Chapter 2 and include low oxygen and elevated levels of iron, manganese, ammonia, and sulfide. At one point it was even speculated that the high iron levels, and high flow rates being planned could result in an iron floe which might be detected all the way to the Similkameen junction, 35 kilometers away. There has been some speculation about possible negative impacts of a withdrawal system on the lake itself. These would include accelerating or encouraging destratification events and altering the overall heat content of the lake. The first withdrawal installation promoted early destratification (Olszewski, 1961), and Livingstone and Schanz (1994) observed that withdrawal operation encouraged stratification in the spring and reduced it in the fall. The withdrawal retains heat by withholding warm and releasing cool water. In a small lake such as Chain Lake, in which surface temperatures as high as 24 °C and bottom temperatures near 16° C during the summer were observed, further increases in water temperature could be detrimental to the trout population. The BC Ministry of Environment approved and working criteria for water quality (Nagpal, 1994) gives upper limits of water temperature for preservation of aquatic life (particularly rainbow trout) as 18-19 °C on a weekly average basis and 22-24 °C as maximum value. A one or two degree increase in water temperature due to the withdrawal, during a summer like 1994 would be a cause for concern. Finally, an operational concern has been raised that the withdrawal should not lower the lake water level. In the event that the lake was not be refilled by the ice-on period it could represent a hazard to the over-wintering of the fish (Mathews, 1994, pers comm). A lower water level implies less water in the lake, and less oxygen for the fish to 81 Chapter 6: Environmental Impact survive the winter. Since winter fish kills have been reported (Murphy, 1987), this concern is valid. These examples illustrate that possible negative environmental impacts must be addressed. For the Chain Lake installation, an Environmental Impact Assessment was provided to the Environmental Protection Branch of the BC Ministry of Environment, Lands, and Parks (Penticton Branch) prior to the commencement of operation (Macdonald, 1994). It addressed possible environmental impacts and described a monitoring program that would be employed during the first year of operation. 6.2 Environmental Monitoring Program A lake monitoring program was already in place prior to the operation of the withdrawal system. A few samples were taken in Hayes Creek during 1993, and a formalized sampling program was established in the spring of 1994. 6.2.1 Program Design Water quality variables are continuous time series and affected by many external factors, primarily the weather. Any sampling program will be a trade off between available resources and the efficiency, or value, of the additional information obtained from them. Proper sampling design attempts to maximize information, within the constraints of sampling limitations1. A series of river studies on the Thompson river (Oguss and Erlebach, 1976), Squamish river (Klieber and Erlebach, 1977), and Okanagan river (Kleiber et al, 1978) addressed the issue of single water samples to represent mean water quality. Though they ta lker evaluated stream samples as first-order autoregressive (Markov) processes (1977, reported in Rechow and Chapra, 1983). This means that the value of a sample is related to a previous value. He estimated that for phosphorus sampling in rivers an interval of 14-28 days resulted in maximum information per sampling. This does not mean that extra samples do not add more information, just that they are not as 'statistically efficient' at gathering data. 82 Chapter 6: Environmental Impact were on much larger rivers, they do validate that suitable interval sampling can adequately quantify nutrient loadings. This monitoring program was designed to observe iron, manganese, ammonia, and dissolved oxygen concentrations in Hayes Creek from the lake outlet to a point 5 km downstream - past the junction of the Hayes and Shinish Creeks (see Figure 2.2). Sampling was performed once during each site visit - about every two weeks. The initial assumption was that the withdrawal would produce reasonably 'steady state' effects that would be detected with such a program. Shorter term events occurring over the course of only days or less could be missed. For example the sharp declines in concentrations of phosphorus, iron, manganese and ammonia observed at the aerator fountain from August 8 to 18 (Figure 4.23) were observed serendipitously as a result of being on-site at that time, and taking daily samples. Regular bi-weekly sampling would have missed this completely. A method of predicting the moments of poorest water quality was not available ahead of time, so the program was based primarily on routine sampling and not on event sampling. This is an acceptable compromise in order to sample throughout the summer season. 6.2.2 Previous Data Collection Few data are available on the water quality in the Hayes or Shinish Creeks. The data that exists in the provincial database from 1973 to 1975 have been summarized by Swain (1985) and more recent data obtained for 1983 reported by Murphy (1987) and for 1988 reported by Weigand (1989b). These data on Hayes Creek are summarized in Table 6.1. It shows that Hayes Creek often has high levels of phosphorus, iron, and manganese. These values are for water flowing from Chain Lake, and so are the result of algae blooms and/or anoxic conditions in Chain Lake. The available water quality data for the Shinish Creek are summarized in Table 6.2. The Shinish Creek is lower in nutrients than Hayes Creek, which is why it was 83 Chapter 6: Environmental Impact diverted into the lake. It does contain high levels of iron (up to 2.4 mg/L), likely a result of ground water infiltration. The rocks in Shinish Creek are stained red - a sign of iron precipitation2. Table 6.1: Summary of Hayes Creek Water Quality Value Number Parameter Period Units High Low Mean of Values Iron - total 1974-1976 (a) mg/L 0.900 0.100 — 2 1988 0>) mg/L 0.700 0.200 8 Manganese - total 1988 mg/L 0.150 0.025 8 Ammonia 1988 mg/L 0.20 = 0.0 6 Phosphorus - total 1974-1976 ug/L 165 20 86 8 1983 (c) |ig/L 478 33 60(d) 10 1988 ug/L 350 = 20 8 Phosphorus - diss'd 1988 ug/L 200 15 9 Phosphorus -SRP 1974-1976 ug/L 128 <3 46 8 1988 ug/L 140 <5 9 Notes: (a) 1974-1976 data collected by BC MOE from Swain, 1985 (b) 1988 data from Weigand (1989b) recorded at outflow, mean values not available (c) 1983 data from Murphy (1987) recorded at outflow (d) 1983 TP: maximum value not used in average Table 6.2: Summary of Shinish Creek Water Quality Parameter Period Units High Value Low Mean Number of Values Iron - total 1976 (a) , mg/L 2.400 <0.1 2 1988 (»>) mg/L 0.550 <0.05 10 Manganese - total 1988 mg/L <0.05 = 0.0 5 Ammonia 1988 mg/L <0.05 = 0.0 4 Phosphorus - total 1976 1988 |ig/L Ug/L 67 N/A 7 N/A 22 5 10 Phosphorus - diss'd 1988 jig/L 100 < 10 11 Phosphorus -SRP 1976 1988 Ug/L ug/L 15 < 10 <3 <5 6 5 Notes: (a) 1976 data collected by BC MOE, reported in Swain, 1985 (b) 1988 data from Weigand (1989b) - mean values not available 2The flow of iron into Chain Lake through the diversion could help to imobilize phosphorus in the sediments. The diversion water is aerated so the iron would be in an Fe3+ state and provide sorption sites to adsorb phopshorus and remove it from the water column. This is interesting since iron addition has been studied as a nutrient inactivation technique (Hall et al, 1993, Cooke et al, 1993). 84 Chapter 6: Environmental Impact 6.2.3 1993 'Before withdrawal' Monitoring Sampling of Hayes and Shinish creeks was limited in 1993. One complete sample set was taken on August 19. At that time only total phosphorus and suspended chl a (as a measure of algae washout from the lake) were measured. A map of the area showing the phosphorus values sampled on August 19 is shown in Figure 5.1. This is during the year before the withdrawal system had ever been run. The sampling indicates that: • the phosphorus values downstream of Chain Lake are naturally high. Much of this total phosphorus may be in the form of algae 'washout' from the lake. For example chl a values were 36 pg/L at the weir, 12 pg/L at Jellicoe Road, and 7pg/L at the junction of the Hayes and Shinish (not shown in the figure). • the Shinish Creek may gain some nutrients between the point of diversion and the junction with Hayes Creek. This is reasonable as Shinish Creek flows through some pasture land prior to the junction with Hayes Creek. The potential exists for agricultural run-off to reach the creek. • the Shinish Creek (and/or groundwater inflow) have a significant dilution effect on phosphorus content. The phosphorus and chl a values are lower downstream of the junction point. 6.2.4 1994 'During Withdrawal* Monitoring Water sampling in Hayes and Shinish Creeks was more extensive in 1994. A map showing the sampled locations in given in Figure 3.3. A complete summary of the data collected at each sampling station is shown in Appendix D, and the data collected at the withdrawal fountain has been presented in Figure 4.4. The results and implications of the sampling program are discussed in section 63. 85 Chapter 6: Environmental Impact 6.3 Environmental Concerns From the perspective of observing of lake anoxia, stratification, and negative downstream impacts, the severe nature of the heat wave and dry period during July 1994 present what is likely to be close to a worst case scenario. This section addresses the concerns that have been raised with respect to the withdrawal operation. 6.3.1 Lake Heat Content Livingstone and Schanz (1994) studied a withdrawal at Liitzelsee, Switzerland and observed increased thermal stability in the spring, and decreased stability in the fall. This application involved a withdrawal 5 times larger than the one at Chain Lake (based on hydraulic residence time). In Chain Lake, heat fluxes carried by surface water and groundwater flows were considered small compared to the heat fluxes at the water surface (i.e. radiation, latent, and sensible heats). For example fluxes through the water surface are on the order of 100 W / m 2 (24 hour average). In contrast to this a withdrawal flow of 50 L/s removing 16 °C bottom water (instead of 22 °C surface water) is the equivalent of a net heat gain of about 5 W/m 2 . Heat gain (H) due to the withdrawal is given as H=Cp p Q AT; where Cp = heat capacity of water=4179 J/kg/°C, p=water density (=1000 kg/m^), Q= flow rate (nvVs), and AT= temperature difference between the top and bottom of the lake. Thus, the heat gain to the lake due to the withdrawal is equivalent to a lake wide temperature increase of 0.017 °C/day. This response is small and would be unobservable in comparison to the variability of environmental forcing. A more likely thermal effect on the lake will be if the withdrawal successfully increases Secchi transparency. Increased light penetration could alter the stratification more severely than would the effect of heat retention by the withdrawal operation. This effect has not been evaluated formally in this work, but is of interest to managers of shallow lakes. 86 Chapter 6: Environmental Impact 6.3.2 Destratification Chapter 4 discussed the physical details of stratification and de-stratification in Chain Lake. From this discussion, and the negligible thermal effects described in section 6.3.1, the effect of the withdrawal on lake destratification is expected to be unobservable in comparison with the meteorological forcing. This is due to the small size of the lake, allowing it to be dominated by the meteorology, and the relatively small size of the withdrawal flows relative to the lake volume. 6.3.3 Water Budget / Lake Level The installation of the fountain-aerator to the culvert pipe provides an hydraulic limit to the flow withdrawal flow. The maximum flow rate possible is approximately 100 L/s. This flow rate is equivalent to about 2.0 cm of lake level per day. If left unattended for two weeks, the level could only fall by about 30 cm. There is no potential for a 'run away' withdrawal if, for example, no caretakers were available on site to adjust the flow or for some reason it was accidentally 'turned up'. Thus, the potential to lower the lake level still exists, but it is relatively small. 6.3.4 Dissolved Oxygen - Aerator-Fountain Figure 6.2 presents the dissolved oxygen content of the withdrawal stream 'inside' and 'outside' the fountain.3 An important benefit of the aerator/fountain is the increase in dissolved oxygen that results from flow through the system. Figure 6.2 indicates that the fountain adds about 1.5 to 2.0 mg/L dissolved oxygen to the withdrawal stream as it enters Hayes Creek. This helps to remove any immediate chemical oxygen demand 3 The dissolved oxygen concentrations never reach zero inside the fountain although zero values were recorded for the water entering the box from the withdrawal pipe. This small increase is due leaks in the seal between the box and the lake bottom, and around the entry point of the control rod which permit surface water to mix with the withdrawn water. Crude flow measurements and DO mass balances indicate that surface water leaks account for about 10 % of the flow through the fountain. For example, on July 23, 1994 the measured flow into the box was 33 L/s and the flow in the fountain was 38 L/s. Note that both of these numbers are probably + 5 L/s in accuracy. 87 Chapter 6: Environmental Impact created by the reducing conditions in the withdrawn water and begins the process or re-aerating the withdrawal stream. The introduction of even a small amount of oxygen could conceivably reduce the overall detrimental effects of the withdrawal dramatically. From an engineering viewpoint, the aerator/fountain incorporates some effective design characteristics. Nakasone (1987) evaluated the factors affecting aeration at weirs. His work provides some guidance to evaluate the withdrawal fountain. He found that the dissolved oxygen increase over a free fall weir was a function of the drop height, the flow rate, and the tail water depth. This design attempts to maximize the drop height within the hydraulic limits of the system. Nakasone showed that split weirs with several channel sections are more efficient than single weirs. This supports the use of several small slits in the fountain pipe. Finally, he observed that up to 95% of the oxygen transfer is achieved in the 'plunge pool' where air bubbles are entrained in the water body (in our case, the pool around the fountain). This supports our current design of a plunge pool. 6.3.5 Dissolved Oxygen - DO Sag Figure 6.3 shows the dissolved oxygen profiles downstream from the withdrawal fountain. Each line plotted in Figure 6.3 represents a different sampling day. Prior to withdrawal operation oxygen levels are high in Hayes Creek. This is also the case in spring when flow rates are highest. L o w oxygen levels would not be expected at this time. During the summer stratification a high D O deficit is observed at the fountain. Following the increase due to the fountain, the water is aerated in Hayes Creek. On July 9 lake surface water is concurrently flowing over the outlet weir and mixing with the withdrawn water about 50 m downstream of the fountain 4. B y the time the water reaches the Beaver Dam station the D O level is above 8 mg/L. On the July 23, and August 15 4 The existing outlet from the lake is a surface weir located at the dam site. During much of the year water flows over this surface weir as well as through the withdrawal pipe. 88 Chapter 6: Environmental Impact sample dates, water is not flowing over the surface weir is not flowing and withdrawal water composes the entire Hayes Creek flow. On these days the initial DO concentration is between 2-3 mg/L. At the Beaver Dam station the DO has risen to about 5 mg/L. Following fall turnover in late August, the withdrawal D.O. concentrations are much higher (5.2, 9.7 and 7.8 mg/L on Aug. 28, Sept. 12, and Oct. 4 respectively). On these days the DO level 'sags' noticeably and then recovers. The result is DO as low as 2 mg/L at the Beaver Dam station on October 4. One explanation for this is the classic DO sag as a result of BOD loading from the withdrawal stream. That is, the BOD load to the stream depletes the existing DO faster than re-aeration can replace it. There is some question whether the sags observed are solely attributable to the withdrawal system. The area between the fountain, and the beaver dam stations is a large swamped area created by beaver dams. Thus the low oxygen levels observed at the beaver dam station could be due to benthic oxygen demand of bottom sediments in this region, and may occur regardless of the operation of the withdrawal pipe. The data set collected is not sufficient to be sure. Overall, dissolved oxygen levels were consistently at or above 2 mg/L at the Beaver Dam station (400 m downstream) and above 6 mg/L at the Jellicoe Road station (1700 m downstream) throughout the summer. This indicates that dissolved oxygen deficits are a localized phenomenon. 6.3.6 Nutrient Loads The effect of nutrient loading on Hayes Creek from increased bioavailable phosphorus was not evaluated by any direct techniques (e.g. periphyton growth study). Aside from the withdrawal, other factors are also introducing change to Hayes Creek. The pasture between the Beaver Dam and Jellicoe Road stations has been actively trenched, drained, and used for grazing in the last three years and frequently contains 60 or more head of cattle. As well; the hill slope on the west side of the valley was logged in 89 Chapter 6: Environmental Impact 1994, likely changing the runoff characteristics and runoff nutrient content. Evaluation of the growth stimulation effect of the withdrawal system would be inseparable from the effect of these changes to the watershed. It is known, however, that Hayes Creek has high levels of 'background' phosphorus (see Table 6.1). It is unlikely that further additions will stimulate further growth. The work of Bothwell (1994, pers comm) implies that when SRP values are high enough to be measured (== 3 pg/L SRP or =1 pg/L PO4) then phosphorus is unlikely to be a limiting nutrient. It is likely that periphyton growth in Hayes Creek is limited by nitrogen rather than phosphorus. Whether the withdrawal has any effect on this limitation has not been investigated. 6.3.7 Ammonia Ammonia represents a potential hazard to fish and other aquatic life if present in high concentrations. Ammonia is present in aquatic systems in an ionized and un-ionized form, with the relative levels determined by the pH and temperature of the system. The ratio of un-ionized ammonia to ammonium ion increases 10 fold for each unit pH change, and 2 fold for each 10 °C increase (Erickson, 1985). Both forms are toxic to aquatic life though the unionized form is significantly more toxic. Thus the maximum criteria are variable with temperature and pH. For example at 15 °C , the maximum allowable criteria ammonia concentration at pH values of 7, 8, and 9 falls from 19.7 to 5.7 to 0.70 mg/L, respectively (Nagpal, 1994). High ammonia in the withdrawal water was observed twice during 1994, once during the July stratification, and again in the fall. Ammonia concentrations at the withdrawal are shown in Figure 6.4. No measured values were higher than 1 mg/L during 1994. BC criterion for ammonia are also shown in Figure 6.4. All observations were below the maximum guideline (i.e. instantaneous value) for all samples, though the last sample set collected was close to the limit. For the summer stratification period, 90 Chapter 6: Environmental Impact measured ammonia was close to, but below the 30 day average criteria. The peaks during the summer were the result of anoxia in the dredged area, and since the anoxia that results in elevated ammonia during the summer stratification is also usually accompanied by lowered pH, this withdrawal flow may not represent an area of concern (i.e. this ammonia is less toxic due to lower pH). In the fall two samples exceeded the 30-day guideline. The high levels in the fall occurred following fall turnover and represent the decay of organic matter in the lake. What appears more interesting is the high ammonia levels that occurred following fall turnover and were accompanied by high pH levels. The ammonia levels over the summer over the watershed are shown in Figure 6.5. The high levels in July and August were due to the anoxia at the lake bottom. This is what we would expect during a long stratification period. The ammonia concentration is attenuated downstream - either by biologic uptake or degradation, or dilution due to ground water joining Hayes Creek. The high values in the fall were evident at the withdrawal and all sampling stations (including the Shinish Creek Diversion). These are probably indicative of high levels of organic decay occurring in the fall5. The fact that this trend was also observed in the Shinish Creek, generally considered a low nutrient stream, indicates that the high ammonia is likely a watershed phenomenon and not a result of the hypolimnetic withdrawal. 6.3.8 I r o n and Manganese The increases of iron and manganese observed at the withdrawal (see Figure 4.4) were also observed at the downstream sampling stations. Plots of iron and manganese concentration for the four stations on Hayes Creek (Fountain, Beaver Dam, Jellicoe Rd., The observation of high ammonia levels on both the Sept and Oct sample days seems to eliminate one time sample contamination as a cause. There is a possibility that some systematic error could account for this, though none could be determined from a review of field and laboratory notes. 9 1 Chapter 6: Environmental Impact and the Fish Barrier) and the Shinish Diversion Station are shown in Figures 6.6 and 6.7, respectively. Much of the iron appears to be oxidized quickly at the fountain or in Hayes Creek. Typical soluble iron levels are 0.3 to 0.4 mg/L which is in the same range as background levels observed prior to the withdrawal operation. The concentrations of iron are attenuated through Hayes Creek. That is, the maximum value is reduced and the duration of the peak is extended over a longer time period. This may be a result of the large water volume stored behind beaver dams between the Fountain and the Beaver Dam stations which produces an effect similar to flood routing does to water flow rate. As well as physical attenuation, it is possible that there are chemical solubilization effects occurring similar to those observed in the lake. For example, there is a slight elevation of iron level at the Beaver Dam station on the last sample day coinciding with the low (2 mg/L) dissolved oxygen measurement. This suggests that iron may be oxidized at the fountain, precipitate and settle to the bottom of Hayes Creek. Later in the season the low oxygen, reducing conditions could allow for re-mobilization of this iron from the large beaver dammed area between the fountain and beaver dam stations. Virtually all of the observed manganese was in soluble form. Delfino and Lee (1971) found that soluble manganese (passing through 0.45 pm filter) was M n 2 + . The manganese observed in Hayes Creek is likely all in reduced, i.e. 2+ state. The oxidation of manganese is a slow phenomenon in natural waters (Stokes et al, 1988), particularly those with some alkalinity which may act as a complexing agent (Delfino and Lee, 1971). The dominant process affecting Mn solubility are adsorption and complexation processes which occur on times scales of hours, similar to iron oxidation. The concentrations of manganese are also attenuated in Hayes Creek and there is a slight increase in the fall, when low oxygen is observed at the Beaver Dam station. An important difference in the characteristics of the two creeks studied is that while iron is ubiquitous in the watershed, manganese is not observed in significant quantities in 92 Chapter 6: Environmental Impact Shinish Creek. Thus the presence of manganese could be used as an indicator of anoxic lake water from the withdrawal pipe. However since manganese is subject to a complex variety of sorption and complexation reactions, it is unlikely that it could be used as a conservative 'tracer' of anoxic lake water. The attenuation curves for iron and manganese may have another implication. These observations could mean that the water volume in storage between the Fountain station and the Beaver Dam station may be much larger than previously thought. That is, instead of a one or two day residence time, this region may have a ten day residence time. This would certainly explain the attenuation observed. As a final note, recall that the Fish Barrier station is located after the junction of the Hayes and Shinish Creeks, and the water flow rate is at least twice the flow in the Hayes Creek (estimated). The manganese data show this dilution effect with concentrations dropping dramatically after the junction. The iron concentration doesn't show dilution though. This indicates that much of the observed iron is from a natural source in the watershed, and not the result of the withdrawal operation. 6.3.9 Hydrogen Sulfide As already mentioned, hydrogen sulfide was detected only once, during a site visit on August 7/8. The detection was by smell and no sampling equipment was on site at that time. No sulfide smell was detected at the Jellicoe Road junction on August 8, 1994. On a return visit on August 14, no sulfide smell was detected. The sulfide smell was estimated to have lasted for a week. Sulfide can be detected by smell in clean water at concentrations of 0.025 to 0.25 Ug/L (Nagpal, 1994). Smell is thus an extremely sensitive qualitative indicator of the presence of sulfide. The sulfide smell detected was aided by the fact that the fountain aerator 'blew-off a significant amount of dissolved H2S. The BC guideline for protection of aquatic life is 2 ug/L. So the detection by smell does not assist in evaluating 93 Chapter 6: Environmental Impact the total sulfide. Part of the reason that the smell was so strong was due to the presence of the fountain aerator. The high aeration effect also 'blew off a considerable amount of sulfide. No smell was detected in the water at the Jellicoe Road station about 1700 m downstream, and no observation was made at the Beaver Dam station on August 8. These observations are consistent with our expectations. Based on his past experience Murphy expected there to be very little sulfide present (personal communication, 1994). Murphy's sediment sampling noted that of six sediment samples taken in February 1985, five were brown coloured with no sulfide smell, and only one was black coloured with sulfide smell (Murphy 1985). This one sample was from a bay in the south east corner of the lake which is wind sheltered and may have more organic sedimentation. It may be that the amount of organic material in Chain Lake is insufficient to generate large amounts of sulfide. This would support the belief that the sediment phosphorus is chemically iron bound, and not a result of organic mineralization. 6.3.10 Turbidity Prior to the installation concern had been raised that the oxidation of iron from the withdrawn water would result in a 'floe' of iron precipitate that would have significant impact on Hayes Creek - potentially extending to the Similkameen. This concern seems somewhat alarmist now, but at that time withdrawal rates of 300 L/s were being considered. On a site visit beginning August 14, the withdrawal water was observed to be slightly turbid in the form of a fine white precipitate. This had been there for a couple of days. Speculation was that this could be a precipitate of sulfur (Hall, 1994, pers eomm). A sample was taken at the time of departure, to be taken back to the environmental lab. However, by this time (August 19) the turbidity had disappeared, and so the water was not successfully analyzed for turbidity. This demonstrates the time scales on which destratification and oxidation events occur in the lake. 94 Chapter 6: Environmental Impact 6.4 Summary The environmental monitoring evaluated the expected environmental impacts on Chain Lake and Hayes Creek. The withdrawal is expected to have a negligible impact on the thermal structure of the lake. De-stratifications are regular occurrences at Chain Lake, and virtually no changes will be noticeable as a result of the withdrawal A more probable thermal effect on the lake will be if the withdrawal increases visibility. Increased light penetration could alter the stratification more severely than the heat retained by the withdrawal. The meteorological conditions produced in 1994 a worse case event from an environmental impact perspective. Dissolved oxygen deficits were observed throughout the summer at the Fountain. The Fountain aerator was able to increase the dissolved oxygen by about 1.5 mg/L. High levels of ammonia, iron, and manganese were observed during the July stratification. The ammonia was quickly consumed, or lost in Hayes Creek, while iron and manganese exhibited a flood routing effect in which the peak levels were attenuated, and the duration increased. After 5 weeks of stratification, sulfide was detected on August 8. Overall, the environmental impacts of the withdrawal are localized within the area of the Hayes Creek between the Fountain and the Jellicoe Road station (about 500 m downstream). 95 7. CONCLUSIONS and R E C O M M E N D A T I O N S : 7.1 Summary and Conclusions From the work presented, and the observations made, conclusions can be made in several categories: Technology of withdrawal: The withdrawal was installed by resident volunteers, without the use of specialized equipment. The system was designed to highlight simple operation. From this experience it may be concluded that: • the gravity driven withdrawal in Chain Lake is a low technology, zero operating cost technique for lake restoration. Operation can be designed to minimize operator involvement, yet provide a simple calibration to enable residents to accurately log flow rates. • the application is suited to sparsely populated regions in the B C interior, where lakes have high value as recreational sites yet low tax bases and intermittent (i.e. recreational) populations. The physical nature of Chain Lake: Observations of the thermal behavior of the lake indicate the lake is polymictic, undergoing several stratifications and destratifications each year. Seiching was observed with a period during summer stratification on the order of 10 hours. From these observations it may be concluded that: • seasonal stratification at Chain Lake is incomplete. Stratification events may range from several days to several weeks in length. These stratifications are formed and removed on time scales of a few days. The prediction of these stratifications in advance is difficult due to this short time scale. 9 6 Chapter 7: Conclusions and Recommendations • the water column may be modeled as a two-layer stratified or a continuously stratified system. Estimates of the seiche period based on these models appear to bound the observed response. This indicates a mixed behavior between two-layer and linearly stratified behavior. • wind set-up is observed in agreement with theoretical expectations. • the dredged hole provides a region of thermal stability and may represent the only true hypolimnion in the lake. • the dredged area may not mix completely with the water column during spring overturn. • the Shinish diversion water can be expected to sink to the bottom of the lake throughout the summer. The oxygen content of the Shinish diversion is not sufficient to satisfy the sediment oxygen demand. Thus, anoxia at the sediment / water region is expected to persist. Short term (first year) lake response: During 1994 Chain Lake experienced excellent water quality through July. This effect was likely due to the calm weather conditions. Partial destratification at the end of July promoted an algae bloom. From the observations it may be concluded that: • the short term response of the withdrawal was limited by the late commencement of withdrawal, and the low flow rates imposed through the summer. • the onset of an algae bloom was delayed in 1994 over 1993. Attributing this to the withdrawal is not yet possible. • inter-year variation is a dominant factor in the lake response • maximum phosphorus levels were about the same as in other years 97 Chapter 7: Conclusions and Recommendations Longer Term Potential: During 1994 the withdrawal removed 30 kg of phosphorus from the lake. Under better operating conditions this could be increased to about 50 kg. From this and the analysis of the lake data set it can be concluded that: • the withdrawal can be expected to result in a net export from the lake of 5 to 20 kg total phosphorus per year. • the net export will deplete the sediment phosphorus pool slowly in a time of between 380 and 1500 years. • operation could be optimized to improve the export of anoxic water during strongly stratified times. Nutrient Loading: The profiles of oxygen and the chemical samples indicate the depletion of oxygen at the sediments and the release of phosphorus, iron, and manganese from the sediments. The nature of these releases indicate that: • an increase of phosphorus concentration occurred at the bottom of the lake during July in both years. This build up was stable for over two weeks in both years in spite of near isothermal conditions in 1993. • large fluxes of phosphorus are moving to and from the sediment surface continuously. The historical data set suggests that the net flux is relatively small compared to individual fluxes. Environmental Impact: The withdrawal water released to Hayes Creek was at times during 1994 low in oxygen, and high in phosphorus, ammonia, iron, and manganese. From the monitoring program it may be concluded that: 98 Chapter 7: Conclusions and Recommendations • the dissolved oxygen of the withdrawal stream was increased slightly during flow through the withdrawal 'box'. This verified that some leakage was occurring though it was estimated to be no more than 10% of the withdrawal flow. • active aeration employing a gravity driven fountain was found to add between 1.5 and 2.0 mg/L oxygen to the withdrawn water. • dissolved oxygen deficits were observed in Hayes Creek during withdrawal operation. However, DO sags during the fall occurred while the withdrawal was fully aerated implying a substantial benthic demand in the slow moving areas of Hayes Creek directly downstream (i.e. within 500 m) of the withdrawal. • ammonia levels increased during the withdrawal of anoxic water to a maximum of 0.8 mg/L. The levels at the fountain release were below B.C. criteria guidelines considered hazardous to aquatic life. The ammonia was quickly attenuated and / or consumed in Hayes Creek. • ammonia levels increased in the withdrawal and at all the stream stations during the fall. This is likely a result of general decomposition of organic matter during this season and is not a result of withdrawal operation. • elevated iron levels were observed downstream of the withdrawal. These were limited to the region between Chain Lake and the Hayes / Shinish junction. Below this junction iron concentrations increased, indicating iron entering Hayes Creek through surface or ground water flows. • iron is ubiquitous in the surface waters of the area, particularly Shinish Creek. Detection of elevated levels of iron is not a good indicator of anoxic-source water from the Chain Lake withdrawal. • manganese is not present in large quantities naturally, and is below M D L in Shinish Creek (at the point of diversion). Detection of manganese may be suitable for use as an indicator of anoxic-source water from the lake. 9 9 Chapter 7: Conclusions and Recommendations • detrimental impacts were small, short-lived, and localized in the region within 500 m of the withdrawal 7.2 Recommendations Recommendations can be made concerning the withdrawal system Chain Lake Withdrawal The following recommendations can be made regarding the withdrawal. • The withdrawal should be started immediately after ice off and operated at a maximum possible rate during the spring. The objective is to remove the mass of water contained inside the dredged hole. Removal of this water may limit the nutrient pool in the lake and potentially reduce the amount of algae in the lake later in the spring. • The withdrawal should be operated as long as possible during the summer and at maximum flows rates possible within the operational constraints of maintaining the lake level. The limit to the withdrawal rate at Chain Lake is hydrologic (i.e. limited by the amount of water entering the lake water in the lake) and not hydraulic (i.e. not by the ability to move water through the withdrawal system). • Secchi measurements, and withdrawal logs should be maintained. These records provide simple, yet valuable information regarding the effectiveness of the withdrawal. Other Withdrawal Candidates Limnological data is time consuming and/or expensive to gather, while the desire for more data by the technical analyst is insatiable! To that end, the following recommendations apply to other projects where a withdrawal is being contemplated. 100 Chapter 7: Conclusions and Recommendations • Collect as much baseline data as possible. Typically the time from the first evidence of a water quality problem to the implementation of a restoration plan often takes several years. Something as simple as a weekly secchi reading over this period would be valuable data. • Confirm the source of nutrient loading. The background work to this study predicted the sediment based internal loading. The internal loading feature is what allows the withdrawal to accelerate phosphorus withdrawal without accelerating water withdrawal. This is crucial since typically lakes with eutrophic summertime conditions tend to have low water flows in the summer as well. 7.3 Further Research Many of the subjects discussed in this thesis have not been fully explored. Further research could be done to: • quantify the rate of release of phosphorus from the sediments. • quantify the seasonal rate of phosphorus sedimentation using sediment traps • quantify the vertical movement of hypolimnetic water with particular emphasis on the mixing effects of internal seiching on vertical transport. • quantify the effect of the morphology on thermally stabilizing the water column. • perform nutrient enrichment experiments to determine the nutrient limitation effects and seasonal variation of nutrient limitation. 101 8. R E F E R E N C E S Ahlgren, I., T. Frisk, and L. Kamp-Nielsen, 1988, Empirical and theoretical models of phosphorus loading, retention and concentration vs. lake trophic state, Hydrobiologia, 170 : 285-303 Atmospheric Environment Service (AES), 1995, Data Download from Princeton Weather Station, Atmospheric Environment Services - Environment Canada, Applications and Services Division, Pacific and Yukon Region Box, G.P., and W.G. Hunter, J.S. Hunter, 1978, Statistics for Experimenters: An Introduction to Design, Data Analysis, and Model Building, John Wiley & Sons Bull, C.J., 1975, Memo to file on the Chain Lake Fish Kill, Fisheries Division, Penticton Office Bull, C.J., 1982, Chemical Rehabilitation of Chain Lake, Fisheries Division, Region 8 Carlson, R.L., 1977, A trophic state indicator for lakes, Limnol. Oceanogr., 22(2) : 361-369 Carlson, R.L., 1995, personal communication. Chapra, S.C., and K.H. Reckhow, 1983, Engineering approaches for lake management; Volume 2: Mechanistic Modelling, Butterworth Publishers Cole, T .M. , and E . M . Buchak, 1993, CE-QUAL-W2: A two-dimensional, laterally averaged, hydrodynamic and water quality model (users manual), US Army Corps of Engineers Waterways Experiment Station Cooke, G.E., E.B. Welch, S.A. Peterson, P. R. Newroth, 1993, Restoration and Management of Lakes and Reservoirs - 2nd edition, Lewis Publishers, Boca Raton, F L Cornett, R.J. and F.H. Rigler, 1984, Dependence of hypolimnetic oxygen consumption on ambient oxygen concentration: fact or artifact?, Water Resources Research, 20(7) : 823 - 830 Davison, W., 1985, Conceptual models for transport at a redox boundary, in Chemical processes in lakes, W. Stumm (ed), J. Wiley & Sons Davison, W., C. Woof, and E. Rigg, 1982, The dynamics of iron and manganese in a seasonally anoxic lake; direct measurement of fluxes using sediment traps, Limnol. Oceanogr., 27(6), 987-1003 Dawson, J., 1993, personal communication Delfino, J.J., and G.F. Lee, 1968, Chemistry of manganese in Lake Mendota, Wisconsin, Environmental Science and Technology, 2 (12): 1094-1100 102 Delfino, J.J., and G.F. Lee, 1971, Variation of manganese, dissolved oxygen and related chemical parameters in the bottom waters of Lake Mendota, Wisconsin, Water Research., 5 : 1207-1217 Demers, E. , and J. Kalff, 1993, A simple model for predicting the date of spring stratification in temperate and subtropical lakes, Limnol. Oceanogr., 38(5) 1077-1081 Demayo, A., A.R. Davis, and M.A. Forbes, 1978, Forms of metals in water, Scientific series No. 87, Inland Waters Directorate, Pacific and Yukon Region, Water Qualilty Branch, Environmental Canada Duthie, and Beckman, 1980, Solar engineering of thermal processes, John Wiley & Sons Ennis, Gordon L. , 1972, The effects of low nutrient water Additions on summer phytoplankton diversity in a small, eutrophic lake, BSc. Thesis, Department of Biology, University of British Columbia EPA, 1988, Methods, 40 CFR Ch.I (7-l-88Edition) Pt. 136, App. B, pp510-512 Erickson, R.J., 1985, An evaluation of mathmatical models for the effects of pH and temperature on ammonia tixicity to aquatic organisms, Water Res., 19(8): 1047-1058 Fallon, R.D., and T.D. Brock, 1980, Planktonic blue-green algae: Production, sedimentation, and decomposition in Lake Mendota, Wisconsin, Limnol. Oceanogr., 25(1) : 72-88 Fischer, H.B., E.J. List, R.C.Y. Koh, J. Imberger, and N.H. Brooks, 1979, Mixing in inland and coastal waters, Academic Press. Inc., Toronto Gibbons, H.L. 1993, interview June 1993 Gordon, J.A., W.P. Bonner, and J.D. Milligan, 1983, Iron manganese, and sulfide trasnsformations downstream from Normandy Dam, Lake and Reservoir Management, in Lake and Reservoir Management, Proc of the 3rd Annual Conference of the North American Lake Management Society, October 18-20, 1983, Knoxville, Tennessee Golterman, H.L., 1995, The role of the ironhydroxide-phosphate-sulfide system in the phosphate exchange between sediments and overlying water, Hydrobiologia, 297 : 43-54 Gorham, E , and F.M. Boyce, 1989, The influence of lake surface area and depth upon thermal stratification and the depth of the summer thermocline, / . Great Lakes Research, 15(2): 233-245 Hakanson, L , and M . Jansson, 1983, Principles of lake sedimentology, Springer-Verlag Harleman, D.R.F., 1982, Hydrothermal analysis of lakes and reservoirs, J. EHydraulics Division - ASCE, 108 (3): 302-325 103 Hall, K.J., T. Murphy, and M Mawhinney, 1993, Iron Treatment of Black Lake, presented at the 13th International Symposium of the North American Lake Management Society, November 30 - December 4, 1993, Seattle, Washington Holtman, H. , L. Kamp-Nielson, & A.O. Stuanes, 1988, Phosphorus in soil, water, ans sediment: an overview, Hydrobiologia, 170 : 19-34 Henderson-Sellers, B., 1984, Engineering Limnology, Pitman Advanced Publishing Program, Boston Hutchinson, G.E., 1957, A treatise on Limnology, Vol 1: geography, physics, and chemistry, John Wiley & sons, New York, 1015 pp Idso, S.B.,1973, On the concept of lake stability, Limnol. Oceanogr., 18(4): 681-683 Imboden, D.M., and A. Lerman, 1978, Chemical models of lakes, in Lakes: chemistry, Geology, Physics, A. Lerman (ed), Springer-Verlag Imboden, D.M. , and R.P. Schwarzenbach, 1985, Spatial and temporal distribution of chemical substances in lakes:modeling concepts, in Chemical Processes in Lakes, W. Stumm (ed), J. Wiley & Sons Kalff, J., 1991, The utility of latitude and other environmental factors as predictors of nutrients, biomass, and production in lkaes worldwide:Problems and alternatives, Verh. Internat. Verein. Limnol., 24 : 1235-1239 Kallio, K., 1994, Effect of summer weather on internal loading and chlorophyll-a in a shallow lake: a modelling approach, Hydrobiologia, 275/276 : 371-378 K C M , 1986, Restoration of Lake Ballinger, phase III final report, Kramer, Chin & Mayo, Inc. Kleiber, P., P.H. Whitfield, and W.E. Erlebach, 1978, Limitations of single water samples in representing mean water quality: II Spatial and tempporal variation in nutrient concentrations in the Okanagan River at Oliver, B.C., Technical Bulletin No. 107, Inland Waters Directorate, Pacific and Yukon Region, Water Qualilty Branch, Environmental Canada Kleiber, P. and W.E. Erlebach, 1977, Limitations of single water samples in representing mean water quality: III Effect of variability in concentration measurements on estimates of nutrient loadings in the Squamish River, B.C., Technical Bulletin No. 103, Inland Waters Directorate, Pacific and Yukon Region, Water Qualilty Branch, Environmental Canada Kortmann, R.W., E. Davis, C.R. Frink, and D.D. Henry, 1982, Hypolimnetic withdrawal: Restoration of Lake Wononscopomuc, Connecticut, in Lake Restoration and Management, Proc. of the Second Annual Conference of the North American Lake Management Society, Oct 26-29, 1982 Vancouver B.C. Lacelle, L .E .H. , 1986, Estimates of Nitrogen and phosphorus loadings to the Chain-Link -Osprey lakes systems from domestic waste sources, Ministry of Environment and Parks, Wildlife Program, Okanagan Sub-Region 104 Lee, G.F., 1970, Factors affecting the transfer of materials between water and sediments, University of Wisconsin, Water Resoureces Center, Eutrophication Information Program, July, 1970 Lee, G.F., W.C. Sonzogni, and R.D. Spear, 1976, Significance of oxic vs. anoxic conditions for Lake Mendota sediment phosphorus release, in Interactions between sediments and fresh water, proceedings of an international symposium held at Amsterdam, the Netherlands, Sept 6-10, 1976, W. Junk Pub. Lijklema, L. and A.H.M. Hieltjes, 1982, A dynamic phosphate budget model for a eutrophic lake, Hydrobiologia, 91 : 227 - 233 Livingstone, D.M., and F. Schanz, 1994, The effects of deep-water siphoning on a small, shallow lake: A long-term case study, Arch. Hydrobiol, 132 : 15-44 Macdonald, R. H., 1994, Chain Lake Water Quality Improvement Project (Hypolimnetic Withdrawal): Information Supplement to the Environmental Protection Program, Penticton Office, B.C. Ministry of Environmet Lands and Parks, May 31, 1994, Mathews, Steve, 1982, Determination of degree of circulation of Shinish Creek water within Chain Lake, Technical Report, Fisheries Branch, Okanagan Sub-Region, Southern Interior Region, BC Ministry of Environment File: 40.3504 Chain Lake Mathews, S., 1992, 1994, pers comm McCarty, P.L., 1970, Chemistry of nitrogen and phosphorus in water, Journal Amer. Water Works Assoc., 62(2): 127-140 McCallum, 1995, An examination of trace metal contamination and land use in an urban watershed, MSc thesis, University of British Columbia Monosmith, S, 1986, An experimental study og the upwelling response of sdtratified reservoirs to surface shear stress, J. Fluid Mech., 171 : 407-439 Mortimer, C.H. 1941 The exchange of dissolved substances between mud and water in lakes I and II, / . Ecol, 29 : 280-329 Mortimer, C.H. 1942 The exchange of dissolved substances between mud and water in lakes I and II, J. Ecol., 30 : 147-201 Mortimer, C.H., 1971, Chemicla exchanges between sediments and water in the great lakes - speculations on probable regulatory mechanisms, Limnol. Oceanogr., 26(2): 387-404 Moss, B., 1988, Ecology of Fresh Waters: Man and Medium 2nd Ed., Blackwell Scientific Publications, London Murphy, T.P., 1985, The effect of a water diversion on a eutrophic lake., NWRI Contribution Series 85-168, Aquatic Ecology Division, National Water Research Institute Murphy, T.P., 1987, Sediment Phosphorus release reduces the effect of the Chain Lake Water Diversion, Lake and Reservoir Management 3 : 49-57 105 Murphy, T.P., and D. Urciuoli, (unpublished draft report =1984), Limnological Observations of Chain Lake, Aquatic Ecology Division, National Water Research Institute Murphy, T.P., I. Gray, C. McKean, P.G. Manning, and A. Moller, 1990, Diatom indicators of fluctuations of eutrophication and water levels in Chain Lake, British Columbia, NWRI Contribution Series 90-46, Aquatic Ecology Division, National Water Research Institute Nakasone, H. , 1987, Study of aeration at weirs and cascades, J. Environmental Engineering, 113 : 64-81 Nagpal, N.K., 1994, Approved and Working Criteria for Water Quality - 1994, Water Quality Branch, Environmental Protection Department, Ministry of Environment, Lands, and Parks, February 1994 Northcote, T.G. , 1967, An investigation of summer limnological conditions in Chain Lake, British Columbia, prior to introduction of low nutrient water from Shinish Creek, Management Report #55, Fish and Wildlife Branch, Department of Recreation and Conservation, Victoria, B.C. Niirnberg, G.K., 1984, Prediction of internal phosphorus load in lakes with anoxic hypolimnia, Limnol. Oceanogr., 29(1): 111-124 Niirnberg, G.K., 1985, Availability of phosphorus upwelling from iron-rich anoxic hypolimnia, Arch. Hydrobiol, 104(4): 459 - 476 Niirnberg, G.K., 1987a, Hypolimnwtic withdrawal as lake restoration technique, ASCE J. of Environmental Engineering , 113(5): 1006-1017 Niirnberg, G.K., 1987b, A comparison of internal phosphorus loads in lakes with anoxic hypolimnia: Laboratory incubation versus in situ hypolimnetic phosphorus accumulation, Limnol. Oceanogr., 32(5): 1160-1164 Niirnberg, G.K., 1988, A simple model for predicting the date of fall turnover in thermally stratified lakes, Limnol. Oceanogr., 33(5): 1190-1195 Niirnberg, G.K., R. Hartley, and E. Davis, 1987, Hypolimnwtic withdrawal in two North American lakes with anoxic phosphorus release from the sediment, Wat. Res., 21(8): 923-928 Oguss, E. and W.E. Erlebach, 1976, Limitations of single water samples in representing mean water quality: I Thompson River at Shaw Spring, British Columbia, Technical Bulletin No. 95, Inland Waters Directorate, Pacific and Yukon Region, Water Qualilty Branch, Environmental Canada Olszewski, P., 1961, Versuch einer Ableitung des hypolimnischen Wassers aus einem See: Ergebnisse des ersten Versuchsjahres (The attempted diversion of hypolimnetic water of a lake: Results of the first years of experiments), Verh. Internal Verein. Limnol., XIV pp 855-861, (Translated by B Genzmer Nov. 1993) 106 Olszewski, P., 1973, Funfzehn jahre experiment auf dem Kortowo-See (Fifteen years of experiments on the Kortowo-Lake), Verh. Internal Verein. Limnol., 18 pp 1792-1797, (Translated by B Genzmer Nov. 1993) Ragotzkie, R.A., 1978, Heat budgets of lakes, in Lakes: chemistry, Geology, Physics, A. Lerman (ed), Springer-Verlag, Reckhow, K.H. , and S.C. Chapra, 1983, Engineering approaches for lake management; Volume 1: Data analysis and empirical modelling, Butterworth Publishers Rhee, G-Yull, 1978, Effects of N:P ratios and nitrate limitation on algal growth, cell composition, and nitrate uptake, Limnol. Oceanogr., 23(1), 10-25 Rich, P.H., and R.G. Wetzel, 1978, Detritus in the lake ecosystem, Amer. Natur., 112 : 57-71 Riley E.T. ,and E.E. Prepas, 1984, Role of internal phosphorus laoding in two shallow, productive lakes in Alberta, Canada, Can. J. Fish. Aquat. Sci., 41 : 845-855 Schindler, D.W., 1985, The coupling of elemental cycles by organisms: evidence from whole-lake chemical perturbations, in Chemical Processes in Lakes, W. Stumm (ed), J. Wiley & Sons Sly, P.G., 1978, Sedimentary processes in lakes, in Lakes: chemistry, geology, physics, A. Lerman (ed), Springer-Verlag Smith, V .H. , 1983, The nitrogen and phosphorus dependence of blue-green algal dominannce in lakes, in Lake Restoration, Protection, and Managment, proceediiings of the 2nd Annual Conference North American Lake Management Soociety, Oct 26-29, 1982, Vancouver B.C. Spigel R.H., and J. Imberger, 1980, The classification of mixed-layer dynamics in lakes of small to medium size, / . Physical Oceanography, 10:1104-1121 Standard Methods for the Examination of Water and Wastewater, 1992, American Public Health Association, 18th Edition Stauffer, R.E., 1987, A comparative analysis of iron, manganese, silica, phosphorus, and sulfur in the hypolimnia of calcareous lakes, Wat. Res., 21(9): 1009-1022 Stefan, H.G., and M.J. Hanson, 1980, Predicting dredging depths to minimize internal nutrient recycling in shallow lakes, Restoration of Lakes and Inland Waters: International Symposium on Inland Waters and Lake Restoration, Portland, Maine, Sept. 8-12, 1980, EPA 440/5-81-010 Stefan, H.G., and M.J. Hanson, 1981, Phosphorus recycling in five shallow lakes, J. Environmental Engineering Division - ASCE, 107 : 713 - 730 Stokes, P.M., P.G.C. Campbell, W.H. Schrader, C. Trick, R.L. France, K J . Puckettt, B. LaZerte, M . Speyer, J.E. Hannaa, and J. Donaldson, 1988, Manganese in the Canadian Environment, National Research Council of Canada, NRCC Associate Committee on Scientific Criteria for Environmental Quality, Publication Number NRCC 26193 107 Stumm, W and E Stumm-Zolinger, 1972, The role of phosphorus in Eutrophication, Water Pollution Microbiology, Wiley-Science, Toronto Swain, L .G. , 1985, Okanagan area, Similkameen sub-basin, Water quality asessment and objectives: Technical Appendix, Resource Quality Section, Water Management Branch, Ministry of Environment, Province of British Columbia Taylor, G.D., 1971, Evaluation of partial diversion of Shinish Creek to Chain Lake (1970), Fish Habitat Improvement Section, British Columbia Fish and Wildlife Thermo Jarrel Ash, 1986, Atomic absorption: Methods manual for flame operation Vol 1 Thorpe, S A . , 1968, On standing internal gravity waves of finite amplitude, J Fluid Mech 32(3): 489 van Orman, E. , 1993, Modelling sediment phosphorus release through analysis of sediment phosphorus profiles, MSc thesis, Michigan technological University Walpole R.E. and R.H. Myers, 1978, Probability and Statistics for Engineers and Scientists 2nd edition, Macmillan Publishing Co. Inc. Water Investigations Branch, 1977, Observations on the water quality in the Chain-Link-Osprey lakes system Princeton B.C. 1973 to 1976: Part I. Water Quality Data for Chain Lake and Discussion of the effectiveness of the Shinish Creek diversion, BC Ministry of the Environment, File: 0290862-Gen Weigand, R.C., 1989a, The Chain Lake Dredging Experiment, Volume I, The Dredging of Chain Lake, SciTech Consultants Weigand, R.C., 1989b, The Chain Lake Dredging Experiment, Volume III, Assessment of the Dredging of Chain Lake: Summary of Physical, Biological, and Sedimentological Data Collection (Figures) Weigand, R.C., and N. Munteau, 1989, The Chain Lake Dredging Experiment, Volume II, Assessment of the Dredging of Chain Lake: Summary of Physical, Biological, and Sedimentological Data Collection Wetzel, R.G., 1975, Limnology, W.B. Saunders and Company, Toronto Yoshida, T., 1982, On the summer peak of nutrient concentrations in lake water, Hydrobiologia, 92 : 571 - 578 108 9. FIGURES Figure 1.1 Morphometric Map of Chain Lake (after Murphy et al, 1990) Figure 1.2 Aerial Photo of Chain Lake Figure 2.1 Comparison of Two-layer and Linear Stratification Parameters -Figure 2.2 Typical Sediment Structure at Chain Lake Figure 2.3 Sediment Composition: Iron, Phosphorus, Water Content, and Age vs. depth (after Murphy et al, 1990) Figure 3.1 Schematic of Withdrawal Chain Lake Withdrawal Figure 3.2(a) Photograph of Withdrawal 'Box' Figure 3.2(b) Photograph of Aerator Fountain Figure 3.3 Stream Sampling Locations Figure 3.4 Extract of Times series and Daily Averaged Data Figure 3.5 Chain Lake Sample Schematic for 94-07-23 Figure 4.1 Historical Meterolrologic Conditions at Princeton Figure 4.2 Meteorlogical Measurements at Chain Lake -1994 Figure 4.3 Daily Average Water Temperatures in Chain Lake Figure 4.4 Hypolimnetic Withdrawal Composition - 1994 Figure 4.5 Temperatures of Shinish Creek and Hypolimnetic Withdrawal Figure 4.6 C T D transect 95-07-07 Figure 4.7 Thermistor Data Extract from July 24 to July 27, 1994 Figure 4.8 Expected Chain Lake Seiche Period and Observed FFT Components-1994 Figure 4.9 Wind Forcing and Isotherm Response July 15 to July 26, 1994 Figure 4.10 CTD Transect Profile Summary Figure 4.11 Lake Stability, Wind Work and Daily Precipitation - 1994 Figure 4.12 Temperature and Dissolved Oxygen Profiles in the Dredged Hole Figure 4.13 Heat Content of Chain Lake 1993 and 1994 Figure 4.14 Water Budget for Chain Lake - 1994 109 Figure 4.15 Dissolved Oxygen Profiles - 1993 Figure 4.16 Dissolved Oxygen Profiles - 1994 Figure 4.17 Chain Lake 1993: Secchi Depth, Chla, and Total Phosphorus Figure 4.18 Chain Lake 1994: Secchi Depth, Chla, and Total Phosphorus Figure 4.19 Chain Lake 1994 Phosphorus Concentrations Figure 4.20 Chain Lake 1994 Iron Concentrations Figure 4.21 Chain Lake 1994 Manganese Concentrations Figure 4.22 Ratio of Iron to Manganese in the Withdrawal Flow - 1994 Figure 4.23 Hypolimnetic Withdrawal Composition August 4 to Sept 1, 1994 Figure 5.1 Flow Test of Withdrawal System 95-05-20 Figure 5.2 Chl a vs. Total Phosphorus in Chain Lake Figure 5.3 Maximum Measured Phosphorus Content in Chain Lake Figure 5.4 Chain Lake Secchi Measurement - Historical Perspective Figure 5.5 Apparent Net Phosphorus Concentration Change - 1994 Figure 5.6 Chain Lake Water and Phosphorus Export - 1994 Figure 5.7 Chain Lake Phosphorus Renewal Time - 1994 Figure 5.8: Lake Stability Correlation to Princeton Air Temperature Figure 5.9: Princeton Mean Air Temperature during summer - 1937 to 1994 Figure 5.10: Results of 'Worm Count' of Princeton Warm Periods Figure 6.1 Hayes and Shinish Creek Total Phosphorus, 1993-08-19 Figure 6.2 Chain Lake Withdrawal Fountain Dissolved Oxygen Levels -1994 Figure 6.3 Hayes Creek Dissovled Oxygen Levels -1994 Figure 6.4 Withdrawal Ammonia Levels - 1994 Figure 6.5 Hayes and Shinish Creek Ammonia Levels - 1994 Figure 6.6 Hayes and Shinish Creek Iron Levels - 1994 Figure 6.7 Hayes and Shinish Creek Manganese Levels - 1994 110 MJX>= M i d Lake Station N O R T H = N o r t h Station Figure 1.1 Morphometric Map of Chain Lake (after Murphy et al, 1990) 111 112 Two- layer Stratification 4 T i . P i h 1 T 2 ' P 2 A p=p 2 -P l g' = g Ap/p 2 Ri = g'h u * 2 Linear or Continuous Stratification ^ A p ^ | Az H N 2 = 8/p 0 * 5 p / 8 z =g/p 0 *A P / A z g' =ll2 N 2 H Ri = g'(H/2) u * 2 P0= ! / 2 (P2 + Pi) Figure 2.1 Comparison of Two-layer and linear Stratification 113 Water Column Water Sampler begins to collect sediments... ~ 5.5 m 1 . - 6 . 2 m 0.4 kg D.O. Prober almost always 0 mg/L oxygen Unconsolidated Sediment j7 kg weight • • • • • • • • • • • • • a Figure 2 . 2 Typical Sediment Structure at Chain Lake 114 911 LU 3 era' Br** 5" ? n ff < e , •3 , to ©• & * 9" o a t ? w 3 O S ^ 2 3 ° ^ B Si ST 3"< f»Oq » ? a g if is ^ o era 5* - ff po • - f f 3 g sn o ^ 2 o 5*1 t?ff 3 s Q 2 a g IJ &f s o a. Figure 3.2(b) Photograph of Aerator Fountain Taken at time of installation. Culvert pipe is 50 cm diameter - flow rate approximately 50-70 L/s. Author (185 cm height) included for scale. 118 6 1 1 i—»• CfQ § 00 9 00 p r o o &. o p co —* -P CD CO s on P" CO p* n 9 CD co 0) 3 5' o" o fi) o 3 (A I "ST 0 Q 0 D Q. (0 O (D 7T # S. 3 CD W o Z3 m s CO 3 O. a s;u; DH (D No ZJ rth End Sample Date: 94-07-23 ations-->> chla (micro-g/L) Dredged Hole Cross Section: S <--> N Met Station Mid Lake _L 8.7 9.5 _ / « -Dredged Hole "577-11.3 47 North End J 7 Peptfo (m) 1-2 •4 f 6 •8 Total P (micro-g/L) Total Solu (micro-g/L) Soluble Re! (micro-g/L) 36 39 "76-19 1.5 1.5 35 48 ~9~ 44 2.4 1.1 7 7 7 Ammonia (mg/L) <0.010 <0.010 <0.010 <0.010 7 Total Fe (mg/L) Soluble Fe' (mg/L) Total Mn (mg/L) Soluble M (mg/L) 0.071 0.128 0.052 0.069 0.041 0.056 0.031 0.039 0.086 0.168 0.052 0.037 "0U43-0.078 0.035 0.056 NOT TO SCALE 7 7 7 Figure 3.5 Chain Lake Sample Schematic 94-07-23 The samples were taken at a time of strong thermal stratification. The concentrations of Fe, and Mn are significantly greater at 4 m depth than at 1 m. The dredged hole contains noticeably higher concentrations of phosphorus, ammonia, iron and manganese than the rest of the water column. 121 Figure 4.1 Meterolrologic Conditions at Princeton (a) 58 year mean of daily average temperaturewith 1993 and 1994 average temperatures overlain (b) average monthly rainfall with 1993 and 1994 rainfall shown 122 24 05-01 06-01 07-01 08-01 09-01 10-01 (a ) D ° t e Figure 4.3 Daily Average Water Temperatures Recorded at met station: (a) 1993 and (b) 1994 124 (a) Apr-01 May-01 Jun-01 Jul-02 Aug-01 Sep-01 Oct-02 9.0 8.0 7.0 (b) A P r " 0 1 May-01 Jun-01 Jul-02 Aug-01 Sep-01 Oct-02 350 300 250 200 150 100 50 0 (ug/L) • Total Phosphorus o Total Soluble Phosphorus - - - O- - - Soluble Reactive Phosphorus - p (c) Apr-01 May-01 Jun-01 Jul-02 Aug-01 Sep-01 Oct-02 Figure 4.4 Hypolimnetic Withdrawal Composition - 1994 (a) temperature and dissolved oxygen, (b) pH, and (c) phosphorus (total, total soluble, and soluble reactive) Question marks indicate that no sample was available for soluble analysis on August 8. 125 Figure 4.4 (continued) Hypolimnetic Withdrawal Composition - 1994 (d) Nitrogen - Ammonia and Nitrate+nitrite, (e) Iron - total and soluble, and Manganese - total and soluble Question marks indicate that no filtered sample was available for soluble analysis on August 8. 126 — — ^ Diversion at Shinish Creek -)K Hypolimnetic Withdrawal 20 o cn a) 15 "D Q_ E 0) (b) 1 0 h 5 h i i i i i i I i m i i i i i—i—i—i—i—i—i—i—i—r Hypolimnetic Withdrawal Diversion Ditch at entry to Lake 1 I I I I 0 8 - 1 8 0 8 - 2 5 0 9 - 0 1 Date ( 1 9 9 4 ) 0 9 - 0 8 Figure 4.5 Temperatures of Shinish Creek and Hypolimnetic Withdrawal, (a) single readings, and (b) continuous time series 94-08-17 to 94-09-12 The diversion water entering the lake is consistently cooler than the withdrawn water. This implies that the Shinish inflow should sink to the bottom of the lake, below the withdrawal depth. 127 ( a ) 2 4 : 0 0 4 3 — 2 1 188.00 6:00 t ime (hour) 12 :00 18 :00 2 4 : 0 0 from North f i x 1 from South 188.25 188.50 day 188.75 (Arrows indicate, time of CTD profiles) 189.00 15 T (deq C) 17 19 15 1 T 7 ( d ' f e C ) 17 T (deq C) 17 19 15 T (deq C) 19 17 Figure 4.6 CTD transect 95-07-07 (a) wind forcing on July 7, (b) series of 5 profiles taken along lake axis, and (c) cross setion of lake showing contours of water temperature derived from the profiles. The effect of the wind has been to mix a surface 'wedge' of water and tilt the isotherm structure. 128 T h e r m i s t o r Time Series \ 0 7 - 2 4 \ ' 0 7 - 2 5 24 hr 209 100 Figure 4.7 Thermistor Data Extract: July 24 to July 27,1994 w o v e , i l s = / m s / t h e s i s• U u" l-"" n p l e- p l t (a) water temperatures at met station thermistors, (b) spectra of the 0.5 m thermistor, and (c) spectra of the 3.5 m thermistor. The surface water shows a definite diurnal response (i.e. 24 hr period) due to daily heating and cooling while the deeper thermistors have a dominant response period of 10 hours as a result of seiching. 129 o e i . Period (hours) —* —* NO N J OJ o c n o c n o c n o i i i i | i i i i | i i i i | i i i r~j i i i i | i i r n i i i i [Windspeed Averaged oyer 2 .40 hours"l A2 196 198 200 202 204 206 Iso therms at South End (based on 2 .40 hour average water t emp) 208 (b) 1000.0 100.0 1 9 6 1 9 8 200 202 . 2 0 4 20§ws indicates thermistoZOjgpth Wedderburn Number - (wind averaged over 2.40 hours/daily averaged water temp) 200 202 Date (Julian day of 1994) 204 206 208 wavefile... wind Jsotherms_wed. pit Figure 4.9 Wind Forcing and Isotherm Response July 15 to July 26, 1994 (a) windspeed (squared), (b) isotherm response at the south end met station, and (c) Wedderburn number associated with the wind event 131 Figure 4.10 CTD Transect Profile Summary (a) May 9, (b) May 23, (c) June 20, and (d) July 7 The dotted line indicates the estimated observed slope of thermal structure (see Table 4.2). The arrows indicate the horizontal position of the profile along the axis of the lake. 132 5 0 4 0 % 3 0 ^2, O CO 2 0 10 (a) 0 2 5 2 0 15 E 10 E r-(bf 5 0 Lake Stability Wind Energy 0 5 - 0 1 0 6 - 0 1 0 7 - 0 1 0 8 - 0 1 0 5 - 0 1 0 6 - 0 1 0 7 - 0 1 Date 0 9 - 0 1 0 8 - 0 1 0 9 - 0 1 wave file = /thermistors/stab_windwork_thesis.plt.. Figure 4.11 Lake Stability, Wind Work and Daily Precipitation -1994 Low wind energy input (i.e. work) is required to allow the establishment of thermal stability. A decrease of stability may be a result of wind mixing (destratifications in May and June), convective cooling (decline of stability in late July), or preciptation (drop in stability accompanying Aug 8 rainfall). 133 94-02-04 Figure 4.12 Temperature and Dissolved Oxygen Profiles at the Dredged Hole -1994 (a) Feb 4, (b) April 22, (c) May 8, and May 23 The dredged hole begins at 6 m depth and extends as deep as 9 m. Thermal and oxygen stratification is evident under ice (plot a), through the spring overturn period (b and c), and into the beginning of overall lake stratification (plot d). 134 600 Figure 4.13 Heat Content of Chain Lake 1993 and 1994 135 07-01 08-01 Diversion Hayes Creek (est) Direct Precip 09-01 10-01 100 75 50 25 0 Withdrawal Spillway (est) Evaporation 07-01 08-01 09-01 10-01 50 25 0 -25 -50 •4 <s\_ 07-01 Storage Rate (left scale) Lake Level (right scale) I--o-08-01 09-01 _J L_ _L 2.45 2 . 4 0 ^ c o 2.35 1 UJ 2.30 10-01 07-01 08-01 09-01 10-01 wavefile = woter_budget_doily.plt Figure 4.14 Water Budget for Chain Lake -1994 (a) surface inflows, (b) outflows, (c) storage rates and level changes, and (d) groundwater and unaccounted for flows 136 1993 1 , , , 1 - 0 1 0 6 - 0 1 0 7 - 0 1 i 0 8 - 0 1 Date 0 9 - 0 1 . 1 0 - 0 1 I 1 1 L 1 1 - 0 1 i ' 1 1 1 1 -C o n t o u r s are yiig/L + + ;' ' + : + 4 • + Z s 2 +51 1 1 1 1 — + + — — • 5,-- + + — — + ; + + + " \ + — "Dredged Kdle below 6.25 fr i —+ symbol indicates sample point i . . . i + i C h l a 2 3 % * (b) 0 1 2 3 1 " (c) Con tou r s are /J-g/L ' 'Dredged "h'o'le 'b'e'low"6.25' fri ' (opprox 0.6% of lake volume) I—+ symbol indicates sample point .2©o332-> Total P Figure 4.17 Chain Lake 1993: Secchi Depth, Chla, and Total Phosphorus (a) Secchi depth measurements, (b) contours of chl a concentrations, and (c) contours of total phosphorus concentration 139 Figure 4.18 Chain Lake 1994: Secchi Depth, Chla, and Total Phosphorus (a) Secchi depth measurements, (b) contours of chl a concentration, (c) contours of total phosphorus concentration 140 1994 Contours in /ig/L + 0 5 -0 1 2 3 0 6 - 0 1 0 7 - 0 1 • 0 8 - 0 1 0 9 - 0 1 1 0 - 0 1 1 1-01 Date Contours in jug/L 1 1 + 1 1 + + +1 1 + •. + • + 1 + + .— — + + — — + + + — + ', i 1 + \ i 1 + ' Dredged hole below 6.25 m + — S o l u b l e . R e a c t i v e P i . . . i . 4 h -o. 5 Q 6 7 ( C ) 9 05- -01 0 6 - 0 1 0 7 - 0 1 0 8 - 0 1 Date 0 9 - 0 1 1 0 - 0 1 1 1 - 0 1 Figure 4.19 Chain Lake 1994 Phosphorus Concentrations (a) Total Phosphorus, (b) Total Soluble Phosphorus, (c) Soluble Reactive Phosphorus 141 1994 (a)9l 0 5 -Tota l Iron i I 01 0 6 - 0 1 0 7 - 0 1 0 8 - 0 1 Date 0 9 - 0 1 1 0 - 0 1 1 1 - 0 1 0 1 Contours in m g / L 2 3 £ 5 Q 6 7 (b)9 £ . 2 0 8 Dredged hole below 6.25 m S o l u b l e Iron 0 5 - 0 1 0 6 - 0 1 0 7 - 0 1 0 8 - 0 1 Date 0 9 - 0 1 1 0 - 0 1 1 1 -01 Figure 4.20 Chain Lake 1994 Iron Concentrations (a) contours of total iron, (b) contours of soluble iron 142 ( a ) 9 1994 Total Manganese -i i i I 05-01 06-01 07-01 08-01 Date 09-01 10-01 -01 0 1 Contours in m g / L \ i + (b)9 05-01 06-01 07-01 08-01 Date 09-01 Soluble Manganese _L 10-01 1 1-01 Figure 4.21 Chain Lake 1994 Manganese Concentrations (a) contours of total manganese, (b) contours of soluble manganese 143 F e / M n Ratio ro c 3 00 co 3 CD CD CD CD CD O ^ f i 00 5? T l 5-H5. S D - C D CD -S=T. O ro o ^. P O . O 3 O O r P g.CD »' © I — CD CO CD co cr o a a* CD o* 2 CD P „ . T 1 & 3 S o ? < a 3 £ . 3 CD 3 CD 2 © &2 £ CD T l 2 ^ 2 p T l ' C 3 CD £ p CD N > c CQ to <D T3 O o (a) 20. 1 6 1 2 8 4 0 Aug-04 •— ' • —• •—— •— Temperature (deg C) o - -- DO (mg/L) _ _ - - o - - - - f - — — r o- - 1 Q —o- — ~~ ~~ ~~ I : 1_ Aug-11 Aug-18 Aug-25 Sep-01 400 * ( ^ g / L ) 300 200 100 0 - • Total Phos. - O Total Soluble Phos. "O- Soluble Reactive Phos. I 1 0 i_ i l l (b) Aug-04 Aug-11 Aug-18 Aug-25 Sep-01 (ug/L) • - NH3 o NOx ( c ) Aug-04 Aug-11 Aug-18 Aug-25 Sep-01 (mg/L) • - Total Fe o — Sol Fe ( d ) A u g - 0 4 Aug-11 Aug-18 Aug-25 Sep-01 0.6 0.4 0.2 i ( m 9 / L ) • ; — T L • Total Mn O Sol Mn ! . ^ o (e) A ug-04 Aug-11 Aug-18 Aug-25 Sep-01 Figure 4.23 Hypolimnetic Withdrawal Composition August 4 to Sept 1,1994 (a) temperature and dissolved oxygen, (b) phosphorus, (c) nitrogen, (d) iron, and (e) manganese During a turnover in the third week of August concentrations of 'anoxia-related' substances declined quickly -i.e. within a few days from Aug 15 to 19. 145 Figure 5.1 Row Test of Withdrawal System 95-05-20 More water leaves the box through the fountain than enters via the withdrawal pipe due to leaks around the control rod, and leaks in the seal between the box and the lake bottom. 146 125 100 (0 1c o (a) 75 50 25 Chla=f(TP) - (1993) • 93-07-20 • 93-08-13 • 93-08-26 X 93-09-26 — chla=1/2 TP ' X v _ 3 • 50 100 150 200 250 TP (ug/L) 300 350 100 75 & 50 o 25 chla=f(TP) - (1994) X ' yi • X 50 100 (b) 150 TP (ug/L) 200 • 94-04-22 • • 94-05-09 • 94-05-22 o 94-06-20 • 94-07-09 A 94-07-23 • 94-08-18 O 94-08-28 X 94-09-14 X 94-10-05 — chla=1/2 TP Figure 5.2 Chl a vs. Total Phosphorus (a) 1993, and (b) 1994 The 1:2 chl a to TP rato is an approximate biomass ratio for algae under phosphorus limiting conditions. Points on the line are likely phosphorus limited. When phosphjorus is in excess, then the points will fall to the right of the line. This data indicate that the system is likely limited by phosphorus part of the time. 147 8H o o ro o o kg of Phosphorus CO o o o o cn o o o o o o era c -CV) < CD a era P £ o CD H-> CD 00 ! <-•' a* CO e Ci CD «-»• o a* o a* o ^3 I-2- K) CD S3, p CO CD ><l o o >-t p CD pr. erg & o ' a CO CO C o co a* o O o a CD Ou a o & o a o P 9H a* - o e. 5 3 " P g 5 CD \ © P a. CO O CO - v l CO CO CD CO CO CO CO IV) CO CO cn CO CO CO CO CO CO CO 4^  350 300 "3 250 I 200 O o. 150 o ff 100 50 (a) 0 01-Apr 01-May 01-Jun 01-Jul Ol-Aug 01-Sep 01-Oct Total Phosphorus TP-15% TP+15% 10.00 5.00 0.00 01-1 -5.00 -10.00 +'ve is net gain to water column Apr ~01-May "O'RJun 01-Jul ' J 01-Aug h-Sep i OJ^Oct -Ve is net loss from water column (b) • net change in phosphorus content Maximum change Minimum change Figure 5.5 Apparent Net Phosphorus Concentration Change -1994 (a) lake content of total phosphorus, and (b) net change in phosphorus content normalized to sediment surface area. These fluxes are small compared to the individual release and sedimentation terms which may be on the order of 10 to 30 mg/m2/d. 150 Figure 5.6 Chain Lake Phosphorus Export - 1994 (a) water withdrawal rate, (b) phosphorus concentration, and (c) daily export 151 1 0 0 0 0 Figure 5.7 Chain Lake Phosphorus Renewal Time - 1994 The removal of water with high a concentration of phosphorus reduces the phosphorus renewal time (i.e. increases phosphorus export) without reducing the water renewal time (i.e. without requiring high flushing rates). 152 0 5 - 0 1 0 6 - 0 1 0 7 - 0 1 0 8 - 0 1 0 9 - 0 1 1 0 - 0 1 t ime (day) of year 0 5 - 0 1 0 6 - 0 1 0 7 - 0 1 0 8 - 0 1 0 9 - 0 1 1 0 - 0 1 t ime (day) of year wave file = /thermistors/stab_airtemp.batch Figure 5.8: Lake Stability Correlation to Princeton Air Temperature (a) 1993, (b) 1994 Above average air temperatures at Princeton (right scale) are co-incident with periods of lake stratification as measured by stability (left scale). This relationship is particularly striking during the May 1993 and July 1994 stratifications. 1 0 - 0 1 /weather/oirtemp_worm.plt.. Figure 5.9: Princeton Mean Air Temperature during summer - 1937 to 1994 The symbols indicate temperature average temperatures that are more than 1 °C above the 58 year mean average temperature. The plot symbols are sized according to their magnitude above mean. For estimating the frequency distribution, a 'summer' period was defined from June 1 to August 15, and all warm periods in this time period were counted. 154 Average number of warm periods per year=4.08 0 10 20 30 4 0 5 0 ( b ) Warm Per iod Durat ion (days) wavefile=/weother/airtemp_wormdist.plt Figure 5.10: Results of 'Worm Count' of Warm Periods (a) warm periods recorded in each summer, and (b) histogram of the length of individual warm periods with fitted log-normal distribution included. The warm period in 1994 was exceeded (in length) on^ once in 58 years. 155 9SI 10 T £ c o c CD O c o O c CD X O T3 CD > O CO w 8 + 6 + 4 + n 2 -1 1 1—\ 1 1 1 1—u 1 I I u I I I I | L 1-Jun 1-Jul 1-Aug 1-Sep 1-0ct Inside Fountain Outside Fountain Saturation D O . Notes: (1) Inside fountain measured at center of withdrawal fountain (2) Outside fountain measured in pool around fountain (within =3 m) (3) Measured D.O. values calibrated to sat'd air, values corrected for elevation. (4) Saturated D.O. values from YSI reference guide on D.O. meter Figure 6.2 Chain Lake Withdrawal Fountain Dissolved Oxygen Levels -1994 The fountain aerator added about 1.5 to 2.0 mg/L to the withdrawan water throughout the summer. 157 DO (mg/L) 1 0 (a) 1000 2000 3000 Approximate Distance Downstream (m) Figure 6.3 Hayes Creek Dissolved Oxygen Levels -1994 (a) prior to withdrawal, (b) during summer stratification, (c) after fall turnover 158 100 -r co X o O e E < 10 + 1 + 0.1 + 0.01 1 -Jun 1-Jul • — Measured Ammonia ~~ 30-day Average' Criteria o - • Maximum Criteria 1-Aug 1-Sep 1-Oct Notes: (1) Guideline values from Water Quality Branch, (Nagpal, 1994) (2) Guideline values calculated from water temperature and pH Figure 6.4 Withdrawal Ammonia Levels - 1994 159 1000 500 Fountain (ug/L) ( a ) Apr-01 May-01 Jun-01 Jul-02 Aug-01 Sep-01 Oct-01 1000 500 Beaver Dam (ug/L) (b) Apr-01 May-01 Jun-01 Jul-02 Aug-01 Sep-01 Oct-02 1000 500 Jellicoe Road (u-g/L) i — — . — . — — . — • - i (c) Apr-01 May-01 Jun-01 Jul-02 Aug-01 Sep-01 Oct-02 1000 500 Shinish Creek ® Diversion (ug/L) I . . - j . u i 1 (d) Apr-01 May-01 Jun-01 Jul-02 Aug-01 Sep-01 Oct-02 1000 500 Fish BarrieiJ (Mg/L) • • — / i (e) Apr-01 May-01 Jun-01 Jul-02 Aug-01 Sep-01 Oct-02 Shinish @ Diversion Figure 6.5 Hayes and Shinish Creek Ammonia Levels - 1994 (a) fountain, (b) Beaver Dam, (c) Jellicoe Road, (d) Shinish Creek @ diversion, and (e) Fish Barrier Ammonia released at the fountain during the stratification is consumed or lost in Hayes Creek. In the fall, elevated ammonia levels are apparent everywhere in the watershed - (even in Shinish Creek). 160 Figure 6.6 Hayes and Shinish Creek Iron Levels - 1994 (a) Fountain, (b) Beaver Dam, (c) Jellicoe Road, (d) Shinish Creek @ diversion, and (e) Fish Barrier Iron concentrations are attenuated in Hayes Creek. The concentrations are higher at the fish barrier due to the inflow from Shinish Creek. 161 0.6 0.4 0.2 (mg/L) Fountain (a) Apr-01 May-01 Jun-01 Jul-01 Aug-01 Sep-01 Oct-01 0.3 0.2 0.1 (mg/L) (b) A P r - 0 1 May-01 Jun-01 Jul-01 Aug-01 Sep-01 Oct-01 0.3 0.2 0.1 0.0 (mg/L) ^Jellicoe Roacj ( C ) Apr-01 May-01 Jun-01 Jul-01 Aug-01 Sep-01 Oct-01 0 3 . (mg/L) . Shinish @ Diversion 0.2 0.1 0.0 (jjj Apr-01 May-01 0.3 > 9 / L ) 0.2 0.1 0.0 Fish Barrier (e) A P r - ° 1 May-01 Jun-01 Jul-01 Aug-01 Sep-01 Oct-01 Chain Lake Beaver Dam Jellicoe Rd Shinish @ Diversion Figure 6.7 Hayes and Shinish Creek Manganese Levels - 1994 (a) Fountain, (b) Beaver Dam, (c) Jellicoe Road, (d) Shinish Creek @ diversion, and (e) Fish Barrier 162 10. APPENDICES Appendix A Quality Assurance / Quality Control Appendix B Summary of Water Quality Data Chain Lake Appendix C Calculation of Heat and Water Budgets Appendix D Summary of Water Quality Data Hayes Creek 163 APPENDIX A - Quality Assurance and Quality Control Table A - l . . . Method Detection Limit Runs for Fe and Mn Table A - 2 . . . Fe and Mn Replicate Analysis (of Same Sample) 1994 Table A - 3 . . . Fe and Mn Triplicate Analysis (of Same Sample) 1994 Table A - 4 . . . Fe and Mn Replicate Sampling and Analysis -1994 Table A - 5 . . . Fe and Mn Triplicate Sampling and Analysis - 1994 Table A - 6 . . . Fe and Mn Spikes and Standard Additions -1994 Table A - 7 . . . Comparison of H N O 3 vs. H2SO4 Sample Preservation for Fe and Mn-1994 Table A - 8 . . . Method Detection Limit Runs for Phosphorus Table A - 9 . . . Phosphorus Replicate Analysis (of Same Sample) 1994 Table A-10.. Phosphorus Triplicate Analysis (of Same Sample) 1994 Table A - l l . . Phosphorus Replicate Sampling and Analysis 1994 Table A-12.. Phosphorus Spikes and Standard Additions - 1994 Table A-13.. A N O V A Table for Phosphorus Replicate Test - Sampled Oct 5,1994 164 Appendix A - Quality Assurance and Quality Control For the discussion of quality control and quality assurance, the following terms are required: Sensitivity is defined as the amount of analyte required to cause an absorbance = 0.0044 (1% absorbance). It indicates the lowest level of analyte that could be expected to incur a positive machine response. In this context it is a function of the machine. Method detection limit (MDL) is the minimum concentration that can be measured and reported with 99% confidence that the analyte concentration is greater than zero (EPA, 1988). It is the level at which a positive response can be taken to be 99% true. M D L is evaluated based on the total procedure required to obtain a final concentration value, (i.e. digestion, filtering, nuetralization, etc) and incorporates all sources of error in the analyitical method. Values below the M D L are not treated as zero since they may in fact contain analyte. Replicates procedures included replicate analysis of the same sample (i.e. repeating the analysis on a second aliquot of a given sample), and replicate sampling and analysis (i.e. taking two separate samples, and analysing each and comparing the analysis. Spikes were included in the iron, manganese, and phosphorus analyses. Recoveries of spiked are considered acceptable if the calculated recovery is between 85% and 115% of the spiked amount, and excellent if between 95% and 105% of the spiked amount. (Standard Methods, 1992). When evaluating replicates, the same standards were employed as for spikes, that is, if the ratio of the two replicates was between 0.85 and 1.15 the replicate was considered acceptable, and if between 0.95 and 1.05, then it was considered excellent. 165 Appendix A - Quality Assurance and Quality Control Sample Locations are: DH Dredged hole at south end of lake M Mid lake station S South end of lake FG Fish barrier on Hayes Creek (~2 km downstream of Chain Lake) Jell. Jellicoe road station on Hayes Creek SH Shinish Creek at point of diversion B.Dam Beaver Dam station (~ 500 m downstream of Chain Lake) Foun. Withdrawal Fountain S.Way Spillway (i.e. existing surface exit from Chain Lake) Sample handling: (R) raw - i.e. unfiltered sample (F) filtered through 0.45 |im cellulose acetate filter Table A-1 Method Detection Limit Runs for Fe and Mn FE Mn Run 1 2 3 1 2 3 1 0.106 0.234 0.071 0.099 0.046 0.057 2 0.126 0.227 0.068 0.110 0.032 0.048 3 0.118 0.233 0.112 0.103 0.031 0.052 4 0.106 0.229 0.059 0.100 0.030 0.047 5 0.099 0.237 0.066 0.099 0.035 0.051 6 0.105 0.243 0.072 0.099 0.043 0.058 7 0.101. 0.238 0.069 0.095 0.042 0.049 Ave 0.109 0.234 0.074 0.101 0.037 0.052 Std Dev 0.0097 0.0055 0.0174 0.0047 0.0065 0.0043 MDL 0.030 0.017 0.055 0.015 0.020 0.014 Notes: all values in mg/L Run 1 = 0.100 ppm std Run 2 = Sample A-7 taken 94-10-05 Run 3 = 0.050 ppm std MDL = (Std Dev) * t ( n - i , i-oc=0.99) where: t is the t statistic for n-1 degrees of freedom at the 99% level eg: for 7 replicates, n-l=6 and t=3.143 source: EPA, 1988 166 Appendix A - Quality Assurance and Quality Control Table A-2 Fe and Mn Replicate Analysis (of Same Sample) 1994 FE Mn Sample Date #1 #2 Ratio #1 #2 Ratio DH 7m (F) 05-22 238 157 1.52 60 30 2.0 Fount (R) 07-22 1108 1128 1.01 335 340 1.01 M-lm 08-28 302 293 1.03 72 70 1.03 M-3m 08-28 264 263 1.00 77 73 1.05 M-5m 08-28 271 281 1.04 83 80 1.04 DH -5m 08-28 284 308 1.08 134 139 1.04 DH -0.15 m 08-28 277 245 1.13 73 69 1.06 Notes: all values in ug/L Ratio = (maximum of #1 and #2)/(minimum of #1 and #2) Table A-3 Fe and Mn Triplicate Analysis (of Same Sample) 1994 FE(ug/L) Mn (ug/L) Sample Date Ave Std Dev Ave Std Dev FG(R) 07-22 920 25. 44 1 DH - 7m 07-22 313 24 113 2.1 Table A-4 Fe and Mn Replicate Sampling and Analysis - 1994 FE Mn Sample Date #1 #2 Ratio #1 #2 Ratio FG(F) 05-08 284 243 1.17 11 13 1.18 DH - 3m (F) 05-09 92 101 1.10 48 10 4.80 DH - 7.5m(F) 05-09 208 190 1.09 87 80 1.09 DH - 5m (R) 07-09 149 159 1.07 59 65 1.10 DH - 7m (R) 07-09 197 189 1.04 82 82 1.00 Fount (F) 08-28 56 61 1.09 22 23 1.05 Notes: all values in ug/L Ratio = (maximum of #1 and #2)/(minimum of #1 and #2) Table A-5 Fe and Mn Triplicate Sampling and Analysis - 1994 Sample Date FE (ug/L) Ave Std Dev Mn (ug/L) #1 #2 SH(F) Fount (R) 08-26 08-28 172 18. 294 7.5 9 2 93 1.2 167 Appendix A - Quality Assurance and Quality Control Table A-6 Fe and Mn Spikes and Standard Additions - 1994 Sample Date Cone F E Spike Recovery (%) Cone Mn Spike Recovery (%) SH(R) 07-22 172 67 97 7 67 94 172 133 114 7 133 101 " 172 200 85 7 200 90 DH-0.15m(R) 09-14 125 100 106 45 100 104 DH-lm (R) 09-14 206 210 89 49 210 99 D H - 3m (R) 09-14 376 300 90 59 300 100 F G (R) 10-04 718 300 107 45 300 98 SH (R) 10-04 134 300 89 4 300 91 D H - 5m (R) 10-05 196 300 92 39 300 99 D H - 7m (R) 10-05 173 300 96 40 300 101 Notes: • sample values in u.g/L • SH (R) 07-22 spikes are a three point standard addition curve 168 Appendix A - Quality Assurance and Quality Control Table A-7 Comparison of H N O 3 vs. H2SO4 Preservation for Fe and Mn-1994 FE Mn Sample Date H2S04 HN03 Ratio H2S04 HN03 Ratio DH 5m (R) 02-02 564 558 1.01 226 224 1.01 DH 7m (R) 02-02 1413 1432 0.99 356 358 0.99 S 5m (R) 04-22 81 93 0.87 26 29 0.90 SH (F) 05-08 156 152 1.03 9 8 1.13 B. Dam (R) 05-08 207 197 1.05 44 30 1.47 B. Dam (R) 05-09 1642 1622 1.01 216 204 1.06 DH 7.5m (R) 05-09 239 208 1.15 104 87 1.20 FG(R) 07-22 893 1066 0.84 44 38 1.16 SH (R) 07-22 172 169 1.02 7 -1 -7 Jell (R) 07-22 784 909 0.86 158 151 1.05 B. Dam (R) 07-22 638 705 0.90 114 109 1.05 Spill (R) 07-22 155 96 1.61 61 28 2.18 Fount (R) 07-22 962 1108 0.87 351 335 1.05 Jell (R) 08-15 810 925 0.88 141 137 1.03 Jell (F) 08-15 343 418 0.82 129 128 1.01 B. Dam (R) 08-15 818 917 0.89 188 185 1.02 B. Dam (F) 08-15 350 384 0.91 172 176 0.98 Spill (R) 08-15 146 . 142 1.03 35 34 1.03 Spill (F) 08-15 68 61 1.11 11 6 1.83 Fount (R) 08-15 1331 1340 0.99 497 468 1.06 Fount (R) 08-15 321 347 0.93 494 449 1.10 DH - 5m (R) 08-18 369 384 0.96 120 123 0.98 DH - 5m (F) 08-18 146 128 1.14 75 70 1.07 DH - 7m (R) 08-18 877 900 0.97 352 355 0.99 DH - 7m (F) 08-18 346 362 0.96 364 365 1.00 FG (F) 10-04 560 541 1.04 43. 43 1.00 SH (F) 10-04 103 106 0.97 4 3 1.33 Jell (F) 10-04 381 363 1.05 114 113 1.01 B. Dam (F) 10-04 368 366 1.01 141 139 1.01 Spill (F) 10-04 139 152 0.91 25 23 1.09 M-lm(F) 10-05 111 131 0.85 20 23 0.87 M - 3 m (F) 10-05 111 116 0.96 20 20 1.00 M - 5 m (F) 10-05 121 110 1.10 22 18 1.22 DH - 5 m (F) 10-05 129 129 1.00 27 26 1.04 DH - 7 m (F) 10-05 134 147 0.91 26 27 0.96 M-0.15m(F) 10-05 107 116 0.92 19 20 0.95 Notes: H2S04 and HN03 values in (tg/L Ratio = (H2S04 preserved sample)/(HN03 preserved sample) 169 Appendix A - Quality Assurance and Quality Control Table A-8 Method Detection Limit Runs for Phosphorus Pho sphorus Run 1 2 3 4 5 6 7 8 Description 20 ug/L DIDW 12 ug/L 6 Ug/L 50 Ug/L A-29 60 ug/L A-29 Replicate # (sample) (sample) 1 30 2.9 14.4 3.7 45.1 38.5 39.9 30.6 2 28 9.2 11.7 18 55.8 27.8 61.3 19.2 3 19 -0.4 16.1 22.2 48.0 52.4 53.3 12.9 4 6.7 14.2 8.4 50.6 38.2 51.2 15.8 5 5.9 20 63.1 56.1 51.7 4.0 6 12.4 73.2 33.6 52.1 19.2 7 12.1 45.1 52.9 44.5 10.4 8 28.6 7.0 Ave 26 5 14 14 54.4 41.0 50.6 14.89 Std Dev 5.9 3.7 1.8 6.6 10.46 11.32 6.82 8.36 t (n-1. l-cc=0.99) 6.965 3.747 4.541 3.143 3.143 2.998 3.143 2.998 MDL (Jig/L) 41.1 13.9 8.2 20.7 32.9 33.9 21.4 25.1 Notes: all values in u.g/L MDL = (Std Dev) * t (n-i, l-oc=0.99) where: t is the t statistic for n-1 degrees of freedom at the 99% level eg: for 7 replicates, n-l=6 and t=3.143 for 8 replicates, n-l=7 and t=2.998 source: EPA, 1988 170 Appendix A - Quality Assurance and Quality Control Table A-9 Phosphorus Replicate Analysis (of Same Sample) 1994 Sample Date Phosphorus. (|ig/L) #1 #2 Ratio FG (R) 05-08 33 42 1.27 Jell (R) 05-08 37 50 1.35 Jell (F) 05-08 22 32 1.45 Jell (R) 06-18 46 62 1.35 BDam (F) 06-18 40 15 2.67 S Way (R) 06-18 27 5 5.40 S Way (F) 06-18 0 15 ... M-4m (F) 06-20 17 21 1.24 Fount (R) 07-07 39 49 1.26 Fount (R) 07-10 45 67 1.49 Fount (F) 07-22 81 68 1.19 DH - 7m (R) 07-23 127 135 1.06 Swamp (R) 08-15 55 41 1.34 Swamp (F) 08-15 115 192 1.67 Fount (R) 08-15 252 234 1.08 Fount (R) 08-16 191 229 1.20 Fount (R) 08-19 140 112 1.25 DH - 5m (R) 08-18 96 118 1.23 DH - 7m (R) 08-18 167 197 1.18 DH - 7 m (R) #2 08-18 238 232 1.03 Notes: Ratio = (maximum of #1 and #2)/(minimum of #1 and #2) Table A-10 Phosphorus Triplicate Analysis (of Same Sample) 1994 Phosphorus (ug/L) Sample Date Ave Std Dev DH - lm (R) 05-09 24.7 4.5 Stewart H20 05-22 30 10.4 S Way 05-22 38.7 10.4 DH - 7m (R) 07-09 61.7 4.6 Fount (R) 07-22 126 10.6 Fount (R) 08-08 315 10.2 Fount (R) 08-28 58.7 9.2 B.Dam (R) 09-12 81.0 7.9 DH 7m (R) #2 09-14 54.0 9.0 Jell (R) 10-04 63.7 9.1 M - 5m (R) 10-04 97.0 22.3 DH - 5m (R) 10-04 111.7 6.7 171 Appendix A - Quality Assurance and Quality Control Table A - l 1 Phosphorus Replicate Sampling and Analysis 1994 Phosphorus (|Xg/L) Sample Date #1 #2 Ratio B Dam 05-08 18 29 1.61 D H - 7m (R) #1 & #2 08-18 167 238 1.43 D H - 7m (R) #1 & #2 08-18 197 232 1.18 D H - 7m (R) #1 & #2 09-14 91 45 2.02 D H - 7m (R) #1 & #2 09-14 27 24 1.13 Notes: Ratio = (maximum of #1 and #2)/(minimum of #1 and #2) Table A-12 Phosphorus Spikes and Standard Additions - 1994 Sample Date Phosphorus (u,g/L) Cone Spike Recovery (%) D H - 7.5 m (R) 05-09 286 40 153 it 286 80 91 " 286 120 99 Jell (R) 09-12 50 40 93 n 50 80 114 IT 50 120 99 SH (R) 10-04 17 60 93 D H - 7m (R) 10-04 116 40 85 ii 116 80 104 it 116 120 83 Table A-13 A N O V A Table for Phosphorus Replicate Test - Sampled Oct 5, 1994 Source of Variability Sum of Squares Deg. of Freedom Mean Square F Statistic Significant (0.05 level) Significant (0.01 level) Main Effect A (grab) 3288.2 3 1096.1 13.41 Yes Yes B (split) 125.5 3 41.8 0.51 — — C (run) 866.7 3 288.9 3.53 Yes — Interactions AB 3471.6 9 385.7 4.72 Yes Yes A C 515.1 9 57.2 0.70 — — BC 724.9 9 80.5 0.99 ~ — Error 2207.0 27 81.7 Total 11198.9 63 Notes: Four samples were taken ('treatment' A - different grabs) Each sample split on site to four sub-samples ('treatment' B - different splits) Each split analysed four times ('treatment' C - different analysis runs) 172 APPENDIX B - Summary of Lake Sample Values Table B - l - Chla and Phosphorus Values - 1993 Table B-2 - Lake Chemistry Values - 1994 173 Appendix B - Summary of Lake Chemistry Table B-1: Chla and Phosphorus Data - 1993 Total Soluble | DATE LOCATION Depth chla Total P Soluble P Reactive P (ml (ug/L) (nan.) (uo/U (nan.) 93-07-03 Met Sin 0.15 32 93-07-03 Met Sin 2 55 93-07-03 Met Sin 4.5 75 93-07-03 North End 0.15 51 93-07-03 North End 2 44 93-07-03 North End 5 49 93-07-20 Met Sin 0.15 38 93-07-20 Met Sin 1 7.8 93-07-20 Met Sin 2 7.1 67 93-07-20 Met Sin 3 5.7 93-07-20 Met Sin 4 4.2 93-07-20 Met Sin 5 4.7 88 93-07-20 North End 0.15 48 93-07-20 North End 2 48 93-07-20 North End 5 42 93-07-20 North End 5 102 93-08-13 Met Sin 0.15 17 56 93-08-13 Met Stn 1 16 71 93-08-13 Met Sin 2 15 94 93-08-13 Met Sin 3 12 .82 93-08-13 Met Stn 4 11 80 93-08-13 Met Stn 5 8 231 93-08-20 North End 0.15 33 89 93-08-20 North End 1 49 100 93-08-20 North End 2 81 119 93-08-20 North End 3 30 71 9308-20 North End 4 9 54 93-08-20 North End 5 6 89 93-08-26 Dredged Hole 0.15 29 102 93-08-26 Dredged Hole 0.5 46 122 93-08-26 Dredged Hole 1 107 218 93-08-26 Dredged Hole 1.5 92 208 93-08-26 Dredged Hole 2 51 124 12 93-08-26 Dredged Hole 2.5 39 95 93-08-26 Dredged Hole 3 30 79 93-08-26 Dredged Hole 5 19 86 93-08-26 Dredged Hole 6 18 80 93-08-26 Dredged Hole . 7 12 85 . 31 93-06-26 Dredged Hole 8 8 332 166 93-09-26 Met Stn 0.15 9 56 36 12 93-09-26 Met Stn 1 10 43 38 11 93-09-26 Met Stn 2 13 68 36 7 93-09-26 Met Stn 3 17 57 40 6 93-09-26 Met Stn 4 28 75 46 12 93-09-26 Met Stn 5 45 227 49 11 93-11-05 Met Stn 0.15 45 93-11-05 Met Stn 1 66 93-11-05 Met Stn 2 38 93-11-05 Met Stn 3 31 93-11-05 Met Stn 4 26 93-11-05 Met Stn 5 30 174 Appendix B - Summary of Lake Chemistry Table B-2: Chem Deptt (m ical Dat i chla (Ufl/L) a - 1994 Total P Total Soluble Soluble P Reactive F (ug/L) (ug/L) i AmmoniE 1 (mg/L) 1 Total Fe Soluble Fe (mg/L) (mo/U Total Mn Soluble Mr (mg/L) (mg/L) 1 94-02-02 Dredged Hole 1 60 0.22 0.097 94-02-02 Dredged Hole 3 50 0.201 0.102 94-02-02 Dredged Hole 5 91 0.588 0.224 94-02-02 Dredged Hole 7 185 1.432 0.358 94-04-22 Met Sin 1 I 2.7 57 19 0.113 0.109 0.036 0.036 94-04-22 Met Sin 3 2.7 27 32 0.125 0.091 0.042 94-04-22 Met Sin 5 3.4 29 24 0.093 0.095 0.029 0.035 94-05-09 Dredged Hole 1 2.9 29 13 0.146 0.087 0.020 0.007 94-05-09 Dredged Hole 3 3.2 26 1 1 0.131 0.092 0.023 0.048 94-05-09 Dredged Hole 5 4.1 30 12 0.147 0.097 0.030 0.016 94-05-09 Dredged Hole 7.5 9.1 286 49 1.622 0.208 0.204 0.087 94-05-22 Dredged Hole 1 1.5 40 16 0.157 0.092 0.031 0.026 94-05-22 Dredged Hole 3 1.8 36 14 0.154 0.115 0.03 0.024 94-05-22 Dredged Hole 5 2 50 17 0.157 0.087 0.028 0.023 94-05-22 Dredged Hole 7 1.8 39 19 0.02 0.238 0.122 0.06 0.035 94-05-22 North End 1 1.5 37 22 0.143 0.087 0.027 0.02 94-05-22 North End 3 1.7 37 20 0.126 0.092 0.024 0.021 94-05-22 North End 5 36 19 0.158 0.122 0.027 0.022 94-06-20 Mid Lake 0.15 4.3 15 24 0.122 0.05 0.019 0.01 94-06-20 Mid Lake 1 4 25 47 <0.010 0.109 0.078 0.013 0.014 94-06-20 Mid Lake 4 7.6 17 3 <0.010 0.128 0.053 0.016 0.008 94-07-09 Dredged Hole 0.15 2.5 94-07-09 Dredged Hole . 1 3.4 94-07-09 Dredged Hole 3 5.8 94-07-09 Dredged Hole 5 22.7 73 56 3.7 0.054 0.154 0.069 0.062 0.069 94-07-09 Dredged Hole 7 67 79 5.1 0.087 0.193 0.117 0.082 0.117 94-07-09 Mid Lake 0.15 4.5 94-07-09 Mid Lake 1 5.7 25 7 3.2 <0.010 0.059 0.037 0.025 0.037 94-07-09 Mid Lake 4 18.3 46 28 3.8 <0.010 0.065 0.041 0.027 0.041 -94-07-09 North End 0.15 8.9 94-07-09 North End 1 8.4 32 28 2.2 0.023 0.056 0.035 0.022 0.035 94-07-09 North End 4 17.7 48 13 2.6 0.016 0.068 0.037 0.028 0.037 94-07-23 Dredged Hole 1 27 14 0.6 <0.010 0.072 0.065 0.047 0.031 94-07-23 Dredged Hole 4 50 17 1.3 <0.010 0.135 0.065 0.059 0.041 94-07-23 Dredged Hole 7 127 75 30.9 0.471 1.07 0.534 0.4 0.382 94-07-23 Mid Lake 0.15 1.4 94-07-23 Mid Lake 1 5.7 36 76 1.5 <0.010 0.071 0.052 0.041 0.031 94-07-23 Mid Lake 2 8.7 94-07-23 Mid Lake 3 11.3 94-07-23 Mid Lake 4 9.5 39 19 1.5 <0.010 0.128 0.069 0.056 0.039 94-07-23 Mid Lake 5 47 94-07-23 North End 1 35 9 2.4 <0.010 0.086 0.052 0.045 0.035 94-07-23 North End 4 48 44 1.1 <0.010 0.168 0.037 0.078 0.056 94-08-18 Dredged Hole 3 17 87 35 0.3 <0.010 0.214 0.075 0.064 0.01 94-08-18 Dredged Hole 5 1 8 96 54 8.8 0.138 0.369 0.146 0.12 0.075 94-08-18 Dredged Hole 7 7 167 111 39.2 0.444 0.877 0.346 0.352 0.364 Dredged Hole 7 7 ' 238 0.61 1.318 0.434 94-08-18 Mid Lake 1 19 119 48 1.3 0.026 0.244 0.077 0.084 0.012 175 Appendix B - Summary of Lake Chemistry Table B-2: Chemical Data - 1994 Depth (ml chla (ug/L) Total P (ug/L) Total Soluble Soluble P Reactive P (ug/L) ftia/L) Ammonia 1 (mg/L) Total Fe Soluble Fe (mg/L) (mg/L) Total Mn Soluble Mn (mg/L) (mg/L) 94-08-18 Mid Lake 3 34 78 1 9 0.6 0.04 0.227 0.066 0.063 0.007 94-08-18 Mid Lake 5 19 108 76 11.9 0.245 0.457 0.18 0.197 0.147 94-08-18 North End 1 47 117 64 1.9 0.047 0.306 0.086 0.098 0.015 94-08-18 North End 3 43 121 50 1.3 0.025 0.312 0.084 0.084 0.009 94-08-18 North End 5 23 90 66 3.9 0.074 0.313 0.114 0.105 0.029 94-08-28 Dredged Hole 0 .15 1 6 44 27 13.8 0.07 0.277 0.057 0.073 0.004 94-08-28 Dredged Hole 5 9 74 63 41.1 0.192 0.284 0.083 0.134 0.05 94-08-28 Dredged Hole 7 : 6 68 76 27.6 0.269 0.33 0.155 0.115 0.044 94-08-28 Mid Lake 1 21 63 41 15.1 0.063 0.302 0.098 0.072 0.01 94-08-28 Mk) Lake 3 25 78 31 23.2 0.023 0.264 0.078 0.077 0.007 94-08-28 Mid Lake 5 1 0 82 52 28.3 0.107 0.271 0.089 0.083 0.01 94-09-14 Dredged Hole 0 15 76 97 10 16 0.829 0.125 0.061 0.045 0.01 94-09-14 Dredged Hole 1 48 94 21 6.6 0.506 0.206 0.044 0.049 0.009 94-09-14 Dredged Hole 3 30 105 16 7.2 0.987 0.376 0.596 0.059 0.01 94-09-14 Dredged Hole 5 38 69 25 10.6 0.649 0.121 0.041 0.042 0.007 94-09-14 Dredged Hole 7 33 91 27 8.7 1.187 0.13 0.063 0.05 0.014 94-09-14 Dredged Hole 7 36 45 24 6.4 0.731 0.126 0.051 0.047 0.01 94-10-05 Dredged Hole 5 55 104 40 13.5 0.915 0.196 0.129 0.039 0.027 94-10-05 Dredged Hole 7 53 116 28 13.8 1.183 0.173 0.134 0.04 0.026 94-10-05 Mid Lake 0. 15 37 112 36 15.2 0.745 0.156 0.107 0.032 0.019 94-10-05 Mid Lake 1 35 76 35 14.6 0.446 0.167 0.111 0.03 0.02 94-10-05 Mid Lake 3 42 83 33 14.1 0.634 0.213 0.111 0.032 0.02 94-10-05 Mid Lake 5 37 79 43 15 0.609 0.179 0.121 0.034 0.022 176 APPENDIX C - Calculation of Heat and Water Budgets C . l Heat Budget Components The predominant heat fluxes to and from to Chain Lake are: Solar Radiation Solar radiation is the incoming direct (beam) and diffuse (scattered) radiation from the sun. It is sometimes referred to as 'short-wave' radiation though it is composed of all wavelengths and includes a significant infra-red component. The E F M Met station includes two pyranometers, one pointing upward, and the other downward, to record incoming and reflected solar radiation. The clear sky theoretical incoming and reflected radiation can be found by the methods of Henderson-Sellers (1984) which are similar to Fischer et al (1979). The EFM-Met station pyranometer appears to record more than the theoretical radiation, yet it saturates at about 745 W/m2, and so the areas under the recorded and theoretical curves (representing total radiation) are similar. It is believed that the location of the downward looking pyranometer results in erroneous recording of reflected radiation due to reflection from the met station raft., For this analysis, the recorded incoming radiation, and the theoretical reflected were used. The net solar radiation is then the difference between the two. Longwave Radiation Long wave radiation is 'black-body' radiation. All matter will radiate long wave (i.e. infra-red) radiation as: Radiation = eoT4 Q 1 where: e = emissivity (0 < e > 1) a = Stefan - Boltzman constant (5.67 e~8 Wm~1K'A) T = temperature of emitting surface (K) 111 Appendix C - Calculation of Heat and Water Budgets __ Several modifications to this equations exist. For example some sources use 'gray-body' type formula to predict long wave radiation (e.g. T V A in CE-Qual, 1986, Fischer at al, 1979). Several of the formula examined gave equivalent results, and with no data to verify there is no benefit in adopting a more sophisticated formula. Emissivities of the water surface are =0.97 (H-S, 1984), Atmospheric long wave emission is primarily from cloud cover, and water vapour. Emmisivity is a function then of cloud cover and vapour pressure (representing moisture content). Typical clear sky values at Chain Lake are =0.80. (H-S, 1984). The temperature of emission is the water temperature in the case of long wave emitted from the water surface, and the air temperature in the case of atmospheric radiation emitted to the water surface. The use of air temperature is not strictly accurate and several formulae exist for converting air temperature to an effective sky temperature, Tsky- Duthie and Beckman (1980) recommend the formula of Swinbank: The long wave emissions to and from the lake are on the order of 300-500 W/m z , with a net long wave emission from the lake typically on the order of 100-200 W/m^. Sensible Heat Sensible heat (H s ens) is the heat lost by conduction to the air, and aided by the convection of the wind at the surface. It is a function of the wind velocity, and the temperature difference between the water and the air. Several formula exist. This application uses the method of Fischer et al (1979): 7\ =0.05527-; .1.5 C.2 a Hsens = CspaCpu(Tw - Ta ) C.3 178 Appendix C - Calculation of Heat and Water Budgets where: Cs - co - efficient = 1.45e~3 pa = air density (« 1.25 kg m - 3) Cp = specific heat of air (1012 Jkg~lK~l) u = wind speed (ms~l) Tw = water temperature (degC) Ta = air temperature (degC) Latent Heat Latent heat is the heat lost from at the surface to evaporate water. It is the heat required to convert liquid water to water vapour. Thus it is the product of the mass of water lost to evaporation and the latent heat of vaporization. Hlaten, = Qp^uca - Q) c 4 where: C, = co - efficient = 1.45e-3 pa = air density (~ 1.25 kg m~3) Lw = latent heat of vapourization (= 2400 kJkg~lK~l) u - wind speed (ms~l) Qo = saturation specific humidity (kgHi0 I kgmois,air) Q= actual specific humidity (kgHj0 I kgmoistair) 179 Appendix C - Calculation of Heat and Water Budgets C.2 Water Budget FLows Surface Water Flows A V-notch weir was used in to measure the diversion inflow. Withdrawal Row Rates The withdrawal flow rates were measured using a calibrated OTT velocity probe, taking measurements inside the withdrawal fountain. Evaporation Flows Evaporation is estimated from the latent heat flux of section C. 1.1. The evaporation rate E, (in m/s of lake level) is given as: ZJ IP _ " latent „ _ where: Hiatent is in W/m2> and P w is the density of F^O. Lake Storage A lake level measurement was taken at each site vistit. The lake level was linearly interpolated between measurements. The rate of change of lake level times the lake area gives the rate of change of storage. Groundwater and Unaccounted Water Any water that was not accounted for by the other fluxes was attributed to 'ground water / unaccounted for'. 180 APPENDIX D - Stream Monitoring Program -1994 Table D-•1 - Fountain Station Table D-•2 - Spillway Station Table D-•3 - Beaver Dam Station Table D-•4 - Jellicoe Road Station Table D-•5 - Hayes Creek above Hayes-Shinish Junction Table D-•6 - Hayes Creek at Fish Barrier Table D-•7 - Shinish Creek at Diversion Table D-•8 - Shinish Creek above Hayes-Shinish Junction 181 Appendix D - Summary of Stream Sampling Program co E 2 E CM CM CO CM T - CM O O O L 0 C M C 0 r - - O ) T - c r j ^ T a ) C M C O C O - ' ^ - ' ^ - C M O O O o o o o o CM i - i - T -co co cn if) if) CO O O i -C D i n S T - S t D S l f l N C J * C O C O C O O O O C O O - : C O - T - < O C M T - T -O O ^ T - o o o o r - o c o o i r - . ^ t - r - . i - c o ^ - c o i - o t o r - i o i f l o n s m w n C M ^ L O r ^ L O t O L 0 C 0 - i - C 0 C 0 c D C O t n h - - C M r * - i n c o c y 5 r * - r * - c o i - CM CO t - ^ j - i f ) h - . i f ) L O i r > c o i - c o c o •P: w m ^ ^ CM » O d o Q eg co co eg in d O V <o CO CO CO CO CO in in io co cd *" """ e g c o o r ^ o c x j c o c o m m c D o c o c g ^ r W O C M O T - W W O i - t - t - T - N r O a . to 3 C 3 3 3 3 C7>  U l O l O O) CL = : = : : 3 • S < < < < < < I O O a. (0 0) CD (A (0 o o o q d d o d o o o o i - o CO CO CO ^ o o T - T - O o d d co .r< co o> to oo If) co O O O i -d d d d CD CO CD f-LO CO CO i-" If) T- CM d o d o O Z X z LO CO f.. TT O) CM CO CO o to N CJ t-i - ° If) CD CO CM CM CD CJ) CO . CO O) N N (M CM CO CM O X 5 CD a. LL rr p — i o o o o CD CO CO CM 9 d T - CM O • ci E 2 £ o S o c co o CM CO CO O) N Q) OJ CM •»— LO W CM ^ -Q. CO < 5 CM CO LO CO CM t CM O CM •<- O 3 3 = ,2 r< < < < co o 182 Appendix D - Summary of Stream Sampling Program O a> i— O (A » , >J to 3 E O O -< o o o o o o o o o o o o o o o o o CO CM h- CD CO h- CO CO CM CN CO o i s s ' t i o c O ' - m s CO O C O t O O i - N O l C D i - W W W N C O d D C O C O o o o o o o o o o O) O N CD CO CM (D O O) CO N T - CO CO CM CM Tj- CM CD 05 CO if) N Q N O) CM CO CM CO IO O) O i - CO t- CO CO 1 - CM N CO r- rn T - r - O L O C D C O C D t o d o d E ° CD T - CM eo -^ *t io CN T - LO oo co to *^ LO c\i T - s in cq N CO S CD LO LO LO LO CO E t-O I C O C O N W I O C O C M C M O T - O W T - W ' -~ > ^ C Z " 5 " 5 D) Ol O. < 5 " » < < c o (0 X 3 £ 3 f o z Z o> P=5 o Q o 1- "fr CO ^j-o o T - O) (O T -O CM CM O) i— ^ T T ° . T O O O O O o o o o ol • f - C O 0 0 „ C O C O C M C O i -C M l O T f ^ C D ' t O C O C O i - c o i M r ; c M c o c M n n o o o o ol m i n i - . , c n o n m c M O N C O ^ O t - C O C O C O n c M n ^ c n c o c o m i o d d d - ^ d o d o o o c o r ' t n c f i s n c o C O C M i - C D N l f ) r - O C O i - CM T - CO 00 o c o i - o j n c o c o r -c o c M ' - c \ i ' - ' - T r i r ) 0 | ••- CM N CO CM CM CM co n CO CM CO N N Ti- o) in cn| CO CO CO CD CM CO if) CO O If) CO T - N if) If) T}- N O) S CO O) r-CO CO CO CM I CO G) O CO co to | CO 1^  CO Tf LO ^ t l CMCOCONCMincOCM C M O r - O C \ I i - C M ^ !=; ^ c "33 cn cn Q. *x\ 183 Appendix D - Summary of Stream Sampling Program co E 2 =d — cn co E S. J 2 ? CO ^ fe o o Q E 3 2 ~ o ~ o CM i - CD CO O) CO oi d o> in •<- Tt t-i N h-E P CM CO CO N O T -< 5 "3 0 1 0 1 8- o 3 3 CO U < < CO O ct I 6 o D o j - r - c s i c o T r r ^ c o - * i - t - C \ | T - r o i f ) ( D r O o o o o o o o o o o o o o o o o o i - CM O) W CO O) CO T— T— CM CO LO o o o o o o o o o o o o o o o c o ^ c f i w O ) i - c o n o CMCOLOCOCOO)COCT>CO COCMCOCOr-COr-CDLO o o o o o o o o o r - O T f r f C D L O O C O C O W W l O O t D i - C D T T i -o o o o (M T - LO CO TT ^ CM r- i - o> a> o> i - CO CO i - O CO LO T— CO LO CO CO CO LO CO TJ- CM 1^ . TJ C J n 6 N C M C O O ^ L O C O C M O L O C M i - C M 3 t - C M C M C M ^ h- N CM LO CO CD O) O co eg s 6 O -r-' LO CM CO K h*' r-' LO O CM CO E < 5 -5 O O) Q. 3 3 0> ir < < CO O 184 Appendix D - Summary of Stream Sampling Program 5 =d o £> co E 3 ? o o o ^ Q o o o o o o o o o o 6 d d o d o t S CO o o o o o o o CO Tf Tf o o o o o o o o o o o o o o o o C O C O C O C O O I O C O N T f i n c o o c o c o r - T t CSl-r--r-OOT-T-T-y-o o o o o o o o o o o o o o o o o o co o o CM r>- o O O - r - O O t - i - r - L O C O L O T - O J C M C M O J T - ^ -T - IT) CO L[) i -1^  IT,I T f L O Tt I O o p - J O C T _ _ E 2 2 c co o h - T - T - C O 00 O O o CM CM CM oi -r^ 1 ~ C O C O 0) O N C O C O S C M O L O C O *~ oi C O C O r-» C M L O 00 C M O i - O C M i - C M i -a. co < 5 •5 3 6> o» £ g < < co O co o o cn cn O ) T -L O O ) o 2 o £5 1= 5 LO 00 i-LO OJ T -O —J Q E CO LO oi oi o d E „ oo oi a to cd O) CJ) Q . r -3 • • " < < 03 O (/> 185 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items