Fate and transport of nutrients from the Begbie Lake wetland system: measuring the impacts of inundation on hydrologically connected aquatic ecosystems.

(1)

by Jesse Sinclair

B.Sc., University of Saskatchewan, 2004

A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of

MASTER OF SCIENCE in the Department of Biology

c

Jesse Sinclair, 2009 University of Victoria

(2)

Fate and Transport of Nutrients from the Begbie Lake Wetland System: Measuring the Impacts of Inundation on Hydrologically Connected Aquatic Ecosystems

by Jesse Sinclair

B.Sc., University of Saskatchewan, 2004

Supervisory Committee

Dr. Asit Mazumder, Supervisor (Department of Biology)

Dr. Rick Nordin, Departmental Member (Department of Biology)

Dr. Max Bothwell, Departmental Member (Department of Biology)

(3)

Supervisory Committee Dr. Asit Mazumder, Supervisor (Department of Biology)

Dr. Rick Nordin, Departmental Member (Department of Biology)

Dr. Max Bothwell, Departmental Member (Department of Biology)

ABSTRACT

The ecological implications of altering the hydrologic regime, or hydroperiod, of a wetland system include major changes to both ecosystem structure and function. Wetland systems are generally sinks of nutrients such as phosphorus, nitrogen and carbon, which are important to water quality. Changes in the hydroperiod such as the inundation or drainage of a wetland system alter wetland function and may switch the system from functioning as a nutrient sink to a source.

In this study, I explored the effects from the introduction of a seasonal inundation regime to the Begbie Lake wetland system. Begbie Lake is hydrologically connected to the Sooke Lake reservoir, the main drinking water supply for Greater Victoria, British Columbia. In 2002, the dam on the Sooke Lake reservoir was raised by 6 m, which led to the seasonal inundation of the Begbie Lake wetland system in 2005 and 2006. In 2005, the Begbie Lake wetland system was inundated for 57 days, from April through May. In 2006, the wetland was inundated for 123 days, from January through May. The seasonal inundation resulted in the release of phosphorus, nitrogen

(4)

and carbon from the wetland system into Begbie Lake and the Sooke Lake reservoir. Nutrients such as phosphorus and nitrogen are especially important in drinking water reservoirs, where increases in concentrations often lead to increased algal biomass and possibly a shift toward cyanobacteria-dominated phytoplankton communities. Organic carbon in source water is correlated with disinfection by-product production during treatment, increases in bacterial biomass within distribution systems, and decreases in the efficacy of water treatment.

The response to inundation of the wetland system, in terms of the release of key water quality nutrients into Begbie Lake was assessed. Water chemistry data collected from 2005 and 2006 were compared to pre-inundation baseline data (2003 and 2004) over the growing season. Phosphorus, nitrogen and organic carbon concentrations above the flooded wetland soils increased significantly during the inundation periods. Following drawdown, the total phosphorus concentration in Begbie Lake increased significantly. Total nitrogen and total organic carbon concentrations in Begbie Lake did not increase over this same period; the microbial mineralization and reduction of nitrogen and organic carbon, as well as the export of these nutrients into the Sooke Lake reservoir, are proposed as influencing the observations.

The inundation of the Begbie Lake wetland system also resulted in the export of large amounts of nitrogen and organic carbon to the Sooke Lake reservoir. The increase in the duration and extent of inundation in 2006 resulted in higher contribu-tions from the Begbie Lake wetland system. The export of nitrogen increased from 3.8 kg in 2005 to 4.4 kg in 2006. Carbon export increased from 77.7 kg in 2005 to 171.1 kg in 2006. The export of high amounts of phosphorus were not observed during the study. While phosphorus concentrations increased above wetland soils, the data suggest that much of the phosphorus remained cycling within Begbie Lake.

(5)

List of Tables

Table 2.1 Correlation of environmental variables and nutrient concentra-tions above the wetland sites (BBL-03 and BBL-04) in 2006. . 48 Table 2.2 Correlation of nutrient concentrations above the wetland sites

(BBL-03 and BBL-04) in 2006. . . 49 Table 2.3 Total export of phosphorus, nitrogen and organic carbon from

the Begbie Lake wetland system during both inundation years. 50 Table 2.4 Predicted change in total phosphorus, total nitrogen and

to-tal organic carbon concentrations in Sooke Lake reservoir at drawdown. . . 51 Table 2.5 Epilimnetic nutrient concentrations for the Sooke Lake reservoir

(SOL-04) during the inundation of the Begbie Lake wetland system. . . 52

Table A.1 Water chemistry data for the Begbie Lake site (BBL-01) for the pre-inundation baseline years (2003 and 2004). . . 92 Table A.2 Water chemistry data of the epilimnion at the Begbie Lake site

(BBL-01) in 2005. . . 93 Table A.3 Water chemistry data of the hypolimnion at the Begbie Lake

site (BBL-01) in 2005. . . 94 Table A.4 Water chemistry data of the epilimnion at the Begbie Lake site

(9)

Table A.5 Water chemistry data of the hypolimnion at the Begbie Lake site (BBL-01) in 2006. . . 96 Table A.6 Water chemistry data at the Begbie Creek site (BBL-02) for

2005. . . 97 Table A.7 Water chemistry data at the Begbie Creek site (BBL-02) for

2006. . . 98 Table A.8 Water chemistry data for the open wetland site (BBL-03) for

the inundation years (2005 and 2006). . . 99 Table A.9 Water chemistry data for the forested wetland site (BBL-04)

(10)

List of Figures

Figure 2.1 Begbie Lake wetland system; the sampling sites during the study period are shown. The wetland sites (03 and BBL-04) were sampled during the inundation period. . . 53 Figure 2.2 The Begbie Lake wetland system. a) The open (BBL-03) and

b) forested (BBL-04) sites are shown. . . 54 (a) BBL-03 . . . 54 (b) BBL-04 . . . 54 Figure 2.3 Begbie Creek as the water levels receded in Sooke Lake reservoir

in 2005. . . 55 (a) Begbie Creek looking north towards Begbie Lake . . . 55 (b) Begbie Creek looking south towards the Sooke Lake reservoir . . 55 Figure 2.4 Precipitation at the Sooke Lake reservoir spillway over the

in-undation years. Data obtained by CRD. . . 56 Figure 2.5 The Sooke Lake reservoir. The water level of the reservoir is

ar-tificially maintained by the dam at the south end of the reservoir. 57 Figure 2.6 Water levels at the Sooke Lake reservoir taken from the staff

gauge at the spillway. The inundation period for the 2005 and 2006 treatment years lasted for 57 and 123 days, respectively. Data obtained from CRD Water Services. . . 58

(11)

Figure 2.7 The Begbie Lake wetland system. Samples were collected at BBL-01 and BBL-02 over the entire study period. Additional samples were collected at the wetland sites (03 and BBL-04) during inundation periods. . . 59 (a) Begbie Lake (pre-inundation years) . . . 59 (b) Begbie Lake during inundation periods . . . 59 Figure 2.8 Hypsometric curve for the Begbie Lake wetland system. The

curve has been transposed to show estimated relief. . . 60 Figure 2.9 Daily water volumes over the Begbie Lake wetland system in

2005. . . 60 Figure 2.10 Daily water volumes over the Begbie Lake wetland system in

2006. . . 61 Figure 2.11 Inundation and drainage of the Begbie Lake wetland system in

2005 and 2006. . . 61 Figure 2.12 Temperature isopleths for Begbie Lake (BBL-01) during the

experimental years: a) 2004, b) 2005 and c) 2006. . . 62 Figure 2.13 Dissolved oxygen isopleths for Begbie Lake (BBL-01) during

the experimental years: a) 2004, b) 2005 and c) 2006. . . 63 Figure 2.14 Concentration of total phosphorus in: a) 2005 and d) 2006;

total nitrogen concentration in: b) 2005 and e) 2006; and total organic carbon concentration in: c) 2005 and f) 2006 above the open (BBL-03) and forested (BBL-04) wetland sites in each inundation year. . . 64

(12)

Figure 2.15 Total phosphorus release in: a) 2005 and d) 2006; total nitro-gen release in: b) 2005 and e) 2006; and total organic carbon release in: c) 2005 and f) 2006 above the open (BBL-03) and forested (BBL-04) wetland sites in each inundation year. Con-centrations were normalized per water column depth at each site. . . 65 Figure 2.16 Total phosphorus concentrations at the centre buoy on Begbie

Lake (BBL-01) during all experimental years. Inundation of the system occurred in 2005 and 2006. . . 66 Figure 2.17 Impact of temperature on total phosphorus concentrations in

Begbie Lake during: a) all experimental years (p = 0.051), and b) inundation years; total phosphorus concentration was not significantly different between inundation years (p = 0.076). c) Water temperature explained 28% of the variation in TP concentration during the inundation years (p < 0.001). . . 67 Figure 2.18 Total nitrogen concentrations at the centre buoy on Begbie Lake

during all experimental years. Inundation of the system oc-curred in 2005 and 2006. . . 68 Figure 2.19 a) Impact of temperature on total nitrogen concentration. The

interaction effect between pre-inundation and inundation years was insignificant (p = 0.61). b) Water temperature had a sig-nificant effect (p < 0.001) on total nitrogen concentrations dur-ing all experimental years, although explained only 21% of the variability. . . 69

(13)

Figure 2.20 Total organic carbon concentrations at the centre buoy on Beg-bie Lake during all experimental years. Inundation of the sys-tem occurred in 2005 and 2006. . . 70 Figure 2.21 Concentration of total phosphorus at the Begbie Creek site

dur-ing the inundation years; inundation periods are highlighted. . 71 Figure 2.22 Concentration of total nitrogen at the Begbie Creek site during

the inundation years; inundation periods are highlighted. . . . 71 Figure 2.23 Concentration of total organic carbon at the Begbie Creek site

(14)

ACKNOWLEDGEMENTS

Funding for this study has been provided through research grants from the Na-tional Science and Engineering Research Council of Canada and the Capital Regional District (CRD) Water Services to Asit Mazumder.

I wish to foremost thank Asit Mazumder for insightful advice and the many oppor-tunities and experience gained in the Water and Aquatic Sciences Research Program. I also wish to acknowledge Rick Nordin and Max Bothwell for their contributions of comments and insight into this work. In addition, I also wish to thank Stewart Irwin, Jack Hull, Laura Kline, Bernie Morris, Maria Roxborough and Mel Goldstein from the CRD for feedback and support throughout this work.

I also would like to thank my colleagues and friends Kevin Rieberger, Leon Gaber, Anita Narwani, Deb Epps and Carly Chunick for their ideas, their ear, their con-structive criticism, their encouragement and their support. There are many people in the Water and Aquatic Sciences Research Program past and present that I would like to thank for their lab and field assistance including Shapna Mazumder, Heather McNally, Kelly Hyland, Jennifer Linton, Kiyuri Naicker, Kelly Field and especially Crystal Lawrence, whose sense of humour and taste in music helped me through many days in the field and many hours in the lab.

Finally I would like to thank Nicole Obee, whose love and encouragement pushed me to finish this goal, my appreciation to you for your time and energy cannot be expressed enough.

(15)

DEDICATION

To my parents, Don Sinclair and Lynn Sinclair, who have always been there to help clear the path for every endeavour I’ve undertaken.

(16)

(17)

1.1 General Introduction

The diversion and impoundment of freshwater for storage and subsequent use have implications on water quality by significantly altering the physical, chemical and biological composition of the lake or reservoir (Baxter 1977). Impoundments and reservoirs are generally constructed for hydroelectric power production, irrigation, recreation, navigation and drinking water. The purpose of each governs the spatial and temporal patterns in the rise and fall of water levels in the impoundments or reservoirs, but the recharge and drawdown of reservoirs are also subject to natural climatic variability.

A large body of research has been completed on the creation or expansion of impoundments in wetland ecosystems. Whole-ecosystem experiments have been per-formed by altering the hydrologic signatures of the system and exploring impacts on food web structure (Paterson et al. 1997), nutrient transport (St. Louis et al. 2004), methyl-mercury production (St. Louis et al. 1996), emissions of carbon dioxide (CO2)

and methane (CH4) (Kelly et al. 1997, Clair et al. 2002) and biogeochemical cycling

(Reddy and Patrick 1975, Bridgham et al. 1998, Moore et al. 2003, Corstanje and Reddy 2004). In Canada, a large portion of work has been completed in the boreal region of northern Ontario on ombrotrophic peat bogs. However, relatively little work has been completed in the complex wetland systems of British Columbia (Price et al. 2005). Further, studies on the impacts of human-induced disturbance on water quality in British Columbian wetland systems are few (Price et al. 2005).

Wetland ecosystems are large sinks of nutrients important to water quality such as phosphorus, nitrogen and carbon that are stored in the soils and sediments of the wetland system (Hemond 1983, Johnston et al. 1984, Craft 2000, Fisher and Acreman 2004) and both aquatic and terrestrial vegetation (Hemond 1983, Hogan et al. 2004, Asada et al. 2005). In addition to acting as sinks for nutrients, wetland systems are

(18)

also functionally important in regulating the transformation of nitrogen and carbon (Reddy and Patrick 1975, Mitsch and Gosselink, 1993, Bridgham et al.1998, Moore et al. 2003). Due to their function in nutrient retention and transformation, wetland systems are often used in nutrient management and removal (Mitsch et al. 2000, Fisher and Acreman 2004, Verhoeven et al. 2006, Kieckbusch and Schrautzer 2007). However, changes in the hydroperiod of a wetland system can alter wetland func-tioning, even turning the system from a nutrient sink to a nutrient source. Gaps in our current understanding of the effects of timing, duration and extent of inundation need to be explored to elucidate the impacts of inundation on water quality.

1.1.1 Wetland Hydroperiod

The hydrologic regime, or hydroperiod, is defined by the change in the water level observed in a wetland system over the year. While the hydroperiod exhibits annual variability, the pattern observed is a major characterization of the wetland system. The hydroperiod is the main factor in the development and maintenance of the wet-land ecosystem structure and function (Mitsch and Gosselink 1993, Richter et al. 1996, Silvertown et al. 1990, Casanova and Brock 2000, Adamus et al. 2001, Price et al. 2005). Alteration of the wetland hydroperiod, whether by anthropogenic or natural disturbance, alter many components of wetland community structure and function including species composition and richness (Bunn and Arthington 2002), nutrient cycling and availability (Olila et al. 1997, Hill et al. 2000, Corstanje and Reddy 2004), soil oxygen availability (Reddy and Patrick 1975, Mitsch and Gosselink 1993), primary productivity (Robertson et al. 2001) and the accumulation or export of organic matter (Olila et al. 1997, St. Louis et al. 2004). Induced changes in wetland function precede changes to the abiotic and biotic structural components, in-cluding both species composition and species richness, which in turn further influence

(19)

wetland functioning (Mitsch and Gosselink 1993). Wetland systems are transition ecosystems, spanning the boundaries of aquatic and terrestrial ecosystems. Transi-tion zones such as wetlands serve as sensitive ecological niches for many species and thus are susceptible to disturbance from slight environmental change (Mitsch and Gosselink 1993).

1.1.2 Mechanisms of Nutrient Release

The release of nutrients from wetland systems following inundation is controlled by two mechanisms: leaching from wetland soils and the decomposition of organic matter such as vegetation and detritus. Soil type and soil-nutrient concentration (Bryant 2007, Mazumder 2003) and vegetation type (Corstanje and Reddy 2004, de Neiff et al. 2006) all influence nutrient release. Water temperature (D’Angelo and Reddy 1999, St. Louis et al. 2003), inundation frequency (Battle and Golladay 2001, de Neiff et al. 2006), timing (Langhans and Tockner 2005), duration (Langhans and Tockner 2005) and areal extent (Moore et al. 2003) have all been recognized to impact these mechanisms and affect the release, mineralization and transport of nutrients from wetland soils to surface waters.

1.1.3 Impacts on Water Quality

Source water reservoirs, created by the impoundment of existing water bodies, are par-ticularly sensitive to environmental change; nutrient loading, contaminant pollution and increases in turbidity all impact water quality and may lead to negative impacts on human health, increased management costs and higher treatment expenditures.

The management of phosphorus in aquatic ecosystems is an important factor in controlling cultural eutrophication (Schindler 1977). Increases in total phosphorus concentration can significantly increase algal biomass, potentially creating blooms of

(20)

taste and odour causing species or cyanobacteria (Downing et al. 2001, Giani et al. 2005). Total phosphorus concentration is also a predictor of other odour-producing compounds in drinking water reservoirs (Davies et al. 2004). Increases in cyanobac-terial biomass have also been correlated with increased total nitrogen concentrations (Giani et al. 2005), an important implication for drinking water reservoirs. Increases in organic carbon in drinking water reservoirs are also important as increased concen-trations of organic carbon react with bromine and chlorine to produce carcinogenic disinfection by-products (Westerhoff et al. 2004) and increase bacterial biomass, or regrowth, within distribution systems (LeChevallier et al. 1996).

1.2 Thesis Objectives

The objective of this study is to add to the comprehensive assessment of potential impacts to water quality due to the expansion of the Sooke Lake reservoir, the main drinking water supply for Greater Victoria, British Columbia. In this study, I address the impacts of wetland inundation on the release of key nutrients affecting water quality and the subsequent transport of these nutrients to Begbie Lake and the Sooke Lake reservoir. I hypothesize that the introduction of the inundation regime will initiate the release of nutrients into Begbie Lake from the wetland system. Also, that years with a longer inundation period and greater areal extent will increase the release and export of nutrients into the Sooke Lake reservoir. While the results of this study are based on a single system, similar studies were used for comparative analysis.

Chapter 2 has been prepared as a stand-alone manuscript. In this chapter, I ad-dress the impacts of phosphorus, nitrogen and organic carbon release and subsequent transport into the water column of Begbie Lake due to the introduction of a seasonal inundation regime. I explore the effect of environmental variables such as water level,

(21)

water temperature and dissolved oxygen on nutrient release at open and forested wet-land sites within the study area and nutrient cycling within Begbie Lake. The export of key water quality nutrients, such as phosphorus, nitrogen and organic carbon from the Begbie Lake wetland system to the Sooke Lake reservoir is also assessed. The inundation of the Begbie Lake wetland system, due to the expansion of the Sooke Lake reservoir, has potential implications on the water quality of the reservoir as well as management implications for the artificial control of water levels on Sooke Lake reservoir to mitigate impacts from wetland inundation. The assessment of these im-plications is governed by the current hydrologic regime. The conclusions of the study and possible management implications are presented in the final chapter.

(22)

Chapter 2 Nutrient Release and Export from

the Begbie Lake Wetland System

to the Sooke Lake Reservoir in

Response to Hydroperiod

(23)

ABSTRACT

Wetland ecosystems are sinks of nutrients such as phosphorus, nitrogen and carbon, which are key factors of water quality. The retention capacity of wetlands for storing and transforming these nutrients is tied to the wetland hydroperiod. Hydroperiod alteration can stimulate the release of nutrients from soils and increase the decom-position of vegetation and detritus. The release of nutrients and other contaminants may alter the water quality of adjacent and hydrologically connected aquatic ecosys-tems. The water quality of drinking water reservoirs is especially important, where increases in nutrient concentrations may increase primary productivity and phyto-plankton species that may cause taste and odour issues. Secondly, increases in dis-solved organic carbon are correlated with the production of disinfection by-products during treatment, increases in bacterial biomass within distribution systems and de-creasing efficacy of water treatment.

In this study, I investigated the response of the Begbie Lake wetland system to the introduction of a seasonal inundation regime. The Begbie Lake wetland system is hydrologically connected to the Sooke Lake reservoir, the main drinking water supply for Greater Victoria, British Columbia. In 2002, the reservoir was expanded by 52.0 x 106m3to meet the demand of a growing population. Total phosphorus, total nitrogen and total organic carbon concentrations in the water column above the wetland were measured over two inundation seasons (2005 and 2006). The response to inundation, in terms of increased nutrient concentrations in Begbie Lake was compared to pre-inundation years over the growing season. While increases in total phosphorus, total nitrogen and total organic carbon were observed above wetland soils, only total phos-phorus concentrations significantly increased in Begbie Lake upon drawdown. Total nitrogen and total organic carbon concentrations in Begbie Lake did not reflect the release. The inundation of the Begbie Lake wetland system resulted in the export

(24)

of large amounts of nitrogen and organic carbon to the Sooke Lake reservoir during both inundation years.

(25)

2.1 Introduction

The hydrologic regime, or hydroperiod, is the main factor in the development and maintenance of the wetland ecosystem structure and function (Mitsch and Gosselink 1993, Richter et al. 1996, Silvertown et al. 1990, Casanova and Brock 2000, Adamus et al. 2001, Price et al. 2005). Alteration of the wetland hydroperiod, whether by anthropogenic or natural disturbance, alter many components of wetland structure and function including species composition and richness (Bunn and Arthington 2002), nutrient cycling and availability (Olila et al. 1997, Hill et al. 2000, Corstanje and Reddy 2004), soil oxygen availability (Reddy and Patrick 1975, Mitsch and Gosselink 1993), primary productivity via nutrient loading (Robertson et al. 2001), accumula-tion or export of organic matter (Olila et al. 1997, St. Louis et al. 2004) and the production of methyl mercury (St. Louis et al. 1996, St. Louis et al. 2004, Hall et al. 2005). Transition zones such as wetlands serve as sensitive ecological niches for many species and thus are susceptible to disturbance from slight environmental change (Mitsch and Gosselink 1993).

Generally, wetland ecosystems are large sinks of nutrients important to water quality such as phosphorus, nitrogen and carbon that are stored in the soils and sediments of the wetland system (Hemond 1983, Johnston et al. 1984, Craft 2000, Fisher and Acreman 2004) and both the aquatic and terrestrial vegetation (Hemond 1983, Hogan et al. 2004, Asada et al. 2005). In addition to acting as sinks of nutrients, wetland systems are also functionally important in regulating the transformation of nitrogen and carbon (Reddy and Patrick 1975, Mitsch and Gosselink, 1993, Bridgham et al.1998, Moore et al. 2003) and other biologically regulated elements such as sulphur. Due to the functions of nutrient retention and transformation, wetland systems are often used in nutrient removal and management (Mitsch et al. 2000, Fisher and Acreman 2004, Verhoeven et al. 2006, Kieckbusch and Schrautzer 2007).

(26)

Changes however, in the hydroperiod of a wetland system can influence wetland functioning turning the system from a nutrient sink to a source.

The release of nutrients from wetland soils is initiated by leaching from inundated wetland soils and the decomposition of organic matter. Water temperature (D’Angelo and Reddy 1999, St. Louis et al. 2003), inundation frequency (Battle and Golladay 2001, de Neiff et al. 2006), timing (Langhans and Tockner 2005), duration (Langhans and Tockner 2005) and areal extent (Moore et al. 2003) have all been recognized to impact these mechanisms and affect the release, mineralization and transport of nu-trients from wetland soils to surface waters and the potential export to hydrologically connected aquatic ecosystems.

In the majority of temperate freshwater ecosystems, phosphorus is the most limit-ing nutrient regulatlimit-ing primary productivity (Dillon and Rigler 1974, Schindler 1977, Kalff 2002). Aquatic ecosystems are sensitive to increases in external loading of phos-phorus from both anthropogenic sources and natural sources. In wetland ecosystems, organic phosphorus is stored in wetland soils, lake sediments, and vegetation (Kadlec 1997, Hogan et al. 2004). Inorganic phosphorus is adsorbed to particles of soil or bound in lake sediments, held in complexes with iron (Fe3+₃ ), aluminum (Al3+₃ ) and calcium (Ca2+₂ ) (Song et al. 2007, Reddy et al. 1999). In addition to being stored in wetland soils, a large amount of phosphorus is bound in the vegetation supported by wetland soils and sediments, such as emergent macrophytes (Reddy et al. 1999). Hy-drologic disturbance to the wetland, such as hydroperiod alteration from inundation, decreases the adsorption capacity of soil particles and increases the hydrolysis of Fe and Al phosphate complexes (Ponnamperuma 1972) leading to the release of phos-phorus into the overlying water column through leaching. As water levels recede, the subsequent decomposition of organic matter is accelerated as oxygen is introduced. The decomposition promotes increased enzyme activity, specifically phosphatase

(27)

ac-tivity, which regulates the mineralization of organic phosphorus (Song et al. 2007). Released from previously inundated wetland soils, inorganic phosphorus is readily transported to adjacent water bodies in surface run-off during precipitation events (Olila et al. 1997, Young and Ross 2001, Kieckbusch and Shrautzer 2007). The management of phosphorus in aquatic ecosystems is an important factor in control-ling cultural eutrophication (Schindler 1977) and is especially important in drinking water reservoirs, where increases in total phosphorus concentration can significantly increase algal biomass, potentially creating blooms of taste and odour causing species or cyanobacteria (Downing et al. 2001, Giani et al. 2005). Total phosphorus con-centration is also a predictor of other odour-producing compounds in drinking water reservoirs (Davies et al. 2004).

Nitrogen is an important component of aquatic ecosystems and is generally the second-most limiting nutrient next to phosphorus (Kalff 2002), and the most limit-ing nutrient in inundated wetland soils (Mitch and Gosselink 1993). In addition to being in high demand in aquatic ecosystems by macrophytes and algae, it is also an important component supporting microbial activity (Mitsch and Gosselink 1993, D’Angelo and Reddy 1999, Kalff 2002), where nitrate (NO−₃) and nitrite (NO−₂) act as electron acceptors under anaerobic conditions during the decomposition of organic matter (D’Angelo and Reddy 1999, Kalff 2002). Within wetlands, nitrogen is readily cycled and transformed by microbial activity (Mitsch and Gosselink 1993). Nitrogen exists as ammonium (NH+₄), NO−₃, NO−₂ or bound in organic forms; under anoxic conditions, the partial or complete denitrification of inorganic nitrogen by micro-bial activity may be released as nitrous oxide (N2O) or nitrogen gas (N2) (Downes

1988, Mitsch and Gosselink 1993, Kalff 2002, Andersen 2004). The retention of ni-trogen in wetland systems is primarily within vegetation, peat and in organic matter within the wetland soils (Hemond 1983). As organic matter decomposes, organic

(28)

nitrogen is readily mineralized into NH+₄ and may diffuse into the water column as wetland soils are inundated (Kieckbusch and Schrautzer 2007). The inundation of wetland soils and the subsequent receding of water levels increases the mineralization of nitrogen (Corstanje and Reddy 2004) and, under specific conditions, stimulates denitrification (Reddy and Patrick 1975, Seitzinger 1994, Olde Venterink et al. 2002, Gergel et al. 2005) or assimilation by wetland vegetation. However, if mineralization rates of organic nitrogen exceed denitrification rates in wetland sediments or macro-phyte nitrogen-limitation becomes saturated, the transport of inorganic nitrogen to hydrologically connected aquatic systems may occur (Hemond 1983). The retention potential of wetland systems is quite significant, as a large portion of nitrogen en-tering wetland ecosystems through precipitation and surface run-off is lost to the atmosphere (Gergel et al. 2005) or assimilated by wetland macrophytes (Laiho and Vasander 2003). Despite this, the eutrophication of aquatic ecosystems by nitrogen loading, specifically inorganic forms such as NO−₃, is readily documented (see Vi-tousek et al. 1997, Camargo and Alonso 2006). While most temperate freshwater systems are phosphorus limited, increases in total nitrogen concentrations may shift phytoplankton community structure. Increases in cyanobacterial biomass have been correlated with both total nitrogen and total phosphorus concentrations (Giani et al. 2005), an important implication for drinking water reservoirs.

Organic carbon is primarily stored in organic matter and peat within wetland soils (Moore and Dalva 2001, Moore et al. 2003, Laiho 2006), while wetland vegetation comprises a smaller carbon pool (Kelly 1997, Asada et al. 2005). Organic carbon is also an important component of the microbial food web, providing an energy source to power many of the biologically-mediated nutrient cycling processes. While wetland systems generally act as a sink of organic carbon, the inundation of soils can lead to increased leaching of organic carbon (Asada et al. 2005) and the subsequent transport

(29)

to the overlying water column (Moore et al. 2003, Corstanje and Reddy 2004, St. Louis et al. 2004). The inundation of wetland soils also stimulates the mineralization of organic carbon and the subsequent release of carbon dioxide (CO2) (Kelly et al.

1997, St. Louis et al. 2003). Under very reducing environments, the CO2 is also

subject to methanogenesis, being reduced to methane (CH4) (Mitsch and Gosselink

1993, Kelly et al. 1997, St. Louis et al. 2003). In addition to increasing the export of organic carbon to hydrologically connected aquatic ecosystems (Moore et al. 2003, St. Louis et al. 1996), both the methanogenesis and mineralization processes can switch a wetland system from being a carbon sink to a carbon source, significantly increasing the emission of CO2 and CH4 after inundation (Kelly et al. 1997). In

wetland systems, dissolved organic carbon is often transported to adjacent and hy-drologically connected aquatic ecosystems, carrying with it heavy metals (St. Louis et al. 2004). As wetland soils are inundated and organic carbon is processed, decompo-sition, mineralization, respiration and hydrologic transport all influence the potential export of carbon to hydrologically-connected aquatic ecosystems. Management of organic carbon in drinking water reservoirs is important as increased concentrations of organic carbon react with bromine and chlorine to produce carcinogenic disinfec-tion by-products (Westerhoff et al. 2004) and may increase bacterial biomass within distribution systems (LeChevallier et al. 1996).

In this study, I investigated the response of the Begbie Lake wetland system to the introduction of a seasonal inundation regime. Changes to the hydrologic regime of wetland systems can change the system from being a sink of key nutrients and other water quality contaminants to a source of these contaminants to hydrologically connected aquatic ecosystems. The introduction of a seasonal inundation regime may influence the rates of decomposition of organic matter in the wetland soils and increase the leaching of phosphorus, nitrogen and organic carbon from wetland soils

(30)

to the overlying water column. The Begbie Lake wetland system is hydrologically connected to the Sooke Lake reservoir, the main drinking water supply for Greater Victoria, British Columbia. In 2002, the reservoir was expanded to meet the demand of a growing population. The wetland system is subject to the seasonal recharge and drawdown of the reservoir; at full stage the Begbie Lake system is flooded and there is the potential for seasonal release and transport of phosphorus, nitrogen and carbon into Begbie Lake and the eventual export to the Sooke Lake reservoir. Increases in hydrologic connection may have negative implications on water quality for connected systems in terms of increased productivity (Paterson et al. 1997, Furey et al. 2004) leading to algal blooms, increased colour and the export of metals.

The objective of the study was to assess nutrient response in Begbie Lake to the seasonal inundation and nutrient release from the wetland soils. The effects of inundation extent, the timing and length of inundation and surface water temperature on nutrient release and response are explored. In addition, the export of phosphorus, nitrogen and organic carbon from the Begbie Lake wetland system to the Sooke Lake reservoir was calculated and the potential increases in nutrient concentrations were estimated.

2.2 Methods

2.2.1 Study Site

Begbie Lake (48.58509◦N, 123.68359◦W) is a small, shallow lake located on southern Vancouver Island in British Columbia, Canada (Figure 2.1). The lake is approxi-mately 0.86 Ha in area with a maximum depth of 7.5 m. Begbie Lake is surrounded by over 4 Ha of lacustrine shore marsh wetland (NWWG 1997) consisting primarily of sedges and grasses that transition into Red Alder (Alnus rubra) and forests of Western

(31)

Hemlock (Tsuga heterophylla) and Western Red Cedar (Thuja plicata) (Figure 2.2); the forested catchment is 291 Ha in area. The lake is classified as warm monomic-tic (Kalff 2002), mixing in the autumn through to the spring; Begbie lake stratifies between June and October. The Begbie Lake wetland system drains into the Sooke Lake reservoir through Begbie Creek (48.5825◦N, 123.6827◦W), an ephemeral creek, which historically flows during periods of increased precipitation and occasionally forms stagnant pools through the dry summer months (Figure 2.3). The Begbie Lake system is situated in the Sooke Lake watershed. The watershed is classified as part of the Coastal Western Hemlock very dry maritime biogeoclimatic sub-zone (Meidinger and Pojar 1991). Annual precipitation over the study period ranged from 1262 mm in 2004 to 1926 mm in 2006 (CRD 2009), with the majority of the precipitation falling between October and April (Figure 2.4). Between 1971 and 2000, the mean annual air temperature for the Sooke Lake reservoir at the dam was 8.8◦C. Climate normals were obtained by CRD data (Werner 2007).

The Sooke Lake reservoir (Figure 2.5), located within the Capital Regional District (CRD), is the drinking water supply source for the growing population of 330,000 consumers residing in Greater Victoria (Statistics Canada 2006). The water level of the Sooke Lake reservoir is artificially maintained by a spillway structure at the south end of the reservoir. The structure was upgraded in 2002 to meet the growing demand for water, raising the dam by 6 m to 186.75 meters above sea level (masl). The capacity of the Sooke Lake reservoir was increased by 52.0 x 106 m3, to a total of 92.7 x 106 m3, flooding an additional 131 Ha of land around the reservoir at full stage. The Begbie Lake wetland system, situated at 185.9 masl, is subject to annual inundation as the reservoir recharges during most winter months. Due to the recharge and drawdown cycles associated with drinking water reservoirs, the Sooke Lake reservoir reached the new full stage level in the spring of 2005, subjecting both

(32)

the Begbie Lake wetland system and Begbie Creek to a seasonal inundation regime. To mitigate nutrient loading into Sooke Lake reservoir due to the decomposition of vegetation, approximately 133 Ha of land in the inundation zone was cleared of substantial amounts of vegetation and slash. The riparian area of Begbie Creek, part of the inundation zone, was also cleared of vegetation. As water levels in the reservoir surpass 185.9 masl, the soils around Begbie Creek are inundated. The soils at the Begbie Creek site are primarily organic, humic-ferric podzols (Bryant 2007). The approximately 4 Ha of wetland around Begbie Lake was not cleared of any vegetation as part of the mitigation strategy.

Typical of drinking water reservoirs, increased demand for water occurs during the summer months; generally, during the winter through spring months (November to April), recharge exceeds drawdown and reservoir levels rise sharply (Figure 2.6). The reservoir exhibits consistent drawdown during the summer months, as consumption rises and inputs to the Sooke Lake reservoir from rain-generated run-off are negligible (Nowlin 2003). The seasonal drawdown and recharge of the reservoir depends upon both annual climate and consumption variability, both of which affect the seasonal variability of the inundation regime.

2.2.2 Sampling Methods

Over the study period, samples were collected from four sampling sites within the Begbie Lake wetland system (Figure 2.1): Begbie Lake 01), Begbie Creek (BBL-02), Begbie Lake open wetland (BBL-03) and Begbie Lake forested wetland (BBL-04). The Begbie Lake site (BBL-01) was situated over the deepest point of Begbie Lake, near the centre of the lake. Samples were collected monthly between January of 2003 and December of 2004 to determine baseline water chemistry and to assess the natural variability present in Begbie Lake for key water quality nutrients including:

(33)

total phosphorus (TP), total nitrogen (TN) and total organic carbon (TOC) before the introduction of the seasonal inundation regime. Baseline data was used to assess the response to hydroperiod alteration during the inundation of the Begbie Lake wetland system. Water samples were collected weekly between April of 2005 and October of 2006. Water samples were collected in triplicate at the Begbie Lake site using a 6 m length of weighted Tygon tubing. An integrated sample of theR

epilimnetic water column to a 5 m depth was collected and stored in opaque 2 L Nalgene polycarbonate containers. Sample containers were rinsed 3 times withR

de-ionized, distilled water in the laboratory and 3 times with sample water on-site. Water temperature, dissolved oxygen, specific conductivity and chlorophyll a were measured on each sampling day using a portable YSI 6600 multi-parameter sonde, which was calibrated using manufacturer-recommended standards in a laboratory environment before each sampling trip. A depth profile of each of the parameters was developed by taking the mean of 3 measurements at 1 m intervals of the entire water column. During stratification, measurements at 0.5 m intervals were taken over the metalimnion. The depth of the water column was recorded using a portable depth sounder.

Water samples were collected weekly at the Begbie Creek site (BBL-02) over the entire study period. Samples were collected in triplicate using a polycarbonate container attached to an extendable pole. The sampler was extended to the centre of Begbie Creek as to not disturb creek sediments. Water samples were stored in opaque 2 L Nalgene polycarbonate containers. Sample containers were rinsed 3 times withR

de-ionized, distilled water in the laboratory and 3 times with sample water on-site. Mean water temperature, dissolved oxygen concentration and specific conductivity were recorded at the creek site using a portable YSI 85 probe attached to the sampler on each sampling day. The depth of the water column at the creek site was recorded

(34)

using a portable staff gauge by wading into the creek after water chemistry samples had been collected.

Water samples were collected weekly at each of the wetland sites (BBL-03 and BBL-04) during the inundation periods in 2005 and 2006. Samples were collected in triplicate using a polycarbonate container attached to an extendable pole, so as not to disturb wetland soils. Water samples were stored in opaque 2 L Nalgene R

polycarbonate containers. Sample containers were rinsed 3 times with de-ionized, distilled water in the laboratory and 3 times with sample water on-site. The depth at each wetland site was recorded using a portable staff gauge. Mean water tempera-ture, dissolved oxygen concentration and specific conductivity were recorded at each wetland site using a portable YSI 85 probe on each sampling day.

All water samples were transported to the field station in 2 L Nalgene poly-R

carbonate containers held in coolers filled with ice. Water chemistry including: total phosphorus (TP), total dissolved phosphorus (TDP), nitrate, nitrite, total nitrogen (TN), dissolved organic carbon (DOC), total organic carbon (TOC), sulphate, pH and chlorophyll a was analyzed from the samples collected at each site. Samples for the analysis of TDP, nitrate, nitrite and sulphate were all filtered using a 0.45 µm nitrocellulose filter prior to storage. Samples for the analysis of TP and TN were not filtered prior to storage. Water samples for nutrient analysis were then stored in acid-washed 250 mL Nalgene polycarbonate bottles in coolers filled with ice and frozenR

within 6 hours until analysis. Dissolved organic carbon samples were filtered through a Whatman GF/F 0.7 µm glass fibre filter, which was ashed at 500R ◦C for 1 hour

to burn off residual carbon. Samples being analyzed for total and dissolved organic carbon were stored at 4◦C in glass vials in the dark; samples were stored with no air space in the vials. Total and dissolved organic carbon samples were analyzed within 7 days of collection. Dissolved organic carbon comprised 93% of total organic carbon

(35)

in the Begbie Lake wetland system (mean: 93%; standard deviation: 0.07%) and thus total organic carbon was used for statistical analysis in the study. Chlorophyll a samples were filtered through Whatman GF/F 0.7 µm glass fibre filter underR

low-light conditions and stored in polycarbonate screw-top test tubes. The test tubes were held in coolers filled with ice and frozen within 6 hours until analysis.

2.2.3 Laboratory Analyses

TDP, TP and TN were analyzed in the laboratory on a Lachat autoanalyzer (Zellweger Analytics, QuickChem 8000). TDP and TP were digested using the potassium per-R

sulfate method before analysis and measured as ortho-phosphate (APHA 1998). TN was digested with an alkaline potassium persulfate aqueous solution before analysis, buffered using 1 M hydrochloric acid and measured as nitrite using the cadmium reduction method (APHA 1998). Ion species such as nitrate, nitrite and sulphate were all analyzed using ion chromatography. Total (TOC) and dissolved (DOC) or-ganic carbon samples were analyzed on a Shimadzu TOC VCPH total oror-ganic carbon analyzer using the combustion catalytic oxidation/NDIR method. All samples were run with replicate and spiked samples to assess reproducibility and sample recovery, respectively. The filters for chlorophyll a analysis were macerated and stored in 95% ethanol for cold extraction at 4◦C for 24 hours. Chlorophyl a was then measured using a Turner Designs fluorometer calibrated to manufacturer-recommended standards.R

2.2.4 Spatial Analyses

Shapefiles of the Sooke Lake watershed and Begbie Lake wetland system were obtained from the Capital Regional District (CRD). All GIS work was completed using ESRI ArcView 3.2. Surface water features, wetland features, 2 m contours and theR

(36)

region in the Begbie Lake wetland system. Spatial data was collected and digitized from air photos collected by the Capital Regional District. The 2 m contour intervals were developed using the DEM interpreted from the air photos. The inundation zone around Begbie Lake and Begbie Creek was delineated at 0.125 m intervals in the GIS. A hypsometric curve (Luo 1998) of the inundated area was developed for calculation of water volumes. The inundated area was predicted at 0.01 m water level intervals between 186.0 and 187.0 masl using non-linear modelling using least-squares estimation. The inundated area for both inundation years was calculated.

2.2.5 Statistical Methods

All statistical analyses were performed using the open-source statistical computing environment, R (http://www.r-project.org).

Inundation Volume

The hypsometric curve produced from the GIS data for the Begbie Lake wetland system was used to calculate a simple water volume model during the inundation periods. Daily water volumes over the inundated wetland system were calculated based on daily water levels at the Sooke Lake reservoir and the hypsometric curve by taking the integral of the area-elevation hypsometric curve. The Begbie Lake system remains saturated for the majority of the year (Bryant 2007); water lost to groundwater is expected to be minimal relative to overland flow. Bryant (2007) also described groundwater flow between Begbie Lake and the Sooke Lake reservoir to be localized. This is further evidenced by the relatively rapid increase in the water table during high precipitation events.

(37)

Nutrient Release and Loading to Begbie Lake

Water chemistry data collected at the Begbie Lake site during all experimental years were separated into three treatment groups for statistical analysis. Baseline data for the Begbie Lake system were developed using samples collected in 2003 and 2004 to determine ambient conditions and variability of TP, TN and TOC before the introduction of the inundation regime. The baseline data are referred to as pre-inundation years throughout the text. Post-pre-inundation water chemistry data from 2005 and 2006 are referred to as inundation years.

Pearson product-moment correlation tests were used to assess the relationship between areal nutrient concentration and water level, water temperature and dissolved oxygen concentration at each of the inundated wetland sites (BBL-03 and BBL-04) in 2006. The data were first assessed to determine that there were linear relationships; TOC concentrations at the inundation sites were log-transformed prior to correlation analysis.

The effect of hydrologic disturbance on nutrient response in Begbie Lake was as-sessed using randomization (permutation) tests. Randomization tests are a robust method of testing hypotheses in single-case experiments as the assumptions of nor-mality, homoscedascity and random sampling need not be valid (Bult and Onghena 2008). The test statistic for the randomization procedure, for which the null hy-pothesis will be tested, can be chosen by the researcher based on the objective of the test (Bult and Onghena 2008). Indeed, a more familiar parametric test such as an ANCOVA or regression may be modified for use where the F -value may be used to test hypotheses (Edgington 1987). The randomization procedure calculates a test statistic for the original data set and subsequently re-orders the original data amongst the treatments, calculating the test statistic after each permutation. The number of iterations (permutations) is dependent upon sample size and number of

(38)

treatments, however the standard error of the p-value generally becomes very small after 5000 permutations (Potvin and Roth 1993). The null hypothesis is rejected if the proportion of the test statistics calculated after each permutation exceeding or equal to the test statistic of the data in its original form (Equation 2.1) is less than the accepted significance level (α = 0.05).

p = k/N (2.1)

where:

k is the number of observed test statistics ≥ the original test-statistic N is the number of permutations

Nutrient Export to the Sooke Lake Reservoir

The export of phosphorus, nitrogen and organic carbon from the Begbie Lake wet-land system was calculated using the created inundation water volume model. Local weighted regression (LOESS) was used to interpolate (Lee and Hogsett 2001) the daily concentrations of total phosphorus, total nitrogen and total organic carbon in Begbie Creek over the study period. LOESS regression methods are computationally intensive and build local polynomial regression models; the premise is that parameter values are best predicted by the local subset of values. The LOESS model weights the values nearest the point of interest highest and those at the further ends of the subset lower. The smoothing parameter (α) is the proportion of data used to build the local polynomial regression; α was set to 0.1 for estimation of daily nutrient concentrations. The export of TP, TN and TOC was then estimated using the daily changes in water volume applied to the interpolated concentrations.

(39)

2.3 Results

2.3.1 Hydroperiod

The expansion of the Sooke Lake reservoir was completed in 2002. Lower than av-erage water levels resulting from a 100-year drought (Nowlin et al. 2004) in 2001 contributed to incomplete recharge of the reservoir over the following four years. The full stage capacity of the reservoir was finally reached in April of 2005. The hydrologic regime for Sooke Lake followed similar patterns for recharge and drawdown during the experimental years from 2003 - 2006. Water consumption over the summer months exhibited similar characteristics in both treatment years as shown by the slope of the drawdown exhibited in the Sooke Lake reservoir (Figure 2.6). The water level required to inundate the wetland region surrounding Begbie Lake is 185.9 masl; the reservoir reached full stage, inundating the wetland system, in both treatment years (2005 and 2006) (Figure 2.7). In 2005, the inundation period lasted from April 1 to May 28, a period of 57 days. In 2006, the inundation period doubled, beginning on January 19 and remained inundated until May 19, a period of 127 days. The inundation of the wetland system began earlier in the season, but exhibited a similar drawdown pattern to 2005.

2.3.2 Inundation Water Volume Model

A hypsometric curve of the area/elevation relationship in the Begbie Lake wetland system was developed. Non-linear least squares estimation was used to produce the curve (Figure 2.8). The area-elevation relationship is described as Equation 2.2

(40)

Using the hypsometric curve, daily water volumes were calculated based on the water levels obtained at the Sooke Lake reservoir spillway. Estimates of the water volume inundating the Begbie Lake wetland systems include some sources of error. First, water lost to underground flow is not estimated using this approach. Bryant (2007) has estimated that much of the wetland system surrounding Begbie Lake re-mains saturated for the majority of the year, the volume lost to underground flow should be minimal compared to overland flow. The second source of error is evapora-tion; in many studies, estimates of evaporation often comprise the residual error term in water budget models (Mitsch and Gosselink 1993). Due to the sheltered nature of the Begbie Lake system and the timing of year, evaporative loss during the inunda-tion period is expected to be small. A third potential source of measurement error is the vertical precision of the GIS data derived from the air photo analysis. The error associated with the contour development from the DEM was not available.

During both inundation years, the Begbie Lake wetland system was inundated (Figure 2.7). In 2005, approximately 31,117 m2 of wetland and organic soils were inundated as the reservoir water level rose. The wetland region was inundated for 57 days; the maximum water volume over the inundated area was 19,806 m3_{(Figure 2.9).}

The inundation period in 2006 lasted for 123 days. The area inundated reached the full extent of the wetland region; approximately 45,670 m2 _{of wetland was flooded.}

The estimated peak volume over the wetland region was 33,800 m3 _{(Figure 2.10).}

Using the inundation water volume model, the daily volume of water transferred between Begbie Lake and Sooke Lake was estimated (Figure 2.11). The estimated volume of water transferred was used to assess the potential export of TP, TN and TOC into Sooke Lake following drawdown.

(41)

2.3.3 Stratification, Surface Water Temperature and Dissolved

Oxygen in Begbie Lake

Temperature profiles were developed on each sampling day using a YSI 6600 multi-parameter sonde. Generally, Begbie Lake (BBL-01) stratifies between May and Oc-tober. Over the experimental years, the length of stratification was similar (Figure 2.12) although epilimnetic depth decreased during the inundation years (2005 and 2006). Surface water temperature was recorded using the temperature profiles cre-ated on each sampling day as the mean water temperature taken at 1 m intervals from the surface to a depth of 5 m. Water temperature between May and October in 2003 ranged from 9.5 to 18.8◦C; in 2004, water temperature over the same period ranged from 11.1 to 17.8◦C. During the inundation years, water temperature was similar, ranging from 9.9 to 16.7◦C in 2005 and from 8.9 to 16.1◦C in 2006. While the surface water temperature appeared to be slightly higher in the pre-inundation years there was no significant difference (F(0.05,3,61) = 2.14, p = 0.104).

In addition to temperature, dissolved oxygen profiles were also developed for each sampling day a YSI 6600 multi-parameter sonde. Monthly dissolved oxygen isopleths show similar patterns over the experimental years (Figure 2.13). The hypolimnion of Begbie Lake shifted to an anoxic state during the stratification period for each exper-imental year. In 2004, the hypolimnion was anoxic up to a depth of approximately 6 m, lasting from June until mixing occurred in September. While the pattern was similar in both 2005 and 2006, the length of hypolimnetic anoxia shifted slightly from mid-June into October and late September, respectively. The hypolimnion extended slightly in the treatment years, to a depth of approximately 5 m.

(42)

2.3.4 Nutrient Release and Loading into Begbie Lake

Nutrient Release Above Wetland

Water samples were collected over the open wetland (BBL-03) and forested wetland (BBL-04) sites during the inundation years. Water samples were analyzed for TP, TN and TOC for each sampling day during the inundation periods (Figure 2.14). The concentration of all key nutrients increased in both inundation years over the open wetland and forested wetland sites during the inundation period.

During the 2005 inundation period (57 days), the mean concentration of TP in-creased from 5.2 µg/L to 15.3 µg/L over the open wetland site; the mean TP concen-tration over the forested wetland site increased from 5.3 µg/L to 25.5 µg/L (Figure 2.14a). Mean TN concentration over the open wetland site increased from 141.1 µg/L to 335.8 µg/L; over the forested wetland site, mean TN concentration increased from 123.2 µg/L to 505.7 µg/L (Figure 2.14b). Mean concentrations of TOC followed a similar pattern to TP and TN; over the open wetland site, increases in TOC from 3.9 mg/L to 5.5 mg/L were observed. At the forested wetland site, the mean concen-tration of TOC increased from 4.1 mg/L to 7.8 mg/L during the inundation period (Figure 2.14c). A pronounced peak in each parameter was observed at the forested wetland site 28 days after the inundation began. Similar, but much weaker signals in TP and TOC were observed at the open wetland site 28 and 32 days after inundation, respectively.

During the 2006 inundation period (127 days), the pattern of TP concentrations at the open wetland site exhibited similar characteristics to 2005; the mean concen-tration of TP increased from 3.6 µg/L to 25.6 µg/L. At the forested wetland site, the mean TP concentration increased from 3.2 µg/L to 21.6 µg/L over the same period (Figure 2.14d). TN concentration increased from 62.6 µg/L to 351.9 µg/L at the open wetland site. The mean concentration at the forested wetland site increased from 65.0

(43)

µg/L to 183.4 µg/L (Figure 2.14e), but peaked on day 117 at 236.2µg/L. The mean concentration of TOC increased at both the open and forested wetland sites, from 3.7 mg/L to 7.6 mg/L and from 3.6 mg/L to 5.6 mg/L, respectively (Figure 2.14f). The mean TOC concentration above the forested wetland site peaked at 6.6 mg/L on day 117 of the 2006 inundation period.

During the inundation periods of both treatment years, the concentration of total phosphorus, total nitrogen and total organic carbon increased during the inundation period. However, as water levels were continuously changing between sampling days, concentrations were normalized by depth to show the concentration per lake surface area above both wetland types in order to remove dilution and concentration ef-fects (Figure 2.15). The relationship between nutrient concentration and water level, surface water temperature and dissolved oxygen concentration were explored using Pearson product-moment correlation (Table 2.1) for water samples collected during the 2006 inundation period. In addition, the relationships between each of the nu-trients during this inundation period were explored to examine correlations between TP, TN and TOC during release (Table 2.2). The Pearson correlation coefficient (r ) was calculated for each pairing.

At the open wetland site in 2006, TP release was negatively correlated with dis-solved oxygen concentration (t(0.05,15) = −2.95, r = −0.61, p = 0.010). TN was

strongly correlated with surface water temperature (t(0.05;15) = 4.77, r = 0.78, p <

0.001), and a weak, negative correlation with dissolved oxygen concentration (t(0.05;15)=

−2.14, r = −0.48, p = 0.049) was observed. A strong, positive correlation was ob-served between TOC and water level (t(0.05;15) = 3.85, r = 0.71, p = 0.002) at the open

wetland site. A significant relationship was found between total phosphorus and total nitrogen (t(0.05;15) = 2.36, r = 0.52, p = 0.032) over the open wetland.

(44)

water temperature (t(0.05;15) = 2.62, r = 0.56, p = 0.019) was observed and a strong,

negative correlation with dissolved oxygen (t(0.05;15) = −4.19, r = −0.73, p < 0.001)

was also seen; both relationships were significant. The concentration of TN at the forested site was positively correlated with surface water temperature (t(0.05;15) =

2.90, r = 0.60, p = 0.011) ; a negative correlation between TN and dissolved oxygen was observed (t(0.05;15) = −2.73, r = −0.58, p = 0.016). A positive correlation

be-tween TOC and water level was also observed at the forested wetland site (t(0.05;15)=

3.17, r = 0.63, p = 0.006). A very strong positive correlation between TP and TN was observed at the forested wetland site (t(0.05;15)= 8.77, r = 0.91, p < 0.001).

Nutrient Loading to Begbie Lake

As nutrients were released from wetland soils, accumulated organic matter and the decomposition of vegetation at both the open and forested wetland sites and enter the water column, there is potential for transfer into Begbie Lake as the water level recedes. The response of each TP, TN and TOC to the inundation regime during treatment years was measured between May and October at the Begbie Lake site (BBL-01). The response to hydroperiod variability, inundation extent and surface water temperature on the key nutrient concentrations was tested using randomization methods written for ANCOVA, ANOVA and linear regression models. TP and TN were found to be correlated with surface water temperature at the wetland sites (BBL-03 and BBL-04); to account for the variation attributed to water temperature of Begbie Lake, mean surface water temperature was used as a covariate in the models. TP concentration in the epilimnion of Begbie Lake ranged between 6.4 µg/L and 12.6 µg/L during the summer of the pre-treatment years (2003 and 2004). In 2005, the first inundation year, TP concentration ranged from 4.8 µg/L to 29.3 µg/L during the same period; during 2006, TP concentration ranged from 8.9 µg/L to 32.4 µg/L

(45)

(Figure 2.16). To test the impact of the altered hydroperiod and water temperature on TP concentration in the epilimnion of Begbie Lake, a randomization procedure was employed using the ANCOVA F -value as the chosen test-statistic. The analysis was run through the procedure to build a distribution of F -values by randomizing the TP concentrations and the associated water temperature values. The interaction effect between the inundation treatments and water temperature as a covariate was assessed based on the developed distribution. The resulting model explains 40.3% of the variation (R2 = 0.403), while the interaction effect was found to be insignificant (F(0.05,2,59) = 4.17; p = 0.051); the very small p-value suggests that the TP

concen-tration is responding to the inundation of the wetland area and being influenced by surface water temperature (Figure 2.17a). The lack of significance is likely attributed to a small loss in statistical power due to the unbalanced distribution of data amongst treatments (Good 1994). The effect of the lengthened inundation and increase in ex-tent in 2006 over the previous inundation year was also tested using the ANCOVA randomization procedure (Figure 2.17b); the two treatments were found to not be significantly different in terms of TP response (F(0.05,1,50) = 4.63, p = 0.076). Finally,

the effect of water temperature on TP response during the inundation years was analyzed using a randomization procedure developed to test regression relationships (Figure 2.17c). I found that surface water temperature had a significant effect on TP response during inundation years only (F(0.05,1,51) = 21.34, p < 0.001); the regression

model explained 28.0% of the variation (R2 = 0.280).

The concentration of TN at the Begbie Lake site (BBL-01) ranged from 145.4 µg/L to 383.6 µg/L over the summer season during the pre-inundation years. In 2005, the TN concentration ranged from 127.2 µg/L to 363.5 µg/L during the same period. Likewise, TN concentrations during the final experimental season ranged from 112.0 µg/L to 375.0 µg/L (Figure 2.18). The effect of surface water temperature and

(46)

hydrologic disturbance was assessed using the ANCOVA randomization procedure for all experimental years. The analysis showed that both the interaction effect between water temperature and hydrologic disturbance as well as the treatment (hydrologic disturbance) itself did not have a significant effect on TN response (F(0.05,2,58) =

0.506, p = 0.512) and (F(0.05,2,58) = 0.857, p = 0.611), respectively (Figure 2.19a).

The effect of water temperature alone on TN concentration over the experimental years was assessed with linear regression using randomization methods. I found that in both the pre-treatment and treatment years, surface water temperature had a significant effect on TN response (F(0.05,1,62)= 17.62, p < 0.001), although the model

explained little of the variation seen in TN concentration over the summer periods (R2 _{= 0.21) (Figure 2.19b).}

TOC was measured at the centre Begbie Lake site over the summer season. The baseline concentration of TOC ranged from 3.4 mg/L to 8.1 mg/L during the pre-treatment years. During the first inundation year, the concentration of TOC ranged from 3.2 mg/L to 7.4 mg/L and during the second inundation year, the concentra-tions observed at the site were similar to previous years; the concentration ranged from 3.4 mg/L to 7.4 mg/L over the summer season. The seasonal pattern of TOC in the epilimnion of Begbie Lake followed a very similar pattern during all experimental years (Figure 2.20). The randomization procedure was employed to assess the effect of wetland inundation on TOC response in the water column of Begbie Lake using an ANOVA test. The F -statistic was used to test the hypothesis that TOC response re-mained similar among all experimental years; the randomization was run with 50,000 permutations. I found that there was no significant change in TOC concentration between all experimental years (F(0.05,2,61) = 0.384, p = 0.681).

(47)

2.3.5 Begbie Creek Water Quality

The inundation of the Begbie Lake wetland system influenced the water quality in terms of increases in TP, TN and TOC following the drawdown of water levels on the Sooke Lake reservoir. In Begbie Creek, the concentration of TP increased substan-tially during the inundation period (Figure 3.9). In 2005, the concentration of TP ranged from 2.7 µg/L to 23.5 µg/L and in 2006, similar values were observed, ranging from 2.9 µg/L to 17.0 µg/L. The concentration of TN similarly increased during the inundation period (Figure 3.10). In 2005, TN concentration ranged from 86.7 µg/L to 389.1 µg/L and in 2006, the concentration of TN ranged from 47.3 µg/L to 258.7 µg/L. The concentration of TOC at the Begbie Lake site also increased markedly during the inundation period of both years (Figure 3.11). TOC concentration in-creased from 2.6 mg/L to 8.9 mg/L during the 2005 inundation period. In 2006, the concentration of TOC similarly increased from 3.4 mg/L to 7.1 mg/L.

2.3.6 Nutrient Export to the Sooke Lake Reservoir

The export of key water quality nutrients was estimated using the water budget model and the interpolated concentrations of TP, TN and TOC at the Begbie Creek site. The estimated export (Table 2.3) of TP in 2005 was estimated to be 245.9 g; in 2006 an estimated 383.0 g was exported during drainage from Begbie Lake to the Sooke Lake reservoir. Considerably higher contributions of TN were estimated. In 2005, approximately 3.8 kg was exported; in 2006 approximately 4.4 kg. TOC export from the Begbie Lake wetland system was substantially higher than TP and TN. In 2005, an estimated 77.7 kg of organic carbon was exported. More than twice that amount was exported in 2006 (171.1 kg). The intensity of inundation, that is, the extent and duration appeared to have a considerable impact on the loading of nutrients into the Sooke Lake reservoir.

(48)

It is apparent from the export calculations that the introduction of a seasonal inundation regime to the Begbie Lake wetland system will most likely result in the seasonal transport of these nutrients in the future. To assess the impact on Sooke Lake reservoir in terms of increase in nutrient concentrations, the estimated changes in TP, TN and TOC concentrations were calculated; export values of these nutrients were used to determine the potential increase assuming full mixing (Table 2.4). The historic full stage volume (40.7 x 106m3) was used as an estimate of volume in the calculation. This volume was chosen as it closely approximates the volume of the reservoir during the drainage of the Begbie Lake wetland system. Based on these estimates, the change in TP concentration was expected to be 0.006 µg/L in 2005 and 0.009 µg/L in 2006. Estimated increases in TN concentration were 0.929 µg/L in 2005 and 0.109 µg/L in 2006. The greatest increases were expected to be in TOC concentrations; in 2005 the estimated increase is 1.91 µg/L and in 2006, 4.21 µg/L. The assumption of full mixing is a simplistic model used to assess the potential increase in nutrient concentrations. Differences in water temperature at the time of mixing will influence the mixing regime. It is likely, given the physical characteristics of the reservoir, that near the mouth to Begbie Creek, localized increases greater than those predicted could occur.

2.4 Discussion

2.4.1 Hydroperiod

The change in the hydroperiod of the Begbie Lake wetland ecosystem was directly influenced by the expansion of the Sooke Lake reservoir following upgrades to the im-poundment structure in 2002. The increase in full stage of the Sooke Lake reservoir by 6 m altered the hydroperiod in the wetland system. Prior to hydroperiod alteration,

(49)

the Begbie Lake wetland system exhibited saturation, that is a water table near the surface, for the majority of the year (Bryant 2007). Due to the hydrologic alteration, the Begbie Lake wetland system is seasonally flooded during times of full stage. Wet-land ecosystems that are periodically flooded are generally more productive wetWet-lands (Mitsch and Gosselink 1993). The interactions between vegetation, organic matter and detritus and the microbial community in the wetland soils are often increased in seasonally flooded or periodically flooded wetland systems as nutrients and oxygen are introduced into the system (Glazebrook and Robertson 1999, Robertson et al. 2001, Olde Venterink et al. 2001, Corstanje and Reddy 2004, Langhans and Tockner 2005, Kieckbusch and Schrautzer 2007).

The seasonal hydrologic pattern observed in the Begbie Lake system is subject to influence from both climate variability and consumption rates via their impact on the water level of the Sooke Lake reservoir. The Sooke Lake reservoir, primarily recharged by precipitation and surface run-off during the winter months between November and March, exhibits drawdown during the summer as water consumption increases. During the inundation years, water withdrawal proceeded at a similar rate, however precipitation variability drove large differences in reservoir recharge and the subsequent differences in the inundation periods in 2005 and 2006. Multiple storm events in January of 2006 (Figure 2.4) drove the rapid recharge, which led to the lengthened inundation period observed. Both climate variability and consumption play a role in the volume of water stored in the Sooke Lake reservoir throughout the year (Nowlin et al. 2004). Both climate variability and consumption similarly impact the thermal regimes and stratification periods (Nowlin et al. 2004). The water level of the Sooke Lake reservoir is the primary driver of the hydroperiod exhibited in the Begbie Lake wetland system. In the face of climate change and increases in population within Greater Victoria (Statistics Canada 2006), it is possible that

Fate and transport of nutrients from the Begbie Lake wetland system: measuring the impacts of inundation on hydrologically connected aquatic ecosystems.

Table of Contents

List of Tables

List of Figures

1.1

General Introduction

1.1.1

Wetland Hydroperiod

1.1.2

Mechanisms of Nutrient Release

1.1.3

Impacts on Water Quality

1.2

Thesis Objectives

Chapter 2

Nutrient Release and Export from

the Begbie Lake Wetland System

to the Sooke Lake Reservoir in

Response to Hydroperiod

2.1

Introduction

2.2

Methods

2.2.1

Study Site

2.2.2

Sampling Methods

2.2.3

Laboratory Analyses

2.2.4

Spatial Analyses

2.2.5

Statistical Methods

2.3

Results

2.3.1

Hydroperiod

2.3.2

Inundation Water Volume Model

2.3.3

Stratification, Surface Water Temperature and Dissolved

Oxygen in Begbie Lake

2.3.4

Nutrient Release and Loading into Begbie Lake

2.3.5

Begbie Creek Water Quality

2.3.6

Nutrient Export to the Sooke Lake Reservoir

2.4

Discussion

2.4.1

Hydroperiod