(Populus Trichocarpa) Nutrition in the Dewatered Lake Aldwell Reservoir on the Elwha River, Washington

Western Washington University Western CEDAR

WWU Graduate School Collection WWU Graduate and Undergraduate Scholarship

Winter 2016

Black Cottonwood (Populus Trichocarpa) Nutrition in the Dewatered Lake Aldwell Reservoir on the Elwha River, Washington

Ezra H. Citron Western Washington University, [email protected]

Follow this and additional works at: https://cedar.wwu.edu/wwuet

Part of the Environmental Sciences Commons

Recommended Citation Citron, Ezra H., "Black Cottonwood (Populus Trichocarpa) Nutrition in the Dewatered Lake Aldwell Reservoir on the Elwha River, Washington" (2016). WWU Graduate School Collection. 466. https://cedar.wwu.edu/wwuet/466

This Masters Thesis is brought to you for free and open access by the WWU Graduate and Undergraduate Scholarship at Western CEDAR. It has been accepted for inclusion in WWU Graduate School Collection by an authorized administrator of Western CEDAR. For more information, please contact [email protected].

BLACK COTTONWOOD (POPULUS TRICHOCARPA) NUTRITION IN THE DEWATERED LAKE ALDWELL RESERVOIR ON THE ELWHA RIVER, WASHINGTON

By Ezra Citron

Accepted in Partial Completion Of the Requirements for the Degree Master of Science

Kathleen L. Kitto, Dean of the Graduate School

ADVISORY COMMITTEE

Chair, Dr. Peter Homann

Dr. Jenise Bauman

Dr. Rebecca Bunn

MASTER’S THESIS

In presenting this thesis in partial fulfillment of the requirements for a master’s degree at Western Washington University, I grant to Western Washington University the non‐ exclusive royalty‐free right to archive, reproduce, distribute, and display the thesis in any and all forms, including electronic format, via any digital library mechanisms maintained by WWU.

I represent and warrant that this is my original work and does not infringe or violate any rights of others. I warrant that I have obtained written permissions from the owner of any third party copyrighted material included in this work.

I acknowledge that I retain ownership rights to the copyright of this work, including but not limited to the right to use all or part of this work in future works, such as articles or books.

Library users are granted permission for individual, research, and non‐commercial reproduction of this work for educational purposes only. Any further digital posting of this document requires specific permission from the author.

Any copying or publication of this thesis for commercial purposes, or for financial gain, is not allowed without my written permission.

Ezra Citron February 18, 2016

BLACK COTTONWOOD (POPULUS TRICHOCARPA) NUTRITION IN THE DEWATERED LAKE ALDWELL RESERVOIR ON THE ELWHA RIVER, WASHINGTON

A Thesis Presented to The Faculty of Western Washington University

In Partial Completion Of the Requirements for the Degree Master of Science

By Ezra Citron February 2016

Abstract

Following dam removal on the Elwha River, WA, the ability of plants to access and uptake nutrients may be an important factor in ecosystem recovery. Foliage was collected in

November 2014 from naturally-established black cottonwood (Populus trichocarpa) saplings growing in the dewatered Lake Aldwell reservoir sediments and adjacent forest. Analysis of variance was used to assess variability among reservoir sediment textures and the adjacent forest with respect to foliar macronutrient (N, P, K, Ca, Mg) and micronutrient (Fe, Mn, Zn,

Cu, Ni) concentrations. Within the reservoir, foliar P, K, and Mn were higher, and foliar Ca was lower, in the fine sediments than in the coarse sediments. All reservoir environments had lower foliar Mg and Zn and higher foliar Mn than the adjacent forest. Correlation analyses indicated many significant correlations between nutrient pairs. Several positive nutrient correlations with N may indicate that N is stimulating protein synthesis, and several negative correlations between nutrient pairs may indicate that cations are competing for plant uptake.

Considering seasonal nutrient fluctuation and senescence-induced nutrient resorption, foliar nutrient values measured in Lake Aldwell cottonwood were generally consistent with those reported in the literature from other populations of non-fertilized Populus. Exceptions were values for foliar Ca, Fe, and Cu, which were comparatively low. The majority of resorption- adjusted growing-season estimates for Lake Aldwell foliar macronutrients were within ranges reported as optimal in fertilized Populus, with the exception of high estimates for Mg in the adjacent forest and for N at all sites. Presently, black cottonwood productivity at Lake

Aldwell does not seem to be constrained by nutrient limitation, but nutrient availability will be critical in determining species-compositional changes at dewatered Lake Aldwell and successional trajectories in other dewatered environments following dam removal. iv

Acknowledgments This project was funded by a Huxley College Small Grant for Graduate Research awarded by Huxley College of the Environment. Travel to and from the 2015 Elwha River Science Symposium was supported by a Ross Travel Grant awarded by Western Washington University. I appreciate the support of these programs. I would like to thank the National Parks Conservation Association for allocating a scholarship for me to attend this Symposium, which provided valuable context within which to interpret the findings of my research. I would like to thank my thesis committee chair, Dr. Peter Homann, for his suggestions, perspective, and gentle guidance throughout this research process and graduate program. His detailed feedback has fostered my development as a responsible scientist and has challenged me to become more efficient and organized in my research process. I would like to thank the other members of my committee, Dr. Jenise Bauman and Dr. Rebecca Bunn, for the insight that they have provided as my project has developed, especially with respect to my statistical design, analysis, and interpretation. I would like to thank Olympic National Park and Jerry Freilich for allowing me to sample at the Lake Aldwell reservoir and Joshua Chenoweth for showing me appropriate Lake Aldwell sampling sites and sharing details about the Elwha River ecosystem restoration project, which helped to focus my research questions. I would like to thank my brothers, Noah Citron and Andrew Relihan, for their enthusiastic assistance with sample collection. Noah, your late-night help processing leaves in the lab was a memorable and appreciated contribution to my project. I would like to thank my parents, Anji and Todd Citron, for supporting me in every way imaginable throughout the duration of this project, including joining me at the lab on occasion to help me complete my chemical analyses. I would like to thank Dr. David Shull, Dr. Ruth Sofield, Fraser Wilkinson, Kyle Mikkelsen, and Erin Macri for providing me access to and technical assistance with analytical instrumentation, for helping me to procure lab supplies, and for helping me to understand the protocols of trace metals analysis. I would like to thank Chhaya Werner and Dr. Rebecca Brown for providing me with data from previous Elwha River studies. Lastly, I would like to thank John Jordy and Sarah Voth for their valuable support throughout this project.

Abstract ...... iv Acknowledgments ...... v List of Figures ...... ix List of Tables ...... ix 1.0 Introduction ...... 1 1.1 Nutrient Limitation in Primary Successional Environments...... 1 1.2 Ecological Setting: The Dewatered Lake Aldwell Reservoir ...... 2 1.3 Nutrients in the Elwha River Ecosystem...... 3 1.4 Study Objectives, Research Questions, and Hypotheses ...... 4

2.0 Background ...... 7 2.1 Plant Nutrition ...... 7 2.1.1 The Importance of Plant Nutrition ...... 7 2.1.2 The Essential Mineral Elements ...... 8 2.1.3 Nutrient Concentrations in Plant Tissue ...... 8 2.1.4 Functions of Nutrients in Plants ...... 10 2.1.5 Nutrient Mobility and Translocation within Plants ...... 11 2.1.6 Leaf Growth, Senescence, and Nutrient Resorption ...... 14 2.1.7 Nutrient Interactions within Plants ...... 17 2.2 Assessment of Plant Nutrition ...... 19 2.2.1 Plant-Tissue Nutrient Concentrations ...... 19 2.2.2 Plant-Tissue Nutrient Ratios ...... 21 2.2.3 Relation of Plant Nutrition to Substrate Nutrient Concentrations ...... 22 2.2.4 Standardized Approaches...... 24 2.2.5 Populus and the DRIS ...... 27 2.3 Primary Succession and the Dewatered Lake Aldwell Reservoir ...... 27 2.3.1 Principles of Primary Succession...... 27 2.3.2 Primary Succession in the Elwha River Ecosystem ...... 29 2.3.3 Substrate and Nutrients in Primary Successional Environments ...... 30 2.3.4 Substrate and Nutrients in Elwha River Dewatered Reservoirs ...... 31

2.4 Gaps in Knowledge ...... 33 2.4.1 Primary Succession in Dewatered Reservoirs ...... 33 2.4.2 Plant Nutrition in the Dewatered Lake Aldwell Reservoir ...... 34

3.0 Methods ...... 36 3.1 Sample Collection ...... 36 3.1.1 Study Site ...... 36 3.1.2 Sampling Sites ...... 36 3.1.3 Sampling Procedures ...... 38 3.2 Sample Preparation ...... 40 3.2.1 Foliar Sample Drying and Homogenization ...... 40 3.2.2 Sediment Sample Drying and Weighing...... 41 3.3 NC Elemental Analysis ...... 41 3.4 ICP-MS Elemental Analysis ...... 42 3.4.1 Chemical Reagents...... 42 3.4.2 Sample Digestion ...... 42 3.4.3 Elemental Analysis ...... 43 3.4.4 Quality Control ...... 45 3.5 Statistical Analysis ...... 46 3.5.1 Analysis of Variance ...... 47 3.5.2 Correlation Analysis ...... 48 3.5.3 Statistical Errors ...... 48 3.5.4 Literature Value Comparisons and Growing-Season Estimates ...... 49

4.0 Results and Discussion, Part I: Within the Reservoir and Adjacent Forest ...... 50 4.1 Variation in Foliar Nutrients among Microenvironments ...... 50 4.2 Foliar Nutrient Correlations with Environmental Variables ...... 54 4.3 Correlations among Foliar Nutrients ...... 59 4.4 Effects of Sapling Age and Size ...... 63

5.0 Results and Discussion, Part II: Literature Value Comparisons ...... 66 5.1 Seasonal Variation in Foliar Nutrient Values ...... 66

vii

5.1.1 Patterns in Literature Values ...... 66 5.1.2 Values Measured in Lake Aldwell Microenvironments ...... 71 5.2 Optimal Foliar Macronutrient Values ...... 72 5.2.1 Growing-Season Concentration Estimates ...... 72 5.2.2 Growing-Season Ratio Estimates ...... 75 5.3 Interpretations and Implications ...... 77 5.3.1 Interpreting Lake Aldwell Reservoir Foliar Nutrients in Context ...... 77 5.3.2 Implications for Ecosystem Rehabilitation ...... 78

6.0 Future Research ...... 80 6.1 Substrate Characterizations ...... 80 6.2 Monitoring Over Time ...... 80 6.2.1 Seasonal Plant-Tissue Analyses ...... 80 6.2.2 Tracking of Plant Nutrition to Assess Succession ...... 81 6.2.3 Stable Isotope Studies ...... 82 6.3 Quantitative Plant Nutritional Assessments ...... 82

7.0 Conclusion ...... 84 7.1 Evaluation of Hypotheses...... 84 7.2 Nutrient Availability and Plant Succession ...... 87

8.0 References ...... 90 Appendix ...... 114

viii

List of Figures Figure 1. Map of sampling sites……………………………………………………………...38 Figure 2. Correlations between sediment moisture and foliar nutrients………………….…56 Figure 3. Correlations between stem density and foliar nutrients…………………………59 Figure 4. Seasonal trends in foliar nutrient values in Populus………………………………67 Figure 5. Optimal and growing-season-estimated foliar nutrient values in Populus……….74 Figure A. Pairwise macronutrient and micronutrient correlation plots………….…………117

List of Tables Table 1. Mean foliar macronutrient concentrations of cottonwood saplings…………….…50 Table 2. Mean foliar micronutrient concentrations of cottonwood saplings……….………51 Table 3. Nutrient concentrations of local forest soils and reservoir sediments….……….…52 Table 4. Statistically significant results from pairwise nutrient-nutrient correlation tests….60 Table 5. Mean dimensions of cottonwood saplings and woody plant stem densities………64 Table 6. Previously-reported Populus foliar nutrient value study descriptions…………68-69 Table A. Certified and measured SRM 1515 macronutrient concentrations………………114 Table B. Certified and measured SRM 1515 micronutrient concentrations………………114 Table C. Certified and measured SRM 1515 non-nutrient-element concentrations………115 Table D. Mean foliar non-nutrient-element concentrations of cottonwood saplings………115 Table E. Statistical results from all pairwise nutrient-nutrient correlation tests……………116

1.0 Introduction

1.1 Nutrient Limitation in Primary Successional Environments

Macronutrient limitation constrains ecosystem productivity in many different terrestrial environments (Vitousek and Howarth 1991, Kimmins 1997). Micronutrient supply

(Silvester 1989, Barron et al. 2009), as well as interactions between macronutrients and micronutrients (Wurzburger et al. 2012, Jean et al. 2013), limit productivity in some ecosystems. In primary successional environments, nutrient availability constrains ecosystem productivity (Vitousek 1999, Walker and del Moral 2009) and determines species composition (Clein and Schimel 1995, Cusick 2001) and successional trajectories (Walker and del Moral 2009).

Primary successional environments arise from natural events such as floods, volcanic eruptions, deglaciations, and fires, and anthropogenic exploits such as mining and deforestation. Following severe disturbances, ecosystems develop in biological complexity, beginning with plant and animal colonization of barren substrates that have little or no surviving biological legacy. Anthropogenic disturbances often exacerbate the effects of natural disturbances; with the increasingly broad-reaching and extensive environmental impacts of human activities, the line between anthropogenic and natural disturbances is often blurred (Walker and del Moral 2003).

As a result of ecosystem disturbances forming new surfaces, primary successional environments are often initially quite infertile and therefore inhospitable to plant growth

(Walker and del Moral 2003). Substrate organic matter content, and availability of associated nutrients N and P, are generally low (Pare et al. 1993, Huang et al. 2013, Zhang et al. 2015), and the accumulation of these nutrients determines the rate of plant recolonization. Biological N2 fixation, generally the principal source of N in such environments, is often limited by low levels of organic P and soil moisture (Walker and del Moral 2009).

Weathering of primary minerals in the substrate releases P to the soil solution (Ippolito et al.

2010). Mineral weathering is generally greatest in the earliest stages of soil formation (Egli et al. 2001); whether P and nutrient cations are made available through weathering depends highly on soil properties such as surface area and specific mineralogical composition

(Augustin et al. 2015). Because steep nutrient gradients often exist in primary successional substrates (Walker and del Moral 2009), spatial variations in nutrient availability may be dictated by small-scale, site-specific substrate characteristics.

1.2 Ecological Setting: The Dewatered Lake Aldwell Reservoir

Newly-dewatered reservoir sediments resulting from the removal of human-made dams are an environment where exposed surfaces may be inhospitable to natural revegetation. If woody plant establishment is inhibited by soil infertility, the accumulation of biomass is hindered and the process of soil formation slowed (Walker and del Moral 2003).

Studying nutrient availability and nutrient limitation in dewatered reservoirs facilitates greater understanding of ecosystem development following dam removal.

The recent removal of two large, long-standing dams on Washington State’s Elwha

River has exposed expansive regions of previously inundated reservoir sediments. Located on the northern coast of the Olympic Peninsula, the Elwha River drains an 833 km2 basin, the majority of which is in Olympic National Park (Duda et al. 2008). The Lake Aldwell reservoir was created in 1913 with the completion of the 33-m-tall Elwha Dam, located 8 km south of the river’s mouth (Duda et al. 2008). Over the following century this reservoir

became the depository for an estimated 4,900,000 m3 of sediment from the Elwha River as it slowed to a stop behind the dam (Warrick et al. 2015). In 1992 the U.S. Congress passed the

Elwha River Ecosystem and Fisheries Restoration Act, authorizing full restoration of the

Elwha River to its previous free-flowing conditions through dam removal and subsequent habitat restoration (Chenoweth et al. 2011). The Act also called for removal of the 64-m-tall

Glines Canyon Dam, located on the Lake Mills reservoir 14 km upriver from Lake Aldwell

(Duda et al. 2008).

The Elwha Ecosystem Restoration Project began in 2011, and the 2011-2012 removal of the Elwha Dam and concurrent drainage of the Lake Aldwell reservoir left behind a 1.5 km2 plain of lake-bottom sediment. Since complete removal of the Elwha Dam in 2012 and the Glines Canyon Dam in 2014, an essential component of ecosystem restoration in both dewatered reservoirs has been the reestablishment of riparian vegetation in order to support the return of anadromous fish stocks and other communities of native fauna. Revegetation of exposed areas has also been critically important in reducing erosion and sedimentation into the river and in minimizing the establishment of invasive exotic plants that could out- compete native species (Chenoweth et al. 2011).

1.3 Nutrients in the Elwha River Ecosystem

With low levels of nutrients, low moisture-retention capabilities, and spatial separation from intact forests, the dewatered reservoir sediments on the Elwha River were predicted to be relatively inhospitable environments for primary succession to occur. Prior to dam removal, submerged fine sediments in Lake Aldwell were low in K, Ca, and B relative to typical Olympic National Forest soils in the area, and those in Lake Mills were low in N,

P, and K (Chenoweth et al. 2011). At Lake Mills, concentrations of C, N, and P were higher in fine than in coarse sediments (Cavaliere and Homann 2012).

The dewatered reservoirs contain terraces with distinctly different sediment textures

(Randle et al. 2015). In the prodelta regions of the reservoirs, erosion and deposition during reservoir drawdown resulted in a layer of fine sediments deposited atop pre-existing coarse sediments (Warrick et al. 2015). Fine sediments contain more moisture than coarse sediments and have more surface area that can be weathered to release nutrients to the soil solution where they are accessible to plants (Brady and Weil 2009, Jones 2012). With varying moisture and nutrient profiles, different sediment textures within the reservoirs constitute distinct microenvironments for plant development. At dewatered Lake Aldwell, black cottonwood (Populus trichocarpa) and Scouler’s willow (Salix scouleriana) have colonized and become established in several sediment textural microenvironments. Differences between these microenvironments with respect to water and nutrient availability may have important implications for plant succession.

1.4 Study Objectives, Research Questions, and Hypotheses

The ability of plants to access and uptake nutrients may be an important factor in

Elwha River ecosystem recovery. The objective of this study is to increase the understanding of plant nutrients in the dewatered Lake Aldwell reservoir by evaluating naturally-established black cottonwood saplings. The study addresses the following research questions and hypotheses:

1) Do foliar nutrient concentrations vary among Lake Aldwell reservoir

microenvironments?

Hypothesis: There will be variation among microenvironments due to

differences in sediment texture. Specifically, there will be an inverse

relationship between foliar nutrient levels and substrate grain size. Because

the prodelta microenvironment is effectively a mixture of sediment textures,

prodelta foliar nutrient concentrations will be intermediate between those

measured in fine and coarse reservoir sediments.

2) To what extent are foliar nutrient concentrations correlated with physical and

biological environmental variables, in particular sediment moisture content and stem

density of woody plants?

Hypothesis: There will be positive correlations with sediment moisture

content because nutrient retention and moisture retention are influenced by the

same physical and chemical controls. There will be negative correlations with

stem density due to competition between plants.

3) To what extent are foliar nutrient concentrations correlated with one another?

Hypothesis: There will be positive correlations between nutrients that

participate in similar functions in plants, because plants require nutrients in

specified ratios and nutrient uptake is a regulated process.

4) How do the foliar nutrient concentrations in this study compare with those measured

in other populations of non-fertilized genus Populus species?

Hypothesis: There will be substantial variation between values from this study

and values from the literature because nutrient concentrations are influenced

by numerous environmental controls. In general, however, nutrient values

from this study will be lower than those in stands on other types of substrates,

where nutrient availability is higher.

5) How do the foliar nutrient concentrations in this study compare with concentrations

optimal for biomass production in fertilized genus Populus species?

Hypothesis: Many values from this study will be lower than those reported as

optimal in fertilized populations of Populus due to nutrient limitation.

2.0 Background

2.1 Plant Nutrition

2.1.1 The Importance of Plant Nutrition

Plant nutrition is the study of the functions of essential mineral elements, referred to as nutrients, in the physiology and biochemistry of plants. This includes exploration of the mechanisms through which plants acquire nutrients from the environment, as well as the plant-tissue distributions and metabolic functions of these nutrients (Epstein 1972, Marschner

1986).

Plant nutrition is important because of the ways in which nutrients influence plant growth, development, and reproduction. Nutrient limitation, which occurs in plants with insufficient nutrient supply, has measurable implications for plant productivity in both the agricultural (Marschner 1986) and ecological (Vitousek 1999, Barron et al. 2009, Jean et al.

2013) realms. Nutrient deficiencies are important to detect not only because they can result in stunted plant growth but because they can make plants more susceptible to disease and pest infestation, putting them at a reproductive disadvantage (Garten 1978).

To support metabolic processes and synthesis of tissue, plants require nutrients in pre- determined ratios (Knecht and Goransson 2004, Ranade-Malvi 2011, Liu et al. 2014).

Relationships between nutrient concentrations within plant tissue affect measurable plant characteristics such as sap sweetness (Wild and Yanai 2015) and biomass productivity

(Leech and Kim 1981, Richards et al. 2010), which influence a plant’s commercial value.

Inter-nutrient relationships within plant tissue may be especially important in determining a plant’s reproductive success (Ranade-Malvi 2011) and may even provide an indication of

overall ecosystem function (Wan et al. 2009). Woody plants with imbalanced nutrient ratios may be more susceptible to parasitic attacks (Flückiger and Braun 2003, Ranade-Malvi

2011), and plants whose nutrient ratios fall outside of specific ranges may exhibit symptoms of nutritional disorders (Teng and Timmer 1993).

2.1.2 The Essential Mineral Elements

In addition to C, O, and H, which account for as much as 95% of plant dry matter

(Marschner 1995), 14 mineral elements have been determined to be essential for vascular plants to complete their lifecycles (Marschner 2012). It is informative to classify the plant nutrients into subgroups. Several nutrient classification criteria are discussed here, as they are informative for understanding the importance of plant nutrient studies. Common criteria that are considered when classifying plant nutrients into subgroups are plant-tissue concentrations, biochemical functions, and mobility within plants.

2.1.3 Nutrient Concentrations in Plant Tissue

The simplest way to order and classify plant nutrients is by the relative average levels at which they are found in the tissue of healthy, productive plants. Macronutrient concentrations are generally on the order of tens to thousands of times higher than micronutrient concentrations. Nitrogen, P, K, Ca, Mg, and S are macronutrients and Cl, B,

Fe, Mn, Zn, Cu, Ni, and Mo are micronutrients. Nutrient concentrations vary widely between plant species (Hagen-Thorn et al. 2004b) and between different types of plant tissue (Langille and Maclean 1976, Shelton et al. 1981, Hagen-Thorn and Stjernquist 2005), and depending on plant age (Francis and Baker 1981), season (Tew 1970, Verry and Timmons 1977), and concentrations of other nutrients present (Coleman et al. 2006, Marschner 2012). Plant-tissue

nutrient concentrations are constrained relatively tightly by physiological plant processes that regulate nutrient uptake (Larcher 1975, Marschner 2012). Even if given access to the same nutrient resources, however, plant species may uptake nutrients in different quantities, which results in relatively wide variation in the concentrations of given nutrients even between closely related species (Garten 1978).

Patterns in nutrient variability at the species level may be explained by environmental variables. In a study in which foliar nutrient data was compiled from over 700 terrestrial plant species from across China, Zhang et al. (2012) classified nutrients based on patterns in their concentrations related to plant taxonomic group and related to the environmental variables precipitation, climate, and latitude. Foliar N, P, K, and Fe were correlated with precipitation, and these nutrients, as well as Mn and Ca, also seemed to be strongly associated with latitude (Zhang et al. 2012).

The majority of forest nutrient studies have focused on the macronutrients, possibly because in trees growing under normal conditions in temperate forest ecosystems, primary production is generally limited by N and P (Vitousek and Howarth 1991, Kimmins 1997,

Seaman et al. 2015), and micronutrient deficiencies are not very common in forests (Fisher et al. 2000). Nevertheless, the importance of micronutrients in forest ecosystems has been recognized and studied in many terrestrial environments (Stone 1990), including temperate forests (Silvester 1989, Hagen-Thorn and Stjernquist 2005, Jean et al. 2013) and tropical forests (Drechsel and Zech 1991, Barron et al. 2009, Wurzburger et al. 2012). The essential micronutrients are all commonly deficient in United States crops (Berger 1962) and much of

what is known about their functions in plants is a result of agricultural research (Marschner

1986).

2.1.4 Functions of Nutrients in Plants

Macronutrients and micronutrients mainly serve different functions in plants, although in some cases there is overlap between the two groups. The macronutrients can be classified generally as structural elements, which are incorporated into organic compounds

(N, P, and S) or cell walls (Ca), and as osmotica, which regulate osmotic processes (K, Ca, and Mg). Both N and S are involved in protein synthesis and become incorporated into amino acids which participate in enzymatic processes within the plant. Phosphorus, in addition to serving these functions, is involved in energy transfer reactions in the plant cell (Jones 2012).

Sulfur, as a structural component of organic compounds, is involved in protein synthesis and metabolic reactions (Marschner 1986).

Of the osmotica, K is required for maintaining proper plant hydration and for storage and subsequent translocation of carbohydrates within plant tissue (Jones 2012). Present primarily in plant leaves and bark, Ca is an important structural component of cell walls

(Larcher 1995) and also regulates cell membrane permeability, serves in the activation of several enzymes, and plays a role in carbohydrate translocation (Jones 2012). Magnesium, a component of chlorophyll, is found mainly in plant leaves and plays an important role in photosynthesis (Larcher 1995) and is involved in protein synthesis and other enzymatic processes as well (Jones 2012).

The micronutrients are present primarily as part of enzyme molecules, facilitating electron transport and enabling protein synthesis and a vast array of other cellular processes

and chemical reactions to occur. Specifically, Cl, Cu, Mn, Fe, and Zn are important for photosynthetic processes, and Mo functions as a cofactor to the nitrogenase and nitrate reductase enzymes which catalyze plant assimilation of N (Jones 2012). In addition to being involved in photosynthesis, Cl is an osmoticum, increasing plant-tissue hydration and raising osmotic pressure within the plant cell (Jones 2012). Carbohydrate metabolism involves Cu and B, and B and Fe are crucial for cellular respiration. Nickel, which is essential for some, but not all, plants, is a component of the urease enzyme (Jones 2012).

Nutrients generally serve similar roles across plant species. Because of this, several studies have suggested a physiological basis for species-level correlation patterns between nutrients. Garten (1978) analyzed patterns that emerged from inter-nutrient relationships in data compiled from over 100 species of both terrestrial and aquatic plants. Nutrients were well categorized into distinct sets which seemed to be related to their physiological functions in plants. Nutrients involved in protein synthesis (N, P, S, Fe, and Cu) formed one set, those that serve structural roles in the plant cell (Ca, Mg, and Mn) formed another set, and those important in the activation of enzymes (K, Mg, and Mn) formed a third. The remaining nutrient considered, Zn, was grouped by itself into a fourth set, possibly related to plant growth (Garten 1978). In a similar study of plant species across China, nutrients were generally grouped together into the same sets (Zhang et al. 2012).

2.1.5 Nutrient Mobility and Translocation within Plants

The nutrient elements exist in the soil mainly in bound forms, complexed by insoluble organic compounds, organic detritus, and humus, or structurally incorporated into minerals and organic compounds themselves (Larcher 1975). A small portion of soil nutrients exist as

dissolved ions in the soil solution, and it is this fraction that plants access (Jones 2012,

Maathuis 2013). The ability to transport these ions against concentration gradients enables plants to accumulate nutrients from relatively dilute soil solutions (Larcher 1975).

Plants tend to uptake cations preferentially over anions, and rates of plant nutrient uptake from the soil solution tend to be highly regulated (Larcher 1975, Reuter et al. 1997).

Nutrient uptake depends on many complex factors including the physical and chemical properties of the element of interest and of the growth substrate (Jones 2012, Marschner

2012, Maathuis 2013), plant-tissue concentrations of both the element of interest and other ionic species, and mechanisms of ionic transport within the plant (Larcher 1975, Marschner

2012).

The rate-limiting step of nutrient uptake is the conduction of ions from the soil solution through the root symplasts (Larcher 1975). Once within the plant, nutrients are transported with varying degrees of ease to the parts of the plant where they are most needed at any given stage in development. Specifically, nutrients are transported toward the leaf through the xylem and away from the leaf through the phloem (Larcher 1975).

Nutrient concentrations are generally highest in the most metabolically active tissue.

This is usually the plant leaf, which has a high proportion of living cells (Shelton et al. 1981).

In a given type of plant tissue, nutrient concentrations vary seasonally (Verry and Timmons

1977), with plant age (Francis and Baker 1981, Shelton et al. 1981, Martín-García et al.

2012), and with plant-growth stage (Jones 2012). Redistribution of nutrients within plant tissue over time, also referred to as nutrient translocation, is a highly-regulated process

(Reuter et al. 1997).

Nutrient mobility refers to the relative ease with which nutrients are transported within the plant (Reuter et al. 1997). Based on temporal measurements of changes in plant- tissue nutrient concentrations, the nutrients have been classified into groups according to this criterion. One general classification categorizes elements as “very mobile” (N, P, K, and

Mg), “slightly mobile” (S), “immobile” (Fe, Zn, Cu, and Mo), and “very immobile” (B and

Ca) (Jones 2012). Nutrients are also commonly classified by their mobility within the phloem specifically, that is, away from the leaf. In one such classification scheme, nutrients are categorized as “mobile” (N, P, K, Mg, S, and Cl), “intermediate” (Fe, Zn, Cu, B, and Mo), or

“immobile” (Ca and Mn) (Marschner 2012), and in another as “phloem-mobile” (N, K, Mg, and Cu) or “phloem-immobile” (Ca, Fe, Mn, and Zn) (McLennan 1990). Another classification describes nutrients as “phloem-mobile” (N, P, and K), “phloem-immobile” (Ca,

Fe, and Mn), and “variably phloem-mobile” (S, Zn, and Cu), and whether or not B and Mg are phloem-mobile is determined entirely by environmental conditions (Reuter et al. 1997).

Although these classification systems group the nutrients slightly differently, the general trend in within-plant nutrient mobility is that the macronutrients are more mobile than the micronutrients (Marschner 2012). Specifically, N, P, and S in organic form, as well as K, are generally mobile, while the micronutrients are less so. A major exception is Ca, which tends not to be translocated once incorporated into the plant cell wall (Larcher 1975,

Reuter et al. 1997).

Differences in translocation mobility can be manifested in the distributional patterns of nutrients within trees. In an investigation of foliar nutrient variability within a specific population of black cottonwood, McClennan (1990) classified nutrients based on their foliar

concentrations in various positions in the canopy. Three sets of nutrients emerged: phloem- mobile elements (N, P, K, S, and Cu) were most concentrated in upper-canopy foliage, elements that tend not to be translocated (Ca, Mn, and to some extent Zn) were highest in lower-canopy foliage, and Fe had a unique foliar distribution pattern and was grouped by itself.

2.1.6 Leaf Growth, Senescence, and Nutrient Resorption

During the growing season, nutrients are transported into young, unfolding plant leaves, where they serve various functions in photosynthetic processes and are incorporated into the foliar tissue itself. As a plant grows and stores energy, nutrients accumulate in the foliage, which is the end of the xylem translocation route (Larcher 1975). As the growing season draws to a close and the plant goes into a growth stage known as senescence, large amounts of N and P (Marschner 2012), as well as other mobile nutrients, are remobilized from foliar tissue and transported into woody tissue before abscission of foliage (Baker and

Blackmon 1977, Reuter et al. 1997). Although phloem-immobile elements are not well reabsorbed from the leaf, some of these nutrients are lost in guttation fluid and leached out of leaves by rain (Reuter et al. 1997). Leaching of K, which tends to occur in young leaves, can result in substantial losses, while smaller quantities of Ca tend to be leached from old leaves in autumn (Sviridova 1960).

Plants reabsorb different nutrients to different extents. In some instances, plants can conserve more than half of their foliar nutrients for future growth by shunting them away from their foliage before abscission. Internal recycling of nutrients may enable 20-yr-old loblolly pine (Pinus taeda) to meet 20 – 30% of its N, P, and K requirements, 16% of Mg and

S requirements, and 7% of Ca requirements. In 5- to 10-yr-old red alder (Alnus rubra) and

Monterey pine (Pinus radiata), recycling of N, P, and K may contribute 50 – 60% of plant requirements of these nutrients (Reuter et al. 1997).

In Populus, reported resorption rates vary. One study, which found large inter-annual variability in resorption of N and P, reported N, P and Mg resorption rates of 29 – 33%, 33 –

38%, and 3 – 48%, respectively (Francis and Baker 1981). In another study, respective resorption rates of N, P, and Cu were 24 – 56%, 38 – 64%, and 6 – 13% and these nutrients were all better reabsorbed from late-abscising than from early-abscising leaves (Killingbeck et al. 1990). In that study, significant accumulation of Ca, Zn, and Fe was measured in senesced leaves. In a study conducted at natural Populus-dominated forest stands, N, P, and

K resorption rates also varied throughout the process of senescence, the observed effect being largest for N (Niinemets and Tamm 2005).

In general, the net outcome of senescence is that the foliar concentrations of the mobile elements, in particular N, P, and K, decrease (Niinemets and Tamm 2005, Liu et al.

2014), while the less mobile elements, especially Ca and Mn, accumulate in the foliar tissue

(Killingbeck et al. 1990, Marschner 2012). Resorption and accumulation trends for Mg, B,

Fe, Zn, and Cu are more variable, as the mobility of these nutrients is more strongly influenced by environmental controls (Baker and Blackmon 1977, Francis and Baker 1981,

Killingbeck et al. 1990, Liu et al. 2014).

Nutrient resorption is an important conservation mechanism (Baker and Blackmon

1977, Killingbeck 1996) and may be especially pronounced in plants growing in nutrient- poor soils (Seaman et al. 2015). It has been shown that plants reabsorb N at higher rates in N-

limited forests (Wei et al. 2015). The degree to which plants have access to specific nutrients in the substrate also influences the timing of leaf senescence and the nutrient redistribution that occurs (Liu et al. 2014). By upregulating the production of plant hormones that promote senescence, plant deficiencies of both N and K can accelerate the process (Marschner 2012).

In combination with high light exposure, deficiencies of Zn, Mg, and K have been shown to cause stress-induced senescence through accumulation of photosynthate in the leaf

(Marschner and Cakmak 1989).

In the context of agriculture, changes in the timing of leaf senescence have several implications for crop quality. Accelerated senescence may mean a shorter growing season, which limits plant growth and can result in crop spoilage and loss of mobile nutrients from plant tissue. Postponement of leaf senescence, which can result in higher concentrations of micronutrients such as Fe, Mn, and Zn, is often desirable (Diestelfeld et al. 2007).

Plant nutrient levels tend to be quite variable throughout the season and particularly during leaf senescence; an understanding of nutrient remobilization patterns, and their implications for plant growth, enables the use of plant-tissue analyses for plant nutritional assessment and deficiency diagnosis (Reuter et al. 1997). Ordering and classifying the nutrients based on mobility characteristics is useful not only for explaining observed changes in foliar nutrient concentrations over the course of a plant’s lifecycle but also for predicting where mineral deficiency symptoms will first occur within a plant. Deficiency symptoms

(and, theoretically, measurably-lower nutrient concentrations) due to inadequate supply of immobile nutrients generally occur in a plant’s newest leaves. Inadequate supply of more mobile nutrients manifests itself in a plant’s oldest leaves (Reuter et al. 1997, Jones 2012).

2.1.7 Nutrient Interactions within Plants

When plants absorb and utilize nutrients differently depending on the quantity of other nutrients present, the nutrients are said to interact. Often complex, interactions are influenced by numerous factors, including climate, substrate parameters such as moisture, nutrient content, and pH, and plant parameters such as age and species (Fageria and Baligar

1999, Fageria 2001, Ranade-Malvi 2011, Marschner 2012).

Positive nutrient interactions occur when increased supply of one nutrient results in increased absorption of another (Fageria 2001). Such interactions are generally manifested as positive correlations between nutrient concentrations within plant tissue. Plant growth is highly dependent upon N availability, so increased N supply can increase a plant’s growth rate and requirements for other nutrients. Nitrogen fertilization has a particularly strong positive effect on P absorption and can also cause increased uptake of Fe, Cu, and Mn, but it does not appear to influence plant uptake of Zn (Fageria 2001).

Negative nutrient interactions occur when nutrient ions with similar chemical properties compete for plant absorption, for transport within plant tissue, or for use in specific plant physiological functions (Fageria 2001, Ranade-Malvi 2011). Negative interactions are common between Ca and Mg, which form cations that have similar chemical charge and electronic configuration (Jones 2012). Such interactions also occur among Cu,

Mn, and Zn (Fageria 2001).

When a particular cation is present in excess in the soil solution, the rate at which a plant uptakes other cations is generally decreased (Epstein 1972). Not only are plant-tissue cation concentrations thus dependent on the availability of particular elements in the growth

medium, but they depend on the presence or absence, and relative concentrations, of other ions (Ranade-Malvi 2011). This has important implications for the validity of plant nutritional assessments, because plant-tissue nutrient concentrations may become elevated far beyond adequacy levels (Fageria 2001). This dynamic relationship between nutrients also means that the sum of cations in plant tissue generally remains constant even though their relative proportions within the tissue may vary as a function of their relative concentrations in the growth medium (Larcher 1975, Ranade-Malvi 2011). While cation-cation and anion- anion interactions are generally understood to be competitive in nature, the basis for cation- anion interactions is less clear (Fageria 2001).

Specific examples of nutrient interactions abound. Increased soil Ca can decrease plant-shoot Zn, Mn, and Fe concentrations (Fageria and Baligar 1999), and increased Cu concentration in nutrient solution can decrease Zn absorption (Brar and Sekhon 1976).

Because P deficiency is a common yield-limiting factor in agriculture (Fageria et al. 1997), the relationship between P and other nutrients, especially Zn, has been investigated intensively (Jones 2012). In Populus, P fertilization experiments have demonstrated P- induced inhibition of Zn and Cu uptake without a measureable change in the soil-extractable concentrations of either micronutrient. Phosphorus-induced Zn deficiency symptoms may be due to augmented physiological requirements for Zn under high P regimes. Zinc deficiency stunts plant growth, which can result in elevated plant-tissue concentrations of N, Ca, and

Mg (Teng and Timmer 1993).

Nutrient interactions have been relatively well studied in the agricultural realm

(Fageria 2001), and much of what is known is a result of such research. Agricultural studies

are generally carried out under controlled field conditions, where experimental parameters such as nutrient inputs and water availability can be manipulated and closely regulated.

Interpretation of interactions, however, is still complicated by variability in the concentrations of nutrients considered sufficient for adequate growth (Mellert and Goettlein

2012) and the fact that plants respond to nutrient additions differently depending on whether they are growing in monocultures or mixtures (Wan et al. 2009, Richards et al. 2010).

In natural, more ecologically-complex environments such as forests, the results of field experiments often do not apply (Aerts and Chapin 2000). Because of the difficulties in studying nutrient interactions and the difficulties inherent in performing experiments on large, long-lived trees (Bradshaw et al. 2000), there is a lack of understanding regarding nutrient interactions in broadleaf trees in forest ecosystems.

2.2 Assessment of Plant Nutrition

2.2.1 Plant-Tissue Nutrient Concentrations

A plant’s nutritional status is an indication of its potential to produce biomass and of its susceptibility to become diseased. Nutritional status is assessed by measuring the nutrient composition of a plant’s tissue and comparing the values to those measured in healthy plants.

Foliar nutrient concentrations, that is, nutrient concentrations in a plant’s leaves, tend to provide the best indication of a plant’s nutritional status (Marschner 1995). Leaves are therefore the standard plant tissue used in plant nutritional assessment, especially with respect to micronutrients, which are generally highest in the leaf (Reuter et al. 1997).

Compared to other types of plant tissue such as stemwood and bark, foliar tissue has been shown to most closely reflect the chemical properties of the growth substrate (Shelton et al.

1981). Foliar nutrient concentrations are strongly influenced by environmental contamination

(Pietrzykowski et al. 2013, Ciadamidaro et al. 2014) and nutrient amendment (Leech and

Kim 1981, Coleman et al. 2006, Elferjani et al. 2013).

Previous work indicates substantial variability in foliar nutrient concentrations between coniferous and deciduous trees (Langille and Maclean 1976, Hagen-Thorn et al.

2004b, Wan et al. 2009, Liu et al. 2014) as well as among deciduous tree species (Radwan and de Bell 1988, Cusick 2001, Hagen-Thorn and Stjernquist 2005, Hagen-Thorn et al.

2006). Foliar nutrients may vary significantly within individual populations of conspecific trees, among leaves from different parts of the same trees (Verry and Timmons 1977,

McClennan 1990, Ågren and Weih 2012), and even between different regions of single leaves (Marschner 2012). In particular, foliar concentrations of some nutrients can vary depending on position in the canopy (Verry and Timmons 1977), likely as a function of competition for sunlight exposure (Francis and Baker 1981, McLennan 1990). Some past

Populus studies have reported nutrient values for leaves sampled from specific canopy strata

(Tew 1970, Coleman et al. 2006, René et al. 2013, Salehi et al. 2013), but the majority of studies either report nutrient values for composite samples of leaves from all canopy positions (Leech and Kim 1981, Shelton et al. 1981) or do not specify the canopy positions from which leaves were sampled.

In Populus it is well established that foliar nutrient concentrations vary significantly depending on the time of year in which sampling takes place (Verry and Timmons 1977); differences are particularly apparent when comparing the nutrient concentrations of presenescent and senesced leaves (Killingbeck et al. 1990, Salehi et al. 2013). Sampling

season differences result in a high degree of variation in literature values even within genus

Populus plants, which complicates nutritional assessment.

2.2.2 Plant-Tissue Nutrient Ratios

Optimal nutrient ratios in plants vary between dissimilar species based on differences in plant physiology and between different populations of similar species based on environmental factors (Ranade-Malvi 2011). Despite this variability, previous studies have shown that foliar nutrient ratios, when utilized appropriately, provide a sensitive indication of plant nutritional status because they provide a measure of nutrient balance (Flückiger and

Braun 2003). This is especially true for ratios between foliar N, P, and K, which have a strong physiological basis (Bailey et al. 1997b). Physiological explanations for ratios between foliar Ca and Mg are less clear, as their uptake is largely dictated by the cationic composition of the soil solution (Larcher 1975) and is not as tightly regulated (Bailey et al.

1997b).

Because Ca and Mg do not exhibit the same seasonal patterns of change as do N, P, and K (Verry and Timmons 1977, Niinemets and Tamm 2005), ratios between Ca and Mg and these other macronutrients vary throughout the season. Foliar Ca/Mg ratios can serve as internal references, however, as they may indicate the degree to which environmental factors, as opposed to nutrient availability, limit plant growth. When growth is limited by multiple nutrients, but when absolute nutrient concentrations are such that nutrient ratios still indicate a state of nutritional balance, it can be useful to consider Ca/Mg ratios in conjunction with ratios between N, P, and K to avoid overlooking nutrient deficiencies (Bailey et al. 1997b).

Although micronutrients are equally important in plant physiological processes,

micronutrient concentrations are much more variable within plants and therefore do not provide a sensitive measure of nutritional status (Ranade-Malvi 2011, Zhang et al. 2012).

2.2.3 Relation of Plant Nutrition to Substrate Nutrient Concentrations

Sediments and soils (collectively referred to as substrates) are composed of particles that can be classified according to grain size along a textural continuum of cobbles, gravels, sands, silts, and clays. With more surface area and interstitial space, finer substrates (those with higher clay and silt content) generally have higher capacities to adsorb and retain nutrient ions on their surfaces and release them to soil solution (Brady and Weil 2009). Finer substrates also generally contain more organic matter (Brady and Weil 2009), which can aid in increasing moisture-holding capacity (Jones 2012). Nutrient and moisture levels in coarser substrates, those composed of cobbles, gravels, and sands, are generally lower (Brady and

Weil 2009).

Plant nutrient availability, or bioavailability, is a measure of the extent to which mineral elements in the growth substrate are accessible to plants through root uptake.

Generally greater in finer sediments (Brady and Weil 2009), nutrient bioavailability is increased in the presence of organic matter, which provides additional exchange sites for nutrient cations (Wichard et al. 2009).

Nutrient bioavailability is often estimated through direct extraction of nutrients from soils and sediments (Marschner 2012) using various chemical agents and extraction methods

(Cusick 2001, Singh and Sharma 2007, Reed et al. 2008, Wichard et al. 2009, Wurzburger et al. 2012, Ciadamidaro et al. 2013, Li et al. 2014). Past studies have explored the relationship between substrate and plant-tissue nutrient levels (Shelton et al. 1981, Hagen-Thorn et al.

2004a, Martín-García et al. 2012, Maisto et al. 2013, Ciadamidaro et al. 2013, Salehi et al.

2013), investigating changes as a function of both substrate and foliar nutrient amendments

(Leech and Kim 1981, Teng and Timmer 1993, van den Driessche 2000, Coleman et al.

2006, Valenciano et al. 2011).

In some cases, sediment-extractable nutrient concentrations have been shown to be indicative of nutrient bioavailability (Singh and Sharma 2007), especially when crop response to nutrient amendment is considered (Houba et al. 1996). In some cases, however, sediment nutrient composition is only weakly correlated with plant nutrient composition

(Garten 1978) and may provide misleading information regarding the actual levels of nutrients that are available for plant uptake (Bailey et al. 1997c).

Plant roots can only take up nutrients in dissolved form, and it is difficult to account for all of the complex biogeochemical interactions that occur between metal ions and any specific growth substrate (Goldberg et al. 2008, Wichard et al. 2009, Jones 2012, Marschner

2012). For example, ectomycorrhizal fungi and arbuscular mycorrhizal fungi, which both associate with black cottonwood (Cusick 2001, Karliński et al. 2010), can influence plant nutrient uptake by facilitating nutrient movement within the substrate (Cusick 2001). In soils contaminated with Cu and Pb from Cu smelting, ectomycorrhizal fungal inoculation of

Populus increased plant-tissue concentrations of both Cu and Pb (Bojarczuk et al. 2015). In metal-contaminated, macronutrient-deficient mine sites, Acacia and Eucalyptus inoculated with ectomycorrhizal fungi accumulated fewer toxic metal ions and more macronutrients

(Jourand et al. 2014).

There is no simple, universal means of predicting plant nutrient uptake and utilization

(Garten 1978) or plant growth response (Jones 2012) based on substrate nutrient levels. The accuracy of analyses that rely on measurements of substrate nutrient composition to assess nutrient bioavailability is therefore inherently limited. Methods that directly measure plant- tissue nutrient composition to assess nutrient bioavailability in the substrate (Maathuis 2013) and to assess plant nutritional status with respect to these elements (Marschner 2012) have been shown to be more reliable than methods based solely on substrate analyses (Richards and Bevege 1972, Beaufils 1973, Binkley 1986).

2.2.4 Standardized Approaches

Variability in foliar nutrient levels complicates plant nutrient assessment and has prompted the standardization of foliar sampling procedures (Luyssaert et al. 2002) and the development of several methods that attempt to standardize interpretations of foliar nutrient values. Among these methods are analyses based on critical values (Jones 2012), critical deficiency concentrations (Marschner 2012), and standard values and sufficiency ranges

(Jones 2012). In some methods, sufficiency concentrations are determined separately for individual nutrients, and nutrient ratio sufficiency ranges are then determined based on these values. Even within a single plant species there is considerable variability in the ratio ranges deemed adequate for sufficient plant growth, so nutrient ratios calculated from nutrient concentrations measured in isolation do not necessarily provide an accurate measure of nutrient balance (Flückiger and Braun 2003).

The diagnosis and recommendation integrated system (DRIS) uses optimal plant- tissue nutrient ratios for plant nutritional assessment (Beaufils 1973). Based on relative

relationships between nutrients, the DRIS is a more reliable approach than methods based on absolute nutrient concentrations (Hallmark and Beverly 1991, Bailey et al. 1997b). Nutrient ratios are somewhat less influenced by plant aging processes, growth dilution, and changes in concentrations of nonstructural carbohydrates (Fageria 2001, Flückiger and Braun 2003,

Marschner 2012, Elferjani et al. 2013). Ratios between some of the nutrients are also less temporally variable over the growing season, remaining relatively constant even as individual nutrient concentrations change (Verry and Timmons 1977). For the diagnosis of P, K and S deficiencies, the DRIS has specifically been shown to be superior to the critical value approach (Bailey et al. 1997a).

In the DRIS, nutrient ratio norms are established for each specific plant species by measuring highly-productive “reference population” plants grown under field conditions

(Filho 2004, Hundal 2007, Maia 2012) with fertilizer amendments (Leech and Kim 1981).

These reference values are established based on foliar nutrient concentrations measured during the summer (generally July through September) (Leech and Kim 1981). In order to avoid variability due to nutrient resorption at the end of the growing season (Hansen 1994),

DRIS analyses are best completed using values from foliage sampled during July and

August, when nutrient concentrations are the most stable (Verry and Timmons 1977,

Coleman et al. 2006).

The assumption in the DRIS is that the concentrations of all nutrients, and therefore all nutrient ratios, are optimal in plants that yield the most biomass (Campion and Scholes

2007, Elferjani et al. 2013). Calculation of standardized DRIS nutrient indices is based on these optimal nutrient ratios (Filho 2004, Campion and Scholes 2007). Nutrient ratios

measured in individual plants or populations of plants are compared against these standards to guide nutrient amendment recommendations (Leech and Kim 1981, Elferjani et al. 2013) and to detect nutrient deficiencies (Beaufils 1971, Walworth and Sumner 1987, Filho 2005,

Serra et al. 2013). Index values near zero reflect nutritional balance, while nutrient excesses and deficiencies are indicated by positive and negative indices, respectively (Walworth and

Sumner 1987.

Nutritional assessments using the DRIS rely on large sample sizes for which reliable coefficients of variation can be calculated in order to determine nutrient indices (Bailey et al.

1997b, Filho 2004, Elferjani et al. 2013). Because of this, the DRIS is not always judged to be the best method for analysis of plant nutrition. In some cases, foliar nutrient ratios are used in conjunction with other analyses. In their nutritional assessment of hybrid poplar, in which they had sampled a relatively small number of tree stands, Martín-García et al. (2012) opted to assess foliar nutrient balance using tree growth measurements, analysis of nutrient ratios and correlations between substrate and foliage, and observations of deficiency symptoms, instead of employing the quantitative DRIS approach.

While some early studies did extend the DRIS to micronutrient assessment (Righetti et al. 1988, Hallmark and Beverly 1991, Rathfon and Burger 1991), it was only within the last decade that the method became more widely implemented for this purpose.

Micronutrient indices have now been reported for several plant species (Filho 2005,

Nachtigall and Dechen 2007a, Srivastava and Singh 2008, Singh et al. 2012, Raghupathi et al. 2013), but DRIS assessments using these values are difficult and often inaccurate. Plant- tissue micronutrient concentrations are highly variable (Martín-García et al. 2012, Zhang et

al. 2012) and their direct relationships to the concentrations of other nutrients often lack a physiological basis (Ranade-Malvi 2011).

2.2.5 Populus and the DRIS

Many industrial-scale poplar (genus Populus) plantations have been established over the past several decades in the Pacific Northwest, where climatic and soil conditions are especially suitable for plant growth (Brown and van den Driessche 2002). Of high commercial value, poplar also have high growth rates and high nutrient demands (Bradshaw et al. 2000, Lteif et al. 2008). Poplar nutrition and productivity, especially of hybrid varieties

(Dunlap and Stettler 1998, Brown and van den Driessche 2002, Martín-García et al. 2012), has been the subject of many studies in both planted and naturally-established settings

(Blackmon et al. 1979, Heilman 1985, Coleman et al. 2006, Sayyad et al. 2006, Singh and

Sharma 2007, Lteif et al. 2008, René et al. 2013).

Macronutrient norms for the DRIS have been established for some species of poplar

(Leech and Kim 1981, René et al. 2013), and several studies have evaluated poplar macronutrition using the DRIS approach (Coleman et al. 2006, Guillemette and DesRochers

2008, Lteif et al. 2008, Elferjani et al. 2013, René et al. 2013). No DRIS micronutrient assessments of poplar have yet been published.

2.3 Primary Succession and the Dewatered Lake Aldwell Reservoir

2.3.1 Principles of Primary Succession

Primary succession is the process by which ecosystems develop in biological complexity following severe disturbances (Walker and del Moral 2003, Wali 1999). Primary succession occurs on floodplains (Walker and del Moral 2003), lava flows (Vitousek 1999),

tephra deposits following volcanic eruptions (del Moral and Wood 1993), recently- deglaciated surfaces (Cusick 2001, Pérez et al. 2014), and reclaimed mining spoils (Wali

1999, Groninger 2007), among many other types of substrates.

Successional trajectories differ between environments based on the type and severity of disturbance, but the mechanisms of primary succession are similar. Newly formed or newly uncovered surfaces, with little or no surviving biological legacy, are initially colonized by mosses and lichens. These initial colonizers modify the habitat, facilitating the colonization and establishment of plant and animal species (Wali 1999). Over time, interspecific competition for resources and patterns of resource supply change. Varying resource-supply trajectories (Barbour et al. 1999) result in the development of complex biotic communities, which generally become increasingly resilient (Wali 1999, Walker and del

Moral 2003).

Succession occurs over wide temporal scales, the measurement of which depends on the species involved. For example, succession of grasses may occur over decades, while tree species change over the timescale of centuries (Walker and del Moral 2003). The speed at which succession proceeds is partly dictated by the size of the recovering area, and plant recolonization is generally controlled by water availability and soil fertility (Walker and del

Moral 2003, Michel et al. 2011). The structure of plant and animal communities is determined by the species present and by their means of dispersal and propagation (Barbour et al. 1999, Groninger 2007). Proximity to intact forests and means of seed dispersal and transport are of particular importance in determining rates of revegetation (Chenoweth et al.

2011, Michel et al. 2011).

2.3.2 Primary Succession in the Elwha River Ecosystem

Dewatered reservoirs of the Elwha River share attributes with primary successional environments. Barren glacial outwash plains in the wake of the retreating Exit Glacier in

Alaska are frequently flooded and are exposed to strong wind action. Near these barren regions, raised terraces comprised of coarse gravels contain scattered seedlings of naturally- established cottonwood and willow (Cusick 2001). These conditions are similar to those found at Lake Aldwell, a floodplain environment lacking mature woody vegetation to provide protection from the wind.

In dewatered environments, successional trajectories are strongly influenced by local geomorphological characteristics, and rates of ecosystem change are highly variable (Doyle et al. 2005, Lafrenz et al. 2013). On the Elwha River, phased reservoir drawdown (Warrick et al. 2015) resulted in prime conditions for the establishment of black cottonwood and

Scouler’s willow, as substrate water availability was high during the period of seed dispersal.

These species have naturally colonized many sites within the dewatered reservoirs. Red alder

(Alnus rubra), a potentially important component of ecosystem recovery due to its N2-fixing capacity, has also established at reservoir sites near the forest border. Both cottonwood and alder disperse their seeds in the wind, but cottonwood seeds, being lighter, are easily carried aloft for several hundred meters (Braatne et al. 1996, Naiman et al. 2010) while wind- dispersed alder seeds are constrained to within about 100 m of mature stands (Harrington

2006).

2.3.3 Substrate and Nutrients in Primary Successional Environments

Previous studies have assessed nutrient availability in primary successional environments. Early in ecosystem development, plant growth is usually limited by low levels of N in the substrate. Over time, accumulation of N and weathering of mineral P brings the two nutrients into balance. In older substrates, as P is lost through mineral weathering, P eventually becomes limiting (Walker and Syers 1976, Vitousek 1999). This pattern has been shown on young volcanic substrates in Hawaii (Vitousek and Harrington 1997, Vitousek

1999). Although plant growth at these sites was limited by N availability, P was found to be relatively more limiting to plant growth in coarser-textured substrates, where less surface area was accessible to weathering (Vitousek 1999).

In postglacial succession in southern Chile, total substrate concentrations of C and N increased over time in a 400-year soil chronosequence, as did foliar C/N and C/P ratios, but foliar N/P ratios stayed constant. Biological N2 fixation was at a maximum in the earliest stages of the chronosequence, where a symbiotic N2-fixing herb was present. Over the chronosequence, rates of N2 fixation decreased (Pérez et al. 2014).

Recent studies have investigated sediment chemistry specifically following dam removal. In post-dam-removal riparian environments, reservoir sediment characteristics are highly variable in general but especially variable with respect to levels of organic matter

(Rohdy 2013). One year after removal of a 10-m-tall dam in Wisconsin, concentrations of extractable K, and total P, K, Ca, and S, were higher in the dewatered sediments, while concentrations of extractable P, as well as total Cu and Mn, were higher in soils buried under the sediments (Wells et al. 2008). A study of dewatered substrate characteristics following

removal of a 14-m-tall dam in Oregon indicated that delayed soil development may inhibit rates of accumulation of plant-available N and Fe (Lafrenz et al. 2013).

The substrate characteristics in primary successional environments can influence plant nutrition. In reclaimed mining sites, foliar macronutrient concentrations in naturally- established trees have been shown to be affected more by physical and chemical substrate characteristics than by mineral fertilization (Pietrzykowski et al. 2013). Smaller-scale environmental characteristics, such as sediment texture and water availability, influence both vegetation structure and soil development and are highly dependent on local site conditions

(Cusick 2001, Burga et al. 2010).

2.3.4 Substrate and Nutrients in Elwha River Dewatered Reservoirs

The basin drained by the Elwha River is composed primarily of sandstone and shale from marine sedimentary deposits in the river’s upper reaches and of glacial and alluvial deposits closer to the river’s mouth (Duda et al. 2008). The river valley bottom is comprised of sandy gravels, while soils in the forested upland areas are composed of clay and rock

(Chenoweth et al. 2011).

Sediment textures in dewatered Lake Aldwell and Lake Mills vary spatially as a result of fluvial processes that occurred both prior to and during dam removal. As the Elwha

River’s competence decreased as it flowed into the dammed reservoirs, suspended coarse sediments were deposited preferentially in the reservoir deltas. Finer material was carried in suspension farther into the reservoirs before deposition, resulting in sediments with higher clay and silt content in reservoir reaches more distant from the deltas (Warrick et al. 2015).

During phased drawdown of both reservoirs, the delta regions advanced northward toward

the dams. Throughout this process, erosion and subsequent deposition of coarser delta sediment resulted in the burial of finer lake-bottom deposits that existed in the reservoirs’ deeper interior regions. In both reservoirs, this resulted in a prodelta area comprised of a layer of coarse sediments deposited on top of fine lake-bottom sediments (Randle et al. 2015,

Warrick et al. 2015).

Submerged fine sediments sampled from the northern end of Lake Mills prior to reservoir drawdown were composed of 6% clay (< 2 μm), 85% silt (2 – 50 μm), 7% fine sand

(50 – 500 μm), and 1% coarse sand (> 0.5 mm). Analogous values for coarse sediments sampled from the reservoir delta above the water level were < 1%, 2%, 87%, and 11%, respectively (Cavaliere and Homann 2012).

The substrate chemistry differs between coarse sediments, fine sediments, and established forests. In the < 0.5 mm fractions of the samples taken from the Lake Mills reservoir prior to dam removal, fine sediments had higher concentrations of C, N, organic P,

Ca-bound P, and Fe-bound P than coarse sediments (Cavaliere and Homann 2012). In submerged fine sediments sampled from Lake Aldwell prior to dam removal, extractable concentrations of K, Ca, and B were lower, and P, Mn, and Cu higher, than those measured in soils of Olympic National Forest stands containing vegetation communities similar to those found in the forests adjacent to the Elwha River.

The dewatered Lake Aldwell sediments likely reflect these pre-dam-removal differences in substrate chemistry. At Manatawny Creek in Pennsylvania, trace metal concentrations did not vary significantly between sediments sampled before and after dam removal (Bushaw-Newton et al. 2002).

Following dam removal at Lake Aldwell, samples from the fine-sediment sites used in the current study were composed of 14% clay, 80% silt, and 6% sand (> 50 μm) (Chhaya

Werner, University of California, Davis, personal communication). Near-surface (1-m) C content was highly variable throughout the dewatered sediments and especially low near

Lake Aldwell’s southern delta (Wing 2014). In fine sediments sampled from plots on the valley walls of both dewatered Elwha River reservoirs, concentrations of N, P, and K were found to be quite low compared to levels normally found in riparian zones, with especially high variability in K concentrations (Werner 2014). The findings from these studies are generally consistent with those from a small-dam-removal study in California, where dewatered reservoir sediments were found to be low in organic matter compared to forest reference sites (Rohdy 2013).

Dewatered Lake Mills sediments contain viable arbuscular and ectomycorrhizal fungi. Arbuscular mycorrhizal spore density was found to decrease with distance from the edge of the adjacent intact forest (Cortese 2014). At dewatered Lake Aldwell, mycorrhizal fungi, which associate with black cottonwood (Karliński et al. 2010), may play a role in facilitating plant uptake of substrate nutrients.

2.4 Gaps in Knowledge

2.4.1 Primary Succession in Dewatered Reservoirs

Dam removal, in itself an approach to ecosystem restoration, causes significant environmental disturbance. Considering the increasing occurrence of dam-removal projects, there is a clear need for scientific assessments of ecosystem responses (Babbit 2002, Poff and

Hart 2002, Shafroth et al. 2002). None of the studies that have investigated the ecological

effects of small-dam removal (Bushaw-Newton et al. 2002, Thomson et al. 2005, Ashley et al. 2006, Velinsky et al. 2006, Rohdy 2013, Cantwell et al. 2014) have specifically assessed plant nutrient availability or plant nutrition. There is a scarcity of knowledge about the ecological impacts of large-dam removal; in two such studies (Wells et al. 2008, Lafrenz et al. 2013), substrate nutrient levels were measured but plant nutrition was not. Past studies have examined impacts of river damming on cottonwood forests downstream (Rood and

Mahoney 1991), but none have investigated cottonwood nutrition in dewatered reservoirs.

2.4.2 Plant Nutrition in the Dewatered Lake Aldwell Reservoir

Analyses completed both prior to and following removal of the dams on the Elwha

River indicate that reservoir sediments are generally low in organic matter as well as several plant-essential elements such as N, K, Ca, and B. Among sediment types, there is considerable spatial variation in nutrient levels (Chenoweth et al. 2011, Cavaliere and

Homann 2012, Werner 2014, Wing 2014, Schuster 2015).

Within the texturally-diverse Lake Aldwell reservoir sediment environment, water and nutrient availability may be highly variable as a function of the physical, chemical, and geomorphological characteristics of the substrates (Calimpong 2014). This could have important implications for plant succession. In the former Lake Mills reservoir, substrate texture and moisture content were found to significantly influence survivorship for several woody species planted in the dewatered sediments, with plant survival being highest in sediments composed primarily of silts and clays (Calimpong 2014). No vegetation studies have measured or compared plant nutrition between the various sediment microenvironments in either reservoir. This study’s nutrient assessment will provide information regarding the

extent to which nutrient availability may impact ecosystem restoration in dewatered sediments of the Elwha River and in other environments following dam removal.

3.0 Methods

3.1 Sample Collection

3.1.1 Study Site

The study site is the dewatered Lake Aldwell reservoir (48°05′40″N 123°33′24″W,

Elevation 61 m), which is cut by the Elwha River and bordered on all sides by second-growth conifer forest (Chenoweth et al. 2011). The area receives an average of 1.4 m annual precipitation (Duda et al. 2008). Variations in sediment texture within the reservoir are the result of several factors, including the manner in which suspended sediments in the Elwha

River were originally deposited, characteristics of reservoir currents, and the rate and characteristics of reservoir drawdown, which influenced erosion and transport of sediment within the reservoir (Randle et al. 2015, Warrick et al. 2015). In the time since the completion of dam removal and reservoir drawdown in early 2012, natural establishment of many native riparian woody plants, including black cottonwood and Scouler’s willow, has occurred at several locations within the reservoir. Many native species have also been seeded or planted throughout the reservoir in order to assist the natural revegetation process and to accelerate plant succession.

3.1.2 Sampling Sites

Sampling took place on November 6 and 7, 2014. No precipitation occurred during sample collection, but wind and moderate precipitation were observed both before and after sampling on the first day. Naturally-established black cottonwood saplings were sampled in the dewatered Lake Aldwell reservoir from seven distinct stands (Figure 1) selected under the guidance of Josh Chenoweth, lead restoration botanist at Olympic National Park. Two of the cottonwood stands were growing on fine-grained sediments, three on coarse-grained 36

sediments, and two on complexly-bedded prodelta substrates, the latter being comprised of a layer of fine-grained sediments deposited on a layer of coarse-grained sediments (Warrick et al. 2015). Cottonwood saplings were also sampled from one site outside of the reservoir sediments at the border of the adjacent established forest (Figure 1). In this study, unless explicitly stated otherwise, the term “reservoir” is used when referring exclusively to the three sediment microenvironments within the dewatered reservoir itself (including all fine, prodelta, and coarse sites), while “Lake Aldwell” is used in the absence of “reservoir” to refer to all four microenvironments in which cottonwood were sampled. Because black cottonwood seeds are dispersed by wind and because seeds are only viable for a short window of time, black cottonwood was not artificially seeded in the reservoir or planted at the sites of these stands (Joshua Chenoweth, Olympic National Park, personal communication), thus ensuring that only naturally-established plants were sampled.

Figure 1. Sampling locations in the dewatered Lake Aldwell reservoir and adjacent forest (left), with close-ups of northernmost (upper right) and southernmost (lower right) sampling sites. Highway 101 and the site of the previous Elwha Dam are labeled for reference (map data: Google, 2016).

3.1.3 Sampling Procedures

From each reservoir site, four cottonwood saplings (Leech and Kim 1981, Rodríguez-

Barrueco et al. 1984, Hagen-Thorn and Stjernquist 2005) were selected haphazardly, using the following avoidance criteria: no visual mechanical or insect damage (Cusick 2001), no visual disease infestations, and no plants growing within 1 m of the border of the stand

(Jones 2012). Plant height, stem diameter at 5 cm above ground, and stem density of surrounding woody plants (based on a variable-density radius from 1 to 5 m) were recorded for each sapling. Reservoir saplings sampled were either 2 or 3 yr old (Joshua Chenoweth,

Olympic National Park, personal communication) depending on the site and were between 1 and 5 m tall. Outside of the dewatered reservoir sediments, only two cottonwood saplings were found for sampling in the adjacent forest. These plants, of unknown age, were between

1 and 2 m tall and were growing among mature trees at the border of the forest adjacent to the reservoir, on the uphill side of the current dam-site access road.

Plant leaves (10 – 20) with petioles intact were sampled from all canopy positions and combined for each individual sapling, yielding one composite sample from each plant.

Visual, subjective observations about leaf tissue color and morphology were recorded.

Yellowing of cottonwood leaf tissue had occurred prior to the time of sampling, and brown regions were visible on the leaves of many of the saplings, particularly those growing at coarse sediment reservoir sites. Substantial leaf litter was observed on the ground, presumably from both the cottonwood saplings and the willow amidst which the cottonwood were growing.

In past studies of autumnal nutrient resorption in Populus (Baker and Blackmon

1977, Francis and Baker 1981, Killingbeck et al. 1990, Salehi et al. 2013), leaves sampled during October and November were senesced. Given that Lake Aldwell cottonwood leaves were sampled in November and that there were already leaves on the ground, and based on visual observations of leaf coloration, it was surmised that leaf senescence was at an

advanced stage in the Lake Aldwell saplings and that extensive resorption of mobile nutrients such as N and P (Baker and Blackmon 1977) had likely already occurred.

Prior to collection of sediment samples at reservoir sites, large cobbles and any existing leaf litter were brushed away from the substrate surface. One sediment sample of

500 to 750 g was then collected from the upper 10 cm of substrate at the base of each sapling and weighed in the field using a portable scale, accurate to ±1 g. No substrate samples were collected at the forest site due to permitting limitations. Leaf and sediment samples were stored in separate plastic bags which were then enclosed within one larger labeled plastic bag for each sampled tree. All samples were refrigerated in a cooler within 4 h of collection and refrigerated below 40 °C within 48 h of collection pending drying (Cusick 2001, Jones

2012).

3.2 Sample Preparation

All laboratory work was completed at Western Washington University. Sample preparation, NC analysis, and sample digestions were carried out through the Department of

Environmental Sciences, and inductively coupled plasma mass spectrometry (ICP-MS) was carried out through the Advanced Materials Science and Engineering Center.

3.2.1 Foliar Sample Drying and Homogenization

Foliar samples were transferred to paper storage bags and oven-dried to constant mass for at least 48 h at 70 °C (Elferjani et al. 2013, Salehi et al. 2013) to stabilize them for storage at room temperature (Reuter et al. 1997). Petioles were removed from leaves by hand and the remaining leaf tissue was homogenized (Hagen-Thorn and Stjernquist 2005, Elferjani et al.

2013, Maathuis 2013) using a Foss Tecator AB 1093 Cyclotec Sample Mill (FOSS, Hillerød,

Denmark) with a 1-mm screen. Prior to introduction of each new sample into the mill, the mill was disassembled, vacuumed, and reassembled. A portion of the new sample was then processed and discarded, and the glass collection vial was vacuumed and then sprayed with forced air. Another portion of the sample was then processed and kept for analysis.

Homogenized leaf tissue samples were transferred to paper envelopes and dried again before elemental analysis (Cusick 2001).

3.2.2 Sediment Sample Drying and Weighing

Sediment samples were transferred to paper storage bags and oven-dried to constant mass for at least 48 h at 80 °C. Each sediment sample was weighed, and sediment dry mass was subtracted from sediment mass at time of collection to determine sediment moisture mass. Percent sediment moisture content was calculated as the quotient of sediment moisture mass and sediment mass at time of collection, expressed as a percentage (Werner 2014).

3.3 NC Elemental Analysis

Dry (70 °C), homogenized leaf samples were analyzed for total N using a Thermo

Electron NC Soil Analyzer Flash EA 1112 Series (Thermo Electron Corporation, Milan,

Italy). A known mass of material of 19 to 21 g was measured into a tin capsule, which was then compressed to seal in the material and to remove any air prior to insertion into an autosampler (Cavaliere 2010). The calibration curve was made from atropine (MP

Biomedicals, Santa Ana, CA), which contained 48.4 g kg-1 N. In the calibration curve, which had an R2 value of 0.999, the mass of N ranged from 0.055 mg N to 0.946 mg N.

3.4 ICP-MS Elemental Analysis

3.4.1 Chemical Reagents

In this study, HNO3 used was of ultrapure grade, certified for trace metals analysis at the parts-per-billion level (VWR International, West Chester, PA), and H2O2 used was of reagent-grade purity (MP Biomedicals, Santa Ana, CA). All glassware was acid-washed with

10% HNO3. High-purity water of 18 MΩ cm from a Thermo Scientific Barnstead water purification system (Thermo Fisher Scientific, Inc., Waltham, MA) was used for all preparation of samples and dilution of calibration standards and for final rinsing of glassware, plastic measuring utensils, and acid-cleaned microwave digestion vessels. The following protocol was carried out to clean microwave digestion vessels prior to each use:

5 mL concentrated HNO3 was added to each vessel, the temperature was ramped to 180

°C over a period of 10 min, and vessels were cooled and vented. Vessel contents were discarded, vessels were disassembled, and all components were rinsed five times with water and dried with forced air.

3.4.2 Sample Digestion

Foliar samples were digested using a microwave-assisted procedure (Maathuis 2013) in a Milestone ETHOS EZ microwave system (Milestone, Inc., Shelton, CT) equipped with

10 high-pressure 100-mL tetrafluoroethylene digestion vessels. A known dry (70 °C) mass of material of 230 to 270 mg was weighed directly into each tared vessel, to which were added

9 mL concentrated (67%) HNO3 and 1 mL H2O2. Vessels were then closed and secured in the digestion oven. Following the “Tea Leaves” digestion protocol (Milestone, Inc., n.d.), temperature was ramped to 180 °C over 5 min and held at 180 °C for 10 min, after which vessels were cooled to room temperature and vented. Each digestate was quantitatively 42

transferred to a tared 50-mL centrifuge tube by rinsing the digestion vessel and vessel lid 6-8 times with water, diluted with additional water to a final volume of 50 mL, and weighed.

Prior to chemical analysis, digestates were diluted to 5% HNO3 and centrifuged for 30 min at

3000 rpm.

3.4.3 Elemental Analysis

Leaf-tissue digests were analyzed for macronutrients (P, K, Ca, Mg) and micronutrients (Fe, Mn, Zn, Cu, Ni, Mo) for this study, as well as several other elements, using an Agilent 7500 Series Inductively Coupled Plasma Mass Spectrometer (Agilent

Technologies, Santa Clara, California, USA) equipped with octopole reaction system and

ASX-510 Autosampler (Teledyne CETAC Technologies, Omaha, NE). A multi-element solution external standard (Inorganic Ventures, Christiansburg, VA) and internal standards

(6Li, 45Sc, 72Ge, 103Rh, 115In, 159Tb, 175Lu, 209Bi) were used for calibration and tuning of ICP-

MS instrumentation for determination of Ag, Al, Ba, Ca, Cd, Co, Cr, Cu, Fe, K, Mg, Mn,

Mo, Na, Ni, V, and Zn. Analyses were run in no-gas mode. External calibration standards were prepared in 5% HNO3 by the serial dilution method at analyte concentrations ranging from 100 mg L-1 to 10 ng L-1 for Ca, Fe, K, Mg, and Na, and ranging from 1 mg L-1 to 0.1 ng

-1 L for the remaining analytes. For determination of P, a KH2PO4 standard (Avantor

Performance Materials, Center Valley, PA) was used for calibration (K concentrations ranging from 100 mg L-1 to 10 ng L-1) with 45Sc as an internal standard, and He-reaction-cell mode was used for the analysis. A calibration blank containing 5% HNO3 completed the calibration curve for each analyte measured (Enamorado-Baez et al. 2013), and Ar was used as a carrier gas for all ICP-MS analyses.

Final masses of digestates (diluted to 5% HNO3 in 50 mL) were corrected for the increase in density (relative to water) caused by the addition of HNO3. Plant-tissue nutrient concentrations ([푛푢푡푟푖푒푛푡]푑푟푦 푙푒푎푓) were calculated based on the following equations, where:

푚푥 = mass of 푥 푣푦 = volume of 푦

[푥]푦 = concentration of 푥 in 푦

휌푧 = density of 푧 퐼퐶푃 푠푎푚푝푙푒 = diluted microwave digestate, as analyzed by ICP-MS In this study:

[퐻푁푂3]푖푛푖푡푖푎푙 = 67% HNO3 by mass

휌([퐻푁푂3]) = interpolated density value @ 22 °C, from Perry and Green (1984)

To determine [푑푟푦 푙푒푎푓]푑푖푔푒푠푡푎푡푒 : 푚 = ([퐻푁푂 ] )(푣 )(휌 ) 퐻푁푂3푑푖푔푒푠푡푎푡푒 3 푖푛푖푡푖푎푙 퐻푁푂3 ([퐻푁푂3]푖푛푖푡푖푎푙) 푚 퐻푁푂3푑푖푔푒푠푡푎푡푒 [퐻푁푂3]푑푖푔푒푠푡푎푡푒 = 푚푑푖푔푒푠푡푎푡푒

푚푑푖푔푒푠푡푎푡푒 푣 = 푑푖푔푒푠푡푎푡푒 휌 ([퐻푁푂3]푑푖푔푒푠푡푎푡푒)

푚푑푟푦 푙푒푎푓 [풅풓풚 풍풆풂풇]풅풊품풆풔풕풂풕풆 = 푣푑푖푔푒푠푡푎푡푒

To determine [푑푖푔푒푠푡푎푡푒]퐼퐶푃 푠푎푚푝푙푒 :

푚퐼퐶푃 푠푎푚푝푙푒 푣 = 퐼퐶푃 푠푎푚푝푙푒 휌 ([퐻푁푂3]퐼퐶푃 푠푎푚푝푙푒)

푣푑푖푔푒푠푡푎푡푒 [풅풊품풆풔풕풂풕풆]푰푪푷 풔풂풎풑풍풆 = 푣퐼퐶푃 푠푎푚푝푙푒

To determine [푛푢푡푟푖푒푛푡]푑푟푦 푙푒푎푓 :

[풅풓풚 풍풆풂풇]풅풊품풆풔풕풂풕풆 [푑푟푦 푙푒푎푓]퐼퐶푃 푠푎푚푝푙푒 = [풅풊품풆풔풕풂풕풆]푰푪푷 풔풂풎풑풍풆

[푛푢푡푟푖푒푛푡]퐼퐶푃 푠푎푚푝푙푒 [푛푢푡푟푖푒푛푡]푑푟푦 푙푒푎푓 = [푑푟푦 푙푒푎푓]퐼퐶푃 푠푎푚푝푙푒 44

Macronutrient values were expressed in g kg-1 and micronutrient values were expressed in mg kg-1.

3.4.4 Quality Control

Reliability of NC analyses was evaluated using lab duplicate samples, sample blanks, and Standard Reference Material (SRM) 1515 “Apple Leaves” (National Institute of

Standards and Technology, Gaithersburg, MD), which contained 22.5 g kg-1 N (Reed 1993).

Observed values of SRM 1515 lab quality control samples, which were interspersed throughout homogenized cottonwood leaf samples, ranged from 22.5 to 23.0 g kg-1 N. For determination of N for five out of six lab duplicate cottonwood samples, relative percent difference was < 1%. Relative percent difference for the sixth pair of lab duplicate samples was > 20%. The N value from this sixth lab duplicate was very consistent with those measured for other plants sampled from the same reservoir site, so the lab duplicate value was accepted in place of the value measured for the original sample.

Reliability of ICP-MS analyses was evaluated using sample digestion blanks and

SRM 1515, certified elements of interest including Al, Ba, Cd, Ca, Cu, Fe, K, Mg, Mn, Mo,

Na, Ni, P, V, and Zn, and non-certified elements of interest including Co and Cr (Reed

1993). All samples, including digestion blanks and SRM 1515, were measured in quintuplicate for validation of analytical precision, and blank-corrected values were reported and used for statistical analysis. Limits of detection and quantification were calculated as three and ten times the standard deviation of the blank, respectively, for each element analyzed. Foliar elemental concentration values determined by ICP-MS were < 10% RSD

(for quintuplicate sample measurements) for all nutrients with the exception of P and Mo, for which values were < 15% RSD and 20% RSD, respectively.

Standard Reference Material 1515 certified values and lab quality control sample values determined by ICP-MS are reported in Appendix Tables A through C. Values measured for nutrients in SRM 1515 quality control samples ranged from 63% to 93% of certified values, while measured values for non-nutrients ranged from 49% to 92% of certified values. All cottonwood samples contained detectable levels of Mo, but the majority of Mo values were below the limit of quantification (Appendix Table B). Because Mo could not be determined with analytical precision, it was not included in any statistical analyses.

Foliar concentration values for non-nutrients that were measured in this study are reported in

Appendix Table D.

To assess potential contamination arising from use of the sample mill, one SRM sample was processed through the mill during homogenization of cottonwood leaf samples.

The Ni concentration determined for the homogenized SRM sample was four times higher than the certified value and five times higher than the value determined for the unmodified

SRM sample (Appendix Table B). It is possible that this sample homogenization procedure may have resulted in inaccurate Ni values, but as foliar Ni concentrations were not reported in the Populus literature, the validity of these values was not assessed further.

3.5 Statistical Analysis

Microsoft Excel (version 2013) was used for all data management in this study, and R

(version 3.1.2; Sarkar 2008, R Core Team 2014) was used for all statistical analyses and plotting.

3.5.1 Analysis of Variance

For each nutrient, a single-factor analysis of variance (ANOVA) was used to compare the four microenvironments (fine-grained and coarse-grained sediments, prodelta, and adjacent forest). The approach of using multiple single-factor ANOVAs is consistent with previous cottonwood foliar nutrient studies (Shelton et al. 1981, McLennan 1990). For statistical analysis, site was the observational unit and the value for a site was determined as the average of values from saplings within the site. The goal of this study was to make inferences that extend beyond these specific sampling locations in order to achieve a broader understanding of foliar nutrient variability between microenvironments within the Lake

Aldwell ecosystem. These specific sampling sites were selected based on their accessibility and because they fit the descriptions of the various microenvironments, but had other suitable sites been available and had resources permitted, other sampling sites could have been selected. Site was therefore considered as a random factor in the ANOVA model.

Levene’s tests were used to check the assumption of homogeneity of variances among microenvironments. If the Levene’s test p-value was < 0.1 for non-transformed data, as was the case for P and Mn, data were log10-transformed to reduce heterogeneity of variances prior to ANOVA. There were too few sampling sites to rigorously test the assumption of normality; for the purpose of ANOVA, it was assumed that the samples came from populations with normally-distributed nutrient concentrations. For nutrients with significant

ANOVA results (p-value < 0.05), Tukey’s honestly significant difference (HSD) pairwise comparisons of treatment means were completed at the α = 0.05 level (Salehi et al. 2013).

3.5.2 Correlation Analysis

Correlation analysis was used to explore relationships between foliar nutrient concentrations and the environmental variables sediment moisture content and woody plant stem density. Twenty correlation tests were performed, one for each combination of nutrient and environmental variable. Correlation analysis was also used to quantify relationships between foliar nutrient concentrations. Forty-five correlation tests were performed, one for each combination of nutrients. All nutrient concentrations were log10-transformed to increase linearity, and Pearson product-moment correlation coefficients (Pearson’s r) and correlation test probabilities (p-values) were calculated and assigned significance at the α = 0.05 level.

3.5.3 Statistical Errors

When conducting multiple statistical tests, several methods are routinely used to control family-wise error rate. Among them are the Bonferroni correction method, in which the family-wise error rate is divided by the number of tests in order to arrive at an adjusted test-wise α level, and the Dunn-Sidak method, in which the adjusted test-wise α level is a reflection of the probability of mistakenly rejecting the null hypothesis at least once for a given family of tests (Gotelli and Ellison 2013). Regarding the adjustment of α, Gotelli and

Ellison (2013) suggest that, rather than systematically increasing the standard of evidence for each test conducted (which effectively penalizes studies involving multiple tests), raw p- values should be reported and interpreted using some degree of common sense.

The ANOVA and correlation tests in this study were performed using an unadjusted test-wise error rate of α = 0.05 with the acknowledgment that doing so involved substantial inflation of the family-wise error rate. Out of the 75 hypotheses tested, 29 were rejected,

which is far more than the 5% alpha error rate. These results clearly indicate that the patterns observed are a product of more than random variability.

3.5.4 Literature Value Comparisons and Growing-Season Estimates

The foliar N, P, K, Ca, and Mg concentrations from this study were compared with the general seasonal patterns in Populus foliage determined by compiling data from the literature (Tew 1970, Verry and Timmons 1977, Blackmon et al. 1979, Francis and Baker

1981, Shelton et al. 1981, Killingbeck et al. 1990, McClennan 1990, Escudero 1992, Cusick

2001, Brown and van den Driessche 2002, Coleman et al. 2006, Martín-García et al. 2012,

Elferjani et al. 2013, Salehi et al. 2013, Ciadamidaro et al. 2014). These specific nutrients were selected for literature value comparisons because very few previous studies reported values for the other nutrients measured in this study.

Growing-season foliar N, P, K, Ca, and Mg concentrations were estimated from the late-season values measured in this study. Estimates were calculated based on repeat foliar macronutrient measurements of 43-yr-old Populus tremuloides, using values from growing- season (July 30) and late-season (September 30) live leaves (Verry and Timmons 1977):

[growing‐season estimate] =

[growing‐season measure](Verry and Timmons 1977) [late‐season measure](this study) ∗ ( ) [late‐season measure](Verry and Timmons 1977)

The estimated growing-season concentrations were qualitatively compared with optimal DRIS values for Populus (Leech and Kim 1981, René et al. 2013).

4.0 Results and Discussion, Part I: Within the Reservoir and Adjacent Forest

4.1 Variation in Foliar Nutrients among Microenvironments

Foliar nutrient concentrations differ between microenvironments for macronutrients

P, K, Ca, and Mg (Table 1) and for micronutrients Mn and Zn (Table 2). Within the reservoir, foliar P and K concentrations are higher at fine than coarse sites, while Ca follows the opposite trend. Foliar Mn is three times higher at fine and prodelta sites than at coarse sites. The forest has higher foliar Ca and lower P than fine sites (Table 1) and higher foliar

Mg and Zn, but lower foliar Mn, than all reservoir microenvironments (Tables 1 and 2).

TABLE 1. Mean foliar macronutrient concentrations of cottonwood saplings from four microenvironments at the dewatered Lake Aldwell reservoir, November 2014.

Macronutrient (g kg-1) Microenvironment n † N P K Ca Mg Coarse Sediment 3 12.9 0.94b‡ 3.41b 13.47a 2.76b Prodelta 2 17.2 1.29a,b 5.82a,b 10.19a,b 2.53b Fine Sediment 2 15.5 1.59a 7.38a 8.17b 2.58b b a,b a a Adjacent Forest 1 12.0 0.72 3.79 15.17 4.52 Standard Deviation (SD)§ 1.9 0.18 0.83 1.3 0.17 p-value Levene's Test 0.172 0.329¶ 0.159 0.805 0.730 ANOVA 0.021*¶ 0.183 0.023* 0.023* 0.002* †n = number of sites per microenvironment ‡for nutrients with significant ANOVA p-values within a column, mean nutrient values not followed by the same letter are different (p-value < 0.05) as determined by Tukey's HSD pairwise comparisons §SD = √(pooled within-microenvironment variance) = √(MSE from ANOVA table) ¶ p-value from statistical test on log10-transformed data *statistically significant ANOVA result (p-value < 0.05)

TABLE 2. Mean foliar micronutrient concentrations of cottonwood saplings from four microenvironments at the dewatered Lake Aldwell reservoir, November 2014.

Micronutrient (mg kg-1) Microenvironment n † Fe Mn Zn Cu Ni Coarse Sediment 3 67.6 211b‡ 245b 4.77 34.0 Prodelta 2 76.0 596a 174b 6.56 45.1 Fine Sediment 2 58.4 635a 149b 5.53 38.4 c a Adjacent Forest 1 76.5 71 487 4.55 23.2 Standard Deviation (SD)§ 13.5 110 36 1.07 11.9 p-value Levene's Test 0.112 0.180¶ 0.685 0.204 0.534 ANOVA 0.003*¶ 0.601 0.006* 0.376 0.544 †,‡,§,¶,*see Table 1

The foliar nutrient levels in Lake Aldwell cottonwood (Tables 1 and 2) are consistent with trends in the nutrient concentrations of local forest soils and sediments (Table 3). The higher foliar P concentrations at fine than at coarse sites are consistent with the greater P concentrations in fine than in coarse sediments of the upriver Lake Mills reservoir (Table 3).

The higher foliar Ca and Mg in the adjacent forest than at fine sites are consistent with

Olympic National Forest forest soils having higher levels of Ca and Mg than submerged fine sediments within the Lake Aldwell reservoir (Table 3). Both foliar (Tables 1 and 2) and substrate (Table 3) P and Mn are higher in the fine sediments than in the forest. In contrast, foliar Zn concentrations are elevated in the adjacent forest relative to the fine sediments

(Table 2), but Zn levels are similar in forest soils and the submerged fine reservoir sediments

(Table 3). Foliar N is no different between fine and coarse sediments in Lake Aldwell, but substrate N is higher in fine than in coarse sediments at Lake Mills (Table 3).

TABLE 3. Nutrient concentrations of local forest soils and reservoir sediments.

Total Organic Ca-bound Fe-bound Substrate Location N ------P ------(mg kg-1) ------Coarse sediment† Mills pre-removal 460 52 257 82 Fine sediment† Mills pre-removal 710 83 315 101

Soluble ------Extractable‡ ------P K Ca Mg Mn Zn Cu -1 -1 -1 --- (mg kg ) ------(cmolc kg ) ------(mg kg ) ------Fine sediment§ Aldwell pre-removal 12.0 19 3.1 0.5 66 0.83 2.8 ¶ Fine sediment Aldwell post-removal 8.5 35 - - - - - Forest (minimum)* Olympic Nat’l. Forest 7.4 61 4.0 0.4 30 0.78 1.1 Forest (maximum)* Olympic Nat’l. Forest 9.2 83 10.0 2.3 31 0.84 1.4 †data from Cavaliere and Homann (2012) ‡ extractants used: NH4F for P; NH4OAc for K, Ca, Mg; DTPA for Mn, Zn, Cu §data from Chenoweth et al. (2011) ¶data from Werner (2014); mean values from the current study’s fine-sediment sites (Chhaya Werner, University of California, Davis, personal communication) *data from Henderson et al. (1989), as cited in Chenoweth et al. (2011)

The adjacent forest site shares similarities with the coarse-sediment reservoir sites in that foliar P is lower in both of these microenvironments than at fine-sediment sites (Table

1). The causes for these lower levels of P, however, are likely different. Phosphorus availability in the coarse reservoir sediments is presumably limited by the rate at which P is weathered from minerals; in the coarser-textured substrates, less surface area is accessible to weathering than in the fine sediments, resulting in lower P availability for plants (Vitousek

1999). Also, fine Elwha River sediments have three times as much oxalate-extractable Fe as coarse sediments (Cavaliere and Homann 2012), and P sorption is highly correlated with oxalate-extractable Fe + Al (Binner et al. 2015). Phosphorus released to the soil solution from the sorbed P pool may have been taken up (Wang et al. 2012, Binner et al. 2015) by cottonwood, potentially explaining higher foliar P levels in the fine sediments. In the forest, 52

competition between plants – as opposed to the physical and chemical characteristics of the substrate – is likely limiting P availability.

Plant growth in early successional environments is generally limited by low levels of

N in the substrate (Walker and Syers 1976). However, in contrast to foliar P, foliar N does not vary significantly between Lake Aldwell microenvironments. This suggests that plants are deriving N from sources other than the substrate. While the growth substrate serves as the principal source of plant P, several atmospheric inputs may be contributing to plant N content. It has been demonstrated that Populus trichocarpa clones inoculated with diazotrophic endophytes – N2-fixing microorganisms that live within plant tissue – exhibit a positive growth response and an increase in total plant N content (Knoth et al. 2014).

Nitrogen fixation may be occurring in these cottonwood, resulting in higher N supply independent of substrate composition.

Increases in N deposition from industrial and agricultural processes in recent decades

(Galloway et al. 2003, Morier et al. 2008) suggest that forests near urban areas are becoming saturated with biologically-reactive N (Chen et al. 2010). In forests of Pacific Northwest

- + Washington, wet deposition of NO3 and NH4 , as well as deposition of biologically-reactive atmospheric N in the form of NH3 and NOx gases, has been increasing since 1980 (Fenn et al.

2003). Wet deposition of N compounds was found to be significantly lower in Olympic

National Park than at urban sites in Washington and Oregon (Fenn et al. 2013). Nonetheless, the Elwha River forest ecosystem may be enriched in biologically-reactive atmospheric N species generated from emissions from marine vessel traffic in the nearby Strait of Juan de

Fuca (Fenn et al. 2013) and from regional industrialized areas.

Direct foliar uptake of biologically-reactive atmospheric N has been shown to account for up to 50% of foliar N content in Populus tremuloides (Nave and Curtis 2011). At

Lake Aldwell, canopy uptake of atmospheric N inputs may constitute a significant source of

- + N to cottonwood. Root uptake of NO3 and NH4 deposited to the substrate from the atmosphere may also be substantial (Fenn et al. 2013, Wei et al. 2014). These potential sources of plant N may help to explain why foliar N was not different among Lake Aldwell microenvironments, whereas foliar P varied significantly.

4.2 Foliar Nutrient Correlations with Environmental Variables

Lake Aldwell reservoir microenvironments are distinctly divided into two clusters with respect to sediment moisture content: a low-moisture group of coarse-sediment sites and a higher-moisture group of prodelta and fine-sediment sites (Figure 2). The foliar concentrations of several nutrients track this trend: sediment moisture content is positively correlated with foliar P, K, and Mn and negatively correlated with foliar Ca and Zn (Figure

2). With the exception of Zn, these results are consistent with the results from ANOVA

(Tables 1 and 2), which indicate that fine sites have higher foliar P, K, and Mn, and lower foliar Ca, than coarse sites.

The relative ordering of sites with respect to sediment moisture content (Figure 2) is consistent with their qualitative classification as coarse, prodelta, and fine sediment sites, indicating that these sampling sites are categorized correctly by their textural characteristics.

Sediment moisture content generally correlates positively with nutrient retention (Brady and

Weil 2009) and may be a good indicator of relative foliar nutrient levels. It is possible that higher foliar K and Mn in the higher-moisture fine sediments are a function of better

retention of these nutrients on the surfaces of clay particles, making them more available for plant uptake. Kubota et al. (1970) found that variability in foliar Mn in deciduous trees was due partly to soil drainage effects, foliar Mn in well-drained soils being significantly lower than in soils that retained more moisture. Lower foliar Mn in the better-drained coarse sediments is consistent with this explanation.

Figure 2. Correlation plots comparing sediment moisture content with log10-transformed mean foliar nutrient concentrations for reservoir cottonwood saplings from n = 7 sites spanning 3 microenvironments within the dewatered Lake Aldwell reservoir sediments, November 2014. Only significant correlations (p-value < 0.05) are shown. 56

Water availability may interact with principles of nutrient translocation to explain some of the foliar nutrient differences observed between Lake Aldwell reservoir microenvironments. The correlation patterns based on sediment moisture serve to classify the nutrients into three distinct groups: those that correlate positively with sediment moisture (P,

K, and Mn), those that correlate negatively (Ca and Zn), and those for which no correlation is observed (N, Mg, Fe, Cu, Ni). These nutrient groupings share little overlap with those proposed by either Garten (1978) or Zhang et al. (2012) but share some overlap with those found by McClennan (1990) (see Sections 2.1.4 and 2.1.5). Specifically, foliar concentrations of Ca and Zn, which were part of the set of immobile elements (McClennan 1990), are negatively correlated with sediment moisture in the Lake Aldwell reservoir (Figure 2). The opposite is true for phloem-mobile elements P and K.

In the coarse sediments, which are dryer than the other reservoir sediments (Figure 2), water stress may have induced the onset of leaf senescence earlier in the season, thereby accelerating the leaf-aging process in the saplings growing there (Marschner 2012). If this is the case, the observed differences in foliar nutrients between these microenvironments may be a result of sampling leaves from different stages of senescence and may not be an indication of differential nutrient availability. Immobilized in the plant cell walls (Larcher

1975), Ca would have accumulated in the foliar tissue and remained concentrated even as foliar P and K were reabsorbed via phloem-loading (Marschner 1986, Jones 2012). These resorption patterns would explain the observed lower levels of foliar P and K, and higher levels of foliar Ca, in the coarse sediments compared to those in the fine sediments (Table 1).

When foliar nutrient levels are plotted as a function of stem density (Figure 3), Lake

Aldwell reservoir sites cluster in the same pattern – with the exception of one coarse site – as observed for sediment moisture content (Figure 2). Both sediment moisture content and stem density are positively correlated with foliar Mn and negatively correlated with foliar Zn, but no other consistent patterns emerge between the two variables.

Most of the foliar nutrients measured are not correlated with stem density. The distribution of cottonwood saplings at Lake Aldwell reservoir sampling sites was determined by conditions present at the time of seed dispersal (Braatne et al. 1996). Later in the successional process, intraspecific competition for nutrients and water (Walker and del Moral

2003) will likely result in competitive exclusion in these cottonwood stands, at which point additional negative correlations between stem density and foliar nutrients, such as that observed for Zn, may be detected. Whether or not nutrients are limiting, at this stage in ecological succession it is still too early for nutrient availability to have had any measurable effect on the density of cottonwood at these sites (Joshua Chenoweth, Olympic National

Park, personal communication).

Figure 3. Correlation plots comparing stem density with log10-transformed mean foliar nutrient concentrations for reservoir saplings from n = 7 sites spanning 3 microenvironments within the dewatered Lake Aldwell reservoir sediments, November 2014. Only significant correlations (p-value < 0.05) are shown.

4.3 Correlations among Foliar Nutrients

Of the 45 pairs of nutrients, 16 pairs are correlated when all 8 sampling sites are considered (Table 4). The majority of significant correlations are positive, including all of those involving N. When only the reservoir sampling sites are evaluated, only 11 pairs of nutrients are correlated. This indicates that the forest site exerts a disproportionate influence on the strength of 5 of the correlations. This is especially the case for correlations between

Mg and Mn and between Mg and Zn, nutrients which are found at different foliar concentrations at the forest site than at all sites within the reservoir (Tables 1 and 2).

TABLE 4. Results from pairwise nutrient-nutrient correlation tests yielding p-value < 0.05 when the forest site is included. Tests were performed on log10-transformed mean foliar nutrient concentrations of cottonwood saplings from the dewatered Lake Aldwell reservoir, November 2014, including (left, n = 8 sites) and excluding (right, n = 7 sites) data from the forest site.

With Forest Site (n † = 8) Without Forest Site (n † = 7) Nutrient Pair Pearson's r p-value Pearson's r p-value N-Cu 0.92 0.001 0.92 0.004 N-Ni 0.90 0.002 0.88 0.008 P-K 0.89 0.003 0.95 0.001 N-P 0.83 0.011 0.79 0.035 P-Mn 0.83 0.011 0.73 0.065‡ Mg-Zn 0.83 0.011 0.36 0.427‡ Cu-Ni 0.80 0.016 0.79 0.033 N-K 0.77 0.025 0.76 0.047 Ca-Zn 0.76 0.030 0.72 0.071‡ K-Mn 0.72 0.045 0.82 0.025 P-Ca -0.91 0.002 -0.89 0.008 Mn-Zn -0.91 0.002 -0.78 0.037 K-Ca -0.85 0.008 -0.86 0.014 Ca-Mn -0.84 0.009 -0.81 0.026 Mg-Mn -0.84 0.010 -0.66 0.111‡ P-Zn -0.83 0.011 -0.74 ‡ 0.059 †n = number of sites ‡indicates non-significant correlation (p-value > 0.05) when forest site is excluded

Most of the correlations observed between foliar nutrients in Lake Aldwell cottonwood are consistent with those reported in previous studies, and many have explanations based on plant physiology. Foliar N, P, and K are all positively correlated with each other (Table 4), consistent with the overall pattern for these nutrients observed by

Zhang et al. (2012) in their analysis of an expansive collection of plant species. Potassium is essential for the activation of enzymes that catalyze protein synthesis (Ranade-Malvi 2011), and P is a component of nucleic acids (Marschner 1986). This could also explain why the concentrations of these nutrients are positively correlated with N, a structural component of both proteins and nucleic acids (Marschner 1986).

Plant growth is highly dependent on N availability; positive correlations between N and P, K, Cu, and Ni (Table 4) could be an indication that N-stimulated growth is causing increased nutrient demand and uptake in order to support the production of new cellular tissue (Fageria 2001). Alternately, higher P (Fageria 2001) or Cu (Ranade-Malvi 2011) in the soil solution may be enhancing plant uptake of N. Another explanation for the strong correlation between foliar N and Cu (Table 4) is that Cu may be catalyzing N2 fixation

(Salisbury and Ross 1969 [as cited in Garten 1978]), which may be occurring in Lake

Aldwell cottonwood through associations with diazotrophic endophytes (Knoth et al. 2014).

The strong positive correlation that exists between Mg and Zn when forest foliar nutrient data is included disappears when that site is excluded from the analysis (Table 4). In other studies that describe nutrient interactions (Fageria 2001, Ranade-Malvi 2011), interactions between Mg and Zn are not reported. Foliar Mg and Zn are both significantly higher in the adjacent forest than in all Lake Aldwell reservoir microenvironments. It seems that this correlation is solely a result of the high forest values and is likely not well explained based on the physiological functions of these nutrients in plants.

Zhang et al. (2012) found positive correlations between the majority of foliar nutrients analyzed, with the exception of Mn, which was negatively correlated with N, P, K

and Ca. Positive correlations observed within Lake Aldwell reservoir sites between foliar P and Mn and between K and Mn (Table 4) are inconsistent with these findings, but the correlation between K and Mn may have a physiological basis. A positive interaction between K and Mn has been reported (Ranade-Malvi 2011), and although Garten’s (1978) analysis did not detect a positive correlation between K and Mn in several plant species, it was found that these nutrients were grouped into a distinct set characterized by their roles in the activation of plant enzymes (Garten 1978). For example, K is involved in the stimulation of over 50 enzymes, including pyruvate kinase, which is of central importance for carbohydrate metabolism. Manganese is a component of superoxide dismutase enzymes, which protect plant tissue from the damaging effects of oxygen free radicals (Marschner

1986).

In Garten’s (1978) study, Mn was also grouped with Ca and Mg in a set of nutrients that serve structural functions in the plant cell: Ca is a component of the cell wall, Mg is the central coordinating ion in chlorophyll, and Mn is a structural element in the membrane system of chloroplasts (Salisbury and Ross 1969 [as cited in Garten 1978]). Given this grouping, the negative correlation between foliar Ca and Mn at Lake Aldwell sites is somewhat surprising; if both of these nutrients serve structural functions in the cell, it seems that they would be positively correlated at any given time of year even as the leaf goes through different stages of growth.

In crop plants, high levels of Zn have been shown to reduce uptake of Mn from the soil solution (Fageria 2001, Ranade-Malvi 2011) which could explain the negative correlation between these nutrients in Lake Aldwell cottonwood (Table 4). The negative

correlation observed between foliar P and Zn in reservoir cottonwood may indicate that higher substrate P at some sites is inhibiting plant uptake of Zn (Ranade-Malvi 2011), as has been observed previously in Populus and which resulted in growth stunting, as well as cupping and scorching of younger leaves (Teng and Timmer 1993). The negative correlation between K and Ca could be due to several factors: K and Ca often compete for the same cation exchange sites in plants, and translocation of Ca to the leaf is reduced with lower levels of K (Ranade-Malvi 2011).

4.4 Effects of Sapling Age and Size

This analysis may be confounded by differences in the ages and sizes of the saplings sampled at the Lake Aldwell reservoir. The fine sites are located on the valley walls, at higher elevation than the other reservoir sites; these sediments were exposed earlier during the process of reservoir drawdown, and the saplings growing there are older. Saplings at the fine sites were 3 yr old, as were the majority of saplings at prodelta sites, while there were both 2- and 3-yr-old saplings at the coarse reservoir sites. The ages of the adjacent forest saplings were not determined. The fine-sediment and prodelta saplings were on average twice as tall as those growing in the coarse sediments (Table 5).

In these young, rapidly-developing plants, small differences in age could mean differences in nutrient utilization (Braatne et al. 1996), which could influence foliar nutrient levels. In plantation-grown eastern cottonwood (Populus deltoides Bartr.), Francis and Baker

(1981) found a consistent decrease in annual height growth over the first four years of plant development, and annual diameter-at-breast-height growth peaked during the second year.

They attributed these decreases in growth to the limited supply of available moisture in the

well-aerated upper 75 cm of soil. Krinard (1979), who compared growth in both pruned and unpruned plantation cottonwood, found that diameter growth peaked in the third year for pruned trees and in the fourth year for unpruned trees but detected minimal differences in height growth in the second through fifth years of plant development.

TABLE 5. Mean dimensions of cottonwood saplings, and mean stem densities of surrounding woody plants, from four microenvironments at the dewatered Lake Aldwell reservoir, November 2014. Plant diameter was measured at 5 cm above ground and stem density of surrounding woody plants was measured based on a variable-density radius from 1 to 5 m.

Plant Height Plant Diameter Stem Density Microenvironment n † (m) (mm) (stems m-2) Coarse Sediment 12 1.8 19 5 Prodelta 8 4.7 24 17 Fine Sediment 8 3.6 19 14 Adjacent Forest 2 1.9 10 -

†n = total number of saplings measured

Differences in plant age, and associated differences in growth rates, may have implications for foliar nutrient accumulation. In plantation cottonwood, live-leaf foliar N and

K concentrations decreased from the first to the fourth year of development, foliar P and Ca showed no significant change, and foliar Mg increased (Francis and Baker 1981). Although

Lake Aldwell cottonwood sapling height was measured in this study, plant growth was not. It could be the case that sediment moisture availability, which is of critical importance for young cottonwood development (Braatne et al. 1996), is influencing these plants’ growth rates. Lower nutrient concentrations, often interpreted as an indication of nutrient limitation, could be a result of growth dilution (Flückiger and Braun 2003) if measured in larger, older

plants. It is difficult to disentangle the possible effects of the variables plant age and height, which were not considered in this study’s statistical analyses.

5.0 Results and Discussion, Part II: Literature Value Comparisons

5.1 Seasonal Variation in Foliar Nutrient Values

5.1.1 Patterns in Literature Values

Substantial seasonal variation in foliar nutrient levels is evident from values reported in previous studies of Populus species (Figure 4 and Table 6). Two factors that significantly influence foliar nutrient concentrations are environmental contamination, such as that arising from mining (Pietrzykowski et al. 2013) and toxic waste spills (Ciadamidaro et al. 2014), and fertilization (Leech and Kim 1981, Elferjani et al. 2013) of substrates in which plants are growing. In order to avoid obscuring natural seasonal trends in foliar nutrient levels, Figure 4 does not include data from plants growing in either contaminated or nutrient-amended substrates.

The seasonal trends (Figure 4) are generally consistent with basic principles of nutrient mobility. Being quite mobile, N, P, and K decrease in concentration after the peak of the growing season as they are reabsorbed into plants’ stemwood and roots (Niinemets and

Tamm 2005). The opposite is true for Ca and Mn, which are not mobile in the phloem and therefore accumulate in the foliar tissue as they are deposited throughout the season (Reuter et al. 1997).

Figure 4. Foliar nutrient concentrations of Populus species as reported in the literature for plants growing in non-fertilized, non-contaminated substrates, and mean foliar nutrient concentrations of cottonwood saplings from four microenvironments at the dewatered Lake Aldwell reservoir, November 2014. Values are arranged left to right in order of sampling month. All literature values were measured in live leaves except for those labeled as senesced, which were measured in freshly-fallen leaves sampled in either October or November. References for all literature values plotted, and descriptions of corresponding plant populations and sampling timeframes, are reported in Table 6.

TABLE 6. Previous studies of Populus in diverse substrate environments, divided by stand type and ordered by first month of sample collection and then by plant age(s). Foliar nutrient values reported from non-fertilized, non-contaminated substrates in these studies are displayed in Figure 4. Table continued on following page.

Stand Type (Plant Age(s) (yr)) Month(s) of Location Sample Collection

Substrate Description(s)

July

Oct.

May

June

Aug. Nov. Reference Sept. Plant Species Sampled Plantation (43) P. tremuloides ("trembling aspen") North Central Minnesota, MN Well-drained fine-sandy loam soil Verry and Timmons (1977)

Plantation (1) P. trichocarpa Torr. & Gray Duhumal and Duparquet, Quebec, Canada × P. balsamifera L. 45% clay soil, 70% clay soil P. balsamifera × P. maximowiczii Henry Elferjani et al. (2013)

Plantation (4, 5, 6, 7) P. deltoides × P. nigra Oklee, MN Fine-sandy loam soil, sandy-clay loam soil Coleman et al. (2006)

Plantation (11) P. deltoides Bartr. ("eastern cottonwood") Stoneville, MS Alluvial silty-loam soil Blackmon et al. (1979)

Plantation (20) P. caspica Bornm. ("Persian poplar") Guilan Province, Iran P. canadensis × ("Triplo") Fine alluvial silty-loam soil P. canadensis × ("I-45/51") Salehi et al. (2013)

Plantation (1) P. trichocarpa Torr. & Gray Qualicum Beach, BC, Canada × P. deltoides Bartr. ex Marsh. Marine-sediment-derived soil P. trichocarpa × P. maximowiczii A. Henry Brown and van den Driessche (2002)

Plantation (4, 6, 9, 12, 15, 16) P. deltoides Bart. ("eastern cottonwood") Lower Mississippi River Valley, AR and MS Alluvial soil, drainage of loessial bluffs Shelton et al. (1981)

Plantation (1, 2, 4) P. deltoides Bartr. ("eastern cottonwood") Stoneville, MS Recently-cleared clayey soil Francis and Baker (1981)

Plantation (3-7, 8-14) P. deltoides Bartr. × P. nigra L. Duero River Basin, Castilla y León, Spain Fluvial terrace soil, alluvial terrain soil Martín-García et al. (2012)

Plantation (4-15) P. tremuloides Mich. × West Kingston, RI - Killingbeck et al. (1990) 68

TABLE 6. Continued from previous page.

Stand Type (Plant Age(s) (yr)) Month(s) of Location Sample Collection

Substrate Description(s)

July

Oct.

May

June

Aug. Nov. Reference Sept. Plant Species Sampled Natural (-) P. nigra L. ("black poplar") Central Spain - Escudero (1992)

Natural (-) P. tremuloides ("trembling aspen") Logan, UT High elevation canyon Tew (1970)

Natural (7) P. trichocarpa ("black cottonwood") Fraser River Watershed, BC, Canada Coastal alluvial soil McClennan (1990)

Natural (19, 48, 73, 85, 100, 105, 108) P. trichocarpa ("black cottonwood") Exit Glacier, Seward, AK Deglaciated floodplain glacial outwash Cusick (2001)

Natural (-) P. alba ("white poplar") Guadiamar River Valley, Seville, Spain Riparian sandy-loam soil Ciadamidaro et al. (2014)

Foliar concentrations of Fe, Zn, and Cu, nutrients with intermediate to low mobility, are especially variable during the summer months, and no consistent seasonal trends are distinguishable (Figure 4). In Populus, Verry and Timmons (1977) reported seasonal foliar accumulation of Fe and Cu, while Killingbeck et al. (1990) reported resorption of Cu and accumulation of Fe only. It is unclear whether the lack of a strong pattern in the broader collection of literature values is due to highly variable nutrient concentrations or to the absence of underlying seasonal patterns.

Verry and Timmons (1977) also found that foliar Mg concentrations remained relatively constant throughout the summer. Being a fairly mobile nutrient, Mg would be 69

predicted to be present at lower foliar concentrations during and after leaf senescence. The lack of a seasonal pattern even in the late-season Mg literature values (Figure 4) is somewhat surprising and may be a consequence of the large degree of variation in the values reported, as well as the fact that the extent of Mg resorption is usually less than that of other nutrients

(Reuter et al. 1997).

For any given nutrient, much variation exists within each sampling month, as shown in Figure 4. Some of this variation may be due to the fact that regional and even site-specific differences in growing season length (Verry and Timmons 1977) can influence the exact timing of plant development and senescence. For example, a 2- to 3-wk difference in leafing date has been observed even in immediately adjacent stands of trees (Tew 1970). Even if

“month” were directly related to plant-growth stage regardless of location, this might be too long a time interval to be able to discern temporal nutrient patterns. With changes in nutrient utilization, and particularly due to resorption, foliar nutrient levels can change drastically even over the course of one month (Verry and Timmons 1977).

Previous work suggests that plant nutrient composition reflects environmental and plant physiological interactions (Garten 1978) and that differences in environmental conditions such as climate and precipitation may explain more foliar nutrient variability than that accounted for by differences due to species (Zhang et al. 2012). Therefore, while some of the variation observed in Figure 4 is likely due to differences in plant characteristics such as species and age, environmental factors and substrate characteristics such as water and nutrient availability are likely driving most of the variability. Considering that the nutrient values compiled here were measured in many different Populus species growing under

relatively diverse climate and nutrient regimes (Table 6), the seasonal patterns that emerge for several of the nutrients are still remarkably clear. That such trends are manifested in this assemblage of values is an indication of strong seasonal nutrient regulation within Populus and also confirms that timing of foliar sampling is of central importance to interpretation of foliar nutrient values.

5.1.2 Values Measured in Lake Aldwell Microenvironments

Foliar concentrations of the majority of nutrients measured in Lake Aldwell reservoir and adjacent forest cottonwood are consistent with what would be expected based on the seasonal trends in foliar nutrient levels reported in other Populus studies (Figure 4). Lake

Aldwell foliar N, P, and K values are lower than the majority of those reported in the literature; as Lake Aldwell cottonwood were sampled late in the season, these low values are consistent with the well-reported seasonal decrease of these generally well-reabsorbed nutrients (Tew 1970, Verry and Timmons 1977, Niinemets and Tamm 2005, Liu et al. 2014).

Foliar Ca and Fe values measured in this study are lower than the majority of the literature values that were measured earlier in the season for live leaves, and, for Ca, substantially lower than the senesced-leaf literature values (Figure 4). Calcium (Tew 1970,

Verry and Timmons 1977, Reuter et al. 1997) and Fe (Verry and Timmons 1977, Killingbeck

1990, Reuter et al. 1997) have been found to accumulate in leaf tissue throughout the season; the fact that foliar concentrations of these nutrients at Lake Aldwell are comparatively so low even at the end of the season indicates that plant uptake of Ca and Fe may be low at Lake

Aldwell. Because of the wide range in foliar Mn and Zn values measured at Lake Aldwell, it is difficult to determine whether or not these values are consistent with the seasonal increases

in these nutrients as reported by Verry and Timmons (1977) and as indicated by the general trend in the plotted literature values (Figure 4).

Seasonal trends explicitly reported for foliar Cu levels are inconsistent in the literature, and no clear trends arise from the array of literature values plotted (Figure 4).

Despite this, that foliar Cu at Lake Aldwell is at the very low end of the range in literature values indicates that plant-available Cu may be low in the Lake Aldwell ecosystem. Seasonal trends in foliar Mg concentrations are also relatively inconsistent in the literature. Foliar Mg values measured in this study, including the high value measured in the forest, are all within the broad range in literature values, indicating that plant-available Mg is likely sufficient in all microenvironments.

5.2 Optimal Foliar Macronutrient Values

5.2.1 Growing-Season Concentration Estimates

Foliar N, P, and K concentrations measured at Lake Aldwell in November appear to be somewhat low and Ca and Mg somewhat high when compared with optimal nutrient ranges reported in the literature (Figure 5). Optimal ranges are derived from measurements of growing-season leaves, however, and this study measured late-season leaves. Foliar nutrient concentrations change naturally, and, for many nutrients, relatively predictably, throughout the season (Figure 4). The late-season values from this study were therefore adjusted to yield growing-season estimates (Figure 5) in order to make rigorous comparisons.

Foliar N, P, and K growing-season estimates are almost entirely encompassed within the optimal ranges (Figure 5). Foliar Ca estimates are within optimal ranges as well, and foliar Mg seems to be unchanged. Estimated foliar P from the forest site is slightly lower

than optimal, while estimated foliar Mg is higher. The two highest estimated N values exceed foliar N concentrations reported in any other studies of either natural (Figure 4) or fertilized

(Figure 5) Populus. This suggests that these N estimates may be unrealistic and indicates that there is uncertainty associated with this method of estimating growing-season nutrient levels.

This method of estimating values has some inherent flaws. Plant age affects nutrient utilization (Jones 2012), so using nutrient values from mature plants to adjust for seasonal nutrient change in young Lake Aldwell saplings may have decreased the accuracy of the estimates. Although past studies (Baker and Blackmon 1977, Francis and Baker 1981,

Killingbeck et al. 1990) have reported seasonal nutrient changes as percentages or percentage ranges of growing-season concentrations, it is possible that it may be more appropriate to estimate nutrient fluctuations based on changes in absolute nutrient content, perhaps using absolute masses of nutrients per unit surface area of leaf (Killingbeck et al. 1990). The use of percent changes, instead of absolute changes, to adjust nutrient values may have resulted in overestimates when reported seasonal nutrient changes were large and underestimates when reported fluctuations were small, as evidenced by the unreasonably high N estimates (Figure

5).

Figure 5. Foliar nutrient concentrations (top) and ratios (bottom) of Lake Aldwell P. trichocarpa saplings (this study) compared with optimal values from fertilized Populus species reported in the literature. For Lake Aldwell, the four points represent different microenvironments; late-season measurements were made in November 2014 and translated to growing-season estimates based on seasonal trends in Verry and Timmons (1977). Optimal upper and lower bounds for foliar nutrients in P. maximowiczii hybrids were determined based on plants that exhibited 90% of the maximum growth rate, measured by diameter at breast height (René et al. 2013); missing values were not reported in the referenced study because they could not be determined with an appropriate level of accuracy. Optimal values for P. deltoides were determined in greenhouse plants based on maximal stemwood production under different fertilization treatments (Leech and Kim 1981). The lowest K/Ca growing-season estimate data point represents over- plotted values from coarse sediment and adjacent forest microenvironments.

5.2.2 Growing-Season Ratio Estimates

Given the various functions of nutrients in plants and the many ways that nutrients are known to interact, inspecting nutrient data through the lens of ratios can be informative.

Because ratios change as a function of the concentrations of not one but two nutrients, however, interpretations of these values can be somewhat subjective.

With the exception of values from saplings in the fine sediments, estimated growing- season foliar N/P and N/K ratios are higher than those reported as optimal (Figure 5) due to the over-estimated N concentrations. In mature leaves, it has been suggested that foliar N/P ratios of less than 14 indicate N limitation (Wei et al. 2015). By this standard, it does not appear that Lake Aldwell cottonwood saplings are N-limited (Figure 5), but since these values may be over-estimated, the interpretation is less clear.

Estimated foliar Ca/Mg ratios, between approximately 2 and 5, are within the optimal ranges (Figure 5). In foliage from several populations of naturally-established Populus, Tew

(1970) measured consistent foliar Ca/Mg ratios of between 13 and 14 throughout the growing season, even as the absolute foliar concentrations of both elements increased. Plant uptake of

Ca and Mg is not as highly regulated as that of N, P, and K, and the physiological implications of foliar Ca/Mg ratios are less clear than those of ratios between the other macronutrients (Bailey et al. 1997b). It seems that the large difference between optimal values (Figure 5) and those reported in natural Populus (Tew 1970) may be due to differences in substrate nutrient composition, Ca being taken up by plants in quantities far exceeding those required for optimal plant function. Due to this lack of regulation in plant uptake of specific nutrients, and considering the relatively large variability in reported

optimal Ca/Mg ratio values, interpretations of these ratios are somewhat ambiguous and may not be a useful parameter in assessing plant nutrition.

Foliar Ca and Mg are highest and foliar P is lowest in the adjacent forest compared to all Lake Aldwell reservoir microenvironments, yielding estimated P/Ca and P/Mg ratios that are lower than optimal (Figure 5). Estimated ratios between these foliar nutrients are also low in the coarse reservoir sediments (for P/Ca, the value is over-plotted in Figure 5). In the forest, estimated foliar P is only slightly lower than optimal; the substantially higher-than- optimal estimated Mg value seems to be driving down the ratio. Like Ca/Mg ratios themselves, the physiological implications of low P/Ca and P/Mg ratios are difficult to determine (Bailey et al. 1997b). Highly variable resorption patterns for Mg (Figure 4), as well as possible Ca uptake in excess of plant requirements (Fageria 2001), also complicate the interpretation.

Nutrient ratio analyses are often cited (Bailey et al. 1997b, Flückiger and Braun 2003,

Ranade-Malvi 2011) as being superior to those based on nutrient concentrations for revealing nutrient imbalances and for assessing nutrient limitation in plants. Conclusions reached in this study based on comparisons between estimated growing-season ratios and optimal ratios are somewhat equivocal, for several reasons. Most notably, for some nutrients there is more inter-species variability between the reported optimal ratios than between the individual nutrient concentrations themselves (Figure 5), making definitive conclusions difficult.

Secondly, there is uncertainty associated with the estimation method. Lastly, it is especially difficult to interpret ratios between nutrients whose concentrations change in different directions over the course of the growing season, such as N and Ca (Verry and Timmons

1977), and whose relationship lacks specific physiological explanations, such as Ca and Mg

(Bailey et al. 1997b).

5.3 Interpretations and Implications

5.3.1 Interpreting Lake Aldwell Reservoir Foliar Nutrients in Context

Optimal foliar nutrient concentrations reported in the literature (Figure 5) are defined empirically, based on values measured in the leaves of the most productive plants (Leech and

Kim 1981). By this definition, plants whose foliar nutrient concentrations differ from optimal are nutrient-limited. Optimal nutrient ranges are calculated in a similar fashion, based on plants whose productivity is ±10% of maximal. In a plantation environment where plants are produced for biomass (Richards et al. 2010), this meaning of optimal, based on a relatively tight window of productivity, is appropriate. Nutrient imbalance may directly impact profit and must therefore be weighed against the financial costs of nutrient amendment.

The high degree of variability in foliar nutrient levels reported in the Populus literature, even for growing-season leaves, indicates that the nutritional composition of most plants is likely sub-optimal. This is especially the case for non-fertilized plants, but given the range in nutrient levels measured in plants growing under various fertilization regimes

(Coleman et al. 2006, Elferjani et al. 2013), this seems to be true for plants in manipulated environments as well. None of the foliar nutrient values measured in Lake Aldwell are below minima reported in other studies of Populus growing in non-contaminated, non-fertilized environments (Figure 4). Based on these comparisons, Lake Aldwell plants seem to be nutritionally similar to other populations. Compared to optimal foliar nutrient levels measured in Populus in manipulated environments (Figure 5), however, Lake Aldwell

cottonwood nutrition may be less than ideal for the production of biomass. Regardless of how nutrient limitation is defined or assessed, the large degree of variability in the literature values complicates direct comparisons.

Many estimates of Lake Aldwell cottonwood growing-season foliar nutrient levels fall slightly outside of the optimal ranges reported in the DRIS literature (Figure 5). This does not indicate that these saplings lack nutrients in quantities sufficient for growth but rather that their growth could potentially be improved with specific nutrient amendments.

The production of woody biomass itself, which serves various ecosystem functions when it eventually becomes large woody debris, may be important for the development of the dewatered riparian environment. For example, woody debris retains sediment (Helfield and

Naiman 2001) and provides shade for plants (Chenoweth et al. 2011). In a study exploring the potential roles that wildlife may play in ecosystem restoration in dewatered Lake Mills, a disproportionate number of bird-dispersed seeds were found to have been deposited on large woody debris (McCaffery et al. 2015), indicating that woody structures specifically may facilitate natural plant regeneration in the Elwha River reservoirs.

5.3.2 Implications for Ecosystem Rehabilitation

In regard to Lake Aldwell reservoir plant nutrition in the short term, comparisons of this study’s values to those reported in the literature yield several noteworthy results. Based on seasonal trends – both those reported explicitly in the literature and those observed in nutrient value plots (Figure 4) – it appears that availability of Ca, Fe, and Cu may all be low at Lake Aldwell. Out of the remaining nutrients studied, no other deficiencies are indicated by analysis of seasonal trends. Based on comparisons between growing-season estimates and

optimal concentrations (Figure 5), it appears that plant-tissue P and K levels may be slightly low in the coarse sediments and adjacent forest, but it is unclear which nutrient may be limiting; the ratio between these nutrients themselves is optimal.

Achieving maximal rates of biomass production is not the end goal of Elwha River ecosystem restoration; in the ecological context, plant nutrition must be viewed through a different lens. Nutrient imbalance or limitation – especially if only slight – has different implications in a recovering riparian forest environment than in a commercial plantation setting. In the Elwha River ecosystem, plant nutrition affects many trophic levels. Depressed rates of plant growth and increased susceptibility to infestation and disease (Flückiger and

Braun 2003, Ranade-Malvi 2011) may inhibit revegetation of riparian areas, as soil development and plant succession may be slowed. This habitat is crucial for the maintenance of animal populations and species diversity (Walker and del Moral 2003) and for the stability of the landscape.

The nutrient content of foliage and twigs of woody plants, as well as nutrients in sedges, grasses, and forbs (Tew 1970), are important because of the multitude of herbivorous animals that depend on these plants for food. In the case of Populus, although grazing animals such as deer may not be able to reach most of the leaves of mature trees, leaves on ground-level suckers provide a substantial amount of biomass for browse (Tew 1970).

Because the nutritional composition of these animals’ diets depends on the browse available to them, nutrient deficiencies in plants can manifest themselves as nutrient deficiencies in animals (Kubota et al. 1970).

6.0 Future Research

6.1 Substrate Characterizations

This study did not determine the grain size distribution of sediment samples, nor were total or extractable nutrient concentrations measured. It was therefore not possible to quantify relationships between foliar nutrient levels and physical or chemical substrate parameters.

Substrate nutrient content and plant nutrient availability are highly variable between environments (Marschner 1986), and steep nutrient gradients often exist even at small scales

(Walker and del Moral 2009). Determination of physical and chemical substrate parameters would clarify and strengthen the relationships observed in this study. Specifically, analysis of correlations between foliar and substrate nutrient concentrations would aid in distinguishing which of the observed patterns may directly reflect the nutrient composition of the growth medium. Many American river systems are affected by anthropogenic impacts such as pollution; the Elwha River ecosystem is located in a pristine forest where these effects are minimal (Duda et al. 2008). This is an ideal environment in which to carry out such analyses, which would enhance understanding of plant nutrient availability and nutrient cycling in a recently dewatered environment.

6.2 Monitoring Over Time

6.2.1 Seasonal Plant-Tissue Analyses

Due to seasonal nutrient resorption, this study’s analysis of foliar nutrient levels did not provide for robust comparisons of plant nutrition among Lake Aldwell microenvironments or between Lake Aldwell cottonwood and other Populus. Foliar tissue should be sampled throughout the year to provide a more sensitive indication of nutritional

status. In conjunction with foliar analyses, analyses of bud, stem, and root tissue would provide a comprehensive indication of tree nutrition at various stages of seasonal growth.

6.2.2 Tracking of Plant Nutrition to Assess Succession

The Lake Aldwell reservoir sediment environment is in the preliminary stages of succession. A temporal, as opposed to strictly spatial, plant nutrient study would be informative. Data collected over the span of years would be useful for quantifying short- and long-term changes in plant nutrition, and associated changes in plant species community structure, in response to dam removal. A study tracking plant nutrient composition over time could address the following questions:

Does year-to-year variability in growing-season nutrient levels decrease over time,

and if so, how long does it take for these levels to stabilize?

Do growing-season nutrient levels approach those found in the adjacent forest, or

those reported to be optimal for biomass production, or some other values?

How long does it take for competitive exclusion to occur, and do changes in plant

density correlate to changes in plant nutrient levels?

In order to ensure consistency of data over time, as well as to enhance current research in the Elwha River ecosystem, plant-tissue and substrate sampling should be carried out in the permanent plots that have already been established within the dewatered Lake

Aldwell and Lake Mills reservoirs (Chenoweth et al. 2011). As these sites change and develop over time, plant and substrate nutrient levels could be assessed alongside other parameters such as floral and faunal species diversities, and relevant correlations should be explored. In order to account for seasonal nutrient variation, plant-tissue and substrate 81

samples should be collected at consistent times during the spring, summer, and fall seasons.

Plant and substrate nutrient levels within the dewatered reservoir sediments should be compared with those in the adjacent forest.

6.2.3 Stable Isotope Studies

In addition to monitoring plant-tissue nutrient levels over time, it would be informative to specifically investigate isotopic enrichment of plant tissue. The Lake Aldwell ecosystem, along with all riparian environments above the Elwha Dam, were deprived of marine-derived nutrients for a century. As increasing numbers of anadromous fish return to the upper reaches of the Elwha River, the 15N/14N ratio of local flora will provide an indication of the extent to which marine-derived-nutrients are contributing to forest nutrition

(Helfield and Naiman 2001). Within such a study, it would be interesting to measure temporal changes in isotopic enrichment with respect to distance from the river in order to determine how far from the river the marine isotopic signature is detectable and over what timescales significant changes in floral 15N/14N are observed.

6.3 Quantitative Plant Nutritional Assessments

Because foliar nutrients in this study were measured in November long after the onset of leaf senescence, and due to the relatively small number of sampling sites, it was not deemed appropriate to calculate nutrient indices and to perform a rigorous, quantitative DRIS nutritional assessment of Lake Aldwell reservoir black cottonwood. Furthermore, the reliability of the DRIS depends on the use of reference values from the same species as that being assessed; such values specific to black cottonwood were not available in the literature.

If foliar sampling were to take place during the peak of the growing season when nutrient values are directly comparable to reported optimal ranges, foliar nutrient values would provide a better indication of the extent to which nutrient limitation may potentially influence ecosystem recovery. The DRIS has been demonstrated to be effective in plant nutritional assessment. If DRIS reference values for black cottonwood were to be published, the DRIS could be used for a robust assessment of plant nutrition in the Lake Aldwell reservoir.

7.0 Conclusion

7.1 Evaluation of Hypotheses

In this study, foliar micronutrient and macronutrient concentrations of Lake Aldwell reservoir black cottonwood were successfully determined and values were compared among sediment textures. Correlations between nutrient concentrations and environmental variables, as well as correlations among nutrient concentrations, were explored, and foliar nutrient values were compared to those reported in other populations of similar species. This study’s hypotheses are restated and evaluated here.

1) Hypothesis: There will be variation among microenvironments due to

differences in sediment texture. Specifically, there will be an inverse

relationship between foliar nutrient levels and substrate grain size. Because

the prodelta microenvironment is effectively a mixture of sediment textures,

prodelta foliar nutrient concentrations will be intermediate between those

measured in fine and coarse reservoir sediments.

In support of this hypothesis, sediment texture appears to have a significant influence on the foliar concentrations of some of the nutrients studied. Specifically, foliar Mn, K, and P are higher, and foliar Ca is lower, in fine than in coarse sediments (Tables 1 and 2). The hypothesized inverse relationship between foliar nutrients and substrate grain size is therefore supported for foliar Mn, K, and P. For all of the nutrients which vary significantly among microenvironments, the prodelta is never placed in a statistically unique group.

Qualitatively speaking, foliar P, K, Ca, Mn, and Zn at the prodelta are between values

measured in the fine- and coarse-sediment microenvironments, indicating that prodelta foliar nutrients may indeed be intermediate.

2) Hypothesis: There will be positive correlations with sediment moisture

content because nutrient retention and moisture retention are influenced by the

same physical and chemical controls. There will be negative correlations with

stem density due to competition between plants.

Correlation analyses indicate positive relationships between sediment moisture and foliar P, K, and Mn concentrations, which supports this hypothesis. The negative relationships observed between sediment moisture and foliar Ca and Zn do not support this hypothesis. The hypothesized relationship between foliar nutrients and woody plant stem density is not observed. Stem density was determined by conditions present at the time of seed dispersal and colonization, not by competition between plants.

3) Hypothesis: There will be positive correlations between nutrients that

participate in similar functions in plants, because plants require nutrients in

specified ratios and nutrient uptake is a regulated process.

In support of this hypothesis, foliar concentrations of N, P, and K are all positively correlated (Table 4). These macronutrients either are involved in protein synthesis or serve structural roles in organic compounds. Although plant availability of these nutrients may vary between sites, these correlation results indicate that the plant-tissue ratios of these nutrients remain relatively constant, meaning that uptake is controlled so as to achieve nutrient balance. Foliar N is also positively correlated with foliar Cu and Ni; no negative nutrient correlations with N are observed (Table 4). This could indicate that higher N supply 85

at some of the sites is causing increased uptake of micronutrients to meet the demands of N- stimulated growth.

4) Hypothesis: There will be substantial variation between values from this study

and values from the literature because nutrient concentrations are influenced

by numerous environmental controls. In general, however, nutrient values

from this study will be lower than those in stands on other types of substrates,

where nutrient availability is higher.

The majority of literature values from unfertilized Populus in naturally-established stands (Table 6) are from secondary successional environments (Cusick 2001) and stands growing on more well-developed soils (McClennan 1990, Ciadamidaro et al. 2014). In these environments, substantial accumulation of organic material and weathering of primary minerals would have resulted in higher nutrient accessibility in the substrate. The other literature values are from Populus in plantations (Table 6), where mechanisms of primary succession do not dictate nutrient availability but where nutrients would nonetheless be higher than in the recently dewatered Lake Aldwell environment due to accumulation and recycling of organic matter from previous generations of trees.

Taking into account seasonal nutrient resorption based on the timing of foliar sampling (Figure 4), foliar N, P, and K levels at Lake Aldwell are consistent with the literature. Although it seems that foliar Ca, Fe, and Cu are lower in Lake Aldwell cottonwood than in other Populus, this hypothesis does not have substantial support.

5) Hypothesis: Many values from this study will be lower than those reported as

optimal in fertilized populations of Populus due to nutrient limitation. 86

Compared to reported optimal values, foliar N, P, and K are all low as measured at

Lake Aldwell in November (Figure 5). The hypothesis lacks strong support, however, because after adjusting for nutrient resorption, the majority of nutrient estimates are encompassed within the range of optimal values. Notable exceptions are foliar estimates from the adjacent forest saplings, where P is lower than optimal and Mg is higher. The two high estimates of foliar N, from fine and prodelta microenvironments, are likely inaccurate as they fall outside of the range of N values reported in the literature for unfertilized Populus

(Figure 4). The validity of the estimation approach is in question, and many of the nutrient ratio comparisons are relatively ambiguous because of the high degree of inter-species variability between reported optimal values.

7.2 Nutrient Availability and Plant Succession

Although total N levels may vary among sediment textures in the Elwha River ecosystem, foliar concentrations measured in cottonwood provide no evidence that plant N availability differs among the four Lake Aldwell microenvironments studied. This indicates that cottonwood are obtaining N from atmospheric sources. Foliar and root uptake of biologically-reactive N from atmospheric deposition may be important means of N acquisition. Lake Aldwell cottonwood may also be obtaining a substantial quantity of N through associative N2 fixation.

Phosphorus is higher in cottonwood foliage at fine-sediment sites than at coarse- sediment and forest sites. This pattern in foliar P is consistent with relative levels of P in substrates sampled prior to dam removal. At the coarse sites, the rate at which P can be weathered from the low-surface-area sediments is likely constraining P availability, whereas

competition between plants is likely limiting P availability in the forest. Due to their proximity to the forest edge, fine sediments may also have higher concentrations of mycorrhizal fungi than coarse sediments. In association with cottonwood, these fungi may be facilitating P acquisition. A significant quantity of forest-soil P may be contained within the bacterial population. In the absence of robust microbial food webs within the reservoir sediments, this reservoir of P is not yet as accessible to cottonwood.

Availability of Mg, Mn, and Zn may differ between Lake Aldwell reservoir sediments and the adjacent forest. This is indicated by differences in foliar concentrations as well as by inter-nutrient correlations involving these nutrients, many of which become much weaker with the exclusion of forest foliar nutrient data. With respect to Mg and Zn, differences detected in cottonwood foliage are in contrast with the relative levels of these nutrients previously measured in Elwha River reservoir sediments and forest soils.

For several of the nutrients analyzed, findings based on direct measurements of cottonwood foliar tissue are consistent with relative nutrient levels measured in Elwha River sediments. For some of the nutrients, however, this strong connection between substrate and foliar nutrient levels is absent, indicating the need to consider other factors that influence nutrient availability, such as atmospheric deposition, N2 fixation, substrate microorganisms, and water availability.

In the recovering Lake Aldwell sediments, ratios between N, P, and K – nutrients of central importance to the synthesis of new plant tissue – are balanced. Low foliar concentrations of Ca, Fe, and Cu indicate that plant availability of these nutrients may be

lower in the Elwha River sediments than on more well-developed substrates, but there is presently no overwhelming indication of substantial nutrient deficiencies.

In the dewatered Elwha River sediments, nutrient availability will change over the course of succession, as will the causes of nutrient limitation. The adjacent conifer forest, at a later successional stage than the reservoir sediments, provides an indication of one possible resource-supply trajectory. As the plant-P-limited coarse sediments are weathered, P availability will increase until it is eventually once again limited by plant competition.

This study of Lake Aldwell black cottonwood provides a snapshot in time of an early- successional plant species equipped to survive in low-nutrient, drought-stressed environments. Following dam removal, other dewatered sediment environments will likely share these basic characteristics. At Lake Aldwell, substrate physical and chemical parameters do not necessarily indicate the degree to which cottonwood can access nutrients.

Other plant species will access nutrients to different extents in different environments as well.

Nutrient availability will depend on site-specific mineralogical conditions and on species-specific physiological characteristics and mechanisms of nutrient acquisition.

Although nutrient availability may not currently limit Elwha River ecosystem productivity to any considerable extent, it will be central in determining the composition of the plant community that develops and the rate at which riparian ecosystem function is restored.

8.0 References

Aerts, R., and F. S. Chapin. 2000. The mineral nutrition of wild plants revisited: A re-

evaluation of processes and patterns. Advances in Ecological Research 30:1–67.

Ågren, G. I., and M. Weih. 2012. Plant stoichiometry at different scales: Element

concentration patterns reflect environment more than genotype. New Phytologist

194:944–952.

Ashley, J. T. F., K. Bushaw-Newton, M. Wilhelm, A. Boettner, G. Drames, and D. J.

Velinsky. 2006. The effects of small dam removal on the distribution of sedimentary

contaminants. Environmental Monitoring and Assessment 114:287–312.

Augustin, F., D. Houle, C. Gagnon, and F. Courchesne. 2015. Long-term base cation

weathering rates in forested catchments of the Canadian Shield. Geoderma 247:12–

23.

Babbitt, B. 2002. What goes up, may come down. Learning from our experiences with dam

construction in the past can guide and improve dam removal in the future. BioScience

52:656–658.

Bailey, J. S., A. Cushnahan, and J. A. M. Beattie. 1997a. The diagnosis and recommendation

integrated system (DRIS) for diagnosing the nutrient status of grassland swards: II.

Model calibration and validation. Plant and Soil 197:137–147.

Bailey, J. S., J. A. M. Beattie, and D. J. Kilpatrick. 1997b. The diagnosis and

recommendation integrated system (DRIS) for diagnosing the nutrient status of

grassland swards: I. Model establishment. Plant and Soil 197:127–135.

Bailey, J. S., R. A. Dils, R. H. Foy, and D. Patterson. 1997c. The diagnosis and

recommendation integrated system (DRIS) for diagnosing the nutrient status of

grassland swards: III. Practical Applications. Plant and Soil 222:255–262.

Baker, J. B., and B. G. Blackmon. 1977. Biomass and nutrient accumulation in a cottonwood

plantation – 1st growing season. Soil Science Society of America 41:632–636.

Barbour, M. G., J. H. Burk, W. D. Pitts, F. S. Gilliam, and M. W. Scwartz. 1999. Terrestrial

Plant Ecology, 3rd Edition. Benjamin Cummings Publishing, Menlo Park, CA.

Barron, A. R., N. Wurzburger, J. P. Bellenger, S. J. Wright, A. M. L. Kraepiel, and L. O.

Hedin. 2009. Molybdenum limitation of asymbiotic nitrogen fixation in tropical forest

soils. Nature Geoscience 2:42–45.

Beaufils, E. R. 1971. Physiological diagnosis: A guide for improving maize production based

on principles developed for rubber trees. Fertilizer Society of South Africa Journal

1:1–31.

Beaufils, E. R. 1973. Diagnosis and recommendation integrated system (DRIS). A general

scheme of experimentation and calibration based on principles developed from

research in plant nutrition. Soil Science Bulletin 1:1–130.

Berger, K. C. 1962. Micronutrient shortages, micronutrient deficiencies in the United States.

Journal of Agricultural and Food Chemistry 10:178–181.

Binkley, D. 1986. Forest Nutrition Management. John Wiley & Sons, New York, NY.

Binner, I., S. Dultz, and M. K. Schenk. 2015. Phosphorus buffering capacity of substrate

clays and its significance for plant cultivation. Journal of Plant Nutrition and Soil

Science 178:155–164.

Blackmon, B. G., J. B. Baker, and D. T. Cooper. 1979. Nutrient use by three geographic

sources of eastern cottonwood. Canadian Journal of Forest Research 9:532–534.

Bojarczuk, K., L. Karliński, T. Hazubska-Przybył, and B. Kieliszewska-Rokicka. 2015.

Influence of mycorrhizal inoculation on growth of micropropagated Populus ×

canescens lines in metal-contaminated soils. New Forests 46:195–215.

Braatne, J. H., S. B. Rood, and P. E. Heilman. 1996. Life history, ecology, and conservation

of riparian cottonwoods in North America. In R. F. Stettler, H. D. Bradshaw, P. E.

Heilman, and T. M. Hinckley (editors), Biology of Populus and its Implications for

Management and Conservation. NRC Research Press, Ottawa, ON, Canada. Pp. 57-

86.

Bradshaw, H. D., R. Ceulemans, J. Davis, and R. Stettler. 2000. Emerging model systems in

plant biology: Poplar (Populus) as a model forest tree. Journal of Plant Growth

Regulation 19:306–313.

Brady, N. C., and R. R. Weil. 2009. Elements of the Nature and Properties of Soils, 3rd

Edition. Pearson Prentice Hall, Upper Saddle River, NJ.

Brar, M., and G. Sekhon. 1976. Interaction of zinc with other micronutrient cations. Plant

and Soil 45:137–143.

Brown, K. R. and R. van den Driessche. 2002. Growth and nutrition of hybrid poplars over 3

years after fertilization at planting. Canadian Journal of Forest Research 32:226–232.

Burga, C. A., B. Krüsi, M. Egli, M. Wernli, S. Elsener, M. Ziefle, T. Fischer, and C. Mavris.

2010. Plant succession and soil development on the foreland of the Morteratsch

Glacier (Pontresina, Switzerland): Straightforward or chaotic? Flora - Morphology,

Distribution, Functional Ecology of Plants 205:561–576.

Bushaw-Newton, K. L., D. D. Hart, J. E. Pizzuto, J. R. Thomson, J. Egan, J. T. Ashley, T. E.

Johnson, R. J. Horwitz, M. Keeley, J. Lawrence, D. Charles, C. Gatenby, D. A.

Kreeger, T. Nightengale, R. L. Thomas, and D. J. Velinsky. 2002. An integrative

approach towards understanding ecological responses to dam removal: The

Manatawny Creek study. Journal of the American Water Resources Association

38:1581–1599.

Calimpong, C. 2014. Elwha River revegetation 2013: A plant performance study. M.S.

Thesis, University of Washington, Seattle, WA.

Campion, J., and M. Scholes. 2007. Diagnosing foliar nutrient dynamics of Eucalyptus

grandis in KwaZulu-Natal, South Africa, using optimal element ratios and the

diagnosis and recommendation integrated system (DRIS). Southern Hemisphere

Forestry Journal 69:137–150.

Cantwell, M. G., M. M. Perron, J. C. Sullivan, D. R. Katz, R. M. Burgess, and J. King. 2014.

Assessing organic contaminant fluxes from contaminated sediments following dam

removal in an urbanized river. Environmental Monitoring and Assessment 186:4841–

4855.

Cavaliere, E. 2010. Elwha River sediments: Phosphorus dynamics under diverse

environmental conditions. M.S. Thesis, Western Washington University, Bellingham,

WA.

Cavaliere, E., and P. Homann. 2012. Elwha River sediments: Phosphorus characterization

and dynamics under diverse environmental conditions. Northwest Science 86:95–107. 93

Chen, F., T. J. Fahey, M. Yu, and Lu Gan. 2010. Key nitrogen cycling processes in pine

plantations along a short urban-rural gradient in Nanchang, China. Forest Ecology

and Management 259:477–486.

Chenoweth, J., S. A. Acker, and M. L. McHenry. 2011. Revegetation and restoration plan for

Lake Mills and Lake Aldwell. Olympic National Park and the Lower Elwha Klallam

Tribe, Port Angeles, WA. Available online at: http://www.nps.gov/olym/

naturescience/elwha-restoration-docs.htm.

Ciadamidaro, L., E. Madejón, M. Puschenreiter, and P. Madejón. 2013. Growth of Populus

alba and its influence on soil trace element availability. Science of the Total

Environment 454–455:337–347.

Ciadamidaro, L., E. Madejón, B. Robinson, and P. Madejón. 2014. Soil plant interactions of

Populus alba in contrasting environments. Journal of Environmental Management

132:329–337.

Clein, J. S. and J. P. Schimel. 1995. Nitrogen turnover and availability during succession

from alder to poplar in Alaskan taiga forests. Soil Biology and Biochemistry 27:743–

752.

Coleman, M., D. Tolsted, T. Nichols, W. D. Johnson, E. G. Wene, and T. Houghtaling. 2006.

Post-establishment fertilization of Minnesota hybrid poplar plantations. Biomass &

Bioenergy 30:740–749.

Cortese, A. M. 2014. Mycorrhizal availability in the basin of Lake Mills and influence on

colonization and growth of Salix scouleriana under drought stress. M.S. Thesis,

Western Washington University, Bellingham, WA.

Cusick, J. A. 2001. Foliar nutrients in black cottonwood and Sitka alder along a soil

chronosequence at Exit Glacier, Kenai Fjords National Park, Alaska. M.S. Thesis,

University of Alaska, Anchorage, AK.

Diestelfeld, A., I. Cakmak, Z. Peleg, L. Ozturk, A. Yazici, H. Budak, Y. Saranga, and T.

Fahima. 2007. Multiple QTL-effects of wheat Gpc-B1 locus on grain protein and

micronutrient concentrations. Physiologia Plantarum 129:635–643.

Doyle, M. W., E. H. Stanley, C. H. Orr, A. R. Selle, S. A. Sethi, and J. M. Harbor. 2005.

Stream ecosystem response to small dam removal: Lessons from the Heartland.

Geomorphology 71:227–244.

Drechsel, P., and W. Zech. 1991. Foliar nutrient levels of broad-leaved tropical trees – a

tabular review. Plant and Soil 131:29–46. van den Driessche, R. 2000. Phosphorus, copper and zinc supply levels influence growth and

nutrition of a young Populus trichocarpa (Torr. & Gray) × P. deltoides (Bartr. ex

Marsh) hybrid. New Forests 19:143–157.

Duda, J. J., J. E. Freilich, and E. G. Schreiner. 2008. Baseline studies in the Elwha River

ecosystem prior to dam removal: Introduction to the special issue. Northwest Science

82:1–12.

Dunlap, J. M., and R. F. Stettler. 1998. Genetic variation and productivity of Populus

trichocarpa and its hybrids. X. Trait correlations in young black cottonwood from

four river valleys in Washington. Trees 13:28–39.

Egli, M., P. Fitze, and A. Mirabella. 2001. Weathering and evolution of soils formed on

granitic, glacial deposits: Results from chronosequences of Swiss alpine

environments. Catena 45:19–47.

Elferjani, R., A. DesRochers, and F. Tremblay. 2013. DRIS-based fertilization efficiency of

young hybrid poplar plantations in the boreal region of Canada. New Forests 44:487–

508.

Enamorado-Baez, S. M., J. M. Abril, and J. M. Gómez-Guzmán. 2013. Determination of 25

trace element concentrations in biological reference materials by ICP-MS following

different microwave-assisted acid digestion methods based on scaling masses of

digested samples. ISRN Analytical Chemistry 2013:1–14.

Epstein, E. 1972. Mineral Nutrition of Plants: Principles and Perspectives. John Wiley &

Sons, Inc., New York, NY.

Escudero, A., J. M. del Arco, I. C. Sanz, and J. Ayala. 1992. Effects of leaf longevity and

retranslocation efficiency on the retention time of nutrients in the leaf biomass of

different woody species. Oecologia 90:80–87.

Fageria, N. K., V. C. Baligar, and C. A. Jones. 1997. Growth and Mineral Nutrition of Crop

Plants, 2nd Edition. Marcel Dekker, Inc., New York, NY.

Fageria, N. K., and V. C. Baligar. 1999. Growth and nutrient concentrations of common

bean, lowland rice, corn, soybean, and wheat at different soil pH on an inceptisol.

Journal of Plant Nutrition 22:1495–1507.

Fageria, V. D. 2001. Nutrient interactions in crop plants. Journal of Plant Nutrition 24:1269–

1290.

Fenn, M. E., R. Haeuber, G. S. Tonnesen, J. S. Baron, S. Grossman-Clarke, D. Hope, D. A.

Jaffe, S. Copeland, L. Geiser, H. M. Rueth, and J. O. Sickman. 2003. Nitrogen

emissions, deposition, and monitoring in the western United States. BioScience

53:391–403.

Fenn, M. E., C. S. Ross, S. L. Schilling, W. D. Baccus, M. A. Larrabee, and R. A. Lofgren.

2013. Atmospheric deposition of nitrogen and sulfur and preferential canopy

consumption of nitrate in forests of the Pacific Northwest, USA. Forest Ecology and

Management 302:240–253.

Filho, F. de A. A. M. 2004. DRIS: Concepts and applications on nutritional diagnosis in fruit

crops. Scientia Agricola 61:550–560.

Filho, F. de A. A. M. 2005. DRIS and sufficient range approaches in nutritional diagnosis of

“Valencia” sweet orange on three rootstocks. Journal of Plant Nutrition 28:691–705.

Fisher, R. F. 2000. Ecology and Management of Forest Soils, 3rd Edition. John Wiley &

Sons, New York, NY.

Flückiger, W., and S. Braun. 2003. Critical limits for nutrient concentrations and ratios for

forest trees – a comment. In (editors), Empirical Critical Loads for Nitrogen –

Proceedings. Swiss Agency for the Environment, Forests and Landscape, Berne,

Switzerland. Pp. 273–280.

Francis, J. K. and J. B. Baker. 1981. Biomass and nutrient accumulation in a cottonwood

plantation – the first four years. US Department of Agriculture Forest Service

Research Note SO-278. Southern Forest Experiment Station, New Orleans, LA.

Galloway, J. N., J. D. Aber, J. W. Erisman, S. P. Seitzinger, R. W. Howarth, E. B. Cowling,

and B. J. Cosby. 2003. The nitrogen cascade. Bioscience 53:341–356.

Garten, C. T. 1978. Multivariate perspectives on the ecology of plant mineral element

composition. The American Naturalist 112:533–544.

Goldberg, S., E. Scalera, and P. Adamo. 2008. Molybdenum adsorption by volcanic Italian

soils. Communications in Soil Science and Plant Analysis 39:693–706.

Gotelli, N. J., and A. M. Ellison. 2013. A Primer of Ecological Statistics, 2nd Edition.

Sinauer Associates, Sunderland, MA.

Gregory, S., H. Li, and J. Li. 2002. The conceptual basis for ecological responses to dam

removal. BioScience 52:713-723.

Groninger, J., J. Skousen, P. Angel, C. Barton, J. Burger, and C. Zipper. 2007. Mine

reclamation practices to enhance forest development through natural succession.

Forest Reclamation Advisory No. 5. The Appalachian Regional Reforestation

Initiative, Pittsburgh, PA.

Guillemette, T., and A. DesRochers. 2008. Early growth and nutrition of hybrid poplars

fertilized at planting in the boreal forest of western Quebec. Forest Ecology and

Management 255:2981–2989.

Hagen-Thorn, A., I. Callesen, K. Armolaitis, and B. Nihlgard. 2004a. The impact of six

European tree species on the chemistry of mineral topsoil in forest plantations on

former agricultural land. Forest Ecology and Management 195:373–384.

Hagen-Thorn, A., S. Armolaitis, I. Callesen, and I. Stjernquist. 2004b. Macronutrients in tree

stems and foliage: A comparative study of six temperate forest species planted at the

same sites. Annals of Forest Science 61:489–498.

Hagen-Thorn, A., and I. Stjernquist. 2005. Micronutrient levels in some temperate European

tree species: A comparative field study. Trees 19:572–579.

Hagen-Thorn, A., I. Varnagiryte, B. Nihlgard, and K. Armolaitis. 2006. Autumn nutrient

resorption and losses in four deciduous forest tree species. Forest Ecology and

Management 228:33–39.

Hallmark, W., and R. Beverly. 1991. An update in the use of the diagnosis and

recommendation integrated system – review. Journal of Fertilizer Issues 8:74–88.

Hansen, E. A. 1994. A guide for determining when to fertilize hybrid poplar plantations. US

Department of Agriculture Forest Service Research Paper NC-319. North Central

Forest Experiment Station, St. Paul, MN.

Harrington, C. A. 2006. Biology and ecology of red alder. In R. L. Deal and C. A. Harrington

(editors), Red Alder - A State of Knowledge. US Department of Agriculture Forest

Service General Technical Report PNW-669. Pacific Northwest Research Station,

Portland, OR. Pp. 21-54.

Heilman, P. E. 1985. Sampling and genetic variation of foliar nitrogen in black cottonwood

and its hybrids in short rotation. Canadian Journal of Forest Research 15:1137–1141.

Helfield, J. M., and R. J. Naiman. 2001. Effects of salmon-derived nitrogen on riparian forest

growth and implications for stream productivity. Ecology 82:2403–2409.

Henderson, J. A., D. H. Peter, R. D. Lesher, and D. C. Shaw. 1989. Forested plant

associations of the Olympic National Forest. US Department of Agriculture Forest

Service Technical Paper R6 ECOL 001-88. Pacific Northwest Region, Portland, OR.

Houba, V. J. G., T. M. Lexmond, I. Novozamsky, and J. J. van der Lee. 1996. State of the art

and future developments in soil analysis for bioavailability assessment. Science of the

Total Environment 178:21–28.

Huang, W., J. Liu, Y. P. Wang, G. Zhou, T. Han, and Y. Li. 2013. Increasing phosphorus

limitation along three successional forests in southern China. Plant and Soil 364:181–

191.

Hundal, H. S., D. Singh, and K. Singh. 2007. Monitoring nutrient status of guava fruit trees

in Punjab, Northwest India through the diagnostic and recommendation integrated

system approach. Communications in Soil Science and Plant Analysis 38:2117–2130.

Ippolito, J. A., S. W. Blecker, C. L. Freeman, R. L. McCulley, J. M. Blair, and E. F. Kelly.

2010. Phosphorus biogeochemistry across a precipitation gradient in grasslands of

central North America. Journal of Arid Environments 74:954–961.

Jean, M.-E., K. Phalyvong, J. Forest-Drolet, and J.-P. Bellenger. 2013. Molybdenum and

phosphorus limitation of asymbiotic nitrogen fixation in forests of eastern Canada:

Influence of vegetative cover and seasonal variability. Soil Biology and Biochemistry

67:140–146.

Jones, J. J. 2012. Plant Nutrition and Soil Fertility Manual, 2nd Edition. CRC Press, Boca

Raton, FL.

100

Jourand, P., L. Hannibal, C. Majorel, S. Mengant, M. Ducousso, and M. Lebrun. 2014.

Ectomycorrhizal Pisolithus albus inoculation of Acacia spirorbis and Eucalyptus

globulus grown in ultramafic topsoil enhances plant growth and mineral nutrition

while limits metal uptake. Journal of Plant Physiology 171:164–172.

Karliński, L., M. Rudawska, B. Kieliszewska-Rokicka, and T. Leski. 2010. Relationship

between genotype and soil environment during colonization of poplar roots by

mycorrhizal and endophytic fungi. Mycorrhiza 20:315–324.

Killingbeck, K. T., J. May, and S. Nyman. 1990. Foliar senescence in an aspen (Populus

Tremuloides) clone – the response. Canadian Journal of Forest Research 20:1156–

1164.

Killingbeck, K. T. 1996. Nutrients in senesced leaves: Keys to the search for potential

resorption and resorption proficiency. Ecology 77:1716–1727.

Kimmins, J. P. 1997. Forest Ecology: A Foundation for Sustainable Management. Prentice

Hall, Upper Saddle River, NJ.

Knecht, M. R., and A. Goransson. 2004. Terrestrial plants require nutrients in similar

proportions. Tree Physiology 24:447–460.

Knoth, J. L., S.-H. Kim, G. J. Ettl, and S. L. Doty. 2014. Biological nitrogen fixation and

biomass accumulation within poplar clones as a result of inoculations with

diazotrophic endophyte consortia. New Phytologist 201:599–609.

Krinard, R. M. 1979. Five years' growth of pruned and unpruned cottonwood planted at 40-

by 40-foot spacing. US Department of Agriculture Forest Service Research Note SO-

252. Southern Forest Experiment Station, New Orleans, LA.

101

Kubota, J., S. Rieger, and V. A. Lazar. 1970. Mineral composition of herbage browsed by

moose in Alaska. The Journal of Wildlife Management 34:565–569.

Lafrenz, M. D., R. A. Bean, D. Uthman. 2013. Soil ripening following dam removal.

Physical Geography 34:124–135.

Langille, W. M., and K. S. Maclean. 1976. Some essential nutrient elements in forest plants

as related to species, plant part, season and location. Plant and Soil 45:17–26.

Larcher, W. 1975. Physiological Plant Ecology, 1st Edition. Springer Verlag. New York,

NY.

Larcher, W. 1995. Physiological Plant Ecology, 3rd Edition. Springer Verlag. Berlin,

Germany.

Leech, R. H., and Y. T. Kim. 1981. Foliar analysis and DRIS as a guide to fertilizer

amendments in poplar plantations. The Forestry Chronicle 57:17–21.

Li, D.-D., Y.-J. Li, J. Liang, C.-Z. Zhao, H.-J. Yin, C.-Y. Yin, X.-Y. Cheng, and Q. Liu.

2014. Responses of soil micronutrient availability to experimental warming in two

contrasting forest ecosystems in the Eastern Tibetan Plateau, China. Journal of Soils

and Sediments 14:1050–1060.

Liu, C., Y. Liu, K. Guo, S. Wang, and Y. Yang. 2014. Concentrations and resorption patterns

of 13 nutrients in different plant functional types in the karst region of south-western

China. Annals of Botany 113:873–885.

Lteif, A., J. K. Whalen, R. L. Bradley, and C. Camiré. 2008. Diagnostic tools to evaluate the

foliar nutrition and growth of hybrid poplars. Canadian Journal of Forest Research

38:2138–2147.

102

Luyssaert, S., H. Raitio, P. Vervaeke, J. Mertens, and N. Lust. 2002. Sampling procedure for

the foliar analysis of deciduous trees. Journal of Environmental Monitoring 4:858–

864.

Maathuis, F. 2013. Plant Mineral Nutrients: Methods and Protocols. Humana Press, New

York, NY.

Maia, C. E. 2012. Fertilization response likelihood for the interpretation of leaf analyses.

Revista Brasileira de Ciência do Solo 36:437–445.

Maisto, G., D. Baldantoni, A. de Marco, A. Alfani, and A. Virzo de Santo. 2013. Ranges of

nutrient concentrations in Quercus ilex leaves at natural and urban sites. Journal of

Plant Nutrition and Soil Science 176:801–808.

Marschner, H. 1986. Mineral Nutrition of Higher Plants, 1st Edition. Academic Press,

London, UK.

Marschner, H., and I. Cakmak. 1989. High light intensity enhances chlorosis and necrosis in

leaves of zinc, potassium, and magnesium deficient bean (Phaseolus vulgaris) plants.

Journal of Plant Physiology 134:308–315.

Marschner, H. 1995. Mineral Nutrition of Higher Plants, 2nd Edition. Academic Press,

London, UK.

Marschner, H. 2012. Marschner’s Mineral Nutrition of Higher Plants, 3rd Edition. Academic

Press, London, UK.

Martín-García, J., A. Merino, and J. J. Diez. 2012. Relating visual crown conditions to

nutritional status and site quality in monoclonal poplar plantations (Populus ×

euramericana). European Journal of Forest Research 131:1185–1198.

103

McCaffery, R., J. McLaughlin, S. Cendejas-Zarelli, K. Sager-Fradkin, K. Jenkins, and J.

Chenoweth. 2015. Roles of wildlife in lakebed restoration following dam removal. In

Abstracts – 2015 Elwha River Science Symposium, Lake Crescent, WA.

McLennan, D. S. 1990. Spatial variation in black cottonwood (Populus trichocarpa) foliar

nutrient concentrations at 7 alluvial sites in coastal British Columbia. Canadian

Journal of Forest Research 20:1089–1097.

Mellert, K. H., and A. Goettlein. 2012. Comparison of new foliar nutrient thresholds derived

from van den Burg’s literature compilation with established central European

references. European Journal of Forest Research 131:1461–1472.

Michel, J. T., J. M. Helfield, and D. U. Hooper. 2011. Seed rain and revegetation of exposed

substrates following dam removal on the Elwha River. Northwest Science 85:15–29.

Milestone, Inc. n.d. Application Note HPR-FO-47: Tea Leaves. Milestone, Inc., Shelton, CT. del Moral, R., and D. M. Wood. 1993. Early primary succession on the volcano Mount St.

Helens. Journal of Vegetation Science 4:223–234.

Morier, I., C. Guenat, R. Siegwolf, J.-C. Védy, and P. Schleppi. 2008. Dynamics of

atmospheric nitrogen deposition in a temperate calcareous forest soil. Journal of

Environment Quality 37:2012–2021.

Nachtigall, G. R., and A. R. Dechen. 2007a. DRIS norms for evaluating the nutritional state

of apple tree. Scientia Agricola 64:282–287.

Nachtigall, G. R., and A. R. Dechen. 2007b. Testing and validation of DRIS for apple tree.

Scientia Agricola 64:288–294.

104

Naiman, R. J., J. S. Bechtold, T. J. Beechie, J. J. Latterell, and R. van Pelt. 2010. A process-

based view of floodplain forest patterns in coastal river valleys of the Pacific

Northwest. Ecosystems 13:1–31.

Nave, L. E., and P. S. Curtis. 2011. Uptake and partitioning of simulated atmospheric N

inputs in Populus tremuloides - Pinus strobus forest mesocosms. Botany 89:379–386.

Niinemets, Ü., and Ü. Tamm. 2005. Species differences in timing of leaf fall and foliage

chemistry modify nutrient resorption efficiency in deciduous temperate forest stands.

Tree Physiology 25:1001–1014.

Pare, D., Y. Bergeron, and C. Camiré. 1993. Changes in the forest floor of Canadian southern

boreal forest after disturbance. Journal of Vegetation Science 4:811–818.

Pérez, C. A., J. C. Aravena, W. A. Silva, J. M. Enríquez, J. M. Fariña, and J. J. Armesto.

2014. Ecosystem development in short-term postglacial chronosequences: N and P

limitation in glacier forelands from Santa Inés Island, Magellan Strait. Austral

Ecology 39:288–303.

Perry, R. H., and D. Green. 1984. Perry’s Chemical Engineers’ Handbook, 6th Edition.

McGraw-Hill, Inc., New York, NY.

Pietrzykowski, M., B. Woś, and N. Haus. 2013. Scots pine needles macronutrient (N, P, K,

Ca, Mg, and S) supply at different reclaimed mine soil substrates – as an indicator of

the stability of developed forest ecosystems. Environmental Monitoring and

Assessment 185:7445–7457.

Poff, N. L., and D. D. Hart. 2002. How dams vary and why it matters for the emerging

science of dam removal. BioScience 52:659–668.

105

R Core Team. 2014. R: A language and environment for statistical computing. R Foundation

for Statistical Computing, Vienna, Austria. Available online at http://www.R-

project.org/.

Radwan, M. A., and D. S. de Bell. 1988. Nutrient relations in coppiced black cottonwood and

red alder. Plant and Soil 106:171–177.

Raghupathi, H. B., T. Sakthivel, and H. Ravishankar. 2013. Preliminary DRIS ratio norms

for diagnosis of nutrient imbalance in Coorg mandarin. Indian Journal of Horticulture

70:33–36.

Ranade-Malvi, U. 2011. Interaction of micronutrients with major nutrients with special

reference to potassium. Karnataka Journal of Agricultural Science 24:106–109.

Randle, T. J., J. A. Bountry, A. Richie, and K. Wille. 2015. Large-scale dam removal on the

Elwha River, Washington, USA: Erosion of reservoir sediment. Geomorphology

246:709-728.

Rathfon, R. A., and J. A. Burger. 1991. Diagnosis and recommendation integrated system

(DRIS) nutrient norms for Fraser fir Christmas trees. Forest Science 37:998–1010.

Reed, S. C., C. C. Cleveland, and A. R. Townsend. 2008. Tree species control rates of free-

living nitrogen fixation in a tropical rainforest. Ecology 89:2924–2934.

Reed, W. P. 1993. US Department of Commerce Certificate of Analysis: Standard Reference

Material 1515 – Apple Leaves. National Institute of Standards and Technology,

Gaithersburg, MD.

106

René, W., B. Côté, C. Camiré, M. Burgess, and J. W. Fyles. 2013. Development and

application of CVA, DRIS, and CND norms for three hybrids of Populus

maximowiczii planted in southern Quebec. Journal of Plant Nutrition 36:118–142.

Reuter, D. J., J. B. Robinson, and C. Dutkiewicz. 1997. Plant Analysis: An Interpretation

Manual, 2nd Edition. CSIRO Publishing, Collingwood, VIC, Australia.

Richards, A. E., D. I. Forrester, J. Bauhus, and M. Scherer-Lorenzen. 2010. The influence of

mixed tree plantations on the nutrition of individual species: A review. Tree

Physiology 30:1192–1208.

Richards, B. N., and D. I. Bevege. 1972. Principles and practice of foliar analysis as a basis

for crop-logging. Plant and Soil 36:109–119.

Righetti, T. L., O. Alkoshab, and K. Wilder. 1988. Verifying critical values from DRIS

norms in sweet cherry and hazelnut. Communications in Soil Science and Plant

Analysis 19:1449–1466.

Rodríguez-Barrueco, C., C. Miguel, and P. Subramaniam. 1984. Seasonal fluctuations of the

mineral concentration of alder (Alnus glutinosa (L.) Gaertn.) from the field. Plant and

Soil 78:201–208.

Rohdy, S. K. 2013. Soil development and vegetation response to removal of a small dam,

Lassen Volcanic National Park, California. M.S. Thesis, Portland State University,

Portland, OR.

Rood, S. B., and J. M. Mahoney. 1991. The importance and extent of cottonwood forest

decline downstream from dams. In S. B. Rood and J. M. Mahoney (editors), The

107

Biology and Management of Southern Alberta’s Cottonwoods – Proceedings of the

University of Lethbridge Conference. Lethbridge, AB, Canada.

Salehi, A., N. Ghorbanzadeh, and M. Salehi. 2013. Soil nutrient status, nutrient return and

retranslocation in poplar species and clones in northern Iran. iForest – Biogeosciences

and Forestry 6:336–341.

Salisbury, F. B., and C. Ross. 1969. Plant Physiology. Wadsworth Publishing, Belmont, CA.

Sarkar, D. 2008. Lattice: Multivariate Data Visualization with R. Springer, New York, NY.

Sayyad, E., S. M. Hosseini, J. Mokhtari, R. Mahdavi, S. G Jalali, M. Akbarinia, and M.

Tabari. 2006. Comparison of growth, nutrition and soil properties of pure and mixed

stands of Populus deltoides and Alnus subcordata. Silva Fennica 40:27–35.

Schuster, J. L. 2015. Vegetation colonization within exposed reservoirs following dam

removal on the Elwha River, Washington. M.S. Thesis, Eastern Washington

University, Cheney, WA.

Seaman, B. J., F. E. Albornoz, J. J. Armesto, and A. Gaxiola. 2015. Phosphorus conservation

during post-fire regeneration in a Chilean temperate rainforest. Austral Ecology

40:709–717.

Serra, A. P., M. E. Marchetti, D. J. Bungenstab, M. A. G. da Silva, R. P. Serra, F. C. N.

Guimãres, V. D. A. Conrad, and H. S. de Morais. 2013. Diagnosis and

recommendation integrated system (DRIS) to assess the nutritional state of plants. In

M. D. Matovic (editor), Biomass Now – Sustainable Growth and Use. InTech Online.

Pp. 129-146. Available online at http://www.intechopen.com/books/

108

biomass-now-sustainable-growth-and-use/diagnosis-and-recommendation-integrated-

system-dris-to-assess-the-nutritional-state-of-plants (accessed 5 December 2014).

Shafroth, P. B., J. M. Friedman, G. T. Auble, M. L. Scott, and J. H. Braatne. 2002. Potential

responses of riparian vegetation to dam removal. BioScience 52:703–712.

Shelton, M. G., L. E. Nelson, G. L. Switzer, and B. G. Blackmon. 1981. The concentration of

nutrients in tissues of plantation-grown eastern cottonwood (Populus deltoides

Bartr.). Mississippi Agricultural and Forestry Experiment Station Technical Bulletin

106. Mississippi State University, Mississippi State, MS.

Silvester, W. B. 1989. Molybdenum limitation of asymbiotic nitrogen fixation in forests of

Pacific Northwest America. Soil Biology & Biochemistry 21:283–289.

Singh, B., and K. N. Sharma. 2007. Tree growth and nutrient status of soil in a poplar

(Populus deltoides Bartr.)-based agroforestry system in Punjab, India. Agroforestry

Systems 70:125–134.

Singh, D., K. Singh, H. S. Hundal, and K. S. Sekhon. 2012. Diagnosis and recommendation

integrated system (DRIS) for evaluating nutrient status of cotton (Gossipium

hirsutum). Journal of Plant Nutrition 35:192–202.

Srivastava, A. K., and S. Singh. 2008. DRIS norms and their field validation in Nagpur

mandarin. Journal of Plant Nutrition 31:1091–1107.

Stone, E. L. 1990. Boron deficiency and excess in forest trees: A review. Forest Ecology and

Management 37:49–75.

Sviridova, I. K. 1960. Results of a study of the leaching of nitrogen and mineral elements by

rain from the crowns of tree species. Doklady Akademii nauk SSSR 133:706–708.

109

Teng, Y., and V. Timmer. 1993. Growth and nutrition of hybrid poplar in response to

phosphorus, zinc, and gibberellic-acid treatments. Forest Science 39:252–259.

Tew, R. K. 1970. Seasonal variation in the nutrient content of aspen foliage. The Journal of

Wildlife Management 34:475–478.

Thomson, J. R., D. D. Hart, D. F. Charles, T. L. Nightengale, and D. M. Winter. 2005.

Effects of removal of a small dam on downstream macroinvertebrate and algal

assemblages in a Pennsylvania stream. Journal of the North American Benthological

Society 24:192–207.

Valenciano, J., J. Boto, and V. Marcelo. 2011. Chickpea (Cicer arietinum L.) response to

zinc, boron and molybdenum application under field conditions. New Zealand Journal

of Crop and Horticultural Science 39:217–229.

Velinsky, D. J., K. L. Bushaw-Newton, D. A. Kreeger, and T. E. Johnson. 2006. Effects of

small dam removal on stream chemistry in southeastern Pennsylvania. Journal of the

North American Benthological Society 25:569–582.

Verry, E. S., and D. R. Timmons. 1976. Elements in leaves of a trembling aspen clone by

crown position and season. Canadian Journal of Forest Research 6:436–440.

Vitousek, P. M., and R. W. Howarth. 1991. Nitrogen limitation on land and in the sea: How

can it occur? Biogeochemistry 13:87–115.

Vitousek, P. M., and H. Farrington. 1997. Nutrient limitation and soil development:

Experimental test of a biogeochemical theory. Biogeochemistry 37:63–75.

Vitousek, P. M. 1999. Nutrient limitation to nitrogen fixation in young volcanic sites.

Ecosystems 2:505–510.

110

Wali, M. K. 1999. Ecological succession and the rehabilitation of disturbed terrestrial

ecosystems. Plant and Soil 213:195–220.

Walker, L. R., and R. del Moral. 2003. Primary Succession and Ecosystem Rehabilitation.

Cambridge University Press, Cambridge, United Kingdom.

Walker, L. R., and R. del Moral. 2009. Lessons from primary succession for restoration of

severely damaged habitats. Applied Vegetation Science 12:55–67.

Walker, T.W., and J. K. Syers. 1976. The fate of phosphorus during pedogenesis. Geoderma

15:1–19.

Walworth, J. L., and M. E. Sumner. 1987. The diagnosis and recommendation integrated

system (DRIS). In B. A. Stewart (editor), Advances in Soil Science. Springer, New

York, NY. Pp. 149–188.

Wan, K., F. Chen, Y. Tao, and S. Chen. 2009. Nutrient elements in leaves of rare and

endangered species in Wuhan Botanical Garden, China. Journal of Plant Nutrition

32:1914–1940.

Wang, Z. H., M. He, T. Wang, and B. Zhu. 2012. Phosphorus sorption-desorption

characteristics of ditch sediments from different land uses in a small headwater

catchment in the central Sichuan Basin of China. Journal of Mountain Science 9:441–

450.

Warrick, J. A., J. A. Bountry, A. E. East, C. S. Magirl, T. J. Randle, G. Gelfenbaum, A. C.

Ritchie, G. R. Pess, V. Leung, and J. J. Duda. 2015. Large-scale dam removal on the

Elwha River, Washington, USA: Source-to-sink sediment budget analysis.

Geomorphology 246:729-750.

111

Wei, H., C. Xu, L. Ma, W. Wang, J. Duan, and L. Jiang. 2014. Short-term nitrogen (N)-

retranslocation within Larix olgensis seedlings is driven to increase by N-deposition:

Evidence from a simulated N-15 experiment in northeast China. International Journal

of Agriculture and Biology 16:1031–1040.

Wei, S., X. Liu, L. Zhang, H. Chen, H. Zhang, H. Zhou, and Y. Lin. 2015. Seasonal changes

of nutrient levels and nutrient resorption in Avicennia marina leaves in Yingluo Bay,

China. Southern Forests: A Journal of Forest Science 77:237–242.

Wells, A. J., N. J. Balster, S. VanWychen, and J. Harrington. 2008. Differences in

belowground heterogeneity within a restoration of a dewatered reservoir in

southwestern Wisconsin. Restoration Ecology 16:678–688.

Werner, C. M. 2014. Early succession in plant communities: Regrowth and invasion in the

drained Elwha River reservoirs. Senior Thesis, Princeton University, Princeton, NJ.

Wichard, T., B. Mishra, S. C. B. Myneni, J.-P. Bellenger, and A. M. L. Kraepiel. 2009.

Storage and bioavailability of molybdenum in soils increased by organic matter

complexation. Nature Geoscience 2:625–629.

Wild, A. D., and R. D. Yanai. 2015. Soil nutrients affect sweetness of sugar maple sap.

Forest Ecology and Management 341:30–36.

Wing, S. 2014. Reservoir sediment carbon along the Elwha River after dam removal. M.S.

Thesis, University of Washington, Seattle, WA.

Wurzburger, N., J. P. Bellenger, A. M. L. Kraepiel, and L. O. Hedin. 2012. Molybdenum and

phosphorus interact to constrain asymbiotic nitrogen fixation in tropical forests. PLoS

ONE 7(e33710):1–7.

112

Zhang, S.-B., J.-L. Zhang, J. W. F. Slik, and K.-F. Cao. 2012. Leaf element concentrations of

terrestrial plants across China are influenced by taxonomy and the environment.

Global Ecology and Biogeography 21:809–818.

Zhang, W., J. Zhao, F. Pan, D. Li, H. Chen, and K. Wang. 2015. Changes in nitrogen and

phosphorus limitation during secondary succession in a karst region in southwest

China. Plant and Soil 391:77–91.

113

Appendix

TABLE A. Certified and measured macronutrient concentration values for Standard Reference Material 1515 “Apple Leaves” determined by ICP-MS, as well as limits of detection and limits of quantification achieved in ICP-MS analysis.

Macronutrient (g kg-1) Sample P K Ca Mg Certified Value† 1.59 ± 0.11 16.1 ± 0.2 15.26 ± 0.15 2.71 ± 0.08

Lab Quality Control (Unmodified) Mean (n = 1 sample) 1.19 12.59 10.7 1.75 Range 0.16 1.28 1.2 0.19 % of Certified Value 75.1 78.2 69.9 64.7

Lab Quality Control (Homogenized) Sample Value (n = 1 sample) 0.92 10.97 9.9 1.54 % of Certified Value

Limit of Detection (3 x SD of blank) - 0.002 0.031 0.001 Limit of Quantification (10 x SD of blank) - 0.007 0.105 0.003 †values from Reed (1993)

TABLE B. Certified and measured micronutrient concentration values for Standard Reference Material 1515 “Apple Leaves” determined by ICP-MS, as well as limits of detection and limits of quantification achieved in ICP-MS analysis.

Micronutrient (mg kg-1) Sample Fe Mn Zn Cu Ni Mo Certified Value† 83 ± 5 54 ± 3 12.5 ± 0.3 5.64 ± 0.24 0.91 ± 0.12 0.094 ± 0.013

Lab Quality Control (Unmodified) Mean (n = 1 sample) 52.7 37 11 4.44 0.7 0.087 Range 10.7 4 1 1.59 0.1 0.123 % of Certified Value 64 69 84 79 76 92

Lab Quality Control (Homogenized) Sample Value (n = 1 sample) 43.9 36 17 2.91 3.8 0.0 % of Certified Value

Limit of Detection (3 x SD of blank) 1.085 0.041 1.252 0.324 0.085 0.116 Limit of Quantification (10 x SD of blank) 3.617 0.138 4.173 1.081 0.284 0.386 †values from Reed (1993)

114

TABLE C. Certified, noncertified, and measured non-nutrient-element concentration values for Standard Reference Material 1515 “Apple Leaves” determined by ICP-MS, as well as limits of detection and limits of quantification achieved in ICP-MS analysis.

Element (mg kg-1) Sample Ag Al Ba Cd Co Cr Na V Certified Value† - 286 ± 9 49 ± 2 0.013 ± 0.002 - - 24.4 ± 1.2 0.26 ± 0.03 Noncertified Value† - - - - 0.09 0.3 - -

Lab Quality Control (Unmodified) Mean (n = 1 sample) 2.00 142.3 36.37 0.01 0.1 0.15 16.7 0.12 Range 6.56 64.9 3.89 0.01 0.0 0.06 3.5 0.03 % of Certified Value - 49.8 74.2 91.7 - - 68.3 44.2

Lab Quality Control (Homogenized) Sample Value (n = 1) 0.38 115.7 32.50 0.03 0.1 0.22 14.8 0.09 % of Certified Value

Limit of Detection (3 x SD of blank) 0.470 2.211 0.116 0.038 0.021 0.078 1.592 0.026 Limit of Quantification (10 x SD of blank) 1.565 7.370 0.388 0.128 0.071 0.258 5.307 0.087 †values from Reed (1993)

TABLE D. Mean foliar non-nutrient-element concentrations of cottonwood saplings from four microenvironments at the dewatered Lake Aldwell reservoir, November 2014.

Element (mg kg-1) Microenvironment n † Ag Al‡ Ba‡ Cd‡ Co‡ Cr Na‡ V Coarse Sediment 3 1.19 43.3 6.58 0.84 4.1 0.73 25.5 0.08 Prodelta 2 3.26 41.6 6.81 0.59 16.3 0.84 25.5 0.08 Fine Sediment 2 3.97 41.3 3.00 0.66 16.2 0.60 22.4 0.05 Adjacent Forest 1 0.77 65.7 4.66 0.89 3.4 0.59 21.2 0.12 § Standard Deviation (SD) 1.51 7.3 0.98 0.21 4.7 0.25 3.4 0.01 †n = number of sites per micronutrient ‡foliar elemental concentration value determined by ICP-MS was < 10% RSD for quintuplicate sample measurements §SD = √(pooled within-microenvironment variance)

115

TABLE E. Correlation coefficients (Pearson’s r, matrices’ upper rights) and p-values (P, matrices’ lower lefts) from correlation tests on log10-transformed foliar nutrient data from cottonwood saplings at the dewatered Lake Aldwell reservoir, November 2014. Values in the top matrix are from tests performed on all nutrient data (n = 8 sites) and values in the bottom matrix are from tests performed on data excluding foliar nutrient values from the forest (n = 7 sites). r logN 0.83 0.77 -0.67 -0.49 -0.32 0.65 -0.57 0.92 0.90 0.011 logP 0.89 -0.91 -0.66 -0.53 0.83 -0.83 0.67 0.70 0.025 0.003 logK -0.85 -0.38 -0.28 0.72 -0.58 0.70 0.53 0.068 0.002 0.008 logCa 0.58 0.40 -0.84 0.76 -0.57 -0.44 0.214 0.077 0.350 0.128 logMg 0.13 -0.84 0.83 -0.36 -0.50 P 0.446 0.176 0.493 0.321 0.754 logFe -0.09 0.40 -0.08 -0.47 0.083 0.011 0.045 0.009 0.010 0.826 logMn -0.91 0.53 0.45 0.141 0.011 0.127 0.030 0.011 0.323 0.002 logZn -0.33 -0.49 0.001 0.069 0.051 0.139 0.382 0.847 0.173 0.420 logCu 0.80 0.002 0.052 0.180 0.277 0.203 0.243 0.266 0.218 0.016 logNi

r logN 0.79 0.76 -0.59 -0.32 -0.24 0.55 -0.42 0.92 0.88 0.035 logP 0.95 -0.89 -0.39 -0.50 0.73 -0.74 0.64 0.57 0.047 0.001 logK -0.86 -0.50 -0.24 0.82 -0.68 0.68 0.48 0.164 0.008 0.014 logCa 0.48 0.34 -0.81 0.72 -0.51 -0.25 0.480 0.382 0.249 0.274 logMg -0.32 -0.66 0.36 -0.24 -0.01 P 0.599 0.259 0.603 0.457 0.483 logFe 0.13 0.36 -0.01 -0.41 0.204 0.065 0.025 0.026 0.111 0.787 logMn -0.78 0.48 0.11 0.343 0.059 0.095 0.071 0.427 0.424 0.037 logZn -0.15 -0.13 0.004 0.124 0.091 0.244 0.613 0.985 0.277 0.741 logCu 0.79 0.008 0.179 0.272 0.588 0.987 0.358 0.823 0.783 0.033 logNi

116

Figure A. Correlation plots for log10-transformed mean foliar macronutrient (top) and micronutrient (bottom) concentrations for cottonwood saplings from four microenvironments (n = 8 sites) at the dewatered Lake Aldwell reservoir, November 2014. 117