UNIVERSITY OF WISCONSIN – LA CROSSE

Graduate Studies

IMPROVING HABITAT: PERIPHYTON AND MACROINVERTEBRATE

COLONIZATION ON LARGE WOOD IN POOL 8 OF

THE UPPER MISSISSIPPI RIVER

A Manuscript Style Thesis Submitted in Partial Fulfillment of the Requirements for the

Degree of Master of Science, Biology: Aquatic Science

Jeffrey Henderson

College of Science and Health

December, 2020

ABSTRACT Henderson, J. G. Improving habitat: periphyton and macroinvertebrate colonization on large wood in Pool 8 of the Upper Mississippi River. M.S. in Biology: Aquatic Science, December 2020, 62pp. (M. Thomsen and R. Haro)

Large wood is an important restoration tool in aquatic ecosystems as it has a variety of geomorphological and ecological benefits. We have limited knowledge regarding the role of large wood in large rivers as most research focuses on streams and small rivers. This research aims to understand the characteristics and dynamics of large wood in Pool 8 of the Upper Mississippi River while examining environmental conditions that influence periphyton accrual and macroinvertebrate colonization on wood. We surveyed naturally occurring large wood along shorelines of reconstructed islands between September and July and deployed wood substrate samplers across varying flow habitats in June that were sampled biweekly until September. The positive net change in wood abundance (+3.2 pieces) and shift in wood characteristics (unattached, dry, and bare without bark) along this flow gradient suggests that flow dynamic is a primary factor influencing wood abundance and mobilization. dry mass (1.1 mg/cm2), abundance (73 individuals), and richness (3 order taxa) were greatest in higher flow habitats where additional habitat substrate was limited, whereas periphyton accrual was similar across flows. The information we provide on use of wood substrates by periphyton and macroinvertebrates across habitat conditions may be further used to link higher trophic level interactions with large wood. Overall, this study supports that large wood is an effective tool in restoration efforts aimed at improving habitat heterogeneity by increasing available substrate.

iii ACKNOWLEDGEMENTS

I would like to thank, Dr. Meredith Thomsen, for providing immense support

from initial experimental design, to fieldwork, and on through data analysis. In addition, I

would like to acknowledge her help during the writing and editing process while

encouraging me to capture my interests and passion. I would like to thank, Dr. Roger

Haro, for his aquatic expertise and assistance during this study, specifically with

methodology, sample processing, and review. I would like to thank my final graduate

committee members of Dr. Kathi Jo Jankowski of the U.S. Geological Survey and Dr.

Ross Vander Vorste for their significant input and help in all parts of this research and the completion of my thesis.

I am also grateful for the other UWL colleagues and outside collaborators who assisted with initial preparation, fieldwork, and sample processing including Alex Oines,

Dr. Molly Van Appledorn, Dr. Colin Belby, Adam Wright, Kurt Grunwald, Dr. Bonnie

Bratina, Vanessa Czeszynski, Dylan Baldassari, Caleb Williams, and David Holmes. A special thank you to Andy Meier of the U.S. Army Corps of Engineers and his interns for the falling and cutting of the wood used for my samplers, as well as Mary Kaufman of the

La Crosse County Facilities Department for permission to obtain this wood from Goose

Island County Park. I would like to thank Stephen Winter of the Upper Mississippi River

National Wildlife and Fish Refuge in partnership with the U.S. Fish and Wildlife

Services for assistance in obtaining Special Use Permits 32579-19-009 and 32579-20-

020. Finally, I would like to express my gratitude to my funding sources that include the

Upper Midwest Environmental Sciences Center, the Upper Mississippi River Restoration

Program, the UWL River Studies Center, and the UWL Graduate Studies RSEL Program.

iv TABLE OF CONTENTS

PAGE

LIST OF TABLES ...... vii

LIST OF FIGURES ...... viii

INTRODUCTION ...... 1

Large Rivers and Large Wood ...... 1

The Foundation of Riverine Food Webs ...... 3

Large Wood in the Upper Mississippi River ...... 4

METHODOLOGY ...... 7

Study Area ...... 7

Objective I: Natural Wood Surveys ...... 9

Objective II: Wood Substrate Samplers ...... 10

Data Analysis ...... 13

RESULTS ...... 16

Hydrographic Context ...... 16

Natural Wood Surveys ...... 18

Wood Substrate Samplers ...... 19

DISCUSSION ...... 38

Hydrographic Context ...... 38

v Natural Wood on Islands ...... 38

Periphyton ...... 41

Macroinvertebrates ...... 43

Wood in Restoration ...... 46

REFERENCES ...... 50

vi LIST OF TABLES

TABLE PAGE

1. Large wood counts from the repeated surveys performed ...... 18

2. ANOVA table for recorded flow velocities ...... 200

3. ANOVA table for periphyton AFDM ...... 21

4. ANVOA table for arthropod dry mass, abundance, and order richness ...... 24

5. ANOVA table for periphyton biomass accrual and arthropod colonization rates ...... 28

6. Total number of individuals sampled ...... 30

7. Total number of molluscs during the colonization period ...... 37

vii LIST OF FIGURES

FIGURE PAGE

1. Location of sampling sites in Lower Pool 8 of the Upper Mississippi River ...... 8

2. Experimental wood substrate sampler ...... 11

3. Individual wood slice, approximately 14 cm in diameter ...... 11

4. Mean daily discharge data at Winona, MN...... 13

5. Mean daily discharge hydrograph at Winona, MN for water year 2019 ...... 13

6. Recorded flow velocities at different dates and flow groups ...... 19

7. Periphyton AFDM at different dates and flow groups ...... 20

8. Arthropod dry mass at different dates and flow groups ...... 22

9. Periphyton biomass accrual rate at different dates and flow groups ...... 23

10. Arthropod percent community composition during the colonization period ...... 32

11. Arthropod percent trophic habit community composition during colonization...... 34

12. Histograms showing frequency of arthropod dry mass across sites ...... 36

viii INTRODUCTION

Large Rivers and Large Wood

Large rivers are dynamic systems that interact extensively with their surrounding

floodplains. River hydrology and flood dynamics play fundamental roles in controlling

their interaction with the surrounding landscape, thereby impacting the exchange of nutrients and organic matter, the creation of habitat, and characteristics of both aquatic and surrounding terrestrial communities (Bridge 2003). Because of the significant interaction between rivers and their floodplains, floodplain forests are ultimately impacted.

Floodplain forests provide a variety of functions and services and are home to large stands of trees. When these stands or trees reach their longevity or are killed by diseases, they are likely to enter the river. Climatic events such as flooding, wind, and strong storms results in this wood being transported into the water. Large wood - defined as any pieces of natural wood larger than 10 cm in diameter and 1 m in length - often consists of broken segments of a tree such as the branches, trunk, stump, and root wad, or entire trees (Kruetzerweiser et al. 2001, Wohl 2014). Changes to both the river and

floodplain have altered all aspects of large wood processes including its movement and

transport, abundance, and ecological influences.

For the past three centuries, there has been a significant loss of large wood in

rivers (Wohl 2014). Wood has been physically removed from rivers to improve

navigation, prevent potential damage to anthropogenic infrastructure from log jams or

1

floating log rafts, and minimize localized flooding (Wohl 2014, Wohl et al. 2016).

Deforestation and other changes within and surrounding the floodplain have decreased the amount of wood entering river ecosystems (Hering et al. 2000, Dossi et al. 2020).

Additionally, changes in river and floodplain connectivity have impacted the transport of

large wood in rivers. As a result, riverine communities, such has habitat heterogeneity,

sedimentation, and hydrology and hydraulics, have been greatly altered (Hering et al.

2000, Wohl 2014, Wohl 2015).

Large wood has a significant ecological role within riverine ecosystems. Wood

provides habitat for both aquatic and terrestrial organisms, including reptiles, amphibians,

fish, birds, and plants, by acting as a place for shelter, refuge, loafing, feeding, and

growth (Lehtinen et al. 1997, Benke et al. 1985, Steel et al. 2003, USBR 2016, Dossi et

al. 2020). Because of its importance in natural river ecosystems, large wood is frequently

integrated in restoration projects for a range of purposes, including habitat improvement

and erosion control (Janvrin 2011). Common structures include buried wood features,

anchored wood structures, suspended or overhung wood, and wood combined with rock.

The benefits of these structures both geomorphologically and ecologically have

been well documented; they decrease erosion, improve habitat, and increase water

movement in the hyporheic zone (Keller and Swanson 1979, Thomas and Stull 2002,

Abbe and Montgomery 2003, Gregory et al. 2003, Roni et al. 2015). However, the effects

of large wood structures have primarily been studied in streams and small rivers, creating

a knowledge gap when understanding the restoration potential of large wood in large

rivers (Lehitnen et al. 1997, Gregory et al. 2003, Angradi et al. 2009, Wohl 2017).

Furthermore, our understanding is limited in terms of how periphyton and

2 macroinvertebrate communities assemble on new pieces of wood substrate over time, and how conditions such as flow velocity influence the process.

The Foundation of Riverine Food Webs

In riverine ecosystems, periphyton is considered an important form of primary production (McCoy 1978, Jones et al. 1998). Periphyton specifically exists as an

assemblage of algae, bacteria, fungi, microbes, and detritus attached to submerged

substrates and benthic structures. Periphyton, freely floating phytoplankton, and aquatic

macrophytes act as the building blocks for the entire ecosystem by feeding primary

consumers, including an array of macroinvertebrate organisms. These in turn serve as a

food source for higher trophic level organisms (Barbour et al. 1999, Pitt and Batzer 2011,

Ruzycki et al. 2015). Together, periphyton and macroinvertebrates are critical

components to riverine communities, and useful indicators of ecosystem health and

condition (McCoy 1978, Poff et al. 2006).

Large wood is a common substrate for periphyton and macroinvertebrates

(Angradi et al. 2009). Periphyton grows on submerged and floating wood, and

macroinvertebrates colonize wood for a variety of reasons. Herbivores (scraper/grazer)

are attracted to wood substrates and structures where periphyton is present because of the

nutrients within the biofilm (Angradi et al. 2009, Coe et al. 2009, Dossi et al. 2020).

Other macroinvertebrates (e.g., collector-gatherers and collector-filterers) tend to

colonize large wood substrate for physical habitat using holes or cracks within the wood

(Benke et al. 2003, Angradi et al. 2009, Coe et al. 2009, Dossi et al. 2020). Thus, large

wood is an important substrate for both periphyton and macroinvertebrate colonization in

riverine ecosystems (Benke et al. 2003, Dossi et al. 2020). Furthermore, large wood can

3 be used as a restoration tool to improve habitat conditions for periphyton and

macroinvertebrates, benefiting other organisms (Benke et al. 1985, Steel et al. 2003,

USBR 2016, Dossi et al. 2020).

Large Wood in the Upper Mississippi River

The Upper Mississippi River (UMR) is one of the nation’s most important large

river ecosystems. The UMR consists of multiple habitats including a main navigation

channel, secondary and tertiary channels, and back water lakes that provide a diversity of

essential habitat for fish, migratory birds, other wildlife, and plants. The UMR also provides immense economic value through its role in commercial navigation and recreation. Since the construction of locks and dams to facilitate navigation, however, habitats along and within the UMR corridor have been significantly altered Locks and dams created large upstream impounded areas that submerged Islands, eroded shorelines reduced habitat for aquatic vegetation, and altered the interaction of the river with its floodplain (Janvrin 2011).

To mitigate these alterations and prevent future degradation, the U.S. Army Corps of Engineers’ Upper Mississippi River program (UMRR) implements habitat restoration and erosion control projects known as Habitat Rehabilitation and Enhancement Projects

(HREPs) are implemented. Large wood is often integrated in these projects in the UMR for habitat improvement, erosion control, bank stabilization, and island formation.

Although large wood structures are theoretically beneficial, very limited monitoring or assessment of them occurs once installed, therefore, their habitat benefits and longevity are not well known. In addition to constructed wood structures, naturally occurring large wood (e.g., not cut lumber) is also prominent in the UMR and can be easily observed

4 deposited along shorelines, sand bars, and restored islands. The abundance, movement,

and ecological influence of naturally occurring large wood throughout the UMR is not

well understood.

Large wood will likely have an increasingly influential role in the UMR as

floodplain forests change (Knutson and Klaas 1998, Angradi et al. 2010, Ramno 2010,

Bouska et al. 2018). Increased abundances of new wood in the UMR is expected as large

stands of silver maples (Acer saccharinum) reach their longevity and green ash (Fraxinus

pennsylvanica) trees are killed by the invasive emerald ash borer in the UMR floodplain

forests. Additionally, the increasing frequency, duration, and intensity of flood events are

predicted to impact the abundance and movement of large wood. Shoreline and island

erosion will contribute more large wood into the UMR, and increased flow velocities and

depth will mobilize and transport large wood to downstream reaches. Thus, it is essential

to understand how the abundance and condition of large wood in the UMR is influenced

by river environmental conditions.

To begin evaluating the ecological role of large wood in the UMR, we designed a

two-part study. First, we characterized the physical attributes and quantified the

abundance of naturally occurring large wood on constructed islands in UMR to assess

what environmental conditions influence large wood mobilization and change in

abundance. Second, we evaluated how environmental conditions (e.g., flow velocities) affect the colonization dynamics of periphyton and macroinvertebrates on wood substrates in the UMR. Studying the use of large wood by organisms at the foundation of the riverine food web across environmental gradients in the river will allow us to improve the process of identifying ideal sites for large wood structures to benefit upper trophic

5 level organisms. Combining the two components will enhance our knowledge regarding the ecology and restoration potential of large wood in the UMR and other large rivers.

6 METHODOLOGY

Study Area

The Lower Pool 8 Islands HREP is located in the lower reach of Pool 8 of the

Mississippi River, between Lock and Dam 7, in Dresbach, MN, and Lock and Dam 8, in

Genoa, WI (Figure 1). The Lower Pool 8 Islands HREP was a multiphase restoration project with Phase I completed in 1989, Phase II in 2000, and Phase III in 2012. New islands were constructed with the overall goal to restore the main channel, side channels, backwater areas, and wetland habitats for waterfowl, nesting birds, turtles, overwintering fish, and aquatic vegetation. To aid the restoration effort, large wood structures were installed along the shorelines (e.g., loafing structures) and off shore (e.g., rock and wood structures) of the newly constructed islands.

As a result of the Lower Pool 8 Islands HREP, there are varying types of aquatic habitats within close proximity of one another, including the main navigation channel, secondary and tertiary channels, and backwater areas. These habitats exhibit variation in flow velocity, depth, aquatic vegetation, and neighboring forest development. Thus, it is an ideal area for comparing the distribution and abundance of natural wood, and for evaluating the colonization of woody substrates across a range of environmental conditions found in the UMR.

7

Figure 1. Location of sampling sites in Lower Pool 8 of the Upper Mississippi River.

Eighteen sites were selected based on flow velocities and the presence of observed natural large wood along shorelines. Wood samplers were deployed at all labelled locations and a subset of these were used for natural wood surveys. The red line indicates the main navigation channel.

8 Objective I: Natural Wood Surveys

Repeated surveys (2019-2020) were conducted to quantify the abundance and turnover of naturally occurring large wood at six sites within the Lower Pool 8 Islands

HREP. Two sites each from Phase I, Phase II, and Phase III were selected, to incorporate a variety of island ages and construction techniques. Furthermore, a 2D Hydraulic Model

(Smith et al. 2011) of Pool 8 was used to ensure that sites spanned a range of flow velocity conditions. The sites used for natural wood sampling were C2 – East, C4 – West,

Stoddard – North, Stoddard – South, Horseshoe – East, and Horseshoe (Figure 1). At each site, a survey was completed in September 2019 and then repeated in July 2020.

Large wood located within a belt transect 60 m long and 10 m wide along each island’s shoreline was counted and characterized using in-stream large wood common metrics

(Wohl 2010). The characterization included diameter, length, presence of root wad, jam characteristics, stability class (unattached, wet, dry, anchored, pinned, buried), and wood status (leaves or needles, branches or limbs, bare, with bark, without bark, or rotten). In addition, each piece of large wood was marked with an aluminum forestry tag, nailed to where the diameter was greatest. Finally, a handheld GPS unit (Garmin GPSMAP66i) was used to record the location of each piece of wood. A forest composition survey was also completed at each site, which included the identification and counting of trees with a diameter at breast height greater than 10 cm present within each transect.

The follow up survey in July 2020 involved the same characterization, plus the relocation of tagged wood from the initial survey. Large wood observed within the transect without a tag was counted as a recruitment (i.e., the addition of new wood), while previously tagged wood within the transect that was not relocated was counted as a

9 loss (i.e., transported elsewhere). A net change of natural wood was calculated by taking

the number of pieces found in September, subtracting the number of pieces lost (not

relocated in July 2020), and then adding the number of new pieces found in July 2020.

The percent net change was calculated by dividing the net change by the total initial

wood.

Objective II: Wood Substrate Samplers

To better understand periphyton accrual and macroinvertebrate colonization on large wood in the UMR, experimental wood substrate samplers were deployed in June

2019 (Figure 2). Samplers were constructed using green ash wood from two standing dead green ash trees killed by the invasive emerald ash borer. The bark was removed from this wood using a draw knife because initial field observations indicated most large wood in the UMR was without bark. The wood was cut into slices roughly 14 cm in diameter and 4 cm in width. A hole was drilled in the center and four notches to delineate sampling areas around the circumference were created on each wood slice (Figure 3).

Each sampler was created using five wood slices that were placed onto a 30.5 cm eye bolt and secured using a lock washer and wing nut. Additional rigging hardware and cable were used to tether the sampler to a buoy. Buoys were “camouflaged” with green and brown tape to minimize disturbance by humans.

10

11

Figure 3. Individual wood slice, approximately 14 cm in

diameter. Center hole is for eye bolt rigging. Four

notches along the slice’s circumference determined

Figure 2. Experimental wood substrate sampler. Each sampling areas. Faces of the wood slices were not sampler is constructed from five wood slices and is sampled, only the outer edges. roughly 20 cm long.

Eighteen sites were chosen within the Lower Pool 8 Islands HREP by using a 2D

Hydraulic Model (Smith et al. 2011) of Pool 8 that predicts continuous surface flow

velocities at a range of discharges. The sites ranged from low flow velocity areas (0.0-0.1

m/s) (i.e., backwater) to high flow velocity areas (1.1-1.9 m/s) (i.e., main navigation

channel). In June 2019, samplers were anchored offshore of island sites near naturally

occurring large wood using a concrete cinder block and rope, in water depths between 1

and 2 m (Figure 1). Samplers floated initially but became waterlogged and sank into the

water within one week, and then hung suspended at roughly 20 cm below the water

surface.

Sampling began three weeks after the initial deployment (one week was given for

the sampler to become fully submerged) and continued at two-week intervals during July,

August, and September, totaling 10 weeks. The process of field sampling remained the

same over the five sampling rounds. First, an observation of aquatic macrophytes was

completed upon arrival at each site; vegetation cover was scored as none (0), partial (1)

or present (2). The water depth and flow velocity were measured using a depth pole and a

HACH FH950 handheld flow meter, respectively. When the sampler was retrieved, the

nut and washer were removed so an individual wood slice could be removed without

disturbing the outer edge. The nut and washer were screwed back onto the bolt and the

sampler was placed back into the water. Using forceps, macroinvertebrates were removed

from the outer edge of the removed slice from two randomly selected sampling quadrants

and placed in 70% ethanol (Hauer and Lamberti 2017). A 3.46 cm2 circle stencil, brush, and distilled water were then used to collect a periphyton sample from the outer edge of a

third randomly selected sampling quadrant. The samples were funneled into amber

12

bottles and placed on ice. Lastly, a flexible tape measure was used to measure the width

and length of the two sampling quadrants from which macroinvertebrates were removed.

Periphyton samples were analyzed for ash free dry mass (AFDM) following

methods of Hauer and Lamberti (2017). Lab processing of macroinvertebrates included

identification for abundance and richness following methods of Hauer and Lamberti

(2017). All macroinvertebrates were identified to the family taxonomic level, while

Trichoptera were identified to the genus or species level because of their diverse trophic

habits. Arthropod body metrics (e.g., total body length, head width, etc.) were measured

using a Nikon SMZ camera scope, NIS Elements Imaging Software (F Package), and

Image-Pro Premier Software (Version 9.2). Length-mass power equations were used to

calculate individual dry mass. The measured arthropod metric (x) was used in a y=axb equation where both (a) and (b) were standards obtained from Smock (1980), Benke et al.

(1999), and Saha et al. (2016). Molluscs (Gastropoda and Myoida) were sorted and

identified for abundance.

Data Analysis

Hydrographic data from the Winona, MN Mississippi River Gauge (USGS

053378500) above Navigation Pool 8 was used to better understand river discharge during the water year 2019 and the colonization period of June 26th – August 29th in

comparison to discharges from previous periods of record (POR) spanning years 1950 to

2019.

Mean recorded flow velocities were used to divide sites into three groups: low

flow, medium flow, and high flow. A mean aquatic vegetation presence was calculated

for each flow group at each sample time point based on the scores recorded in field

13 observations. Daily colonization rates were calculated for periphyton biomass accrual, arthropod dry mass, and arthropod abundance for each site and sampling time point by dividing metric by the number of days between each sampling date.

Data analysis was conducted using JMP 15.2.1 (SAS Institute). Differences in mean recorded flow among flow groups were compared with a repeated measures

ANOVA and post hoc Tukey HSD tests. All variables analyzed with ANOVA were transformed (log or root) to improve normality and equality of variances, and models were assessed with residual normal quantile plots. A chi squared test was used to evaluate the relationship between flow groups and the abundance of aquatic vegetation during the colonization period.

Data from the wood characterization component was combined across all sites to document the dominant wood status and stability class of natural large wood located on islands. Chi squared tests were performed to analyze mean wood abundance, mean wood length, and the net change in wood across sites grouped by flow rate.

Repeated-measures ANOVAs with post hoc Tukey HSD tests were used to evaluate the effects of sample date, flow group, and their interaction on the biotic response variables of periphyton AFDM, arthropod dry mass, abundance, and richness, and the accrual and colonization rates for these same measures. Where interaction terms were significant, one-way ANOVAs were used to evaluate differences among flow groups on each date.

Total macroinvertebrate abundance compared to date and flow group was analyzed at the order taxa level and furthermore at the family and genus (Trichoptera) taxa levels. Chi squared tests were completed to confirm significances differences in total

14 macroinvertebrate abundance across flow groups for each identified invertebrate.

Arthropod community composition in relation to trophic habits (i.e., functional feeding groups) from Poff et al. 2006 were compared between date and flow group. Histograms were additionally constructed to represent individual arthropod dry mass frequency across the different flow groups.

15 RESULTS

Hydrographic Context

Due to significant and record setting flooding during Spring 2019, the discharge

of water year 2019 was 229.5% greater than the typical discharge for the period of record

(POR; Figure 4a). The specific discharge during the colonization period (July –

September) on our wood substrate samplers was also greater than that same time period over of the POR (Figure 4b). The discharge generally decreased during the colonization period (Figure 5). River discharge during our initial natural wood survey in September

2019 was 1448.6 m3/s and during our follow up survey in July 2020 river discharge was

1273.1 m3/s.

We grouped study islands into three flow groups as follows: Teardrop/Grassy, C8

– East, C8 – West, C2 – East, Snake Tongue – West Fork, and Snake Tongue – Top were

grouped as low flow (0.0-0.1 m/s), sites; C4 – West, C2 – West, W2, Cant Hook – North,

and Cant Hook – South were grouped as medium flow (0.1-0.3 m/s), and sites; Horseshoe

– West, Horseshoe – Extension, Horseshoe – Tip, Horseshoe – East, and Horseshoe were

grouped as high flow (0.3-0.6 m/s; Figure 1).

The presence of aquatic vegetation differed predictably between flow groups

(X2=10.7, P<0.05), declining as flow increased. Low flow sites had a high abundance of

aquatic vegetation (mean value=2), medium flow sites had partial (mean value=1), and

high flow sites did not have any vegetation present (mean value=0).

16

(a) (b) 10000 ) -1 s 3

1000 Discharge (m

100 P.O.R. WY 2019 P.O.R. for CPCP

Time period

Figure 2. Mean daily discharge data for USGS Gage 05378500 at Winona, MN. (a) The mean daily discharge (m3/s) for the period of record (P.O.R.) using data spanning water years (WY) 1950-2019 and for water year 2019. (b) The mean daily discharge covering dates spanning the colonization period (CP) of the wood substrate sampler experiment for the P.O.R. and for WY 2019.

6000

) 5000 -1 s

3 4000

3000

2000

Discharge (m 1000

0 100 200 300 Water Day Figure 3. Mean daily discharge (m3/s) hydrograph for USGS Gage 05378500 at Winona,

MN for water year 2019. The red lines and double arrow bound the colonization period for the wood substrate samplers.

17

Natural Wood Surveys

A total of 94 pieces of large wood were found during the two island surveys

(September: 65, July: 29). Across all sites 54% of the wood was bare without bark, followed by rotted (28%) and bare with bark (14%). Stability class was primarily unattached and dry (65%) and unattached and wet (26%). 86% of the wood did not have a root wad attached and 97% were not part of a jam.

There was no significance difference in wood abundance across flow groups

(X2=4.09, P>0.05). There was also no significance difference in wood length (m) among flow groups (X2=0.26, P>0.05). At low flow sites there was a loss of 4 pieces and a gain of 9, a loss of 3 pieces and a gain of 12 at medium flow sites, and a loss of 3 and gain of 8 at high flow sites (Table 1.) Overall, there was a mean gain of 3.2 pieces of large wood at each site (3.2 pieces/60 meters). There was no significant difference between net change of wood and flow groups (X2=1.68, P>0.05).

Table 1. Large wood counts from the repeated surveys performed in September 2019 and

July 2020 in Lower Pool 8 of the UMR. Net change is the difference in wood abundance from the two surveys. Mean length (m) and diameter (cm) values were obtained from all

94 pieces.

Wood Survey Wood Size Mean Mean Initial Recaptured New Net Flow Length Diameter Wood Wood Wood Change (m) ± SE (cm) ± SE Low 29 25 9 +5 (17%) 4.0 ± 0.8 19.8 ± 1.9 Medium 20 17 12 +9 (45%) 5.5 ± 0.8 29.8 ± 3.1 High 16 13 8 +5 (31%) 4.1 ± 0.5 16.7 ± 1.1

18 Wood Substrate Samplers

Mean recorded flow velocities confirmed our low, medium, and high flow groups.

Flows were greatest for the high flow group, followed by the medium flow group. The

low flow group had the lowest flows which were 12.5% of the flow seen in high flow

sites and 50% of the flow seen in medium flow sites. (Figure 6, Table 2). Flow velocities

also decreased significantly over time (Figure 6, Table 2). July 17th had the greatest

recorded flows (mean value=0.32 m/s), followed by August 1st (mean value=0.26 m/s).

August 15th (mean value=0.18 m/s) and August 29th (mean value=0.14 m/s) had similar low flows. These changes mirror the overall decrease in discharge during the colonization period (Figure 5).

0.7 High Flow 0.6 Medium Flow 0.5 Low Flow 0.4 0.3 0.2 0.1 Mean Recorded Flow (m/s) 0 14-Jul 24-Jul 3-Aug 13-Aug 23-Aug 2-Sep Date (day-month)

Figure 4. Recorded flow velocities (m/s) at different dates and flow groups during the

sampling period (July 17th – August 29th). Post-hoc Tukey tests indicate High Flow >

Medium Flow > Low Flow (P<0.0001), and July 17th > August 1st > August 15th and

August 29th (P<0.001).

19

Table 2. ANOVA table for recorded flow velocities (m/s) and date and flow.

Treatment df F P Date 3,34.58 37.41 <0.001* Flow 2,13.04 58.18 <0.001* Date*Flow 6,34.53 2.28 0.059

Periphyton AFDM deferred significantly at the different sampling dates, but flow

and the interaction of date and flow did not (Figure 7, Table 3). Sampling date August

29th had the greatest periphyton AFDM (mean value=3.2 mg/cm2), while dates July 17th,

August 1st, and August 15th were all similar (mean value=2.1 mg/cm2; Figure 7).

4 High Flow 3.5 Medium Flow 3 Low Flow )

2 2.5 2

(mg/cm 1.5 1

Mean Periphyton AFDM 0.5 0 14-Jul 24-Jul 3-Aug 13-Aug 23-Aug 2-Sep Date (day-month)

Figure 5. Periphyton AFDM (mg/cm2) at different dates and flow groups during the sampling period (July 17th – August 29th). Post-hoc Tukey tests indicate August 29th >

July 17th, August 1st, and August 15th (P<0.001)

20

Table 3. ANOVA table for periphyton AFDM (mg/cm2) and date and flow.

Treatment df F P Date 3,39 14.97 <0.001* Flow 2,13 1.05 0.376 Date*Flow 6,39 0.91 0.499

Flow had significant effects on arthropod dry mass, arthropod abundance, and

arthropod order richness, but sampling date and the interaction of date and flow did not

(Figure 8, Table 4). Post-hoc Tukey tests indicate that arthropod dry mass was

significantly different across each flow group. High flow sites had a mean dry mass of

1.1 mg/cm2, medium flow sites had a mean dry mass of 0.19 mg/cm2, and low flow sites

had a mean dry mass of 0.05 mg/cm2 (Figure 8a). High flow sites had the greatest abundance (mean value=73), while medium (mean value=12) and low flow (mean value=5) sites were both similar (Figure 8b). Abundance in the high flow group was 5.9 times greater than the abundance in the medium flow group and 14.5 times greater than the abundance in the low flow group. Arthropod richness at the order level was significantly different between high flow (mean value=3) and low flow (mean value=1.5;

Figure 8c).

21

2 High Flow (a) Medium Flow ) 2 1.5 Low Flow

1

0.5 Mass (mg/cm

Mean Arthropod Dry 0 14-Jul 24-Jul 3-Aug 13-Aug 23-Aug 2-Sep Date (day-month)

120 (b) 100 80 60 High Flow 40 Medium Flow Abundance

Mean Arthropod Arthropod Mean 20 Low Flow 0 14-Jul 24-Jul 3-Aug 13-Aug 23-Aug 2-Sep Date (day-month)

4 3.5 (c) 3 2.5 2

Richness 1.5 High Flow 1

Mean Arthropod Medium Flow 0.5 Low Flow 0 14-Jul 24-Jul 3-Aug 13-Aug 23-Aug 2-Sep Date (day-month)

Figure 6. (a) Arthropod dry mass (mg/cm2) at different dates and flow groups during the sampling period (July 17th – August 29th). Post-hoc Tukey tests indicate High Flow >

22 Medium Flow > Low Flow (P<0.001). (b) Arthropod abundance at different dates and flow groups. High Flow > Medium Flow and Low Flow (P<0.001). (c) Arthropod order richness at different dates and flow groups. High Flow > Low Flow (P<0.001).

23

Table 4. ANVOA table for arthropod dry mass (mg/cm2), arthropod abundance, and arthropod order richness and date and flow.

Arthropod Dry Mass Arthropod Abundance Arthropod Richness Treatment df F P df F P df F P Date 3,39 1.53 0.222 3,39 0.75 0.527 3,39 2.48 0.076 Flow 2,13 24.51 <0.001* 2,13 27.55 <0.001* 2,13 6.81 0.001* Date*Flow 6,39 1.21 0.322 6,39 0.98 0.450 6,39 0.52 0.787 24

Sampling date had a significant effect on periphyton biomass accrual rate,

arthropod dry mass colonization rate, and arthropod abundance colonization rate (Figure

9, Table 5). Flow had a significant effect on arthropod dry mass colonization rate, but not

on periphyton biomass accrual rate or arthropod abundance colonization rate (Figure 9,

Table 5). The interaction of date and flow had a significant effect on arthropod dry mass

colonization rate, but not on periphyton biomass accrual rates or arthropod abundance

colonization rates (Figure 9, Table 5).

Periphyton biomass accrual rates were greater between June 26th – July 17th

(mean value=0.11 AFDM/day) and August 15th – August 29th (mean value=0.10 mg of

AFDM/day) than between July 17th – August 1st (mean value=-0.02 mg of

AFDM/cm2/day) and August 1st and August 15th (mean value=-0.01 mg of

AFDM/cm2/day; Figure 9a).

Arthropod dry mass colonization rates were greater between June 26th – July 17th

(mean value=0.02 dry mass/day) and July 17th – August 1st (mean value=0.009 dry mass/day) than the proceeding dates (Figure 9b). Dry mass colonization at high flow sites

(mean value=0.02 dry mass/day) were greater than low flow sites (mean value<0.0001 dry mass/day; Figure 9b). In addition, dry mass colonization rates between June 26th –

July 17th were significantly greater within the high flow group (0.03 dry mass/day) than

the low flow group (0.003 dry mass/day; Figure 9b).

Arthropod abundance colonization rates were similar to dry mass colonization

rates. The period between June 26th – July 17th had the greatest abundance colonization

rate (mean value=1.2 individuals/day; Figure 9c). This rate was 14 times greater than the

25 mean colonization rate of the remainder of the colonization period (July 17th – August

29th; mean value=0.13 individuals/day).

0.2 High Flow (a) 2500 /s) 3 0.15 Medium Flow 2000 0.1 Low Flow 1500 0.05 1000 0 -0.05 500 River Discharge (m

Periphyton Biomass Biomass Periphyton -0.1 0

Accrual Rate (AFDM/day) 14-Jul 24-Jul 3-Aug 13-Aug 23-Aug 2-Sep Date (day-month)

0.05 High Flow 2500 /s) 0.04 Medium Flow 3 (b) 2000 0.03 Low Flow 0.02 * 1500 0.01 1000

mass/day) 0 500 -0.01 Arthropod Mass Dry Arthropod River Discharge (m Colonization Rate Colonization Rate (dry -0.02 0 14-Jul 24-Jul 3-Aug 13-Aug 23-Aug 2-Sep Date (day-month)

26 5 High Flow (c) 2500 /s) 4 3 Medium Flow 2000 3 Low Flow 2 1500 1 1000 0 500 (individuals/day)

Colonization Rate Colonization Rate -1 River Discharge (m Arthropod Abundance Abundance Arthropod -2 0 14-Jul 24-Jul 3-Aug 13-Aug 23-Aug 2-Sep Date (day-month)

Figure 7. (a) Periphyton biomass accrual rate (AFDM/day) at different dates and flow

groups during the colonization period (June 26th – August 29th) underlaid with river

discharge (m3/s). Post-hoc Tukey tests indicate June 26th – July 17th and August 15th –

August 29th > July 17th – August 1st and August 1st – August 15th (P<0.001). (b)

Arthropod dry mass colonization rate (dry mass/day) at different dates and flow groups.

Post-hoc Tukey tests indicate High Flow > Low Flow (P=0.0013), and June 26th – July

17th and July 17th – August 1st > August 1st – August 15th and August 15th – August 29th

(P=0.021). Asterisk defines pair wise difference for sampling date July 17th (P=0.033).

(c) Arthropod abundance colonization rate (individuals/day) at different dates and flow groups. Post-hoc Tukey tests indicate June 26th – July 17th > July 17th – August 1st,

August 1st – August 15th, and August 15th – August 29th (P=0.001).

27

Table 5. ANOVA table for periphyton biomass accrual rate (AFDM/day), arthropod dry mass colonization rate (dry mass/day),

arthropod abundance colonization rate (individuals/day) and date and flow.

Arthropod Dry Mass Arthropod Abundance Periphyton Biomass Accrual Rate Colonization Rate Colonization Rate Treatment df F P df F P df F P

Date 4,52.9 28.01 <0.001* 4,52.85 5.22 0.001* 4,52.57 5.34 0.001*

Flow 2,13.17 0.14 0.871 2,13.8 5.16 0.021* 2,13.49 0.59 0.567

Date*Flow 8,52.89 1.4 0.220 8,52.82 2.32 0.033* 8,52.54 1.46 0.195 28

When combining and molluscs there were 1,959 macroinvertebrates sampled during the colonization period. The macroinvertebrates that were significantly different between flow groups were Amphipoda, Chironomidae, Simuliidae, Baetidae,

Caenidae, Heptageniidae, Gastropoda, Cheumatopsyche, Hydropsyche, Potamyia flava,

and diversa (Table 6).

29

Table 6. Total number of individuals (N=1,959) sampled from the wood substrate

samplers during the colonization period (June 26th – August 29th). Asterisks (*) indicate a chi squared significance across flow groups. The darker shade of color represents high

abundant taxa compared to the overall sample.

Macroinvertebrate Taxa Total # of individuals Low Medium High Order Family Genus Flow Flow Flow Amphipoda* 8 18 2 Coleoptera Elmidae Stenelmis 0 2 2 Diptera Chironomidae* 63 100 114 Diptera Simuliidae* 0 4 193 Ephemeroptera Baetidae* 8 4 242 Ephemeroptera Caenidae* 1 30 3 Ephemeroptera Heptageniidae* 17 41 64 Gastropoda* 102 12 0 Myoida 2 0 0 Plecoptera Pteronarcyidae Pteronarcys 0 0 1 Trichoptera Hydropsychidae Cheumatopsyche* 4 17 16 Trichoptera Hydropsychidae Hydropsyche* 20 33 826 Trichoptera Hydropsychidae Potamyia flava* 0 0 7 Trichoptera Hydroptilidae Hydroptila 1 0 0 Trichoptera Lype diversa* 0 3 0

30 Overall, all flow groups were dominated by Diptera, Ephemeroptera, and

Trichoptera during the colonization period (Figure 10). The low flow group primarily consisted of 50% Diptera, 22% Ephemeroptera, and 20% Trichoptera (Figure 10).

Amphipoda (8%) were also collected (Figure 10). A shift in community composition was highlighted on sampling date August 1st when diptera became the most dominant arthropod (60%) for the remainder of the colonization period at low flow sites. The

medium flow group primarily consisted of 37% Diptera, 31% Ephemeroptera, and 23%

Trichoptera (Figure 10). Amphipoda (8%) and Coleoptera (1%) were also collected

(Figure 10). Sampling date August 29th shows a shift in community composition as

Diptera became the most dominant arthropod (67%). The high flow group primarily

consisted of 57% Trichoptera, 23% Diptera, and 20% Ephemeroptera (Figure 10).

Amphipoda (<1%), Coleoptera (<1%), and Plecoptera (<1%) were also collected (Figure

10). After sampling date August 1st, Diptera percent community decreased and

Trichoptera (57%) and Ephemeroptera (32%) became most prominent.

31 17-JUL 1-AUG 100.00 100.00 80.00 80.00 60.00 60.00 40.00 40.00 20.00 20.00 Percent Community Percent Community Percent 0.00 0.00 Low Medium High Low Medium High Flow Flow Flow Flow Flow Flow Flow Group Flow Group

15-AUG 29-AUG 100.00 100.00 80.00 80.00 60.00 60.00 40.00 40.00 20.00 20.00 Pecent Community Percent Community Percent 0.00 0.00 Low Medium High Low Medium High Flow Flow Flow Flow Flow Flow Flow Group Flow Group

Figure 8. Arthropod percent community composition (N=1,843) during the colonization period (June 26th – August 29th). Colors represent the different arthropod orders sampled on each date.

Arthropods characterized as collector-gatherers are Stenelmis, Chironomidae,

Baetidae, and Caenidae. Hydropsyche, Cheumatopsyche, and Potamyia flava are characterized as collector-filterers. Herbivores (scraper/grazer) included Heptageniidae,

32 Hydroptila, and Lype diversa. Simuliidae is a predator and Amphipoda and Pteronarcys

are both shredders (detritovore).

Initial colonization was diverse in terms of trophic habits, but became dominated

by collector-gatherers and collector-filterers after sampling date August 1st (Figure 11).

The low flow group primarily consisted of 58% collector-gatherers, 19% collector-

filterers, and 15% herbivores (Figure 11). Shredders (8%) were also collected but only on

the last sampling date (August 29th) (Figure 11). The medium flow group primarily consisted of 48% collector-gatherers, 22% collector-filterers, and 19% herbivores (Figure

11). Shredders (8%) and predators (2%) were also collected (Figure 11). Shredders were

not found on the last sampling date (August 29th) and predators were not found after the

second sampling date (August 1st). The high flow group primarily consisted of 57%

collector-filterers, 24% collector-gatherers, and 15% predators (Figure 11). Herbivores

(4%) were also collected (Figure 11). After sampling date August 1st predator trophic

composition decreased to 2% and collector-gatherer composition increased to 34%.

33 17-JUL 1-AUG 100.00 100.00 80.00 80.00 60.00 60.00 40.00 40.00 20.00 20.00 0.00 0.00 Low Medium High Low Medium High

Flow Flow Flow Community Trophic Percent Flow Flow Flow Percent Trhophic Community Trhophic Percent Flow Group Flow Group

15-AUG 29-AUG 100.00 100.00 80.00 80.00 60.00 60.00 40.00 40.00 20.00 20.00 0.00 0.00 Low Medium High Low Medium High

Percent Trophic Community Trophic Percent Flow Flow Flow Flow Flow Flow Percent Trophic Community Trophic Percent Flow Group Flow Group

Figure 9. Arthropod percent trophic habit (functional feeding group) community composition (N=1,843) during the colonization period (June 26th – August 29th). Colors represent trophic habit (Poff et al. 2006).

34 Average dry mass of individuals, a representation of individual invertebrate size,

increased from low to high flow sites. Forty-six percent of arthropods sampled from low

flow sites (N=121) had a dry mass between 0.0001 mg/cm2 and 0.001 mg/cm2 (Figure

12a). 44% of arthropods sampled from medium flow sites (N=251) had a dry mass

primarily between 0.001 mg/cm2 and 0.01 mg/cm2 (Figure 12b). 50 % of arthropods in high flow sites (N=1840) had a dry mass primarily between 0.001 mg/cm2 and 0.01

mg/cm2 and 38% had a dry mass between 0.01 mg/cm2 and 0.1 mg/cm2.

35 Histogram - Low Flow 800 (a) 700 600 500 400 300 Frequency 200 100 0

Arthropod Dry Mass (mg/cm2)

Histogram - Medium Flow Histogram - High Flow

800 (b) 800 (c) 700 700 600 600 500 500 400 400 300 300 Frequency 200 Frequency 200 100 100 0 0

Arthropod Dry Mass (mg/cm2) Arthropod Dry Mass (mg/cm2)

Figure 10. Histograms showing frequency (abundance) of arthropod dry mass (mg/cm) for low flow sites (a), medium flow sites (b), and high flow sites (c). Bin 0.0001 = (0.0-

0.0001), bin 0.001 = (0.00001-0.001), bin 0.01 = (0.001-0.01), bin 0.1 = (-0.01-0.1), and bin 1 = (0.1-1).

Molluscs included both Gastropoda (N=114) and Myoida (N=2; Table 7). Of those sample sizes low flow sites accounted for 102 Gastropoda and 2 Myoida. Medium flow sites contributed 12 Gastropoda. Neither Gastropoda nor Myoida found in high flow

36 sites. Mollusc abundance differed significantly across dates (X2=21.79, P<0.05) and

across flow groups (X2=167.4, P<0.05).

Table 7. Total number of molluscs (N=116) during the colonization period (July 26th –

August 29th). Chi squared tests indicates differences across dates (X2=21.79, P<0.05) and

flow groups (X2=167.4, P<0.05).

Total # of Individuals Date Low Flow Medium Flow High Flow th July 17 9 0 0 st August 1 29 0 0 th August 15 37 6 0 th August 29 29 6 0

37 DISCUSSION

Hydrographic Context

During our study, river discharge was greater than discharges in previous years.

Our flow groups were based on recorded flow velocities, and aquatic vegetation observations showed strong patterns across high, medium, and low flow sites. Aquatic vegetation is important for a variety of reasons (e.g., habitat, refuge, food, and water quality; Moore et al. 2010). The lack of aquatic vegetation at high flow sites may correlate to a lack of overall substrate. Large wood therefore plays a critical role in providing a stable surface for organisms to colonize in higher-flow areas without aquatic vegetation.

Natural Wood on Islands

Our results show an overall positive net change in wood across all flow groups during the period of September 2019 to July 2020. The time between the two sampling dates include the spring season. It is during this season the river floods, increasing river

discharge and depth, along with the likelihood of wood transport and deposition. Water

discharge and depth slowly recede (excluding the occasional precipitation event or

fluctuation from dam operations) during the time from July to September. Because of this it is predicted that large wood pieces are less likely to be mobilized and transported later in the year. The positive net change across the sample sites suggest two possibilities: 1) large wood is being transported from other places within the river and being deposited further downstream (i.e., the lower reach of Pool 8), or 2) more large wood is entering

38

Pool 8. Other factors such as annual variation, shoreline elevation, duration of inundation,

distance from dam, distance from main navigation channel, characteristics of the riparian

community, etc. need to be examined to fully understand the cause of the positive net

change in large wood along island shorelines.

Flow is major factor in wood recruitment and change (Gurnell et al. 2002); however, our findings do not show any differences in the abundance of large wood across sites in contrasting flow environments. The size of large wood pieces also greatly influences wood recruitment and mobility (Gurnell et al. 2002). The size of large wood informs us on how likely the wood will become jammed or stuck within the channel (e.g., minimizing movement) and dictates the water depth required to initiate mobilization

(e.g., larger pieces require greater depths; Gurnell et al. 2002). We did not find any differences in large wood size across our sample sites. The lack of correlation between

flow and either large wood abundance or size is surprising and suggests other factors

such as increased water depth during elevated water events, relative to shoreline elevation

play more critical roles in wood movement. Our sample sites (N=6) and surveyed large

wood pieces (N=94) limit the scope and statistical power of understanding of large wood

recruitment and change throughout the UMR but provide the foundation for future wood

dynamic studies.

More specifically, by examining the characteristics of naturally occurring large

wood we can improve our overall understanding of wood in the UMR. The large wood

surveyed on islands was dominantly bare without bark (i.e., no branches or bark). Our

‘wood status’ characterization provided us with a general idea of the age of the wood,

based on principles put forth by Gurnell et al. (2002) and Wohl et al. (2010). These

39

authors suggest that the absence of branches and bark indicates the wood is not of recent

arrival and has been damaged by previous movements or decay but not so old that it is

considered rotten. They further suggest that the dominant ‘stability class’ of unattached

and dry highlights that the wood did not originate from the island nor that the large wood

pieces were physically placed there. Only three percent of the large wood pieces surveyed

were included in jams, indicating that the recruitment of large wood from previous pieces is minimal along island shorelines. The lack of root wads on large wood pieces surveyed

(14%) suggests that most wood is of smaller size and more easily transported than larger pieces with roots wads attached (Gurnell et al. 2002, Wohl et al. 2010).

These large wood characterizations are important as we have gained a sense of what large wood in the UMR is like in regard to size, age, decay status, and location along shoreline. We have also gained insight to where wood is coming from. Most of the wood is transported to and captured by these islands and not a product of forests on the islands themselves. Thus, these island features have a natural ability to recruit wood and provide at least temporary woody substrate. This understanding can be used to improve the restoration potential of large wood structures used in restoration projects by identifying how naturally occurring large wood corresponds with different types of ecological benefits. For example, Dossi et al. 2020 found wood age to play a critical factor in macroinvertebrate colonization on wood. Macroinvertebrate density and diversity were greater on rotten (aged) wood than fresh (new) wood or artificial

(concrete) substrates (Dossi et al. 2020). Furthermore, our finding that naturally occurring large wood was equally abundant across all of the sites we sampled suggests that wood placement can occur in a variety of ecological contexts and still mimic the

40 effect of natural wood. Finally, the consistent increase in the number of pieces sampled in

each of our sites is congruent with the expectation that large wood is becoming

increasingly abundant on the shorelines of UMR islands, at least within lower pool

reaches.

Periphyton

Periphyton plays a critical role in riverine food webs. Our results indicate that

wood provides additional substrate for periphyton colonization in river environments. In

addition, our results parallel findings from Biggs and Stokseth 1996. Biggs and Stokseth

1996 found periphyton AFDM on benthic substrate was similar across all flow velocities

(mean value=0.0037 AFDM/day); however, at the end of their colonization period (77

days) the low flow sites (<0.3 m/s) had a greater periphyton AFDM than high flow sites

(>0.7 m/s;). Our periphyton AFDM and periphyton biomass accrual rates during our

study were similar across flow groups (mean value=2.3 mg/cm2, 0.045 AFDM/day).

Although accrual rates were greater during our study our finding parallels the initial trend

found by Biggs and Stokseth 1996. This indicates that periphyton is abundant throughout

the UMR and not resource-limited between flow habitats and that flow directly is not a

major factor affecting periphyton accrual on wood.

Previous research has related flow velocity to overall periphyton accrual (Biggs et

al. 1998, Biggs and Stokseth 1996). Periphyton accrual on substrate is predicted to be

greater in lower flow habitats (Biggs et al. 1998). The similarity in periphyton AFDM across flow groups in our study could be a result of herbivory from grazing and scraping macroinvertebrates. Our low flow sites were dominated by Gastropoda, primarily grazing herbivores, which could reduce periphyton AFDM. Because of their significant

41 abundance and trophic habit, we believe periphyton accrual in lower flow habitats was likely influenced or modulated by their presence. This concept is supported at low flow sites between June 26th – August 15th Gastropoda abundance increased and periphyton

AFDM decreased (Figure 7, Table 7). After August 15th Gastropoda abundance decreased and periphyton AFDM increased (Figure 7, Table 7).

Our medium flow sites had fewer Gastropoda individuals than our low flow sites but had an increased flow velocity stress. This habitat was therefore more suitable for

Diptera and Ephemeroptera that feed on periphyton. These macroinvertebrates are not as substantial herbivores as Gastropoda (Biggs et al. 1998, Poff et al. 2006). Again, periphyton accrual may have been controlled by the lack of effective herbivores and increase in flow velocity stress at medium flow sites. Our high flow sites were exposed to the greatest flow velocity stress and were not dominated by periphyton grazing macroinvertebrates hence why periphyton accrual was similar to the other sites. Initially it seemed as if flow did not directly influence periphyton accrual but when examined conjointly with macroinvertebrate trophic habit the similarity in periphyton AFDM across flow groups could be explained.

An increase in overall periphyton for all three flow groups during our last sampling date (August 29th) was found. We did not observe any significant changes in flow velocity or river discharge to contribute the increase to; however, there are several other factors that influence periphyton accrual on large wood, such as light and nutrient availability and water temperature, that were not analyzed in our study. Overall, our results show that periphyton accrual is not simply related to flow but rather reflects top- down effects of macroinvertebrate communities on wood as well.

42 Macroinvertebrates

Arthropod dry mass, abundance, and richness were all greatest at high flow sites throughout the colonization period indicating high flow exposure habitat is more suitable for arthropods on large wood. Studies on macroinvertebrate colonization on benthic substrates in various rivers globally have found similar results further supporting greater invertebrate richness and diversity in higher flow habitats (Degani et al. 1993, Nelson and Lieberman 2002). Degani et al. 1993 reported highest taxa abundance (26 species) in higher flow velocities while Nelson and Lieberman 2002 acknowledged greatest abundance (911 macroinvertebrates) at their higher flow site. A follow up study is needed to determine the maximum flow velocity before a negative response occurs on arthropod colonization (i.e., too much flow could prevent colonization) is observed.

Other studies have more specifically explored macroinvertebrate colonization on large wood in comparison with other substrate types. Angradi et al. 2009 examined macroinvertebrate colonization between woody substrates in the UMR and other benthic substrate along shorelines and found varying community composition and structure between substrate types and varying flow habitats (Angradi et al. 2009). Although limited in scope this study provides further support that large wood is a suitable substrate in large rivers. Additionally, macroinvertebrate density and richness on wood substrate was greater than on surrounding substrates in the middle Mississippi River (USACOE 2004).

Although not examined in our study, this study suggests large wood is just as effective and possibly better than other substrates in promoting macroinvertebrate assemblage

(USACOE 2004).

43 A primary difference between our high flow sites and our low flow sites was the

presence of aquatic vegetation, an additional substrate for macroinvertebrates to colonize.

Because of the limited substrate in high flow areas, arthropods may have colonized our wood substrate samplers for refuge or shelter. Another difference between our sites was the amount of water exposed to the samplers. Samplers deployed in high flow habitats

were potentially exposed to a greater volume of water. Drift downstream is a common

occurrence that happens to all aquatic macroinvertebrates (Boon 1988, Brittain and

Eikeland 1988, Castro et al. 2013). As a result, our wood substrate samplers deployed in

higher flow areas were exposed to more arthropods drifting by. The combination of the

lack of additional substrate and the increased exposure to drift, may have contributed to a

greater arthropod dry mass and abundance in high flow habitats than in lower flow

habitats. There are several other factors that influence macroinvertebrate colonization that

were not examined in our study but should be considered for future studies. These

include environmental conditions such as water temperature, nutrient exposure, and

turbidity. Additionally, biological factors such as life history and predation would

influence arthropod abundances and colonization.

Our findings suggest arthropod colonization rates were less influenced by flow

and more influenced by date (time). While we did not find differences in arthropod

abundance colonization rates across flow groups, we did find arthropod dry mass

colonization rates were greater at high flow sites rather than low flow sites. This supports

the idea that the lack of habitat heterogeneity in high flow areas makes large wood an

appealing substrate for arthropods in return increasing the rate of colonization. A

previous study of macroinvertebrates on wood determined colonization peaks at

44 approximately 2 – 6 weeks (Dossi et al. 2020). Our sampling date of July 17th,

encompassing the initial colonization period of June 26th – July 17th, was at the 3-week mark and had the greatest colonization rate across all flow groups. After July 17th –

August 1st (5 weeks into the colonization period), colonization rates across flow groups

began to decrease. This again supports the idea that large wood is a suitable substrate

when present.

Our mollusc data is opposite from our arthropod data in regard to flow. Gastropod

abundance was greatest at low flow sites, which supports existing literature that many of

these macroinvertebrates prefer lower flow velocities (Pryon and Brown 2013). In

addition, many Gastropods tend to inhabit areas with a greater abundance of aquatic

vegetation which acts as an additional food source and habitat for snails (Pryon and

Brown 2013). This also supports the correlation of our low flow sites having a high

abundance of aquatic vegetation.

Our findings more specifically reveal which macroinvertebrates are most commonly found on large wood. Diptera, Ephemeroptera, and Trichoptera were dominant across all flow groups, similar to the other large wood findings of Angermeir and Karr 1984, Dudley and Anderson 1992, Dossi et al. 2020. These arthropods have trophic habits of collector-gatherers, collector-filterers, herbivores (scraper/grazer), and even predators. Of the dominant trophic habits found, arthropods classified as collector- gatherers and herbivores most likely colonized our wood substrate samplers because of the periphyton availability. The arthropods with trophic habits of collector-filterers and predators most likely colonized the wood for refuge or habitat based on their preferred habit (e.g., burrowers, sprawlers, clingers, etc.). Lype diversa, Hydropsyche, and

45 Cheumatopysche have all been previously documented to be associated with large wood

in rivers (Phillips 1994, Merrit et al. 2019, USACOE 2004). The presence of these

arthropods on our wood substrate samplers helps to confirm our sampling method is

effective in evaluating arthropod colonization of large wood in the UMR.

Our study spanned the months June – August and seasonality may result in

differences in both periphyton accrual and arthropod colonization. Without periphyton,

arthropod abundance on large wood would most likely be lower; however, we cannot determine in the current study whether arthropods seek large wood primarily for a habitat

substrate or for a potential periphyton food source. Acknowledging that arthropods

colonize large wood for both reasons is critical for improving the use of large wood in

restoration projects.

Wood in Restoration

Combining our findings from the natural large wood surveys with our periphyton

and macroinvertebrate findings we can better improve our use of large wood in

restoration projects. When using large wood structures in restoration efforts, it is

important that the overall objective be clearly defined. If the goal of the project is

geomorphologically related (e.g., bank stabilization or erosion control), then the

placement of large wood may differ from a project that was ecologically related (e.g.,

overall habitat improvement). Although the primary purpose of this study was to better

understand the role of large wood ecologically, our findings may highlight additional

benefits to large wood when used with other objectives (e.g., large wood structures

integrated for erosion control will protect shorelines while creating habitat for

organisms).

46 In regard to wood dynamics, our research and findings suggest inundation may be a more important determinant of large wood abundance and mobility than flow rate. As water depth increases, more wood will become mobilized initiating the transport of wood and increasing the abundance of wood being deposited on islands, along shorelines, and to downstream reaches. As river discharge and depth increase from flood events along with dam operations an increasing positive trend in the net change of wood may become more evident and significant. The increase in wood recruitment along island shorelines of reconstructed islands creates additional habitat benefits.

To maximize the ecological benefits of large wood structures integrated in restoration projects the wood should be aged and without bark; however, storing wood and manually removing bark would not be cost effective, therefore letting the river and physical weathering naturally age large wood structures is recommended. Recent findings support this as wood status (e.g., fresh versus rotten) and wood species (e.g., hard versus soft) are found to be important factors in influencing periphyton and increasing macroinvertebrate composition (Dossi et al. 2020). Whether placing large wood structures along the shorelines of newly created islands or along channel margins, it is important to note initial colonization rates were the greatest. The first three to five weeks may be critical for community colonization and dictate future assemblage and development. Further research will be needed to explore large wood colonization over longer time frames.

The location of the large wood structure itself is as equally as important.

Integrating large wood structures in high flow areas increases habitat heterogeneity. As a result, a more abundant macroinvertebrate community is found which acts as a food

47 source for higher trophic level organisms. Although adding large wood structures into high flow areas will increase community diversity, caution is needed. These structures are more vulnerable to degradation and damage and could create potential risks and hazards if installed in navigation channels. If added to high flow areas carefully engineering should be taken to make sure they are secure and can withstand increased river discharge and depth.

Perhaps a more specific examination of trophic interaction should be accounted for when integrating large wood structures in restoration projects. For example, a study of bluegill sunfish (Lepomis macrochirus) in the Pool 7 of the UMR found their diets to primarily consist of chydorids, chironomids, and gastropods (Dewey et al. 2006). These fish are a prey for largemouth bass (Micropterus salmoides), an organism of high management interest. Large wood structures could be placed in low to medium flow habitat areas where chironomids and gastropods are found to colonize wood substrate most abundantly. This placement could result in more food for bluegills, increasing the interest of largemouth bass. The size of arthropods may influence trophic interactions as well. Our findings suggest high flow sites contain larger sized arthropods (Figure 12).

This may be more favorable for upper trophic level organisms. Other trophic and diet studies, combined with our macroinvertebrate findings, could be used to determine the placement of large wood structures integrated in future restoration projects. This process could be applied to both aquatic and terrestrial organisms linking an entire community around the presence of large wood.

Our findings help to solidify the overall importance of large wood in ecosystems while also providing insight to better improve the use and site selection of large wood

48 structures in restoration projects. Not only does large wood influence geomorphology and create habitat for organisms, it acts as an effective substrate for periphyton and macroinvertebrate colonization in the UMR. This naturally occurring substrate will continue to have influential role in the UMR and should be considered a suitable tool for restoration efforts. These efforts should take into account river hydraulics, wood characteristics, and trophic interactions before integrating large wood structures into restoration projects.

49 REFERENCES

Abbe, T. B., & Montgomery, D. R. (1996). Large woody debris jams, channel hydraulics and habitat formation in large rivers. Regulated Rivers: Research & Management, 12, 201–221. https://doi.org/10.1002

Angermeire, P. L., & Karr, J. R. (1984). Relationships between woody debris and fish habitat in a small warmwater stream. Transactions of American Fisheries Society, 113, 716-726. https://doi.org/10.1577/1548-8659

Angradi, T. R., Bolgrien, D. W., Jicha, T. M., Pearson, M. S., Taylor, D. L., & Hill, B. H. (2009). Multispatial-scale variation in benthic and snag-surface macroinvertebrate assemblages in mid-continent US great rivers. Journal of the North American Benthological Society, 28, 122-141. https://doi.org/10.1899/07-149.1

Angradi, T. R., Taylor, D. L., Jicha, T. M., Bolgrien, D. W., Pearson M. S., & Hill, B. H. (2010). Littoral and shoreline wood in mid-continent great rivers (USA). River Research and Applications, 26, 261-278. https://doi.org/10.1002/rra.1257

Barbour, M., Gerritsen, J., Snyder, B., & Stribling, J. (1999). Rapid bioassessment protocols for use in streams and wadable rivers: Periphyton, benthic invertebrates and fish. http://water.epa.gov/scitech/monitoring/rsl/bioassessment

Benke, A. C., Henry, R. L., Gillespie, D. M., & Hunter, R. J. (1985). Importance of Snag Habitat for Production in Southeastern Streams. Fisheries, 10, 8–13. https://doi.org/10.1577/1548-8446(1985)010<0008:IOSHFA>2.0.CO;2

Benke, A. C., Huryn, A. D., Smock, L. A., Wallace, J. B. (1999). Length-mass relationships for freshwater macroinvertebrates in North America with particular reference to the Southeastern United States. Journal of The North American Benthological Society, 18(3), 308-343. https://doi.org/10.2307/1468447

50

Benke, A. C., & Wallace, J. B. (2003). Influence of wood on invertebrate communities in streams and communities. In: Gregory, S. V., Boyer, K. L., Gurnell, A. M. eds. The ecology and management of wood in world rivers. American Fisheries Society Symposium, 37, 149-177. https://www.srs.fs.fed.us/pubs/ja/2010/ja_2010_benke_001.pdf

Biggs, B. F., & Stokseth, S. (1996). Hydraulic habitat suitability for periphyton in river. Regulated Rivers: Research and Management, 12, 251-261. Biggs, B. F., Stevenson, R. J., & Lowe, R. L. (1998). A habitat matrix conceptual model for stream periphyton. Archiv für Hydrobiologie, 143(1), 25-56. https://doi.org /10.1127/archiv-hydrobiol/143/1998/21

Boon, P.J. (1988), The impact of river regulation on invertebrate communities in the U.K. Regulated Rivers: Research and Management, 2, 389-409. https://doi.org/10.1002/rrr.3450020314

Bouska, K. L., Houser, J. N., De Jager, N. R., & Hendrickson, J. (2018). Developing a shared understanding of the Upper Mississippi River: the foundation of an ecological resilience assessment. Ecology and Society, 23(2), 6. https://doi.org/10.5751/ES-10014-230206

Bridge J. S. (2003). Rivers and floodplains: Forms, processes, and sedimentary record. Wiley-Blackwell.

Brittain, J. E., & Eikeland, T. J. (1988). Invertebrate drift — A review. Hydrobiologia, 166, 77–93. http://doi.org/10.1007/bf00017485

Castro, D. M. P., Hughes, R. M., Callisto, M. (2013). Effects of flow fluctuations on the daily and seasonal drift of invertebrates in a tropical river. Annales de Limnologie - International Journal of Limnology, 49, 169-177. http://doi.org/10.1051/limn/2013051

Coe, H. J., Kiffney, P. M., Pess, G. R., Kloehn, K. K. & McHenry, M. L. (2009). Periphyton and invertebrate response to wood placement in large pacific coastal rivers. River Research and Applications, 25, 1025–1035. Coe, H. J., Kiffney, P. M., Pess, G. R., Kloehn, K. K., & McHenry, M. L. (2009). http://doi.org/10.1002/rra.1201

51

Degani, G., Herbst, G. N., Ortal, R., Bromley, H. J., Levanon, D., Netzer, Y., Harari, N., & Glazman, H. (1993). Relationship between currently velocity, depth and the invertebrate community in stable river system. Hydrobiologia, 263, 163-172. http://doi.org/10.1007/BF00006267

Dewey, M., Richardson, W., & Zigler, S. (2006). Patterns of foraging and distribution of bluegill sunfish in a Mississippi River backwater: Influence of macrophytes and predation. Ecology of Freshwater Fish, 6, 8-15. http://doi.org/10.1111/j.1600- 0633.1997.tb00137.x

Dossi, F., Leitner, P. & Graf, W. (2020). Age matters: substrate-specific colonization patterns of benthic invertebrates on installed large wood. Aquatic Ecology, 54, 741–760. http://doi.org/10.1007/s10452-020-09772-y

Dudley, T., & Anderson, N. H. (1982). A survey of invertebrates associated with wood debris in aquatic habitats. Taken from: Technical Paper 6419. Corvallis, OR: Oregon Agricultural Experiment Station. Melanderia. 39. https://andrewsforest.oregonstate.edu/publications/1739

Phillips, E. C. (1994) Habitat preference and seasonal abundance of Trichoptera larvae in Ozark streams, Arkansas. Journal of Freshwater Ecology, 9(2), 91-95. https://doi.org/10.1080/02705060.1994.9664435

Gregory, S. V., Meleason, M. A., Meleason, & D. J, Sobota. (2003). Modeling the Dynamics of Wood in Streams and Rivers. American Fisheries Society Symposium, 37, 315-335. https://andrewsforest.oregonstate.edu/publications/3057

Gurnell, A. M., Piegay, H., Swanson, F. J., & Gregory, S. V. (2002). Large wood and fluvial processes. Freshwater Biology, 47, 601–619. https://doi.org/10.1046/j.1365-2427.2002.00916.x

Hauer, R. F., & Lamberti, G. A. (2017). Methods in stream ecology. Elsevier Inc.

Hering, D., Kail, J., Eckert, S., Gerhard, M., Meyer, E. I., Mutz, M., Reich, M., & Weiss, I. (2000). Coarse woody debris quantity and distribution in Central European Streams. International Review of Hydrobiology, 85(1), 5–23. https://doi.org/10.1002/(SICI)1522-2632(200003)85:1<5:AID-IROH5>3.0.CO;2- X

52 Jones, J. I., Moss, B., & Young J. O. (1998). Interaction between periphyton, nonmoluuscan invertebrates, and fish in standing freshwaters. In: Jeppesen E., Søndergaard M., Søndergaard M., Christoffersen K. (eds) The Structuring Role of Submerged Macrophytes in Lakes. Ecological Studies (Analysis and Synthesis), 131.

Keller, E.A., & Swanson, F.J. (1997). Effects of large organic material on channel form and fluvial processes. Earth Surface Processes, 4, 361-380. https://andrewsforest.oregonstate.edu/publications/565

Knutson, M. G., Klaas, E. E. (1998). Floodplain forest loss and changes in forest community composition and structure in the Upper Mississippi River: A wildlife habitat at risk. Natural Areas Journal, 18, 138–150. https://www.sciencebase.gov/catalog/item/505400abe4b097cd4fcfa4ff

Kreutzweiser, D. P., Good, K. P., & Sutton, T. M. (2001). Large woody debris characteristics and contributions to pool formation in forest streams of the Boreal Shield. Canadian Journal of Forest Research, 35, 1213-1223. https://doi.org/10.1139/x05-053

Lehtinen, R. M., Mundahl, N. D. & Madejczyk, J. C. (1997). Autumn use of woody snags by fishes in backwater and channel border habitats of a large river. Environmental Biology of Fishes, 49, 7-19. https://doi.org/10.1139/x05-053

Pitt, D., & Batzer, D. (2011). Woody debris as a resource for aquatic invertebrates in stream and river habitats of the Southeastern United States: A review. Proceedings of the 2011 Georgia Water Resources Conference held April 2011, at the University of Georgia. https://smartech.gatech.edu/handle/1853/46046

Poff, N., Olden, J., Vieira, N., Finn, D., Simmons, M., & Kondratieff, B. (2006). Functional trait niches of North American lotic : Traits-based ecological applications in light of phylogenetic relationships. Journal of The North American Benthological Society, 25, 730-755. https://doi.org/10.1899/0887-3593

McCoy, G. A. (1978). Water quality: briefing paper on periphyton. United States Geological Survey. https://water.usgs.gov/admin/memo/historical/qw78.10Historical.txt

53 Merritt, R. W., Cummins, K. W., & Berg, M. B. (2019). An introduction to the aquatic insects of North America, Fifth Edition. Hunt Publishing Company.

Moore, M., Romano, S. P, & Cook, T. (2010). Synthesis of Upper Mississippi River System submersed and emergent aquatic vegetation: Past, present, and future. Hydrobiologia, 640, 103-114. https://doi.org/10.1007/s10750-009-0062-9

Nelson, S. M., & Lieberman, D. M. (2002). The influence of flow and other environmental factors on benthic invertebrates in the Sacramento River, U.S.A. Hydrobiologia, 489, 117–129. https://doi.org/10.1023/A:1023268417851

Pyron, M., & Brown, K. M. (2015). Chapter 18: Introduction to mollusca and the class gastropoda. Ecology and General Biology, Fourth Edition. 381-421.

Romano, S. P. (2010). Our current understanding of the Upper Mississippi River System floodplain forest. Hydrobiologia, 640, 115–124. https://doi.org/10.1007/s10750- 009-0063-8

Roni, P., Beechie, T., Pess, G., & Hanson, K. (2015). Wood placement in river restoration: fact, fiction, and future direction. Canadian Journal of Fisheries and Aquatic Sciences, 72, 466–478. https://doi.org/10.1139/cjfas-2014-0344

Ruzycki, E., Axler, R., Henneck, J., & Erickson, J. (2015). Lake Superior periphyton surveillance. Minnesota’s Lake Superior Coastal Program. https://www.dnr.state.mn.us/waters/lakesuperior/index.html

Saha, C., Pramanik, S., Chakraborty, J., Parveen, S., & Aditya, G. (2016). Abundance and body size of the invasive snail Physa acuta occurring in Burdwan, West Bengal, India. Journal of Entomology and Zoology Studies, 4(4), 490-497.

Smith, T. J. II. (2011). Simulation of Spatial and Temporal Trends in Hydrodynamic Conditions of Upper Mississippi River Pool 8. University of Iowa. https://doi.org/10.17077/etd.8iow7dlw

Smock, L. A. (1980). Relationships between body size and biomass of aquatic insects. Freshwater Biology, 10, 375-383. https://doi.org/10.1111/j.1365- 2427.1980.tb01211.x

54 Steel, A.E., Richards, W.H., & Kelsey, K.A. (2003). Wood and wildlife: benefits of river wood to terrestrial and aquatic vertebrates. American Fisheries Society Symposium, 37, 235-247.

Thompson, D. M., & Stull, G. N. (2002). The development and historic use of habitat structures in channel restoration in the United States: The grand experiment in fisheries management. Géographie physique et Quaternaire, 56(1), 45-60. . https://doi.org/10.7202/008604ar

United States Army Corps of Engineers (USACOE). (2004). Final Report: Evaluation of Macroinvertebrate use of Woody Structure and Surrounding Substrate in the Open Portion of the Upper Mississippi River.

United States Bureau of Reclamation and United States Army Engineer Research and Development Center (USBR and USERDC). (2016). National Large Wood Manual: Assessment, Planning, Design, and Maintenance of Large Wood in Fluvial Ecosystems: Restoring Process, Function, and Structure.

Wohl, E., Cenderelli, D. A., Dwire, K. A., Ryan-Burkett, S. E., Young, M. K. 4, & Fausch, K. D. (2010). Large in-stream wood studies: a call for common metrics. Earth Surface Processes and Landforms, 35, 618-625. https://doi.org 10.1002/esp.1966

Wohl, E. (2014). A legacy of absence: Wood removal in US rivers. Progress in Physical Geography: Earth and Environment, 38, 637–663. https://doi.org/10.1177/0309133314548091

Wohl, E. (2015). Of wood and rivers: bridging the perception gap. Wiley Interdisciplinary Reviews: Water, 2, 167–176. https://doi.org/10.1002/wat2.1076

Wohl, E., Bledsoe, B. P., Fausch, K. D., Kramer, N., Bestgen, K. R., and Gooseff, M. N., (2016). Management of large wood in streams: An overview and proposed framework for hazard evaluation. Journal of the American Water Resources Association, 52(2), 315‐ 335. https://doi.org/10.1111/1752-1688.12388

Wohl, E. (2017.) Bridging the gaps: An overview of wood across time and space in diverse rivers. Geomorphology, 279, 3–26. https://doi.org/10.1016/j.geomorph.2016.04.014

55