MICROBIAL EXTRACELLULAR ENZYMES AS INDICATORS OF RIPARIAN
AND UPSTREAM FOREST COVER IN HEADWATER STREAMS

A
THESIS
SUBMITTED TO THE SCHOOL OF GRADUATE STUDIES
of
BLOOMSBURG UNIVERSITY OF PENNSYLVANIA

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR
THE DEGREE OF MASTER OF SCIENCE

PROGRAM IN BIOLOGY
DEPARTMENT OF BIOLOGICAL AND ALLIED HEALTH SCIENCES
BY
Braeden Gonzales

BLOOMSBURG, PENNSYLVANIA
2023

Abstract
Headwater streams flowing through forested landscapes receive much of their
energy from terrestrial sources, consisting of coarse, fine, and dissolved organic matter
(OM). Much of this OM is broken down and used by bacteria and fungi within stream
biofilms. These microorganisms produce extracellular enzymes to facilitate the
catabolism and uptake of carbon and nutrients from OM that cannot be directly
transported across cell membranes. The extracellular enzymatic activities (EEAs) of these
different enzymes can potentially indicate the relative importance of the various organic
compounds fueling these systems. Because spatial cover and composition of surrounding
forests influence the quantity and type of OM fueling these systems, I hypothesized that
changes in upstream and adjacent forest cover would be reflected in biofilm EEAs. I
tested this hypothesis by sampling 46 headwater streams, in both Pennsylvania and New
York, throughout the upper Delaware River Basin. I sampled epilithic biofilms for
biomass, nutrient content, and the activities of seven extracellular enzymes. For each
stream, I also measured total nitrogen, total phosphorus, chromophoric dissolved organic
matter (CDOM), and both the above-stream canopy cover and upstream forest cover.
This allowed me to investigate potential linkages between biofilm EEAs, stream nutrient
and OM properties, and both riparian canopy cover and upstream land cover. Using
proportions of nitrogen- and phosphorus-acquiring enzymes and multiple regressions
approach, I explored relative nutrient and carbon limitations. Above-stream canopy
cover, CDOM, and total phosphorus were found to be predictors of phosphorus vs.
nitrogen limitation, while chlorophyll a, biofilm phosphorus, and a proportion phenol
oxidase to β-D-1,4-glucosidase activity, which indicates relative recalcitrance, were

found to be predictors of carbon vs. nutrient limitation. Above-stream canopy cover was
found to indicate both relative phosphorus vs. nitrogen and relative nutrient vs. carbon
limitation, as well as drive a nutrient spiraling effect correlated with water column total
phosphorus concentrations. Overall, this study demonstrated that enzymes can be used to
model OM dynamics in headwater streams. These dynamics have the potential to further
improve the prediction of the effects forest and canopy cover have on freshwater streams.

Acknowledgements
I would like to thank my committee members, Drs. Steven Rier, Thomas Klinger,
and Lauri Green, for their valuable input, both as professors and as mentors, with special
thanks to Dr. Steven Rier for his role as my thesis advisor, mentor, professor, fieldwork
manager, and pizza provider. I would also like to thank both Mitchell Liddick and Hanna
Martin for their parts in sample and data collection and for their time spent in the field
and laboratory. Without their help, the dataset used for this project would not have been
nearly as large or thorough. I would like to thank Dr. Stefanie Kroll, Tanya Dapkey, and
the rest of those involved at the Academy of Natural Sciences, Drexel University, for
their help while they work towards completing their own project. Lastly, I would like to
thank my professors, friends, and family who have helped me with this project and
subsequent analysis, whether it was through assistance with editing, providing
encouragement and optimism, or furthering my knowledge with their own advice and
experience. The completion of this project would not have been possible without them.
This project was funded in part by the Academy of Natural Sciences of Drexel
University, the National Fish and Wildlife Foundation’s Delaware Watershed
Conservation Fund (Grant #68902), and the Pennsylvania Department of Environmental
Protection’s Growing Greener Fund (Grant #2001912203287).

Table of Contents
Title Page

Abstract

Approval Page

Acknowledgements

Table of Contents

Introduction and Background

Methods

Planning and Field Methods

Laboratory Methods

Statistical Analysis

Results

Correlation Analysis

Vector Analysis

Multiple Regression Analysis

Discussion

References

Appendices

Appendix I: Raw Data

Appendix II: List of Figures and Tables

Appendix III: R Code

Introduction and Background
The world is currently experiencing changes in temperatures and precipitation
patterns due to anthropogenic climate change. The Intergovernmental Panel on Climate
Change (IPCC) found that from 2006 to 2015, the average annual surface temperatures
were 0.87 °C higher than they were from 1850 to 1900. Temperatures may further
increase by 0.2 °C every decade due to anthropogenic sources alone (Masson-Delmotte et
al. 2018, Shukla et al. 2019). Consequently, average stream temperatures have been
rising globally between 0.009 and 0.077 °C annually for the past century (Kaushal et al.
2010), and evidence predicts that colder water streams may begin to rise by as much as 3
°C in the coming years (Mohseni et al. 2003). Rising temperatures can alter ecosystem
processes such as stream metabolism, causing shifts towards heterotrophy (Song et al.
2018). Models indicate that for every 1 °C increase in global stream temperatures, up to
0.0194 Pg of CO2 could be released into the atmosphere (Song et al. 2018). These
increases could also change the composition of microbial communities (Benner and
McArthur, 1988). Higher temperatures have impacted aquatic ecosystems and the
organisms that inhabit them. As temperatures continue to increase due to global change,
more drastic changes can and will occur in these delicate aquatic ecosystems.
Climate change can also alter seasonal precipitation patterns, which can lead to
droughts in some areas and flooding in others (Wehner, 2013, Cai et al. 2014). The IPCC
predicts increases in heavy precipitation in several regions and the probability of drought
and precipitation deficits in other regions following a global temperature increase of 1.5
to 2 °C (Masson-Delmotte et al. 2018). These changes in precipitation can cause
variability in flow that have the potential to damage streams and their ecosystem

functions (Dycus et al. 2015, Baruch et al. 2022, Patil et al. 2022). Not only could the
presence of droughts and floods cause damage to human life, changes in these patterns
can also lead to potential taxonomic shifts (Borba et al. 2020, Burcher et al. 2008,
Mohseni et al. 2003). Taxonomic shifts can alter the community structure of aquatic
ecosystems, which can influence species interactions, ecosystem structure, and ecosystem
function, and can result in a loss in biodiversity. One consequence of taxonomic shifts, a
reduction in biodiversity, can cause systems to undergo species reduction and a loss of
taxa richness, as well as facilitate range shifts (Benner and McArthur, 1988, Mohseni et
al. 2003, Burcher et al. 2008, and Borba et al. 2020). For example, a study found that as
waterway temperatures increase, warm water fish will have access to more thermally
suitable habitats, increasing their ranges, while colder water fish will lose their own
habitats, forcing them to move northwards (Mohseni et al. 2003). This could lead to
ecosystem cascades as food webs and resource availability change in response.
Warming stream temperatures could disproportionally impact the balance
between autotrophy and heterotrophy. As carbon either enters the system through
allochthonous inputs, such as leaf litter, or through autochthonous sources, such as algal
production, it plays a key role in a stream’s metabolism. As global temperatures rise, it is
predicted that streams around the world will begin to shift asymmetrically towards
heterotrophy (Song et al. 2018). Most of this increase in CO 2 output is due to ecosystem
respiration being skewed by increasing temperatures as it is a temperature sensitive
process, which includes heavy influence from the microbial communities within streams
(Lock et al. 1984, Peters and Lodge, 2009, Song et al. 2018).

Changes in global precipitation patterns can also lead to phenomena such as
stream browning (Monteith et al. 2007). Stream browning is the darkening of water when
the concentration of chromophoric dissolved organic matter (CDOM) increases
(Weyhenmeyer et a., 2016). CDOM refers to dissolved particles of organic matter (OM)
that tint waters a browner color, and is usually dominated by humic substances
(Thurman, 2012). Previous studies have shown accelerated browning of freshwater
streams and lakes globally due to increased precipitation events, which wash CDOM into
streams via super surface flow (Monteith et al. 2007). Potential consequences of stream
browning include an excess of greenhouse gases entering Earth’s atmosphere via
enhanced microbial respiration, a decrease in general water quality, and a decline of instream productivity through limited light penetration (de Wit et al. 2016).
Climate change resiliency is the capacity for ecosystems to resist the effects of
global climate change and is a key consideration for managers strategizing ways to
protect streams in the face of climate change (Pörtner et al. 2022). Resilient ecosystems
are less likely to be adversely affected by increased global temperatures and altered
precipitation patterns (Walsh et al. 2005). Managers who are interested in maintaining
and protecting streams can utilize resilience factors that would aid conservation. The best
way to do this would be to look for pockets of this resiliency to changing conditions
within these systems, then employ them to impart degrees of resistance to climate change
(Pörtner et al. 2022). Many attributes, like forest and riparian cover, shading, and OM
inputs, affect how well a stream ecosystem can resist climate change by buffering against
the warming of the air, surface, and waters (Vannote et al. 1980, Walsh et al. 2005).

Forest cover, both upstream and adjacent, is likely to be one of these key factors that
improves resiliency of streams (Choi et al. 2021, Cui et al. 2021).
Headwater streams (first and second order streams) are smaller streams with
unique properties when compared to other, larger streams. These streams are typically
groundwater fed, have heavier relative canopy and forest cover, and tend to be cooler,
smaller, and less productive than larger streams (Vannote et al. 1980). Forested
headwater streams tend to have better temperature regulation and more stable
hydrographs due to their size (Vannote et al. 1980, Walsh et al. 2005, Somers et al.
2013). On average, forested headwater streams tend to be cooler during base flow and
after storm events (Vannote et al. 1980, Somers et al. 2013). A higher temperature
stability was also found, where water temperatures varied by only 2 °C in forested
systems, while that value could vary by as much as 10 °C for systems that have had their
tree cover removed (Somers et al. 2013).
Trees in the upstream watershed and in the riparian zone are critical for climate
resiliency. In addition to the comparison of temperatures between forested and altered
stream systems, gaps in forest canopy along riparian zones alone contribute to differences
in temperatures within stream systems (Swartz et al. 2020). The presence of a riparian
zone that is more densely vegetated, as well as an intact upstream forest, also allows for
better filtering of pollution and prevention of run-off from entering the streams. (Walsh et
al. 2005, Swartz et al. 2020). In addition to stabilizing flow and playing a vital role in
temperature regulation and system stability, forest cover in headwater streams also blocks
light from entering the system when compared to wider streams (Vannote et al. 1980).
Dense canopies, which come with heavier and more intact forests, can better block

sunlight, preventing photosynthetically active radiation (PAR) from reaching the stream,
reducing the abundance of algae. In addition to temperature regulation and reduction, this
also leads to these streams having less in-stream productivity, leading to more
heterotrophy in headwater streams (Vannote et al. 1980). Instead, these streams will need
to rely on allochthonous carbon sources for their OM.
Surrounding forests contribute a substantial portion of the carbon that fuels
headwater streams. These allochthonous inputs tend to include a wide range of materials,
from leaf litter to fallen trees, as well as dissolved organic matter (DOM) inputs from
groundwater and soil (Webster and Meyer, 1997). For example, Bear Brook, a secondorder stream in the northeastern U.S., was found to have allochthonous inputs from the
surrounding forest that were as high as 99%, with 44% of that from leaf litter, and 47% of
the total energy delivered to the steam in the form of DOM (Fisher and Likens, 1973).
Another study found that 62% of the benthic organic matter (BOM) of a stream in a
deciduous forest consisted of twigs and branches, and that the litter resided in the stream
year-long, producing OM for the stream consistently (Abelho and Graça, 1998). Even
smaller sections of riparian zones can influence the amount of OM entering a stream. A
previous study found that in blackwater streams in South Carolina, the riparian wetland
forest, which consists of only 6% of the watershed, contributed 10.2% of the organic
carbon entering the stream via detritus, and another 63% from its soil (Dosskey and
Bertsch, 1994). These carbon inputs are the primary source for energy, and a dominating
percentage tends to be in the form of DOM. This only begins to illustrate the importance
forests have in fueling a headwater stream’s carbon cycle, and how that role could be
affected by changes in forest cover.

Extracellular enzymes produced by biofilm microorganisms often reflect the
composition and availability of OM and nutrients (Fisher and Likens, 1973, Dosskey and
Bertsch, 1994, Abelho and Graça, 1998). Biofilms consist of algae, bacteria, protozoa,
and fungi encased in polysaccharide matrices attached to substrata (Lock et al. 1984,
Peters and Lodge, 2009). Bacteria, fungi, and other microbes produce these extracellular
enzymes, releasing them into the biofilm matrix to break down OM and nutrients that
cannot be directly transported across cell membranes (Pohlan et al. 2010). Table 1 lists
the functions of the seven common classes of extracellular enzymes. As these enzymes
are key players in the processing of OM in streams, their extracellular enzymatic
activities (EEAs) can be used as predictors for metabolic rates and OM types, quantities,
and qualities (Sinsabaugh et al. 2010, Zhang et al. 2018, Pastor et al. 2019). Extracellular
enzymes are also accurate predictors of OM decomposition rates within streams
(Sinsabaugh et al. 1994, Moorhead and Sinsabaugh, 2000). By investigating the presence
and activities of these extracellular enzymes, the processing of OM within these streams
can be applied to factors such as riparian canopy cover and upstream forest cover. These
findings could be incorporated into a better understanding of small-scale resiliency to
climate change and the importance of managing forests to better minimize the overall
effects of climate change.

Table 1: Project-specific enzymes, their letter designation, enzyme code, and specific purpose
Enzyme
Code
EC
Purpose
3.1.3.1–2

Decomposition of
phosphomonoesterase

Pastor et al.
2019

Alkaline phosphatase

PHOS

β-Nacetylglucosaminidase

NGASE 3.2.1.52

β-D-1,4-glucosidase

GLU

3.2.1.21

Phenyl oxidase

PHEN

1.10.3.2

Transformation of phenolic
molecules (ex. lignin)

Pastor et al.
2019

Cellobiohydrolase

CEL

3.2.1

Degradation of cellulose

Pastor et al.
2019
Pastor et al.
2019
Caruso G.,
2010

Hydrolyse N-acetylglucosaminecontaining oligosaccharides and
proteins
Last step of cellulose
decomposition, decomposition of
cellobiose or small oligomers
containing β-D-glucose linkages

β-Xylosidase

XYL

3.2.1.37

Last step of hemicellulose
decomposition, decomposition of
xylobiose or
xylooligosaccharides

Leucyl aminopeptidase

LAMP

3.4.11.1

Hydrolysis of leucine, nitrogenacquiring

Zhang et
al. 2018
Yan et al.
2021

As previous studies have shown, completely forested streams offer at least some
degree of resiliency to climate change (Walsh et al. 2005, Somers et al. 2013). Despite
extensive research on the influence and effects of forests on stream systems, there has
been little work that relates upstream and adjacent riparian forest cover to patterns in
EEAs within the microbial communities. The amount of forest necessary to maintain
healthy forested headwater streams, from the microbial perspective, can be examined by
investigating any potential relationships between stream EEAs and forest cover.
Measuring extracellular enzymes in stream biofilms provides a window into the patterns
and magnitude of OM utilization in streams, which can in turn improve links between
this utilization of OM and forest cover.

Links between EEAs produced by microbes in the biofilm (Table 1) and abovestream canopy cover as well as links between EEAs produced by microbes in the biofilm
(Table 1) and upstream and riparian forest cover can provide more insight into ecosystem
function. As some characteristics of forests function as buffers against deviation from the
system’s conditions, a link between the extracellular enzymes and the attributes of
upstream and adjacent forests can be used as a predictor of OM utilization and resilience
to climate change. Overall, the possible solution to mitigate climate change’s effects on
streams is the presence and preservation of local and upstream forests. Investigation into
the ecosystem functions of microbial communities within streams and how they relate to
forests can become a valuable tool in the monitoring and protection of headwater streams
through biofilm analysis in the face of the effects of global warming. For this project,
discerning how forests determine ecosystem function in the context of how biofilm
microbial communities process OM is the primary goal, done primarily through
examination of the EEAs of seven extracellular enzymes utilized by biofilm microbial
communities (Table 1), above-stream canopy cover and watershed forest cover
measurements, and various water and stream quality parameters, both chemical and
biological. Changes in forest cover, both upstream and adjacent, is hypothesized to be
reflected in the EEAs found in a stream’s biofilm.

Methods
Planning and Field Methods
I sampled 46 stream sites in the upper Delaware River Basin, with sites in
Pennsylvania and New York (Figure 1). Each site consisted of a 100 m sample reach. I
delineated the upstream watershed from the bottom of the reach, recorded with a GPS
unit, using ModelMyWatershed (Aufdenkampe et al. 2009). This allowed me to
determine upstream land use, such as forest cover, wetlands, agricultural land, and
developed urban land. National Land Cover Database codes 21 through 24 were
combined to form “percent developed land,” 41 through 43 were combined to form
“percent forest,” and 90 and 95 were combined to form “percent wetlands” (Dewitz et al.
2021). Data and samples were collected, with landowner permission, during either the
summer of 2021 or 2022.

Figure 1: All 50 selected sample locations within the northern Delaware River Basin,
including four sites which were unable to be sampled due to dry conditions.

I used an Insta 360 ONE X2 3D camera to measure above-stream canopy cover.
Starting at 0 meters, I took a 3D image every 20 meters in the center of the stream along
the reach, resulting in six images representing the canopy cover of the reach. I analyzed
each image using the software package ImageJ (1.54d). I first oriented images into a
sphere shape using the camera’s native photo-editing software, with the center of the sky
as the center of the image. I used an ellipse selection to crop each image to ensure that
only the riparian vegetation and above-stream canopy cover was included in the analysis,
with the pixel height and width of the selection noted. I then converted the image to
binary black and white. I generated a histogram using ImageJ’s data analysis tools to
measure the number of black pixels, which was then used to find the percentage of sky
visible/canopy cover using the following equation:

% 𝐶𝑎𝑛𝑜𝑝𝑦 𝐶𝑜𝑣𝑒𝑟 = 1 −

∗ 100

where x = black pixels, h = height of the ellipse selection, and w = width of the ellipse
selection (Ecological Forester, 2011). I then averaged these six canopy estimates.
I collected unfiltered water grab samples in 125 mL acid-washed bottles in a riffle
near the bottom of each reach. Each unfiltered water sample was either frozen on-site
using dry ice or kept on ice, and then I placed them into a freezer once I returned from the
field. I measured field pH, conductivity, temperature, chromophoric dissolved organic
matter (CDOM), and alkalinity readings using a Eureka Manta+35 sonde. I calculated the
average depth of the stream by measuring depth using a meterstick at a random position
in the stream every two meters and then averaging the depths.

Along each 100 m reach, I collected 10 rocks. I selected fully submerged rocks
based on their position along the 100 m reach, with 1 rock per every 10 m. I scraped each
rock for biofilms using a toothbrush and a rubber template fit with a gasket delineating a
11.3 cm2 area to sample on the rocks. These 10 scrapings were composited into a single
site sample for later biofilm analysis. This sample was either frozen on-site using dry ice
or kept in ice until it was able to be placed into a -80 °C freezer.
Laboratory Methods
I thawed frozen composite samples in a running warm water bath, and then
brought each sample to 100 mL using deionized water. I introduced an acid-washed stir
bar to ensure proper mixing and took sub-samples using a clipped tip pipette. I placed six
1.0 mL sub-samples into labeled 2.0 mL microcentrifuge tubes for analysis of EEAs and
two 1.0 mL samples into labeled 1.5 mL microcentrifuge tubes for TN/TP analysis. I
filtered 5.0 mL of sample through a glass microfiber filter (GF/f), and then added the
sample to a 15 mL Falcon tube for later chlorophyll a analysis. I added 20 mL of sample
to pre-weighed aluminum weigh boats, and then I placed those into a 105 °C drying oven,
for ash-free dry mass (AFDM) analysis. I then immediately placed all composite samples
and sub-samples back into a -80 °C freezer.
I analyzed EEAs (Table 1) following modified procedures developed by Rier et
al. (2007) for hydrolytic enzymes and Sinsabaugh et al. (1994) for phenol oxidase. To
test hydrolytic enzymes, I vortexed aqueous composite samples of periphyton for 15
seconds to ensure homogeneity, drew out 0.8 mL of sample with a clipped pipette, and
deposited into a labeled 2.0 mL microcentrifuge tube. Following thawing in a dark area to
prevent light interference, I added 0.4 mL of the sample’s respective substrate (Table 2)
17

to each tube. I then quickly capped the tubes and placed them onto a tube mixer, which
rotated them for 30 minutes (0.5 hours) in a dark chamber. I then added 0.4 mL of
carbonate buffer (pH = 10) and immediately centrifuged at 13,600 rpm for 10 minutes.
Following this, I placed 200 μL of each sample into a 96 well plate in triplicate along
with a blank prepared using deionized water following the same procedure. I prepared
standards following similar methods. Using a Thermo Scientific Fluroskan Ascent
fluorometer, I measured the fluorescence (excitation wavelength 355 nm, emissions
wavelength 460 nm) of each sample and corrected using the blank measurement and the
standard curve, with dilutions performed if a sample measured above the maximum
detectable limit (Rier et al. 2007).
Table 2: Specific substrates to be used for the enzyme assays
Enzyme
Code
Substrate
Alkaline phosphatase
PHOS
4-MUF-phosphate
β-N-acetylglucosaminidase NGASE 4-MUF β-N-acetylglucosaminide (MUFlcNAc)
β-D-1,4-glucosidase
GLU
MUF-β-D-glucoside
Phenyl oxidase
PHEN
L-3,4-Dihydroxyphenylalanine (L-DOPA)
Cellobiohydrolase
CEL
4-MUF-β-D-cellobioside
β-Xylosidase
XYL
MUF-β-D-xyloside
Leucine aminopeptidase
LAMP
L-Leucine-7-amido-4-methylcoumarin
MUF = methylumbelliferyl
To test the activity of phenol oxidase (PHEN), I vortexed aqueous samples of
periphyton for 15 seconds to ensure homogeneity. I then drew out 0.5 mL of sample with
a pipette and deposited it into a labeled 2.0 mL microcentrifuge tube. I also used 0.5 mL
of deionized water as a control. I added 0.5 mL of 2.5mM L-DOPA to each tube before
vortexing again. I placed the tubes in a tube mixer for 1 hour, centrifuged them for 10
minutes, and then added sample to disposable cuvettes. I read samples on a Thermo
Scientific Spectronic Genesys 2 UV-Vis spectrometer at 460 nm (Sinsabaugh et al.

1994). I calculated the stream’s value on my “carbon quality index” (CQI) by using
log (𝑃𝐻𝐸𝑁

/𝐺𝐿𝑈

In addition to biofilm enzyme analyses, I measured total nitrogen and total
phosphorus (TN and TP, respectively) in water samples using a Seal AQ1 discrete
analyzer (EPA, 1993). I created standard curves for both nitrite-nitrate and phosphorus
using stock solutions. I analyzed 24.0 mL of collected sample per site by adding 5.0 mL
of an oxidizing agent (32 g of K2O8S2 and 40 mL 3N NaOH in 500 mL), autoclaving at
120 °C for 1 hour, adding 0.2 mL of NaOH to the sample, and then measuring on a Seal
Analytic AQ1 (EPA, 1993).
I analyzed biofilm nitrogen and total phosphorus (MatN and MatP, respectively)
using a Seal AQ1 discrete analyzer (EPA, 1993). I created standard curves for both
nitrite-nitrate and phosphorus using stock solutions. I added 1.0 mL of collected sample
per site to 23.0 mL of deionized water. I analyzed the diluted samples by adding 5.0 mL
of an oxidizing agent (32 g of K2O8S2 and 40 mL 3N NaOH in 500 mL), autoclaving at
120 °C for 1 hour, adding 0.2 mL of NaOH to the sample, and then measuring on a Seal
Analytic AQ1 (EPA, 1993).
I measured AFDM on biofilm samples by combusting 20 mL aliquots of
composite samples, which I had dried at 105 ° for at least 24 hours. After drying, I
weighed each sample. Following this, I placed the aluminum weigh boats into a Fisher
Scientific Isotemp muffle furnace at 500 °C for 1 hour. After combustion and cooling, I
massed each sample again, with the difference between the dry mass and combustion
mass recorded as the site’s AFDM (Rice at al. 2017).

I measured chlorophyll a from biofilm suspensions that were retained on GFF
filters and stored at -80 °C until analysis. Under dark conditions, I added 5.0 mL of 90%
ethanol to each Falcon tube. Once the ethanol solution was added, I placed the tubes into
a ~80°C water bath for 5 minutes. Following the extraction, I placed the tubes in a dark
refrigerator overnight. The next day, I measured absorbances of the extraction solution at
both 750 and 665 nm using a Thermo Scientific Spectronic Genesys 2 UV-Vis
spectrometer before and after acidification to correct for phaeopigments (Lind, 1985,
Wetzel and Likens, 2000, Biggs et al. 2000).
Statistical Analysis
I used the statistical software package R (2022.12.0 Build 353) to analyze the
initial relationships between EEAs, upstream land and above-stream canopy cover, and
other environmental variables using a Pearson correlation analysis that accounted for
experiment-wise error. I standardized EEAs for biomass by dividing them by AFDM.
Where necessary, data were log10(x + 1) transformed to meet the assumptions of the test.
Variables that I found to be correlated were run against each other in a linear regression
model to evaluate significance. I conducted similar analyses between EEAs and
environmental variables, such as water quality values and land cover, as well as between
environmental variables.
To investigate relationships between environmental variables and enzymeinferred relative nutrient limitation, I conducted a vector analysis, as described by
Moorhead et al. (2013). I plotted proportions of nitrogen- and phosphorus-acquiring
enzymes against each other to determine relative limitation in headwater streams.

𝐺𝐿𝑈/(𝐺𝐿𝑈 + 𝑃𝐻𝑂𝑆) and 𝐺𝐿𝑈/(𝐺𝐿𝑈 + 𝑁𝐺𝐴𝑆𝐸 + 𝐿𝐴𝑀𝑃) represented proportions of
phosphorus- and nitrogen-acquiring enzymes, respectively (Figure 2).

GLU/(GLU+NGASE+LAMP)

Proportions of N-acquiring vs.
P-acquiring Enzyme Activity

0.75

0.50

0.25

0.00
0.0

0.2

0.4

0.6

GLU/(GLU+PHOS)

Figure 2: Proportions of N-acquiring enzymatic activity plotted against proportions of
P-acquiring enzymatic activity (Moorhead et al. 2013). Three example points have had
lines plotted to represent the angle and length of the resultant vector following
calculation, shown left to right: (17.45°, 0.189), (24.68°, 0.354), and (30.85°, 0.644).
The angles of the resultant vectors represented relative phosphorus vs. nitrogen
limitation in the sampled headwater streams, while the lengths of the vectors represented
relative carbon vs. nutrient limitation. I calculated vector angle as 𝑉𝐴𝑛𝑔𝑙𝑒 =
arctan (𝑥, 𝑦), which was measured in degrees (Moorhead et al. 2013). I calculated the
vector length as 𝑉𝐿𝑒𝑛𝑔𝑡ℎ =

𝑥 + 𝑦 , which was measured as a relative, unitless

distance (Moorhead et al. 2013). Following my calculation of the vector angles and
lengths, I again utilized a scatterplot correlation analysis approach with linear regression
modeling to look for relationships between vector measurements and environmental
variables.
I used Akaike information criterion (AICc) to evaluate an array of multiple
regression models predicting vector angles and lengths as a function of several potentially

relevant environmental variables. I created a “master” multiple regression model for both
the vector angle and vector length; Vector Angle Master = log

𝑇𝑃 + log

𝑇𝑁 +

𝐶ℎ𝑙𝐴 and Vector Length Master = log

𝑇𝑃 + log

𝑀𝑎𝑡𝑃 +

log

𝑀𝑎𝑡𝑁: 𝑃 + log

log

𝐶ℎ𝑙𝐴 + log

%𝐶𝑎𝑛𝑜𝑝𝑦 + log

%𝐹𝑜𝑟𝑒𝑠𝑡 + log

𝐶𝐷𝑂𝑀 + 𝐶𝑄𝐼 respectively. I

included variables that were not found to be directly correlated to one another in the
initial correlation analysis. Then, utilizing the “dredge” function in the MuMIn R
package, I found the top 10 models for each dependent variable, which I then ranked by
AICc. I created a similar model for the stream’s CQI value using the same variables as
both the vector angle and vector length models.

Results
Several correlations between EEAs and environmental variables were found,
indicating relationships between factors influencing OM dynamics in headwater streams
(Table 3). Activities of XYL, CEL, NGASE, and PHOS were shown to be positively
correlated. Activities of GLU, PHEN, and NGASE were also shown to be positively
correlated. GLU and CEL were negatively correlated with one another. PHEN and
NGASE were shown to be positively correlated, as well as NGASE and PHOS. The
strongest relationship found was between XYL and CEL, with a correlation of 0.518 (p <
0.001).
Strong correlations between percent above-stream canopy cover and both LAMP
and PHOS were found, with correlation values of 0.385 (p = 0.00816 and -0.478 (p <
0.001), respectively. These correlations show a concurrent increase in activity of
nitrogen-acquiring enzymes and decrease in activity of phosphorus-acquiring enzymes as
above-stream canopy cover increases. A similar relationship is seen between water
column total phosphorus and both LAMP and PHOS, with correlation values of 0.357 (p
= 0.0148) and -0.344 (p = 0.0192) respectively. A negative correlation was found
between the percentage of forest cover in the upstream watershed and PHEN, with a
value -0.398 (p = 0.00619). A negative correlation was also found between the
percentage of wetlands in the upstream watershed and XYL, with a value of -0.358 (p =
0.0146). PHOS and CDOM were also found to negatively correlate, with a value of 0.363 (p = 0.0131).

Table 3: Significant (p < 0.05) correlations between extracellular enzymatic activities
and environmental variables in headwater streams
% Canopy
% Forest
% Wetland
TP (μg/L)
XYL

-0.358

GLU

PHEN

-0.398

CEL

NGASE

LAMP

0.385

0.357

PHOS

-0.478*

-0.344

MatNAFDM

MatPAFDM

Molar N:P

CDOM

XYL

0.391

0.499

GLU

0.337

PHEN

0.439

0.315

CEL

NGASE

0.383

0.348

LAMP

0.376

0.317

PHOS

0.395

-0.363

XYL

GLU

PHEN

CEL

NGASE

PHOS

XYL

0.518*

0.375

0.399

GLU

0.296

-0.374

0.351

PHEN

0.296

0.404

CEL

0.518*

-0.374

NGASE

0.375

0.351

0.404

0.492

PHOS
0.399
0.492
X
Correlation values between XYL (β-Xylosidase), GLU (β-D-1,4-glucosidase), PHEN
(phenyl oxidase), CEL (cellobiohydrolase), LAMP (leucyl aminopeptidase), NGASE
(β-N-acetylglucosaminidase), and PHOS (alkaline phosphatase) activities and
environmental variables, including total phosphorus (TP), biomass-adjusted biofilm
nitrogen and phosphorus (MatNAFDM and MatPAFDM), and chromophoric dissolved
organic matter (CDOM). Values have been log10(x + 1) to meet the assumptions of the
test. Values with an asterisk (*) denote p-values less than 0.001. Correlations which
were not significant are denoted by a dash (-).

Several positive correlations between both biomass-adjusted periphytic nitrogen
and phosphorus and various EEAs were also found. LAMP was found to correlate with
available biofilm nitrogen, with a value of 0.376 (p = 0.0100), while PHOS was found to
correlate with available biofilm phosphorus with a value of 0.395 (p = 0.00658). NGASE
was also found to positively correlate with both biofilm nitrogen and phosphorus, with
values of 0.383 (p = 0.00870) and 0.348 (p = 0.0178). LAMP and the molar nitrogen to
phosphorus ratio within the biofilm was also found to positively correlate with a value of
0.317 (p = 0.0316).
Vector Analysis
The relationship between the angle of the resultant vectors and the percentage of
above-stream canopy cover was found to be significant (p < 0.001), showing an increase
in relative phosphorus vs. nitrogen limitation in headwater streams as they become
progressively more shaded (Figure 3a). A similar significant relationship (p = 0.00631)
was observed with water column total phosphorus (Figure 3b). Negative relationships
between resultant vector lengths and both biofilm nitrogen (p = 0.00326) and biofilm
phosphorus (p = 0.0127) were also found to be significant (Figure 4). Similar
relationships are also seen with both chlorophyll a (p = 0.00829) and AFDM (p =
0.00320) (Figure 5a and 5b).

Figure 3a & 3b: Relationships between a stream’s vector angle and a.) percentage
canopy cover (p < 0.001) and b.) water column total phosphorus (p = 0.00631) were
found to be significant, with the shaded region defining the true regression line to 95%
confidence.

Figure 4a & 4b: Relationships between a stream’s vector length and a.) biofilm
nitrogen (p = 0.00326) and b.) biofilm phosphorus (p = 0.0127) were found to be
significant, with the shaded region defining the true regression line to 95% confidence.

Figure 5a & 5b: Relationships between a stream’s vector length and a.) chlorophyll a
(p = 0.00829) and b.) AFDM (p = 0.00320) were found to be significant, with the
shaded region defining the true regression line to 95% confidence.

Multiple Regression Analysis
Following correlation analysis, several multiple regression models were evaluated
and ranked by AICc values. These models examined the relationship between both the
previously calculated vector angles and lengths and various predictor variables, such as
upstream forest cover and water column CDOM.
Multiple regression models for relative phosphorus vs. nitrogen limitation,
indicated by a stream’s vector angle value, were significantly related to the ratio of
nitrogen to phosphorus in the biofilm, both total water column nitrogen and phosphorus,
and chlorophyll a (Table 4). The most parsimonious model describes relative phosphorus
vs. nitrogen limitation being mostly influenced by the molar ratio of nitrogen to
phosphorus within the biofilm and water column total phosphorus (Table 4). The top 3
models for relative phosphorus vs. nitrogen limitation are also listed. These models
describe positive influence from molar nitrogen to phosphorus, positive influence from
water column total phosphorus, and negative influence from chlorophyll a on a stream’s
vector angle value (Eq. 1, 2, and 3).

Table 4: Top 10 multiple regression models using a stream’s vector angle (relative
phosphorus vs. nitrogen limitation) as the dependent variable, including model rank,
variables included within the model, AICc and relative AICc, R 2, the model’s p-value,
and the residual error of the model.
Model
Relative
Residual
Model Variables
AICc
R2
p-value
Rank
AICc
Error
1
MatN:P, TP
355.3
1
0.2591 0.00158
10.8
2
TP
356.2
1.003
0.2049 0.00158
11.06
3
ChlA, MatN:P, TP
357.8
1.007
0.2596 0.00517
10.93
4
MatN:P, TN, TP
357.9
1.007
0.2593 0.00521
10.93
5
ChlA, TP
358.6
1.009
0.2054 0.00713
11.19
6
TN, TP
358.6
1.009
0.2049 0.00723
11.19
7
ChlA, MatN:P, TN, TP
360.5
1.015
0.2598 0.01328
11.06
8
ChlA, TN, TP
361.1
1.016
0.2054 0.02064
11.32
9
MatN:P
362.1
1.019 0.09598 0.03615
11.79
10
MatN:P, TN
362.9
1.021
0.1267 0.05433
11.73

Eq. 1: 𝑣𝐴𝑛𝑔𝑙𝑒 = −3.838 + 10.20(log
Eq. 2: 𝑣𝐴𝑛𝑔𝑙𝑒 = 5.170 + 20.69(log
Eq. 3: 𝑣𝐴𝑛𝑔𝑙𝑒 = −3.736 − 1.022(log
18.96(log

𝑀𝑎𝑡𝑁: 𝑃) + 18.76(log

𝑇𝑃)

𝑇𝑃)
𝐶ℎ𝑙𝐴) + 10.20(log

𝑀𝑎𝑡𝑁: 𝑃) +

𝑇𝑃)

Multiple regression models for relative carbon vs. nutrient limitation, indicated by
a stream’s vector length value, were significantly related to chlorophyll a, CQI value,
biofilm phosphorus concentration, CDOM concentration, watershed forest cover, and
above-stream canopy cover (Table 5). The most parsimonious model describes relative
carbon vs. nutrient limitation being mostly influenced by the concentration of chlorophyll
a, a stream’s CQI value, and biofilm phosphorus (Table 5). The top 3 models for relative
carbon vs. nutrient limitation are also listed. These models describe positive influence
from CDOM concentrations, positive influence from biofilm phosphorus, negative
influence from chlorophyll a, and negative influence from a stream’s CQI value on a
stream’s vector length value (Eq. 4, 5, and 6).
28

Table 5: Top 10 multiple regression models using a stream’s vector length (relative
carbon vs. nutrient limitation) as the dependent variable, including model rank, variables
included within the model, AICc and relative AICc, R 2, the model’s p-value, and the
residual error of the model.
Model
Relative
Residual
Model Variables
AICc
R2
p-value
Rank
AICc
Error
1
ChlA, CQI, MatP
-20.2
0.5300 5.08e-7
0.1793
1
2
ChlA, CQI
-19.8
0.4991 3.50e-7
0.183
1.02
3
CDOM, ChlA, CQI, MatP -18.6
0.5401 1.47e-6
0.1796
1.02
4
ChlA, CQI, Forest, MatP
-18.3
0.5372 1.66e-6
0.1801
1.02
5
ChlA, CQI, Forest
-18.2
0.5088 1.26e-6
0.1833
1.03
6
CDOM, ChlA, CQI
-17.9
0.5053 1.46e-6
0.184
1.05
7
Canopy, ChlA, CQI, MatP -17.7
0.5312 2.14e-6
0.1813
1.06
8
Canopy, ChlA, CQI
-17.6
0.5023 1.65e-6
0.1845
1.08
9
ChlA, CQI, MatP, TP
-17.6
0.5306 2.20e-6
0.1814
1.14
10
ChlA, CQI, TP
-17.3
0.5004 1.78e-6
0.1849
1.17
Eq. 4: 𝑣𝐿𝑒𝑛𝑔𝑡ℎ = 0.8472 − 0.3901 (log
0.1487(log

𝐶ℎ𝑙𝐴) − 0.2724(CQI) +

𝑀𝑎𝑡𝑃)

Eq. 5: 𝑣𝐿𝑒𝑛𝑔𝑡ℎ = 1.008 − 0.4026 (log
Eq. 6: 𝑣𝐿𝑒𝑛𝑔𝑡ℎ = 0.7712 + 0.04781(log
0.2633(𝐶𝑄𝐼) + 0.1590(log

𝐶ℎ𝑙𝐴) − 0.2676(𝐶𝑄𝐼)
𝐶𝐷𝑂𝑀) − 0.3704 (log

𝐶ℎ𝑙𝐴) −

𝑀𝑎𝑡𝑃)

Linear regression modeling for a stream’s CQI yielded no significant relationships
with either water column total phosphorus or nitrogen, biofilm phosphorus, the molar
ratio of biofilm nitrogen and phosphorus, chlorophyll a, CDOM concentration, and both
upstream forest cover and above-stream canopy cover.

Discussion
Several relationships between canopy cover, biofilm EEAs, and stream
measurements were found in this study. As above-stream canopy cover increased, (a)
enzymatic activity indicative of allochthonous carbon increased and (b) relative nutrient
limitations increased. Above-stream canopy cover was found to be a major influence on
both in-stream relative phosphorus limitation and relative nutrient vs. carbon limitation,
as well as enzyme production within the microbial communities. Above-stream canopy
cover was also shown to influence a nutrient spiraling effect, facilitated by water column
total phosphorus. Overall, canopy cover was found to be an indicator of OM utilization
and processing in headwater streams and a driver for nutrient uptake.
The process of breaking down and utilizing OM within headwater streams, from a
microbial lens, is heavily interconnected. Several EEAs and groups of EEAs were found
to be related to one another, including nutrient-acquiring enzymes. These relationships
were found to be negatively or positively linked to one another as well. This
interconnectedness reflects the range and complexity of some of the OM available to
streams and the microbial communities within them (Tian et al. 2017, Vaughn et al.
2021). In addition to chemical make-up of the available OM, different microbes within
the biofilms have been shown to vary throughout the stages of decomposition of leaf litter
and other inputs, with fungi leading initially and bacteria becoming more dominant in the
latter stages for example (Baldy et al. 1995, Hieber et al. 2002, Pascoal et al. 2004). This
shows a sophisticated network of OM processing and carbon and nutrient utilization.
The negative correlation between PHEN and the percentage of upstream forest
cover was unexpected and may have been driven by algae production directly influencing
30

activities. This is the opposite of what would be expected with increased amounts of
forest in the drainage basin, since more cover would lead to more OM containing
phenolic compounds (Min et al. 2015). These results are most likely due to algal
production, which is influenced by open-canopy systems, enhancing the activity of
PHEN (Rier et al. 2014). This occurrence acts as evidence of the influence of forest cover
in OM dynamics in headwater streams, as more available PAR in systems with less forest
cover was shown to influence the processing of phenolic OM.
As a headwater stream becomes more densely covered by the riparian forest
canopy and riparian zone, nitrogen-containing OM becomes more readily available for
biofilm microbial communities due to a relative abundance of nitrogen within decaying
leaf litter (Tian et al. 2017). This is shown by a decrease in the activity of phosphorusacquiring enzymes while the activity of nitrogen-acquiring enzymes increases. This
occurs as the above-stream canopy cover density increases. The stream can then begin to
experience greater relative phosphorus vs. nitrogen limitation due to this abundance of
available nitrogen within the allochthonous inputs of leaf litter, (Tian et al. 2017). This is
also supported by a positive relationship between biomass-adjusted available nitrogen
and both NGASE and LAMP. The activities of PHOS and NGASE in the periphyton also
have a direct relationship. However, PHOS and LAMP appear to have an inverse
relationship in the context of above-stream canopy cover but are not significantly related
to one another. XYL activity was also shown to have a negative linear relationship with
the percentage of upstream wetlands, potentially due to upstream wetland’s influence on
pH levels, nutrient availability, concentration of dissolved organic carbon, and substrate
availability (Leibowitz et al. 2018).

Above-stream canopy cover can act as a driver of relative phosphorus limitation
in streams, likely due to allochthonous inputs of more nitrogen-rich OM (Tian et al.
2017). This is shown by the positive linear relationship between a stream’s above-stream
canopy cover and vector angle. A similar relationship between a stream’s water column
total phosphorus and vector angle shows increasing relative phosphorus limitation
alongside increasing concentrations of water column total phosphorus. This provides
evidence of nutrient spiraling within headwater streams, especially as these stream’s
canopy covers become denser, as above-stream canopy cover and water column total
phosphorus are significantly related. As a stream acquires more phosphorus, organismal
growth occurs. As organisms continue to grow and abundance rises, the demand for
phosphorus increases. Eventually, as phosphorus is lost due to sedimentation and growth
continues to increase, phosphorus limitation occurs as the system can no longer supply
the necessary nutrients (Newbold, 1992). Above-stream canopy cover is likely a factor
that encourages this spiraling effect and allows a link to be made between OM dynamics
within these headwater streams and their forest cover. This link could be further isolated
and investigated to find possible solutions to nutrient balancing as it relates to canopy
cover.
Relative nutrient limitation within the system increases as biofilm nitrogen and
phosphorus concentrations increase. As more available nutrients give way to excess
organismal growth, nutrient limitation occurs as demand for these nutrients surpasses the
available supply (Newbold, 1992). This is shown by negative linear relationships
between a stream’s vector length and biofilm nitrogen and phosphorus concentrations.
This is likely due to an effect similar to the nutrient spiraling effect already seen in these

streams (Newbold, 1992). An abundance of carbon in these streams, in the form of
biomass, also causes nutrients to become more limited, leading to greater relative nutrient
limitation in streams with higher concentrations of chlorophyll a and AFDM values. This
is shown by negative linear relationships between both chlorophyll a and AFDM with
vector length. Both of these findings indicate the presence of this nutrient spiraling effect
(Newbold, 1992).
The positive relationship between the biofilm molar nutrient ratio and the vector
angle indicates greater relative phosphorus limitation, most likely due to an increase in
more relatively nitrogen rich OM entering these streams (Tian et al. 2017).
Concentrations of chlorophyll a and water column total phosphorus, as well as the molar
ratio of nitrogen to phosphorus in the biofilm, were found through a ranked multiple
regressions analysis to best describe the main influences of relative phosphorus vs.
nitrogen. The positive relationship between water column total phosphorus and vector
angle describes the previously seen spiraling effect as described by Newbold (1992).
Chlorophyll a had a negative influence on the vector angle, indicating that as there is less
biofilm biomass, phosphorus becomes less limited. This is likely due to fewer organisms
taking up available phosphorus. Above stream canopy cover was found to be positively
related to the molar ratio of nitrogen to phosphorus and to the water column total
phosphorus, demonstrating that above-stream canopy cover can be seen as a driver for
relative phosphorus vs. nitrogen limitation in headwater streams.
Concentrations of biofilm phosphorus, CDOM, and chlorophyll a were found
through a ranked multiple regressions analysis to best describe the main influences of
relative carbon vs. nutrient limitation in headwater streams. More chlorophyll a indicates

higher concentrations of biomass, which in turn means higher concentrations of carbon.
Likewise, the relationship between the limitation and the CQI value is likely due to the
recalcitrance of OM within the stream, with higher CQI values indicating harder-tobreak-down OM. This OM leads to this reduced relative carbon limitation as the stream is
able to utilize the carbon rich OM for longer. The positive influence from biofilmavailable phosphorus is due to the nutrient spiraling effect seen in these headwater
streams (Newbold, 1992). However, as biofilm-available phosphorus is both positively
correlated with relative carbon vs. nutrient limitation and negatively correlated with
above-stream canopy cover, lack of canopy cover can be viewed as a driver for increased
relative carbon limitation. The inverse shows that as above-stream canopy cover
increases, relative carbon limitation decreases, and in turn, relative nutrient limitation
increases. This indicates that above-stream canopy cover acts as a driver for nutrient
limitation and can be used as a potential indicator for in-stream water quality and nutrient
content.
CDOM and phosphorus-related variables were shown to be negatively correlated,
and CDOM, via linkages with PHOS, was shown to increase as above-stream canopy
increases. The EEAs of PHOS could then be seen as a potential linkage between abovestream canopy cover and CDOM, as facilitated by the activity of PHOS. As above-stream
canopy cover increases, more terrestrial OM becomes available to enter the stream
following rain events (Monteith et al. 2007). As CDOM in a stream increases, that system
could experience higher relative carbon vs. nutrient limitation.
In conclusion, several environmental variables, specifically above-stream canopy
cover, were found to heavily influence microbial extracellular enzymes found within

stream periphyton. These relationships can be used to model OM dynamics in headwater
streams in the future. By examining both microbial communities and tree cover, there is
potential to further improve predicting the effects of climate and anthropogenic change
on freshwater streams. Specifically, patterns associated with the presence of forest and
canopy cover can be applied to streams by land managers to help create or maintain
balance within the OM processing within headwater streams.

References
Abelho M, Graça MAS. 1998. Litter in a first-order stream of a temperate deciduous
forest (Margaraça Forest, Central Portugal). Hydrobiologia: Kluwer Academic
Publishers. p. 147-152.
Appling AP, Hall RO, Yackulic CB, Arroita M. 2018. Models for estimating aquatic
photosynthesis and respiration. Journal of Geophysical Research. 22:21.
Aufdenkampe A, Arscott D, Kerlin S, Daniels M, Ehrhart M, Gill S, Cheetham R,
McFarland M, Tuhinanshu T, Boyer D, Robbins A, Evans B, Tarboton D,
Horsburgh J, Mayorga E, Haag S, Shokoufandeh A, Dietrich N, Staudt C,
Damiano S, Brook H. 2009. Model My Watershed. WikiWatershed.
wikiwatershed.org/model/.
Baldy V, Gessner MO, Chauvet E. Bacteria, fungi and the breakdown of leaf litter in a
large river. Oikos 1995, 74, 93–102.
Baruch EM, Ruhí A, Harms TK, Sabo JL. 2022. Flow variation at multiple scales filters
fish life histories and constrains community diversity in desert streams. 13(6).
Benner R, McArthrur JV. 1988. Effects of temperature on microbial utilization of
lignocellulosic detritus in a thermally impacted stream. Microbial Ecology.
16:323-330.
Biggs BFJ, Kilroy C. 2000. Stream periphyton monitoring manual. National Institute of
Water and Atmospheric Research, Christchurch, New Zealand.
Borba GC, Costa FRC, Espírito-Santo HMV, Leitão RP, Dias MS, Zuanon J. 2020.
Temporal changes in rainfall affect taxonomic and functional composition of
stream fish assemblages in Central Amazonia. Freshwater Biology. 66(4):12.
Burcher CL, McTammany ME, Benfield EF, Helfman GS. 2008. Fish assemblage
responses to forest cover. Environmental Management. 41:336-346.
Cai W, Purich A, Cowan T, van Rensch P, Weller EJ. 2014. Did climate change-induced
rainfall trends contribute to the Australian millennium drought? Journal of
Climate. 27(9):24.
Caruso G. 2010. Leucine aminopeptidase, beta-glucosidase and alkaline phosphatase
activity rates and their significance in nutrient cycles in some coastal
Mediterranean sites. Marine Drugs. 8(4):25.

Choi Y, Lim C-H, Chung HI, Kim Y, Cho HJ, Hwang J, Kraxner F, Biging GS, Lee WK, Chon J et al. 2021. Forest management can mitigate negative impacts of
climate and land-use change on plant biodiversity: Insights from the Republic of
Korea. Journal of Environmental Management. 288:15.
Cui F, Wang B, Zhang Q, Tang H, De Maeyer P, Hamdi R, Dai L. 2021. Climate change
versus land-use change—what affects the ecosystem services more in the foreststeppe ecotone? Science of The Total Environment. 759:14.
Collins M, Knutti R, Arblaster J, Dufresne J-L, Fichefet T, Friedlingstein P, Gao X,
Gutowski WJ, Johns T, Krinner G, et al. 2013: Long-term Climate Change:
Projections, Commitments and Irreversibility. In: Climate Change 2013: The
Physical Science Basis. Contribution of Working Group I to the Fifth Assessment
Report of the Intergovernmental Panel on Climate. Cambridge University Press,
Cambridge, United Kingdom and New York, NY, USA.
de Wit HA, Valinia S, Weyhenmeyer GA, Futter MN, Kortelainen P, Austnes K, Hessen
DO, Räike A, Laudon H, Vuorenmaa J. 2016. Current Browning of Surface
Waters Will Be Further Promoted by Wetter Climate. Environmental Science &
Technology Letters. 3(12):430–435.
Dewitz J, U.S. Geological Survey. 2021, National Land Cover Database (NLCD) 2019
Products (ver. 2.0, June 2021): U.S. Geological Survey data release.
Dosskey MG, Bertsch PM. 1994. Forest sources and pathways of organic matter transport
to a blackwater stream: A hydrologic approach. Biogeochemistry. 24:1-19.
Dycus JC, Wisniewski JM, Peterson JT. 2015. The effects of flow and stream
characteristics on the variation in freshwater mussel growth in a Southeast
US river basin. Freshwater Biology, 60, 395-409.
Ecological Forester. 2011 Mar 19. Canopy Closure from Digital Photos Using ImageJ.
The Ecological Forester. ecologicalforester.blogspot.com/2011/03/canopyclosure-from-digital-photos.html.
EPA 353.2 Version 2. 1993. Standard Methods 4500-NO 3F (18th, 19th, 20th).
EPA 365.1 Version 2. 1993. Standard Methods 4500-P B, F (18 th, 19th, 20th).
Fisher SG, Likens GE. 1973. Energy flow in bear brook, New Hampshire: An integrative
approach to stream ecosystem metabolism. Ecological Monographs. 43:421-439.

Pörtner H-O, Roberts DC, Poloczanska ES, Mintenbeck K, Tignor M, Alegría A, Craig
M, Langsdorf S, Löschke S, Möller v, et al. 2022. Summary for Policymakers:
Climate Change 2022: Impacts, Adaptation and Vulnerability. Contribution of
Working Group II to the Sixth Assessment Report of the Intergovernmental Panel
on Climate Change. In: IPCC, editor.: IPCC. p. 3-33.
Hieber M, Gessner MO. Contribution of stream detrivores, fungi, and bacteria to leaf
breakdown based on biomass estimates.
Hill BH, Elonen CM, Herlihy AT, Jicha TM, Serenbetz G. 2017. Microbial ecoenzyme
stoichiometry, nutrient limitation, and organic matter decomposition in wetlands
of the conterminous United States. Wetlands Ecology and Management.
26(3):425–439.
Kaushal SS, Likens GE, Jaworski NA, Pace ML, Sides AM, Seekell D, Belt KT, Secor
DH, Wingate RL. 2010. Rising stream and river temperatures in the United States.
Frontiers in Ecology and the Environment. 8(9):461-466.
Kroll S. 2020. Evaluating headwater biodiversity, vulnerability, and potential resilience
to inform conservation in the Delaware Basin. Academy of Natural Sciences of
Drexel University: National Fish and Wildlife Foundation. p. 43.
Leibowitz SG, Wigington PJ, Schofield KA, Alexander LC, Vanderhoof MK, Golden
HE. 2018. Connectivity of Streams and Wetlands to Downstream Waters: An
Integrated Systems Framework. JAWRA Journal of the American Water
Resources Association. 54(2):298–322.
Lind OT. 1985. Handbook of common methods in limnology. Dubuque, Iowa:
Kendall/Hunt Pub. Co.
Lock MA, Wallace RR, Costerton JW, Ventullo RM, Charlton SE. 1984. River epilithon:
toward a structural-functional model. Oikos. 42:10-22.
Min K, Freeman C, Kang H, Choi S-U. 2015. The Regulation by Phenolic Compounds of
Soil Organic Matter Dynamics under a Changing Environment. BioMed Research
International. 2015:1–11.
Mohseni O, Stefan HG, Eaton JG. 2003. Global warming and potential changes in fish
habitat in U.S. Streams. Climate Change. 59:389-409.
Monteith DT, Stoddard JL, Evans CD, de Wit HA, Forsius M, Hogasen T, Wilander A,
Skjelkvale BL, Jeffries DS, Vuorenmaa J, et al. 2007. Dissolved organic carbon
trends resulting from changes in atmospheric deposition chemistry. Nature
450:537–540

Moorhead DL, Sinsabaugh RL. 2000. Simulated patterns of litter decay predict patterns
of extracellular enzyme activities. Applied Soil Ecology. 14(1):71–79.
Moorhead DL, Rinkes ZL, Sinsabaugh RL, Weintraub MN. 2013. Dynamic relationships
between microbial biomass, respiration, inorganic nutrients and enzyme activities:
Informing enzyme-based decomposition models. Frontiers in Microbiology, 4.
Newbold, JD. 1992. Cycles and Spirals of Nutrients. The Rivers Handbook. I.
Hydrological and Ecological Principles. P. Calow and G. E. Petts. Oxford
(England), Blackwell Science Publishers: 379-408.
Odum HT. 1956. Primary production in flowing water. Durham, N. C.: Duke University.
p. 102-117.
Ogden NH. 2017. Climate change and vector-borne diseases of public health
significance. FEMS Microbiology Letters. 364:1-8.
Pascoal C, Cássio F. 2004. Contribution of fungi and bacteria to leaf litter decomposition
in a polluted river. Appl. Environ. Microbiol. 70, 5266–5273.
Pastor A, Freixa A, Skovsholt LJ, Wu N, Romaní AM, Riis T. 2019. Microbial organic
matter utilization in high-arctic streams: Key enzymatic controls. Microbial
Ecology. 78:539-554.
Patil R, Wei Y, Pullar D, Shulmeister J. 2022. Effects of change in streamflow patterns
on water quality. Journal of Environmental Management. 302:113991.
Peters JA, Lodge DM. 2009. Littoral Zone. Encyclopedia of Inland Waters, 79–87.
Pohlon E, Marxsen J, Küsel K. 2010. Pioneering bacterial and algal communities and
potential extracellular enzyme activities of stream biofilms. FEMS Microbiology
Ecology. 71(3):364–373.
Raymond PA, Hartmann J, Lauerwald R, Sobek S, McDonald C, Hoover M, Butman D,
Strigl R, Mayorga E, Humborg C, et al. 2013. Global carbon dioxide emissions
from inland waters. Nature. 503:355-359.
Rice EW, Baird RB, A.D.Eaton. 2017. Standard methods for the examination of water
and wastewater. 23rd edition. American Public Health Association, American
Water Works Association, Water Environment Federation, Washington, DC.
Rier ST, Kuehn KA, Francoeur SN. 2007. Algal regulation of extracellular enzyme
activity in stream microbial communities associated with inert substrata and
detritus. Journal of the North American Benthological Society. 26:11.

Rier ST, Shirvinski JM, Kinek KC. 2014. In situ light and phosphorus manipulations
reveal potential role of biofilm algae in enhancing enzyme-mediated
decomposition of organic matter in streams. Freshwater Biology, 59(5), 1039–
1051.
Romanı ́ AM, Sabater S. 2000. Influence of algal biomass on extracellular enzyme
activity in river biofilms. Microbial Ecology. 40:9.
Shukla PR, Skea J, Calvo Buendia E, V. M-D, Pörtner HO, Roberts D, Zhai P, Slade R,
Connors S, van Dieman R, et al. 2019. Summary for policymakers: Climate
change and land: An IPCC special report on climate change, desertification, land
degradation, sustainable land management, food security, and greenhouse gas
fluxes in terrestrial ecosystems. In: IPCC, editor.: IPCC. p. 1-36.
Sinsabaugh RL, Follstad Shah JJ. 2011. Ecoenzymatic stoichiometry of recalcitrant
organic matter decomposition: The growth rate hypothesis in reverse.
Biogeochemistry. 102:12.
Sinsabaugh RL, Hill BH, Follstad Shah JJ. 2009. Ecoenzymatic stoichiometry of
microbial organic nutrient acquisition in soil and sediment. Nature.
462(7274):795–798.
Sinsabaugh RL, Manzoni S, Moorhead DL, Richter A. 2013. Carbon use efficiency of
microbial communities: stoichiometry, methodology and modelling. Elser J,
editor. Ecology Letters. 16(7):930–939.
RL, Osgood MP, Findlay S. 1994. Enzymatic models for estimating decomposition rates
of particulate detritus. Journal of the North American Benthological Society.
13:160.
Sinsabaugh RL, Van Horn DJ, Follstad Shah JJ, Findlay S. 2010. Ecoenzymatic
stoichiometry in relation to productivity for freshwater biofilm and plankton
communities. Microbial Ecology. 60:885-893.
Somers KA, Bernhardt ES, Grace JB, Hassett BA, Sudduth EB, Wang S, Urban DL.
2013. Streams in the urban heat island: spatial and temporal variability in
temperature. Freshwater Science. 32(1):309–326.
Song C, Dodds WK, Rüegg J, Argerich A, Baker CL, Bowden WB, Douglas MM, Farrell
KJ, Flinn MB, Garcia EA, et al. 2018. Continental-scale decrease in net primary
productivity in streams due to climate warming. Nature Geoscience. 11:415-420.

Swartz A, Roon D, Reiter M, Warren D. 2020. Stream temperature responses to
experimental riparian canopy gaps along forested headwaters in western Oregon.
Forest Ecology and Management. 474:12.
Thurman EM. 2012. Organic geochemistry of natural waters. Springer Science &
Business Media.
Tian D, Yan Z, Niklas KJ, Han W, Kattge J, Reich PB, Luo Y, Chen Y, Tang Z, Hu H,
Wright IJ, Schmid B, Fang J. 2017. Global leaf nitrogen and phosphorus
stoichiometry and their scaling exponent. National Science Review, 5(5), 728–
739.
Trenberth KE, Fasullo JT. 2009. Global warming due to increasing absorbed solar
radiation. Geophysical Research Letters. 36:1-5.
Masson-Delmotte V, Zhai P, Pörtner HO, Roberts D, Skea J, Shukla PR, Pirani A,
Moufouma-Okia W, Péan C, Pidcock R, et al. 2018. Summary for policymakers:
Global warming of 1.5°c. An IPCC special report on the impacts of global
warming of 1.5°C above pre-industrial levels and related global greenhouse gas
emission pathways, in the context of strengthening the global response to the
threat of climate change, sustainable development, and efforts to eradicate
poverty. In: IPCC, editor.: IPCC. p. 1-24.
Vannote RL, Minshall GW, Cummins KW, Sedell JR, Cushing CE. 1980. The river
continuum concept. Canadian Journal of Fisheries and Aquatic Sciences. 37:130137.
Vaughn DR, Kellerman AM, Wickland KP, Striegl RG, Podgorski DC, Hawkings JR,
Nienhuis JH, Dornblaser MM, Stets EG, Robert. 2021 Oct 7. Anthropogenic
landcover impacts fluvial dissolved organic matter composition in the Upper
Mississippi River Basin.
Walsh CJ, Roy AH, Feminella JW, Cottingham PD, Groffman PM, Morgan II RP. 2005.
The urban stream syndrome: Current knowledge and the search for a cure. Journal
of The North American Benthological Society. 24(3):706-723.
Webster JR, Meyer JL. 1997. Stream organic matter budgets. Journal of the North
American Benthological Society. 16(1):11.
Wehner MF. 2013. Very extreme seasonal precipitation in the NARCCAP ensemble:
Model performance and projections. Climate Dynamics. 40:22.
Weyhenmeyer GA, Müller RA, Norman M, Tranvik LJ. 2015. Sensitivity of freshwaters
to browning in response to future climate change. Climatic Change, 134(1-2), 225–
239.

Yan Z, Kang E, Zhang K, Li Y, Hao Y, Wu H, Li M, Zhang X, Wang J, Yan L, et al.
2021. Plant and Soil Enzyme Activities Regulate CO2 Efflux in Alpine Peatlands
After 5 Years of Simulated Extreme Drought. Frontiers in Plant Science. 12.
Ylla I, Peter H, Romaní AM, Tranvik LJ. 2013. Different diversity–functioning
relationship in lake and stream bacterial communities. FEMS Microbiology
Ecology. 85(1):95-103.
Zhang R, Zhou J, Song Z, Huang Z. 2018. Enzymatic properties of β-n
acetylglucosaminidases. Applied Microbiology and Biotechnology. 102:93-103.

4.522604071

0.192648439
0.360295213

4.651956283
17.37542926
52.24437573
0.613556673
12.06357522
4.304317286
3.006796261
0.218559237
0.514929425
0.559392932
22.99855408

XYL

GLUC

PHEN

CEL

NGASE

LAMP

CQI

Paq

Naq

VLength

VAngle

0
2.93
60.95

70
0
0
4.95
63.38
36.47
0.15
0
2.717782632
17.13982301
13.9644962
0.273451327
9.938816746

% Canopy cover

% Open water

% Barren land

Watershed (km2)

% Forest (total)

% Developed (total)

% Agricultural (total)

% Wetlands (total)

MatP (μg/cm2)

MatN (μg/cm2)

MatN:P

AFDM (cm2)

MatP (μg/cm2/AFDM)

8.18

TP (μg/L)
430.26

8.735

CDOM

TN (μg/L)

62.67961165

MatN (μg/cm /AFDM)

0.27899115

1148.06

11.73

15.455

43.82425308

19.35673652

0.503539823

5.013211676

22.06725664

9.74688768

20.9

18.15

0.177539823

0.273451327

AFDM (cm )

0.503539823

28.1332413

0.40856586

27.2950426

5.115469735

15.86456366

0.588541027

322.5290543

11.81639681

Chlorophyll a (μg/cm2)

49.52018525

62.12443304

PHOS

2021

Year

HW Pohopoco Creek UNT

Kepers Run

Site Name

616.02

26.66

32.02

61.69727047

22.958887

0.178318584

5.950435864

11.00176991

4.093996222

5.88

0.94

3.03

85.15

3.83

4.55

0.862336283

0.178318584

36.8804581

0.820352485

0.656191255

0.492332445

3.951565985

4.279767906

8.655689794

0.121085959

97.55837509

24.68853499

1.541819528

25.45753038

2021

Halfway Brook

338.93

17.06

58.82115869

18.91381225

0.351327434

6.886335218

20.66548673

6.644941117

2.05

9.3

2.05

84.31

5.29

1.978300885

0.351327434

27.97553194

0.618637352

0.54634834

0.290199354

1.484953442

16.31919327

12.29784873

0.337689687

51.17816398

34.46449062

1.737415752

84.29694041

2021

Dry Brook

462.38

15.29

47.72214182

15.26707652

0.917256637

6.921459833

43.77345133

14.00382727

0.25

4.44

7.22

85.15

13.32

1.45

4.539946903

0.917256637

20.2025935

0.448564111

0.420967277

0.154907435

6.406943745

23.33213298

14.66540094

12.48119753

176.9911504

27.62489534

13.8104849

150.7067339

2021

Bouchoux Brook

377.02

31.28

18.06

49.35626536

12.20972514

0.360176991

8.95096917

17.77699115

4.397662063

11.26

10.39

77.15

2.78

1.19

0.557982301

0.360176991

26.73222713

0.523699723

0.467725921

0.235571354

5.566907073

2.052934513

19.19799776

1.979788165

103.9556456

18.67386041

4.370713215

60.59664543

2021

Berry Run

154.63

13.15

71.6

52.18836565

10.31319926

0.319469027

11.20505379

16.67256637

3.29474773

0.65

99.33

1.36

1.115964602

0.319469027

24.05265334

0.89459326

0.816916883

0.364614739

2.549854736

1.847333451

4.525130147

2.28930715

72.50239898

28.4339331

2.750870737

49.54956578

2021

Beltzville Lake UNT

452.13

17.42

40.41087613

16.66601672

0.439380531

5.369082922

17.75575221

7.322723274

0.64

3.73

10.84

84.79

2.12

0.608707965

0.439380531

29.44369716

0.844665209

0.735567546

0.41521043

5.518968907

3.941191032

7.76128118

0.430902749

179.6566798

32.55258053

1.722226667

45.84761888

2021

Appenzell Creek

253.84

9.24

5.45

54.74733096

15.87969275

0.248672566

7.634041459

13.61415929

3.94884395

1.71

4.91

2.18

84.87

5.89

2.51

2.18120354

0.248672566

17.24299425

0.747060693

0.713484944

0.221447316

1.322513482

2.295359646

12.43513698

0.087632496

48.51263461

36.68214749

1.449215103

128.965141

2021

Abe Wood Brook

Appendix I: Raw Data (Table 6)

1.949670973

0.147981327
0.321977385

0.719453038

11.98515316

34.65188186

3.397440708

0.062388909

2.891233962

0.388921331

0.775990063

0.867997915

26.61974023

XYL

GLUC

PHEN

CEL

NGASE

LAMP

CQI

Paq

Naq

VLength

VAngle

0
3.27
88.84

6.606275599
0.299557522

0.17

12.68

89.02

6.29

0.72

3.38

3.562884667

12.63716814

7.853832919

0.2

17.81442333

% Open water

% Barren land

Watershed (km2)

% Forest (total)

% Developed (total)

% Agricultural (total)

% Wetlands (total)

MatP (μg/cm2)

MatN (μg/cm2)

MatN:P

AFDM (cm2)

MatP (μg/cm2/AFDM)

33.42

TP (μg/L)

345.66

59.92

CDOM

TN (μg/L)

63.18584071

MatN (μg/cm /AFDM)

% Canopy cover

279.25

22.75

2.68

113.3707533

37.99951057

33.96106195

11.38303923

7.36

1.84

1.95

0.329716814

0.2

1.141327434

AFDM (cm )

0.299557522

24.68358814

0.354355626

18.45993596

3.995404248

14.18791976

1.006163339

159.3986566

8.634843422

Chlorophyll a (μg/cm2)

49.71605546

18.8312413

PHOS

2021

Year

Wynkoop Brook UNT

Vandermark Creek

Site Name

252.14

12.8

0.005

59.34035088

13.79425406

0.378318584

9.525451009

22.44955752

5.218622663

16.37

83.09

2.99

1.876849558

0.378318584

9.684823649

0.000841898

0.000829899

0.000141631

4043.522105

20.47315304

6.511491563

0.75024931

90.62799872

0.022413133

2.905701416

158.2281499

2021

Transue Run

277.23

7.47

2.66

29.91150442

8.224061168

0.12

8.053517062

3.589380531

0.98688734

20.97

78.89

3.94

0.14

0.334366667

0.12

18.03174436

0.728697355

0.692907502

0.225563802

1.946771236

5.714256

25.10040161

12.89334933

1.009851

44.26719333

2021

Swamp Run

446.21

19.91

7.08

56.31095406

24.04158055

0.125221239

5.186370375

7.051327434

3.010516503

1.15

23.45

1.33

72.78

3.04

0.608707965

0.125221239

25.28952286

0.995749883

0.900317893

0.425376917

3.453893584

1.315880861

41.04915236

11.88489204

0.830676342

16.05478112

2021

Shingle Brook

255.65

21.33

55.07

39.36969697

17.82595274

0.219026549

4.890384198

8.62300885

3.904356905

3.28

13.18

4.73

77.16

3.07

1.67

0.431168142

0.219026549

37.9271339

1.063586795

0.83894991

0.653743007

2.727225107

1.827898525

5.564470678

0.011394336

105.0218573

38.50868673

1.378463363

20.39624425

2021

Rocky Run

148.49

12.09

6.76

55.19266055

14.60812001

0.289380531

8.366053925

15.97168142

4.227305526

3.05

10.34

6.78

75.99

12.13

1.71

0.126814159

0.289380531

26.09584297

0.624520989

0.560857003

0.274710554

31.8107655

2.126582655

17.9760033

0.371540224

816.7182002

25.67427056

2.285692743

67.78508201

2021

Oquaga Creek

773.77

10.31

5.35

40.65508685

11.63897711

0.534955752

7.734526596

21.74867257

6.226337756

0.11

32.03

67.86

2.35

0.557982301

0.534955752

23.88768635

0.487735802

0.445956843

0.197506219

7.878071924

4.495597876

11.73003327

2.063570726

102.8894338

13.06023032

3.954238289

53.06543599

2021

Middle Creek

709.01

15.64

54.17

64.75578649

17.42680764

0.936725664

8.228002275

60.65840708

16.32413796

1.04

4.01

11.48

81.45

7.59

0.66

0.74

6.695787611

0.936725664

17.45107559

0.189081182

0.180378413

0.056703805

81.84255082

17.35947327

25.97207587

14.09530231

780.4670007

9.536200826

17.94779581

158.6394773

2021

Little Beaver Kill

532.35

54.03

71.10086455

29.99041328

0.767699115

5.249598503

54.5840708

23.02361373

24.97

3.8

70.29

11.45

0.06

0.48

7.456672566

0.767699115

22.21056573

0.126289896

0.116919298

0.047739035

20.24725886

23.86301333

21.87626018

26.58724237

122.6143512

6.05584944

40.14933917

120.7973534

2021

Laundry Brook

1.401734513

0.516681199
0.733299568

0.779258997

4.662777581

99.69079859

0.171375811

2.858812979

2.318136165

21.38013166

0.254980328

0.473872671

0.538117344

28.28370498

XYL

GLUC

PHEN

CEL

NGASE

LAMP

CQI

Paq

Naq

VLength

VAngle

0
2.73
58.67

0.8

0.19

6.13

61.13

34.23

1.57

0.11

0.424778761

8.920353982

46.5

0.134513274

3.157894737

% Canopy cover

% Open water

% Barren land

Watershed (km2)

% Forest (total)

% Developed (total)

% Agricultural (total)

% Wetlands (total)

MatP (μg/cm2)

MatN (μg/cm2)

MatN:P

AFDM (cm2)

MatP (μg/cm2/AFDM)

19.01838

TP (μg/L)

546.687

16.83

CDOM

TN (μg/L)

66.31578947

MatN (μg/cm /AFDM)

3.98

0.304353982

878.064

60.324

41.33

72.74862385

12.06605505

0.241150442

13.35036496

17.54336283

2.909734513

8.57

10.58

16.75

0.786247788

0.134513274

AFDM (cm )

0.241150442

35.16843696

0.897043877

4.008869558

5.554417876

3.441526844

0.378214749

99.15769272

24.73457699

Chlorophyll a (μg/cm2)

23.13745133

13.6240354

PHOS

2022

Year

East Mongaup

Forest Hills Run

Site Name

371.152

16.11974

3.195

65.04142012

1.420118343

0.149557522

101.4142857

9.727433628

0.212389381

2.43

1.04

6.82

84.17

12.04

0.13

0.177539823

0.149557522

16.85352081

0.688330054

0.658765653

0.199564722

3.786487626

1.75789767

2.471969322

0.343349853

30.92014074

8.165916224

1.217714454

32.75271976

2022

Drakes Creek

799.413

37.13488

53.905

76.13114754

5.114754098

0.215929204

32.95879121

16.43893805

1.104424779

8.95

1.56

8.99

77.23

6.69

0.17

0.47

1.192053097

0.215929204

45.95752107

0.444640098

0.309110015

0.319618234

31.14821643

4.576689263

15.20059469

0.540689086

275.6157373

8.848523894

1.115274336

18.83614159

2022

Decker Creek

208.91

24.45333

16.94

59.52231604

5.84800965

0.366814159

22.53748232

21.83362832

2.145132743

4.02

12.97

5.3

75.36

21.18

0.06

1.5

3.804424779

0.366814159

36.47633275

0.441485543

0.354999627

0.262459043

2.658827403

27.93951448

9.947355752

0.422187611

55.44301098

20.8524295

1.847828909

58.59779351

2022

Balls Creek

268.366

13.58343

17.395

86.06284658

7.541589649

0.239380531

25.26890756

20.60176991

1.805309735

1.42

7.32

2.82

87.87

10.22

0.26

1.597858407

0.239380531

38.8300805

0.391913254

0.305303908

0.245734658

2.082101087

29.72875785

8.723166962

0.350988791

35.18498774

16.89878938

1.205570501

51.86965192

2022

Abe Lord Creek

847.54

6.04

51.1890971

9.990777764

0.259734513

11.34519145

13.29557522

2.5949498

52.73

46.87

1.34743412

0.583345133

0.259734513

27.32990058

0.976709717

0.86768719

0.448420352

21.47400315

2.682012389

2.814734985

0.11145692

774.0697302

36.0468295

1.264888201

44.33941829

2021

Yankee Run

238.13

9.6

12.78

30.71090047

22.20007977

0.280088496

3.063174047

8.601769912

6.217986945

0.18

12.08

87.15

4.96

0.6

0.253628319

0.280088496

36.94816332

0.286104177

0.228648638

0.171974999

19.9921775

13.64637947

2.366041652

0.176676283

94.89284572

4.746498761

1.10212

22.85343599

2021

White Oak Run

413.11

24.88

21.03

21.17431193

5.821307068

0.578761062

8.054200863

12.25486726

3.369145861

0.22

4.66

3.54

85.69

4.81

1.32

0.67

5.529097345

0.578761062

21.3440529

0.449935052

0.419074754

0.163761723

5.772600275

11.25588354

10.89121499

4.12180249

92.22731635

15.97673699

5.687753923

81.58413805

2021

Whitaker Brook

587.61

22.04

6.76

480

170.412453

0.011946903

6.236968744

5.734513274

2.035900987

11.45

19.87

68.68

2.18511793

0.152176991

0.011946903

33.09963563

1.009688517

0.845838475

0.5513875

9.788585318

0.738519764

1.62391082

126.8791982

12.96195457

0.54668944

10.54593156

2021

WB Mongaup River

2.222626549

1.539235398

10.48501947

224.4375733

0.253249558

XYL

GLUC

PHEN

CEL

0.330376882
0.553137939

21.40554664

0.29624442

0.297385062

0.419760207

44.88990794

CQI

Paq

Naq

VLength

VAngle

16.88

2.23

0.67

1.550442478

22.9380531

32.7592955

0.256637168

6.04137931

% Developed (total)

% Agricultural (total)

% Wetlands (total)

MatP (μg/cm2)

MatN (μg/cm2)

MatN:P

AFDM (cm2)

MatP (μg/cm2/AFDM)

34.23624

TP (μg/L)

1382.142

21.11

CDOM

TN (μg/L)

89.37931034

MatN (μg/cm /AFDM)

79.16

% Forest (total)

0.16

6.65

0.71

% Open water

% Barren land

% Canopy cover

Watershed (km2)

1.22

2.105115044

632.042

18.65605

14.48

61.11298482

7.042158516

0.524778761

19.21592775

32.07079646

3.695575221

2.99

9.71

5.53

79.01

28.3

2.916725664

0.256637168

AFDM (cm )

Chlorophyll a (μg/cm2)

0.524778761

30.84892467

0.644290667

1.910705349

19.12920195

12.20182183

5.6764059

19.09595917

LAMP

1.049364012

74.1017166

38.7823882

NGASE

78.6059351

24.90811799

PHOS

2022

Year

Sherman Creek

Sloat Brook

Site Name

368.423

23.72867

20.94

76.21374046

5.221374046

0.231858407

32.32080201

17.67079646

1.210619469

0.37

0.42

29.18

69.26

5.36

0.03

0.49

2.308017699

0.231858407

70.05160461

0.458164156

0.156313538

0.430674438

5.257723968

33.219249

1.805912684

0.098120354

34.11877599

6.489267257

1.251306195

8.578418879

2022

Randall Creek

1680.234

31.69993

13.91

70.58823529

5.915966387

0.157964602

26.42045455

11.15044248

0.934513274

11.35

0.22

86.91

4.06

0.95

1.192053097

0.157964602

31.02178602

0.505746724

0.433410479

0.260643637

10.47151065

5.491843245

2.22840472

0.183519764

61.84028148

5.905574041

0.935269617

16.7520826

2022

Pine Kill

3768.801

28.07663

0.33

60.34865293

7.75911252

0.837610619

17.22222222

50.54867257

6.499115044

1.79

1.84

13.85

78.86

4.83

0.3

1.52

10.98210619

0.837610619

44.87948059

0.396001972

0.280604062

0.279426059

8.479994387

67.82106463

12.92654159

1.382441298

267.0860433

31.49601652

3.752666667

81.22080236

2022

Mill Brook

175.093

31.69993

11.21

75.68147527

12.31181894

0.527876106

13.61134454

39.95044248

6.499115044

3.86

7.29

12.28

70.86

8.95

0.33

4.33

1.902212389

0.527876106

45.89316344

0.924289727

0.643304244

0.663680005

1.994432451

37.20838174

13.33101357

2.107161062

181.7891033

91.14828791

4.042554572

46.18941593

2022

Little Equinunk Creek UNT

431.969

19.74304

1.79

68.27436823

7.104693141

0.490265487

21.27874564

33.47256637

3.483185841

1.12

4.42

3.09

90.78

10.21

0.35

0.05

4.489221239

0.490265487

31.83111031

0.441636017

0.375216805

0.232926427

1.7001038

28.33060593

14.48468909

0.900012979

43.71468174

25.71294867

2.246130973

84.6779115

2022

Laurel Creek

1521.363

12.13411

14.935

74.1686747

13.87951807

0.146902655

11.83258929

10.89557522

2.038938053

0.07

0.55

32.46

65.85

6.88

0.304353982

0.146902655

36.81880696

0.901313689

0.721532886

0.540145037

3.813170528

0.821233658

3.549255457

0.367148083

43.18157586

11.3243233

1.061116224

9.641014749

2022

Jonas Creek

105.187

7.42382

8.67

41.25329429

3.162518302

0.302212389

28.88412698

12.46725664

0.955752212

0.85

0.64

1.91

96.6

5.59

0.126814159

0.302212389

30.79954429

0.769344374

0.660839097

0.393932041

11.8534748

1.943666106

7.012240708

0.604738643

206.8450794

17.45016401

1.934305605

26.84723304

2022

Hawk Run

205.369

16.48207

60.52

87.53623188

8.695652174

0.183185841

22.29047619

16.03539823

1.592920354

0.43

1.18

3.44

92.91

4.87

0.04

1.16

2.409469027

0.183185841

56.6451301

0.512340172

0.281696403

0.427948114

10.52113325

19.13506832

2.054471976

0.66340177

87.42936347

8.309880826

1.487623599

11.1080826

2022

Fuller Creek

1.662927434

0.546790784
0.609770688

1.767130383
10.45270088
63.25301205
0.280083776
4.934057817
17.78775829
6.051355796
0.168674147
0.315082233
0.357390237
28.16164538

XYL
GLUC
PHEN
CEL
NGASE
LAMP
CQI
Paq
Naq
VLength
VAngle

0
2.04
83.67

87
0.03
0.28
6.09
75.37
21.7
0.91
0.47
2.038938053
19.11504425
20.75892857
0.272566372
7.480519481

% Canopy cover
% Open water
% Barren land
Watershed (km2)
% Forest (total)
% Developed (total)
% Agricultural (total)
% Wetlands (total)
MatP (μg/cm2)
MatN (μg/cm2)
MatN:P
AFDM (cm2)
MatP (μg/cm2/AFDM)

48.00478

TP (μg/L)

1851.99

4.66

CDOM

TN (μg/L)

70.12987013

MatN (μg/cm /AFDM)

0.18

2.257292035

955.724

45.46847

12.24

65.58072289

6.939759036

0.183628319

20.925

12.04247788

1.274336283

6.56

0.44

8.9

0.40580531

0.272566372

AFDM (cm )

0.183628319

41.88306901

0.819024086

20.32815073

2.443285428

6.619520944

0.536282006

287.8771724

14.16150324

Chlorophyll a (μg/cm2)

11.73780531

51.51708555

PHOS

2022

Year

Westcolang Creek

Willsey Brook

Site Name

559.491

76.26652

5.1

65.13661202

9.530054645

0.485840708

15.13433814

31.6460177

4.630088496

0.73

21.79

3.84

70.13

4.32

0.02

0.17

1.242778761

0.485840708

53.77252043

0.63697125

0.376445311

0.513830031

13.0976854

55.72656168

24.63860059

1.683493805

635.4622028

48.51713746

3.301185841

45.90540413

2022

WB Mongaup River

606.393

25.17799

68.06486486

10.03243243

0.245575221

15.02278325

16.71504425

2.463716814

1.24

10.08

7.65

75.71

1.74

3.77

1.166690265

0.245575221

53.6023117

0.511020826

0.303232814

0.411329728

10.87756051

16.02882451

5.142659587

0.466748083

100.2239045

9.213821829

1.695931563

13.18626549

2022

UE Branch Callicoon UNT

3898.204

135.32631

30.76

61.40101523

9.015228426

0.348672566

15.08108108

21.40884956

3.143362832

2.98

24.83

6.16

62.18

5.6

2.44

0.076088496

0.348672566

50.17867549

0.335571417

0.214898458

0.257734027

68.65055118

22.45347923

15.19373923

0.944965192

707.4314959

10.30481888

1.622186431

29.67755752

2022

Tributaryof Shehawken

622.011

29.52595

27.62

74.06726825

15.2780968

0.539380531

10.73472018

39.95044248

8.240707965

0.82

7.34

3.53

87.53

7.2

0.27

0.01

4.996477876

0.539380531

49.06844896

0.959675868

0.628738305

0.725028217

0.684735787

55.67767525

12.82175103

4.19259115

79.43277535

116.0050006

5.583073746

43.99566962

2022

Tarbell Brook

501.168

41.12051

25.92

32.92786421

3.46251768

0.312831858

21.05742297

10.30088496

1.083185841

2.3

0.16

5.34

90.66

2.74

0.07

1.547132743

0.312831858

35.47474325

0.543372374

0.442506932

0.315342911

3.578442478

4.870007847

6.391332153

0.22719882

31.98635249

8.938624189

1.152979351

19.40710324

2022

Spackmans Creek

Appendix II: List of Figures and Tables
Table 7: Figures used throughout this document
Figure 1
Map showing all 46 sample locations within the Delaware River Basin.
Figure 2
Plot of proportions of N-acquiring enzymatic activity plotted against
proportions of P-acquiring enzymatic activity.
Figures 3a & 3b Plots of positive linear relationships between the angle of vectors derived
from an analysis of nitrogen- and phosphorus-acquiring and both
percentage canopy cover and water column total phosphorus.
Figures 4a & 4b Plots of negative linear relationships between the length of vectors derived
from an analysis of nitrogen- and phosphorus-acquiring and both biofilm
nitrogen and biofilm phosphorus.
Figures 5a & 5b Plots of negative linear relationships between the length of vectors derived
from an analysis of nitrogen- and phosphorus-acquiring and both
chlorophyll a and AFDM.
Table 8: Equations used throughout this document
Equation 1
𝑣𝐴𝑛𝑔𝑙𝑒 = −3.838 + 10.20(log 𝑀𝑎𝑡𝑁: 𝑃) + 18.76(log 𝑇𝑃)
Equation 2
𝑣𝐴𝑛𝑔𝑙𝑒 = 5.170 + 20.69(log 𝑇𝑃)
Equation 3
𝑣𝐴𝑛𝑔𝑙𝑒 = −3.736 − 1.022(log 𝐶ℎ𝑙𝐴) + 10.20(log 𝑀𝑎𝑡𝑁: 𝑃)
+ 18.96(log 𝑇𝑃)
Equation 4
𝑣𝐿𝑒𝑛𝑔𝑡ℎ = 0.8472 − 0.3901 (log 𝐶ℎ𝑙𝐴) − 0.2724(CQI)
+ 0.1487(log 𝑀𝑎𝑡𝑃)
Equation 5
𝑣𝐿𝑒𝑛𝑔𝑡ℎ = 1.008 − 0.4026 (log 𝐶ℎ𝑙𝐴) − 0.2676(𝐶𝑄𝐼)
Equation 6
𝑣𝐿𝑒𝑛𝑔𝑡ℎ = 0.7712 + 0.04781(log 𝐶𝐷𝑂𝑀) − 0.3704 (log 𝐶ℎ𝑙𝐴)
− 0.2633(𝐶𝑄𝐼) + 0.1590(log 𝑀𝑎𝑡𝑃)
Tables 9: Tables used throughout this document
Table 1
Enzymes of interest, their letter designation, enzyme code, and specific
purpose
Table 2
Specific substrates to be used for the enzyme assays
Table 3
Correlation values and linear regression p-values between EEAs and
environmental variables
Table 4
Ranked vector angle multiple regression models
Table 5
Ranked vector length multiple regression models
Table 6
All raw data collected during the study
Table 7
Figures used throughout this document
Table 8
Tables used throughout this document
Table 9
Tables used throughout this document

Appendix III: R Code
Scatterplot Correlation
library(assertthat)
library(tidyverse)
library(ggplot2)
library(GGally)
library(plotly)
library(vegan)
#import data_set.txt from "scatterplot matrix correlations" folder
data<-data_set_bm
#add log10(x+1) transformed variables to dataframe
data<-data%>%mutate(LXYL=log10(XYLbm+1),
LGLUC=log10(GLUCbm+1),
LPHEN=log10(PHENbm+1),
LCEL=log10(CELbm+1),
LNGASE=log10(NGASEbm+1),
LLAMP=log10(LAMPbm+1),
LPHOS=log10(PHOSbm+1),
LForest=log10(forest+1),
LCanopy=log10(Canopy_Cover_percent+1),
LcDOM=log10(cDOM+1),
LWetland=log10(1+percent.wetland),
LDeveloped=log10(development+1),
LmatP=log10(biofilm_P_ug_cm2+1),
LmatN=log10(biofilm_N_ug_cm2+1),
LmatRatio=log10(molar_NtoP+1),
LmatPAFDM=log10(biofilm_P_AFDM+1),
LmatNAFDM=log10(biofilm_N_AFDM+1),
LTP=log10(TP+1),
LTN=log10(TN+1),
LAFDM=log10(AFDM_cm2+1),
LChlA=log10(Chlor+1))
attach(data)
#enzymes vs. enzymes
enzymes<-data.frame(LXYL,LGLUC,LPHEN,LCEL,LNGASE,LLAMP,LPHOS)
enzymes_matrix<-ggpairs(enzymes)+theme_classic()
graphEnzymes_matrix<-ggplotly(enzymes_matrix)
graphEnzymes_matrix
#enzymes vs. envvar
enzymes_envvar<data.frame(LXYL,LGLUC,LPHEN,LCEL,LNGASE,LLAMP,LPHOS,LCanopy,LForest,LWe
tland,LTP,LmatNAFDM,LmatPAFDM,LmatRatio,LcDOM)
enzymes_envvar_matrix<-ggpairs(enzymes_envvar)+theme_classic()
graphenzymes_envvar_matrix<-ggplotly(enzymes_envvar_matrix)
graphenzymes_envvar_matrix

#envvar vs. envvar
envvar_envvar<data.frame(LCanopy,LForest,LWetland,LTP,LmatNAFDM,LmatPAFDM,LmatRatio,L
cDOM)
envvar_envvar_matrix<-ggpairs(envvar_envvar)+theme_classic()
graphenvvar_envvar_matrix<-ggplotly(envvar_envvar_matrix)
graphenvvar_envvar_matrix
#linear regression blank
investigate<-lm(XXX ~ XXX , data = data)
summary(investigate)

Vector Analysis
library(tidyverse)
library(REdaS)
library(assertthat)
library(ggplot2)
library(GGally)
library(plotly)
library(vegan)
library(Rcpp)
library(data.table)
#import vector.txt from "vector analysis" folder
data<-vector
attach(data)
#x = P-acquiring enzymes
P<-Paq
x<-c(P)
#y = N-acquiring enzymes
N<-Naq
y<-c(N)
data<-tibble(x,y)
ggplot(data,aes(x,y))+
geom_point()+
xlab("GLU/(GLU+PHOS)")+
ylab("GLU/(GLU+NGASE+LAMP)")+
theme(panel.grid.major = element_blank(), panel.grid.minor =
element_blank())+
labs(title = "N-acquiring vs. P-acquiring Proportions")+
theme(plot.title = element_text(hjust = 0.5))
#add lengths and angles to dataframe
data<-data%>%mutate(length=sqrt((x^2)+(y^2)),angle=rad2deg(atan2(x,y)))
print(data, n = 46)
write.table(data, file="vector_analysis.csv",sep=",",row.names=F)
#saved .csv to files, added to vector.txt
#reimport, reload dataset
data<-vector
attach(data)
#plot lengths and angles against environmental variables
envvar_limits<data.frame(Vlength,Vangle,Canopy_Cover_percent,percent.wetland,forestTP
,TN,biofilm_N_AFDM,biofilm_N_ug_cm2,biofilm_P_AFDM,biofilm_P_ug_cm2,mol
ar_NtoP,cDOM,AFDM_cm2,Chlor)
envvar_limits_matrix<-ggpairs(envvar_limits)+theme_classic()
graphenvvar_limits_matrix<-ggplotly(envvar_limits_matrix)
graphenvvar_limits_matrix

#regression analysis
#vector angle blank
investigate_Vangle<-lm(Vangle ~ XXX, data = data)
summary(investigate_Vangle)
#vector length blank
investigate_Vlength<-lm(Vlength ~ XXX, data = data)
summary(investigate_Vlength)

Multiple Regression Analysis
library(tidyverse)
library(MuMIn)
#import "MRdata" with headings
data<-MRdata
attach(data)
LogData<-data%>%mutate(LXYL=log10(XYLbm+1),
LGLUC=log10(GLUCbm+1),
LPHEN=log10(PHENbm+1),
LCEL=log10(CELbm+1),
LNGASE=log10(NGASEbm+1),
LLAMP=log10(LAMPbm+1),
LPHOS=log10(PHOSbm+1),
LCQI=log10(RECAL+1),
LForest=log10(forest+1),
LCanopy=log10(Canopy_Cover_percent+1),
LcDOM=log10(cDOM+1),
LWetland=log10(1+percent.wetland),
LDeveloped=log10(development+1),
LmatP=log10(biofilm_P_ug_cm2+1),
LmatN=log10(biofilm_N_ug_cm2+1),
LmatRatio=log10(molar_NtoP+1),
LmatPAFDM=log10(biofilm_P_AFDM+1),
LmatNAFDM=log10(biofilm_N_AFDM+1),
LTP=log10(TP+1),
LTN=log10(TN+1),
LChlA=log10(Chlor+1),
Vangle=Vangle+0,
Vlength=Vlength+0)
attach(LogData)
options(na.action = "na.fail")
vAngle_lm<-lm(Vangle ~ LTP + LTN + LmatRatio + LChlA, LogData)
vAngle_D<-dredge(vAngle_lm,trace=TRUE,rank="AICc",REML=FALSE)
vAngle_D #displays all multiple regression combinations ranked by AICc
#top 10 multiple regression models for VAngle:
vAngle_lm1<-lm(Vangle ~ LmatRatio + LTP, LogData)
summary(vAngle_lm1)
vAngle_lm2<-lm(Vangle ~ LTP, LogData)
summary(vAngle_lm2)
vAngle_lm3<-lm(Vangle ~ LChlA + LmatRatio + LTP, LogData)
summary(vAngle_lm3)
vAngle_lm4<-lm(Vangle ~ LmatRatio + LTN + LTP, LogData)
summary(vAngle_lm4)
vAngle_lm5<-lm(Vangle ~ LChlA + LTP, LogData)
summary(vAngle_lm5)

vAngle_lm6<-lm(Vangle ~ LTN + LTP, LogData)
summary(vAngle_lm6)
vAngle_lm7<-lm(Vangle ~ LChlA + LmatRatio + LTN + LTP + LTN, LogData)
summary(vAngle_lm7)
vAngle_lm8<-lm(Vangle ~ LChlA + LTN + LTP, LogData)
summary(vAngle_lm8)
vAngle_lm9<-lm(Vangle ~ LmatRatio, LogData)
summary(vAngle_lm9)
vAngle_lm10<-lm(Vangle ~ LmatRatio + LTN, LogData)
summary(vAngle_lm10)
Vlength_lm<-lm(Vlength ~ LTP + LmatPAFDM + LChlA + LCanopy + LForest +
LcDOM + LCQI, LogData)
Vlength_D<-dredge(Vlength_lm,trace=TRUE,rank="AICc",REML=FALSE)
Vlength_D #displays all multiple regression combinations ranked by AICc
#top 10 multiple regression models for VAngle:
Vlength_lm1<-lm(Vlength ~ LChlA + LCQI + LmatPAFDM, LogData)
summary(Vlength_lm1)
Vlength_lm2<-lm(Vlength ~ LChlA + LCQI, LogData)
summary(Vlength_lm2)
Vlength_lm3<-lm(Vlength ~ LcDOM + LChlA + LCQI + LmatPAFDM, LogData)
summary(Vlength_lm3)
Vlength_lm4<-lm(Vlength ~ LChlA + LCQI + LForest + LmatPAFDM, LogData)
summary(Vlength_lm4)
Vlength_lm5<-lm(Vlength ~ LChlA + LCQI + LForest, LogData)
summary(Vlength_lm5)
Vlength_lm6<-lm(Vlength ~ LcDOM + LChlA + LCQI, LogData)
summary(Vlength_lm6)
Vlength_lm7<-lm(Vlength ~ LCanopy + LChlA + LCQI + LmatPAFDM, LogData)
summary(Vlength_lm7)
Vlength_lm8<-lm(Vlength ~ LCanopy + LChlA + LCQI, LogData)
summary(Vlength_lm8)
Vlength_lm9<-lm(Vlength ~ LChlA + LCQI + LmatPAFDM + TP, LogData)
summary(Vlength_lm9)
Vlength_lm10<-lm(Vlength ~ LChlA + LCQI + TP, LogData)
summary(Vlength_lm10)