Research Proposal Overview

The following are the aims of the research project:

• 1. Which mixtures of urban stressors (elevated water temperatures, low oxygen levels, and variable contaminant mixtures) lead to synergistic (multiplicative) or additive sub-lethal effects on stream biota?
• 2. What urban contaminant mixtures exacerbate or negate sub-lethal effects?
• 3. What are the primary sources, in the landscape, of these contaminants?

General Overview

The following is exploratory analyses to explore data collected in Ellerbe Creek during the 2021-23 synoptic surveys.

Conceptual Model

Major Milestones

• Water sampling
• Chemical analyses of water samples (ions, nutrients, trace metals, indicator compounds)
  • Ions + Nutrients
  • Trace Metals
  • Indicator compounds
• Ecotoxicology assays:
  • Zebrafish assays (mortality, deformities, behavior; basal metabolism)
  • Wild caught fish (same as above, mosquito fish)

Remaining Questions:

• Sites to select for sampling? Timing?
• Right ecotox assays?
• Target more indicator compounds for source attribution?

Initial Data Results

The following slides show some preliminary analyses on water chemistry.

Overview of Sampling Sites

We will target 10 samples, as a minimum. Possibly from existing samples that have been collected or through new sample collection at the same sites.

Map of Sites

Here’s a simple map of where the sampling locations are located.

Timing of Sampling

PCA of All Sites: No Metals

Bi-Plots

NOTE: Metals, AMPA, and Glyphosate are removed, given missing data for a sampling date. They will be added later

• Main branch clusters into two, based on the WWTP point-source: Sites upstream vs. downstream
• Tributary chemical mixtures are MORE spread relative to the main branch sites upstream of the WWTP

Indicator Compound Exploration

Let’s explore the three organic indicator compounds that were measured throughout Ellerbe Creek

• Sucralose: artificial sweetener, used as an indicator of treated and untreated waste water
• Diphenyl guanidine (DPG): rubber additive used in tires, used as an indicator of road runoff
• Glyphosate + AMPA: the active ingredient of RoundUp (pesticide) and it’s major degradation product, used as an indicator of road runoff.

Also will show some results from 3 trace metals:

• Cu
• Ni
• Pb
• Zn

Indicator Compounds: Tributaries - Exclude Effluent Channel

• All data is normalized by compound (centered), to enable cross comparison of tributaries
• Metals not measured in 2021-09-25 yet

Sucralose: Main Branch

• Check out the impact of the effluent!
• Brown line is where effluent from N. Durham WWTP is released into Ellerbe Creek

DPG: Main Branch

• Note, y-axis is log-transformed to visualize all data points

AMPA: Main Branch

• Check out both the seasonality and dilution effect of effluent!
• Glyphosate was not detected in samples, thus only AMPA is shown
• Data is not yet available for the 2022-06-16 samples

Indicator Compounds: Tributaries - Effluent Channel Included

• All data is normalized by compound (centered), to enable cross comparison of tributaries
• Metals not measured in 2021-09-25 yet
• Effluent concentrations are included (in both normalization and this graph)

Additional Slides

The following slides are a deeper dive into the PCA analyses.

PCA Exploration

We will look at a few different PCAs.

• Tributaries Only
• All Sites

For now we are not including AMPA and Glyphosate due to some concerns with the QA/QC of the data. E.coli for similar reasons will not be included quite yet, except for PCAs of specific dates.

PCA of Tributaries: No Metals Part 1

• Metals are excluded, since they are missing for a few dates of sampling. They will be added later.

Each PC’s Contribution

First, let’s look at the loadings. Looks like the PC1 has a decent amount of the variation. PC2 and PC3 are relatively close to each other.

PCA of Tributaries: No Metals Part 2

Variable Loading to Each PC

Okay let’s see what variable is loading most to which PC.

Possibly PC1 is driven by wastewater, like before. Not quite sure for PC2 and PC3.

PCA of Tributaries: No Metals Part 3

PCA Graphs

Bi-Plots

• There’s some interesting separation of the 9-25-2021 date from the other two. Might be worth diving into further!

PC1 vs. PC2

PC1 vs. PC3

PCA of All Sites: No Metals Part 1

Each PC’s Contribution

First, let’s look at the loadings. Looks like the PC1 has a decent amount of the variation. PC2 and PC3 are close to each other.

PCA of All Sites: No Metals Part 2

Variable Loading to Each PC

Okay let’s see what variable is loading most to which PC.

Possibly PC1 is driven by waste water and PC2 by road runoff? Not quite sure about PC3.

PCA of All Sites: No Metals Part 3

PCA Graphs

Bi-Plots

• The main branch clusters into two clusters. The ones that are centered are all upstream of the effluent release point. The ones that are to the right side (PC1) are all downstream of the wastewater treatment plant.
• There is a wider relative spread in the tributary sites relative to the cluster of main branch sites of the WWTP.

PC1 vs. PC2

PC1 vs. PC3

Other Experimental Analyses

Additional code, not shown in the HTML, exists to explore the data further. It can be seen only on the RMD file, currently.