Microcystin_Tox

This repository summarizes and analyzes aquatic toxicity data for microcystins to support water quality criteria development for California freshwater species. Two analyses are implemented following the US EPA (1985) Guidelines for Deriving Numerical National Water Quality Criteria for the Protection of Aquatic Organisms and Their Uses (Stephan et al. 1985).

Repository Contents

File / Folder	Description
CCC.qmd	Criterion Continuous Concentration (CCC) derivation — step-by-step EPA 1985 method in base R
SSD.qmd	Species Sensitivity Distribution (SSD) construction — log-normal fits across four species scopes using ssdtools
Data/ToxSummary.csv	Compiled microcystin toxicity dataset (see below)

Data: ToxSummary.csv

Each row is a single toxicity test result. Key columns:

Column	Description
source	Source database or review (e.g., Mehinto)
Species	Common name with scientific binomial in parentheses
reference	Full literature citation
life_stage	Life stage tested (e.g., Embryonic, Juvenile, Adult)
endpoint	Biological endpoint (Growth, Development, Reproduction, Mortality, Growth + Development, Behavior, Histology, Subcellular)
congener	Microcystin congener tested (e.g., Microcystin-LR)
exp_days	Exposure duration in days
chronic_acute	Whether the exposure is chronic or acute
noec_ug_l	No Observed Effect Concentration (µg/L)
loec_ug_l	Lowest Observed Effect Concentration (µg/L)
other_dose_ug_l	Other reported dose values (µg/L)
doses	Full dose series tested
group	Broad taxonomic group (e.g., Chordata, Crustacea, Insecta)
ca_resident	Yes if the species has an established California population
ca_native	Yes if the species is native to California
ca_native_genus	Yes if the genus is native to California
ca_relevant	Yes if the species is considered relevant to California conditions (residents + widely used non-native test species)
note	Flags such as Invasive

Analyses

CCC Derivation (CCC.qmd)

Implements the EPA FAV/FACR fallback method to produce a Criterion Continuous Concentration:

1. Load data and parse genus from species names
2. Filter to apical endpoints (Growth, Development, Reproduction, Mortality, Growth + Development)
3. Apply ca_relevant == "Yes" scope filter
4. Chronic side — compute Chronic Values (√NOEC×LOEC), Species Mean Chronic Values (SMCV), and Genus Mean Chronic Values (GMCV)
5. Acute side — restricted further to Mortality endpoint; compute Acute Values, SMAV, GMAV
6. Fit EPA's log-triangular extrapolation to the four lowest GMAVs to produce the Final Acute Value (FAV)
7. Compute the Final Acute-Chronic Ratio (FACR)
8. CCC = FAV / FACR

Note: The dataset does not meet EPA's 8-family minimum data requirement on either the chronic (2 genera / 2 categories) or acute (6 genera / 3 categories) side. The resulting CCC (~0.03 µg/L) is a worked demonstration of the calculation mechanics, not a defensible regulatory criterion.

SSD Construction (SSD.qmd)

Fits log-normal SSDs across four species-scope variants using ssdtools:

Variant	Filter	N species
All Species	none	maximum available
California Relevant	ca_relevant == "Yes"	subset comparable to CCC scope
California Residents	ca_resident == "Yes"	established CA populations only
California Native Genus	ca_native_genus == "Yes"	CA-native genera

Data preparation steps applied before fitting: apical endpoint filter, LOEC/2 substitution for missing NOECs, acute values divided by 10 (fixed ACR), and per-species median collapse.

A fifth variant (ca_native == "Yes") is attempted but fails due to insufficient species for reliable log-normal fitting.

Reproducing the Analyses

The analyses are written as Quarto (.qmd) documents. To render them:

1. Install R and Quarto.
2. Install required R packages:

   install.packages(c("tidyverse", "ssdtools", "ggplot2"))

   CCC.qmd uses only base R — no additional packages needed.
3. Render from the terminal:

   quarto render CCC.qmd
   quarto render SSD.qmd

   Or open in RStudio and click Render.

Output is written as self-contained HTML and Word (.docx) files.

Acknowledgments

Code and document text in this repository were developed with assistance from Claude (Anthropic), an AI assistant. All code, analytical methods, and written content were reviewed and verified for accuracy by Leah Thornton Hampton prior to publication.

Contact

Leah Thornton Hampton Southern California Coastal Water Research Project (SCCWRP) leahth@sccwrp.org

For questions about the dataset, analysis methods, or California-specific scoping decisions, please reach out via email.