Distribution fits & goodness of fit
Distributions are fit by maximum likelihood on the µg/L scale; the SSD curve and HC values
are AICc-weighted model averages (Burnham & Anderson), matching ssdtools.
Lower AICc and Anderson-Darling (AD) indicate better fit. CIs are non-parametric bootstrap
(species resampled with replacement).
Data preview
Most sensitive species
Per-species value distribution (rough baseline)
Every test record for each species, pooled across all chemicals and shown as a probability density (so species with very different record counts stay comparable). Because species are tested on different sets of chemicals, this does not control for chemical potency — treat it as a rough eyeball baseline; the pairwise shared-chemical comparison below is the rigorous one.
Pairwise sensitivity ratios
Each cell is the geometric-mean sensitivity ratio SR(row → column) over the chemicals
the two species share: SR < 1 means the row species is more sensitive (lower
threshold) by a factor of 1/SR. n is the number of shared chemicals; cells
based on fewer than 5 are flagged as low-confidence.