Assessment plots and statistical analysis
This report provides details of the assessment of mercury concentrations in plaice muscle at station North Sea (DOGGBK).
-
Timeseries metadata. Key information about the timeseries.
-
Assessment plot. The fitted trend (solid line) with pointwise 90% confidence bands (grey shaded areas). The trend is plotted on the log scale, with the axis labels back-transformed for ease of interpretation. The points are the annual medians of the individual log concentration measurements, with solid circles denoting uncensored values and < denoting less-than values. The assessment criteria are indicated on the right hand side of the plot and by horizontal dashed lines (if within the range of the data).
-
Trend with data. The fitted trend with pointwise 90% confidence bands. The points are the individual concentration measurements, with + denoting uncensored values, d or q denoting values below the limit of detection or quantification and < denoting other less-than values.
-
Auxiliary data. The individual concentration measurements with supporting information: mean length of the individuals in the sample; dry weight (%) of the sample; lipid weight (%) of the sample.
-
Assessment (related compounds). Assessment plots for all the compounds in the same chemical group measured in plaice muscle at this station.
-
Data (related compounds). A scatter plot matrix of all the individual concentration measurements for all the compounds in the same chemical group in plaice muscle at this station.
-
Statistical analysis. This tab summarises the fits of models with different amount of smoothing. The final choice is the model with the minimum AICc (with the caveat that a linear model is always preferred to a mean model if there are sufficient years of data). The tab also gives the contrast between the fitted values at the start and end of the timeseries, and over the most recent twenty years. Finally, the tab assesses status by comparing the fitted value at the end of the timeseries to the assessment criteria.
Assessment plot
Trend with data
Auxiliary data
Statistical analysis
Trend assessment
Analysis of variance
Df AIC AICc Log lik Deviance Chisq Chi df Pr(>Chisq)
mean 1 -5.743991 -2.143991 5.871995 -11.74399 NA NA NA
linear 2 -5.348201 2.651799 6.674101 -13.34820 1.604211 1 0.20530749
smooth (df = 2) 3 -6.647175 10.019491 8.323588 -16.64718 3.298974 1 0.06932318
Change in log concentration
Year start Year end Fit start Fit end Change Std error t Pr(>|t|)
overall 2014 2021 3.081327 3.362424 0.2810971 0.2082132 1.350044 0.2257031
last 20 years 2014 2021 3.081327 3.362424 0.2810971 0.2082132 1.350044 0.2257031
Status assessment
Conc fitted Conc ref Log ratio Std error t Pr(>t)
QSsp 28.85905 20 -0.3666914 0.1276684 -2.872218 9.858276e-01
MPC 28.85905 500 2.8521844 0.1276684 22.340566 2.630774e-07
Interpretation
There is no significant temporal trend in the time series (p = 0.2053).
Concentrations at the end of the time series were not significantly below the QSsp (p = 0.9858). Concentrations could therefore have harmed marine life. Concentrations were significantly below the MPC (p < 0.0001), so were unlikely to have harmed human health.