2020-07-07 P&E Agenda - Item 3 - Attachment 1 - TNC Project Management Plan

the benchmark. This scoring will be consistent with those allocated for the ecological and socio-economic target
indicators.

Evaluation steps

Summarise
Initially for each performance indicator, where applicable, the data will be compiled and summarised returning
mean, standard deviation, standard error, coefficient of variation (CV) and a simple calculation of sample size (n)
to understand if increased sampling effort is required. Metrics will also be visualised by plotting time series plots
that visualise how the metric has varied through time.

Score
Following summarising the data a score will be calculated in reference to a benchmark. This benchmark will be
either a pre-defined benchmark in reference to a known condition i.e. 50 shellfish per m2 that defines a viable
population density similar to that found in natural shellfish ecosystems (Gillies et al. 2017). If no pre-determined
benchmark exists, the indicator will be measured against a benchmark defined by the mean of the value of the
indicator as measured at control sampling locations. The results will then be standardised by providing % of the
measure compared to the benchmark and categorised into one of six score categories (see Figure 8).

Trend
The trend for each indicator will be assessed by fitting the raw data for each site as a linear model. The significance
of the t-value, that assesses whether the coefficient is significantly different to zero will first be examined. If the
metric is not significantly different to zero, the slope will then automatically be assigned to the ‘Stable’ category.
If slope is significant the value of the slope coefficient is then categorized into one of 6 categories (see Figure 8).

Additional analysis
Where necessary additional analyses will be undertaken to statistically assess differences of metrics between sites
i.e. impact versus control. These analyses will be univariate (Analysis of variance - ANOVA) or multivariate analyses
(permutational multivariate analysis of variance) where appropriate.

The data will be evaluated using the R language and environment for statistical computing alongside RStudio
which allows scripts of code to be created ensuring standard analysis workflows and evaluation products




65 | Bring Back the Fish MER plan 2020 – 2022