Effect of sampling interval
What we see through the window requires knowledge of the
processes - Catch 22
think about the processes
design a pilot sampling scheme
use pre-existing data
see what it tells us
refine if necessary
Differences due to the putative (proposed) cause cannot be distinguished
from other factors that might produce variations in the effect
Good monitoring designs severe tests
Reduce the possibility of confounding
Effects of measurement scale
SPACING TOO BIG - tends to measure noise since cannot
capture small scale variability
SUPPORT TOO BIG - smoothes out variability
EXTENT TOO SMALL - variability appears as trend
in cyclical processes, need to be careful of
Accuracy and Precision
Which has high precision and which high accuracy?
What can we do to overcome low accuracy and precision?
The spread of a group of points
Standard deviation of a sample
Assessment Endpoints Environmental Flow Impacts
There are too many things that we could monitor (13 broad groups of endpoints
Assessment Endpoints). Initial pruning based on
How good is the theoretical / empirical support for the effect?
Is it relevan
Original (null) hypothesis is rejected when the probability (pr or p) of
observing the sample data (or more extreme data than this) under the
null hypothesis is less than the significance level ()
It is common to set = 0.05
Paired or Block Designs
Group locations into blocks with similar background conditions
Blocks might be close in space or time or some other factor (e.g.
elevation, geology )
Comparisons are made within blocks, then averaged among
Steps for Desig
Types of Monitoring
Types of Monitoring
Surveillance monitoring / condition monitoring
Impact assessment / intervention monitoring
Attributing causality to the human activity
What are cause
Scale of Replication
Replication must match the scale of the impact being assessed!
Broad study design types in intervention monitoring
Has the intervention caused a change?
Control sites, before data, continuous designs, all reduce the likelihood of rea
Measurement and Assessment Endpoints
Suter (1990) [in the context of ecological risk assessment] defines.
Formal expressions of the actual environmental value of concern to stakeholders
Examples are: The security of a water supply
Hypothesis Testing Comparison of Two Samples
Flow regulation does not reduce channel depth
Choose test statistic:
t statistic (for testing difference in mean observed/expected)
Choose a priori significance level