2.3 Field Methods
49
Analysis of trends is most often done on an annual basis using one of several
approaches that incorporate seasonal effects (Helsel and Hirsch 1992, Lawrence
et al. 2004). These approaches are most effective if multiple samples are collected for each season. Weekly sampling provides a sufficient number of samples to account for within-season variability and is likely to enable a trend to be
detected with fewer years of monitoring data than data collected at longer intervals. Annual or quarterly sampling is less expensive than weekly sampling but
cannot account for within-season variability and, relative to weekly sampling,
can substantially increase the length of time needed to detect a trend (Murdoch
and Shanley 2006). However, annual or quarterly, as opposed to weekly, sampling may free up resources to monitor more sites to obtain a better picture of
regional patterns. Thus, the intended eventual use of the data is important for
making sampling decisions, as is the length of time one is willing to wait before
being able to document with statistical certainty that a change has taken place.
Because LTM of surface water chemistry is usually based on sampling at a
constant frequency, most samples are typically not collected during high-flow
periods. However, long-term trends in stream chemistry may first become apparent during high flows. An approach for separate trend analysis of high, medium,
and low flows has been developed (Murdoch and Shanley 2006). This method
uses annual or grouped years of data to develop concentration-discharge relationships that enable concentrations to be predicted for various flow conditions
throughout the year. An annual value can then be derived for upper-, medium-,
or low-flow ranges so that long-term trends can be determined for each specific flow range. This type of approach requires that (1) flow is monitored for the
stream site of interest, (2) the solute of interest is statistically related to flow, and
(3) sufficient data are available to develop the concentration-discharge relations.
2.3 FIELD METHODS
The methods outlined here are appropriate for analysis of stream and lake waters
with low ionic-strength and associated with forested and alpine watersheds in
lands that are sensitive to acidification, toxicity, or nutrient enrichment impacts
from atmospheric deposition. Because stream and lake waters in the most highly
sensitive areas can be extremely dilute (and therefore easily contaminated),
great care must be taken in all phases of sample collection and analysis to ensure
that samples are not contaminated during collection or processing so that data
will be of sufficient quality to support the intended assessment purposes. Each
of the important aspects of field sampling is discussed here, with an explanation
of the reasons why certain steps should be taken, or avoided, in the sampling
program. The intent is to provide a general understanding of sampling issues.
The specific, step-by-step instructions to the field personnel can be developed
49
Analysis of trends is most often done on an annual basis using one of several
approaches that incorporate seasonal effects (Helsel and Hirsch 1992, Lawrence
et al. 2004). These approaches are most effective if multiple samples are collected for each season. Weekly sampling provides a sufficient number of samples to account for within-season variability and is likely to enable a trend to be
detected with fewer years of monitoring data than data collected at longer intervals. Annual or quarterly sampling is less expensive than weekly sampling but
cannot account for within-season variability and, relative to weekly sampling,
can substantially increase the length of time needed to detect a trend (Murdoch
and Shanley 2006). However, annual or quarterly, as opposed to weekly, sampling may free up resources to monitor more sites to obtain a better picture of
regional patterns. Thus, the intended eventual use of the data is important for
making sampling decisions, as is the length of time one is willing to wait before
being able to document with statistical certainty that a change has taken place.
Because LTM of surface water chemistry is usually based on sampling at a
constant frequency, most samples are typically not collected during high-flow
periods. However, long-term trends in stream chemistry may first become apparent during high flows. An approach for separate trend analysis of high, medium,
and low flows has been developed (Murdoch and Shanley 2006). This method
uses annual or grouped years of data to develop concentration-discharge relationships that enable concentrations to be predicted for various flow conditions
throughout the year. An annual value can then be derived for upper-, medium-,
or low-flow ranges so that long-term trends can be determined for each specific flow range. This type of approach requires that (1) flow is monitored for the
stream site of interest, (2) the solute of interest is statistically related to flow, and
(3) sufficient data are available to develop the concentration-discharge relations.
2.3 FIELD METHODS
The methods outlined here are appropriate for analysis of stream and lake waters
with low ionic-strength and associated with forested and alpine watersheds in
lands that are sensitive to acidification, toxicity, or nutrient enrichment impacts
from atmospheric deposition. Because stream and lake waters in the most highly
sensitive areas can be extremely dilute (and therefore easily contaminated),
great care must be taken in all phases of sample collection and analysis to ensure
that samples are not contaminated during collection or processing so that data
will be of sufficient quality to support the intended assessment purposes. Each
of the important aspects of field sampling is discussed here, with an explanation
of the reasons why certain steps should be taken, or avoided, in the sampling
program. The intent is to provide a general understanding of sampling issues.
The specific, step-by-step instructions to the field personnel can be developed