We present a statistical model relating measurements of water quality to associated errors in laboratory methods. Estimation of the model allows us to correct trends in water quality for long-term and short-term variations in laboratory measurement errors. An illustration of the bias correction method for a large national set of stream water quality and quality assurance data shows that reductions in the bias of estimates of water quality trend slopes are achieved at the expense of increases in the variance of these estimates. Slight improvements occur in the precision of estimates of trend in bias by using correlative information on bias and water quality to estimate random variations in measurement bias. The results of this investigation stress the need for reliable, long-term quality assurance data and efficient statistical methods to assess the effects of measurement errors on the detection of water quality trends.
Additional publication details
|Publication Subtype||Journal Article|
|Title||Correction of stream quality trends for the effects of laboratory measurement bias|
|Series title||Water Resources Research|
|Publisher||American Geophysical Union|