A note on recognizing autocorrelation and using autoregression

The presence of autocorrelation in the analysis of a variable sampled seque ntially at regular time intervals appears to be unknown to many agricultura l meteorologists despite abundant documentation found in the traditional me teorological and statistical literature. It follows that the statistical co nsequences as well as methodological alternatives are also unknown. Through an example using paired radiometer observations, this note discusses recog nition of autocorrelation as well as the importance of testing ordinary lea st squares regression parameters in the presence of autocorrelated residual s. An autoregression example is presented as one alternative way to analyze the given dataset. (C) 1999 Elsevier Science B.V. All rights reserved.