STATISTICAL LINEAR MODELS IN THE 
COLLECTION AND ANALYSIS OF 
ECOLOGICAL DATA 
Saul B. Saila, Martin A.M. Hyman and Ernesto Lorda 
Marine Experiment Station 
Graduate School of Oceanography 
University of Rhode Island 
Kingston, Rhode Island 02881 
ABSTRACT 
The problems of environmental monitoring and baseline studies are 
considered. The importance of sampling is emphasized, and items to be 
considered are outlined. Time series analysis and linear models are discussed as 
two kinds of statistical methodology applied to environmental impact analysis 
and monitoring. Application of both these approaches is discussed and the 
reader is referred to comprehensive treatment in various references. While time 
series is mentioned only briefly, linear models are dealt with at length. The use 
of a simple linear model is illustrated by an example which relates to deciding 
where to establish monitoring stations along a cross-sectional area of a 
hypothetical estuary which is of interest. A scheme for collection of the data is 
presented along with the general analysis of variance table for this particular 
model. 
INTRODUCTION 
There seems to be no limit to the demands for. more and more data 
concerning the problems of prediction and protection of the marine and 
estuarine environments. From the volume of data being generated in some 
studies, the ultimate goal in monitoring and baseline establishment appears to 
be to measure everything, everywhere, continuously. It should be recognized 
that even if this virtually infinite amount of data were gathered, there is no 
guarantee that it would lead to complete understanding or predictive inferences 
from a given system. Thus, any environmental monitoring or baseline study 
should be practical and feasible within reasonable time and cost constraints. 
However, the data gathered must be accurate, pertinent to the problems at 
hand, concise and purposefully collected. 
337 
