Environmental monitoring, assessment and prediction of change

Abstract

Abstract Estimating the present status, detecting and estimating trends, identifying problem areas, and testing for compliance with standards are some common goals of environmental monitoring programmes. Issues of concern when setting these programmes are what to measure, how and when to make these measurements, and how to interpret the data. The role of statistics in addressing these issues is discussed and illustrated using three examples. The first is the upstream/downstream Niagara River monitoring program which is used to: (1) estimate the loads of contaminants to the river and from the river to Lake Ontario; (2) compare the concentrations of various contaminants with water quality standards. The monitoring programme for the Fraser River provides the second example. The Fraser River has, in contrast to the Niagara River, a complicated hydrological cycle which needs to be considered when interpreting water quality changes. The third is the before and after time series experiments for environmental impact assessment. In these examples, parametric and non‐parametric methods are used to model seasonality and trends. Techniques for dealing with censored data are also discussed.