While everyone recognises the importance of the environment, there is frequent debate about how serious some issues are and what actions are appropriate. Much of the debate at every level from global to local reflects a lack of knowledge, and this in turn reflects difficulties in measuring environmental systems. Statistics has a key role in making decisions in the presence of such uncertainty.
Global warming is a large scale example of this. The first hint of warming was published in the 1970s, based on long term measurements at a number of sites in the northern hemisphere. However the data was not always reliable, coming from many weather stations of varying quality, the analysis difficult with a spatial as well as a temporal aspect and the temperature change small, leaving the overall findings questionable. Now with more data and more refined statistical analysis, the reality of global warming is generally accepted and the debate has moved on to effects and actions.
At the local level, the problems are the same but less dramatic. Questions of whether an environment has been affected involve detecting change amid the natural variation. This leads to a need to quantify and understand natural variation and the setting of quantitative guidelines for what is an acceptable impact.
A common problem in environmental measurement is the high cost. For instance sampling at less accessible locations, such as on the seabed or in an exhaust stack, can make data collection expensive. Some environmental variables such as odour are inherently difficult to measure. Biological measures such as species counts or population estimates and epidemiological variables are also challenging. Planning and the optimisation of data collection are essential to keep costs reasonable while ensuring that the critical questions can be answered.
The amount of data required is determined by the variability and the efficiency of the analysis, both of which are statistical issues. Involvement of statisticians in the planning of data collection can lead to significant savings as well as giving an assurance that it will be fit for its purpose. The statistician also provides an objectivity that is critical when considering emotional environmental issues. Accredited Statisticians follow a professional code of ethics that emphasises this independence.
Data Analysis Australia has collaborated with a number of clients in this area. Where possible in these studies the first step involved statistical analysis of existing data to gain an understanding of natural variation and measurement errors in a cost-effective way. These analyses led to the design of optimal sampling methods to contain costs while achieving the required precision. Environmental projects Data Analysis Australia has been involved in include monitoring run-off from mine sites, analysing relationships between emissions and odour, and a baseline study of benthic (bottom dwelling) communities in Cockburn Sound.
For further information on our services in this area please contact Data Analysis Australia at daa(at)daa.com.au or phone 08 9468 2533.