Environmental Monitoring

Environmental Monitoring
   



A good hypothesis will define clearly:

  • the types of impacts;
  • the consequences;
  • the variables and units of concern;
  • the degree of response that matters and that would require a management response.


Failure to be specific about these will limit the ability to achieve useful or usable results.

It is as though you are telling a "story",

  • make sure it is comprehensible;
  • get the logic right and be clear about your assumptions and the basis for the argument.


There has been some discussion as to whether hypotheses are relevant in ecological studies. The answer must be "yes, always", if only because the process of deriving an hypothesis is critical to good design. An hypothesis that is useful will require specific things to be tested.

To say "my hypothesis is that urban development causes environmental degradation" is essentially useless. At one level a truism and at another untestable: what constitutes "urban development" and how do you measure "environmental degradation". As a first step the characteristics that you might be concerned about that are indicative of urban development could be:

  • volume of runoff;
  • rate of runoff ("flashiness");
  • pollutants (etc., etc.).


Environmental degradation might be related to:

  • loss of biota (eg. loss of particular pollution sensitive species);
  • change in biota (eg. change in number of genera/species; x% reduction in a particular species);
  • eutrophication (eg. Number of days with algal blooms; or Chlorophyll A greater than X);
  • loss of recreational amenity (eg. Number of days not suitable for swimming);
  • public risk (etc., etc.).


Good reasoning sorts out these issues so you say something useful and testable. Hypothesis formulation has the benefit of keeping you honest; even if you subsequently cannot develop an "ideal" design you still need to be very clear about what you are interested in and why. Conceptual, and other, models can be useful to clarify thinking:

  • draw a picture of the system;
  • diagram the relationships of interest;
  • identify what you "know", based on data, and what are guesses, hunches or prejudices etc.;
  • think about how your experiment might perturb the system;
  • resolve the relevant scales for the study area and parameters of concern.


Different kinds of models will help you see the "problem" in different ways.

What is a "null" hypothesis?


   
 

Hosted by Keysoft Pty Ltd