Preface
Ecology is sexy. Teaching ecology is therefore the art of presenting a fascinating
topic to well-predisposed audiences. It is not easy: the complexities of modern
ecological science go well beyond the introductory chapters taught in high schools
or the marvellous movies about ecosystems presented on TV. But well-predisposed
audiences are ready to make the effort. Numerical ecology is another story. For some
unclear reasons, a majority of ecology-oriented people are strangely reluctant when
it comes to quantifying nature and using mathematical tools to help understand it. As
if nature was inherently non-mathematical, which it is certainly not: mathematics is
the common language of all sciences. Teachers of biostatistics and numerical
ecology thus have to overcome this reluctance: before even beginning to teach the
subject itself, they must convince their audience of the interest and necessity of it.
During many decades ecologists, be they students or researchers (in the academic,
private or government spheres), used to plan their research and collect data with few,
if any, statistical consideration, and then entrusted the “statistical” analyses of their
results to a person hired especially for that purpose. That person may well have been
a competent statistician, and indeed in many cases the progressive integration of
statistics into the whole process of ecological research was triggered by such people.
In other cases, however, the end product was a large amount of data summarized
using a handful of basic statistics and tests of significance that were far from
revealing all the richness of the structures hidden in the data tables. The separation
of the ecological and statistical worlds presented many problems. The most important were that the ecologists were unaware of the array of methods available at the
time, and the statisticians were unaware of the ecological hypotheses to be tested and
the specific requirements of ecological data (the double-zero problem is a good
example). Apart from preventing the data to be exploited properly, this double
unawareness prevented the development of methods specifically tailored to ecological problems.
The answer to this situation is to form mathematically inclined ecologists.
Fortunately, more and more such people have appeared during the recent decades.
The result of their work is a huge development of statistical ecology, the availability
v
Ecology is sexy. Teaching ecology is therefore the art of presenting a fascinating
topic to well-predisposed audiences. It is not easy: the complexities of modern
ecological science go well beyond the introductory chapters taught in high schools
or the marvellous movies about ecosystems presented on TV. But well-predisposed
audiences are ready to make the effort. Numerical ecology is another story. For some
unclear reasons, a majority of ecology-oriented people are strangely reluctant when
it comes to quantifying nature and using mathematical tools to help understand it. As
if nature was inherently non-mathematical, which it is certainly not: mathematics is
the common language of all sciences. Teachers of biostatistics and numerical
ecology thus have to overcome this reluctance: before even beginning to teach the
subject itself, they must convince their audience of the interest and necessity of it.
During many decades ecologists, be they students or researchers (in the academic,
private or government spheres), used to plan their research and collect data with few,
if any, statistical consideration, and then entrusted the “statistical” analyses of their
results to a person hired especially for that purpose. That person may well have been
a competent statistician, and indeed in many cases the progressive integration of
statistics into the whole process of ecological research was triggered by such people.
In other cases, however, the end product was a large amount of data summarized
using a handful of basic statistics and tests of significance that were far from
revealing all the richness of the structures hidden in the data tables. The separation
of the ecological and statistical worlds presented many problems. The most important were that the ecologists were unaware of the array of methods available at the
time, and the statisticians were unaware of the ecological hypotheses to be tested and
the specific requirements of ecological data (the double-zero problem is a good
example). Apart from preventing the data to be exploited properly, this double
unawareness prevented the development of methods specifically tailored to ecological problems.
The answer to this situation is to form mathematically inclined ecologists.
Fortunately, more and more such people have appeared during the recent decades.
The result of their work is a huge development of statistical ecology, the availability
v