JPR Advance Access originally published online on July 15, 2006
Journal of Plankton Research 2006 28(10):965-967; doi:10.1093/plankt/fbl022
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
COMMENT |
Reply to Horizons Article ‘Castles built on sand: dysfunctionality in plankton models and the inadequacy of dialogue between biologists and modellers’ Flynn (2005). Shiny mathematical castles built on grey biological sands
AZTI, Herrera Kaia Portualdea, Z/G, 20110 Gipuzkoa, Spain
* Corresponding Author: xirigoien{at}pas.azti.es
Received June 20, 2006; accepted in principle June 28, 2006; accepted for publication June 30, 2006; published online July 15, 2006
Communicating editor: R.P. Harris
Two discussion papers (Harte, 2002
; Flynn, 2005) give me the opportunity to debate two common views about biology: (i) biology is a grey science and (ii) biologists only look for complexity, not large unifying theories (that can be used in models).
- Biology is not a grey science—all you need to know about how an organism works, what it does and even how it got to be here today are written inside the organism in a very precise digital code called DNA (and we are learning to read it). There are two reasons why biology is usually considered as a grey science. There is one that exact science practitioners tend to accept: that the expression of the instructions is flexible in relation to the physical environment, phenotypic plasticity. The other is one that modellers tend to profoundly dislike: that except for clones no two organisms are equal. The latter one is the main reason why physicists and chemists tend to consider biology as a grey science. Whereas two atoms of hydrogen are exactly equal, no two Calanus finmarchicus are the same. However, this feature that ‘exact’ scientists dislike so much is not greyness—it even has a mathematical name: variability. Variability in the genetic code is the basis of evolution—without it we would still be a soup of amino acids, and it is also what allows populations to adapt to new situations. Individuals that do things differently are the colonizers of new niches. However, for some reason, variability is something modellers prefer to ignore. Although they probably exist, personally I do not know of a single model in marine science that includes some stochastic variability in what the organisms can do. The possibility of some outliers being important for the population seems to be rather unsettling for people with a mathematical training. However, 99.99% of fish die in the first few days after birth; it is the outliers and not the average that determine the population outcome.
- Biologists are not always looking for further complexity, at least not all of them. Contrary to Harte’s (Harte, 2002) perception, whereas physics is still looking for the big unifying theory that will integrate quantum mechanics and general relativity, in biology we have had such a unifying theory for a while now. It is called Darwinian evolution (Darwin, 1859). Again, contrary to Harte’s unfortunate use of ‘Newton vs Darwin’, evolution is an extremely simple, universal and predictive theory. The theory is actually so good at explaining what we observe that since its original proposition most of the work that has been carried out has been to fill in the details. I accept that, perhaps because it has been so long, most present-day biologists do not realize they are working in the frame of evolutionary theory, and often they do not even bother checking whether their hypothesis fits with the general theory (sometimes it does not). Anyway, I suspect that a significant percentage of the physicists who spend their time colliding particles similarly do not really understand how they contribute to a unifying theory and are rather looking for a new particle to name (in biology, we call this taxonomy, and we respect it because Linnaeus classification work was a necessary step to develop evolutionary theory). In any case, if in physics there are scientists trying to look at the large picture, the same kind of scientists do exist in biology and ecology, for a Hawkins we do have a Dawkins. When modellers with physical and mathematical backgrounds complain that biologists in their university are only concerned by the complexity of a particular bug (Harte, 2002), the problem is not with biology but with those particular biologists; they should look for biologists with a larger view of the picture.
Coming back to Flynn’s comments on the dialogue between modellers and plankton biologists (Flynn, 2005), there are some very basic points where the big picture can help. As an example, evolution tells us that in the long term, the growth and mortality of a population must be equal, otherwise the population will become extinct. However, most of the plankton models I know do not get the numbers right yet (Anderson, 2005)—if the peak is more or less at the proper time everything is fine, the numbers are usually kept at the appropriate level with a ‘closure’ term that is a nice euphemism for not knowing what is wrong in the model. In the words of a modeller, present-day models are qualitative not quantitative. It should not be difficult to understand that for a biologist, it is difficult to trust a model that can fill the planet with a single bug in a few cycles. Not knowing what is wrong is perfectly acceptable, but not worrying about it is a problem.
It is also interesting to observe how little use is made of the basic evolutionary rules when modelling biological systems (or at least plankton). However, the use of evolutionary principles can transform what might look like very complex behaviours (feeding and migrations), a Pandora’s Box that modellers hesitate to open (Flynn, in press), into something that can be modelled on the basis of very basic rules (e.g. Fiksen, 2000; Fiksen et al., 2005).
Another example is size spectra: it is >30 years now that we know there is a relation between size and abundance in the ocean and we even know that the slope of that relation expressed in logarithmic terms is –1 (Sheldon et al., 1972). This is a well-known feature of plankton communities. But how many models actually check whether the plankton communities they simulate fulfil this basic rule? Not many to my knowledge, although there are now groups modelling on a size basis.
How have we reached this dialogue inadequacy raised by Flynn (Flynn, 2005)? In my view, the fault is mainly on the biologists’ side, not because of greyness or always looking for further complexity, but because of their training most biologists (such as myself) end up as mathematical illiterates. This deficiency in training results in biologists who have to review a model being so impressed with anything that moves on the screen that they never question what it really does and are unable to distinguish between technological (programming) demonstrations and real modelling. As a result, biologists now complain that many models do not reproduce reality, but this is normal, many of them never intended to reproduce reality. They were just technological demonstrations of what could be done. If biologists had done their review job properly, such models should have found their way into the methodological sections and not as results.
How to solve this lack of dialogue? On one side, there are a few present-day polymaths who are able to handle both mathematics and biology: I know a few and I suspect they missed their real vocation, they are biologists with a mathematics vocation or vice versa. This type of scientist is a ‘rara avis’ that we should protect to optimize their production. However, for the rest of us mortals, such an extensive knowledge is not possible. Flynn has tried to train mathematicians as biologists, and it does not work, as there are too many things to learn in biology. He is right, and I have tried myself. I have also tried to train biologists as mathematicians, and this is not easy either. There is much more to the mathematics than what a biologist thinks.
In my opinion, for most of us, the solution is a bicephalous modeller, a mathematician and a biologist sitting together and collaborating. However, both sides have also to learn that not all ‘biologists’ and ‘mathematicians’ are the same, the mathematician has to look for a biologist with a relatively large view of the system, and the biologist has to look for a real modeller and not just a programmer. In the long term, the biologist’s educational itinerary should include more mathematics. But, Harte and Flynn’s comments indicate that, as strange as it may seem, the real training deficiency for biologists might be in evolution, as it is obvious that most biologists are not able to communicate to modellers the basic rules of life.
 |
REFERENCES |
|---|
 TOP REFERENCES |
|---|
Anderson, T. (2005) Plankton functional type modelling: running before we can walk? J. Plankton Res., 27, 1073–1081.
Darwin, C. (1859) The Origin of Species by Means of Natural Selection or the Preservation of Favoured Races in the Struggle for Life. J. Murray, London.
Fiksen, Ø. (2000) The adaptive timing of diapause – a search for evolutionarily robust strategies in Calanus finmarchicus. ICES J. Mar. Sci., 57, 1825–1833.
Fiksen, Ø., Eliassen, S. and Titelman, J. (2005) Multiple predators in the pelagic: modelling behavioural cascades. J. Anim. Ecol., 74, 423–429.
Flynn, K. (2005) Castles built on sand: dysfunctionality in plankton models and the inadequacy of dialogue between biologists and modellers. J. Plankton Res., 27, 1207–1210.
Flynn, K. (2006) Reply to Horizons article ‘Plankton functional type modelling: running before we can walk’ Anderson (2005): II. Putting trophic functionality into plankton functional types. J. Plankton Res., 28, 873–875.
Harte, J. (2002) Toward a synthesis of the Newtonian and Darwinian worldviews. Phys. Today, 55, 29.
Sheldon, R. W., Prakash, A. and Sutcliffe, W. H. Jr. (1972) The size distribution of particles in the ocean. Limnol. Oceanogr., 17, 327–340.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||