10 

 Modeling 



J. L. Tiwari, R. J. Daley. J. E. Hobbie, 

 M. C. Miller. D. W. Stanley and J. P. Reed 



MODELING IN THE AQUATIC PROGRAM 

 OF THE TUNDRA BIOME 



The mathematical modeling of whole ecosystems was one of the 

 scientific tools that all parts of the U.S. IBP studies were to use. It is 

 obvious that a model of an ecosystem that was biologically correct would 

 be of great aid in predicting the effect of changes, such as a temperature 

 decrease or an increased rate of nutrient input. Yet no large-scale model 

 had been constructed when IBP began; the personnel were inexperienced 

 and it was not entirely clear how to proceed. To some scientists, the 

 modeling was a tool to be learned about; to others, it was already suspect 

 because of past failures or abuses. Some scientists looked at the whole 

 U.S. IBP as an experiment in the use of models; they asked the questions 

 "Can a large-scale mathematical model be constructed?" and "Is this 

 approach a useful one for a detailed study of an ecosystem?" 



The experiences of the aquatic program were by no means unique; 

 other groups that used modeling reached the same conclusions. Yet many 

 of these conclusions have not been published and there are programs 

 beginning each year that appear to be making some of the same mistakes 

 that we did. In the hope of at least making people aware of the problems, 

 we present here some of our experiences and general conclusions. 



The first important step is to assemble a group of ecologists who are 

 used to thinking at the whole-system level. This is relatively easy in 

 limnology, for studies of lakes or watersheds lend themselves to simple 

 input-output models and the studies of cycling of nutrients are further 

 advanced in lakes than in soils, for example. A majority of the scientists on 

 the project must be able to think like this. Some specialists, such as 

 taxonomists and chemists, are necessary but every attempt should be made 

 to attract scientists who understand more of the ecosystem than their own 

 specialty. One reason for the need for this type of scientist is that there are 

 many questions of judgment that arise during modeling, such as deciding 

 what are the important processes and organisms that have to be in the 

 model. Another reason for having this type of scientist is that modeling 

 often requires different types of data than might be collected usually. For 

 example, zooplankton counts may have to be transformed into carbon 

 amounts, yet the factors for the transformation are not constants and must 

 be separately measured. 



407 



