PART IX — TERRESTRIAL ECOSYSTEMS 



Figure IX-4 — A MODEL VALIDATION STUDY 



Grams/ m- 



40 ■— 



36 



32 



28 



24 



20 



12 



4 - 



i FIELD DATA 



■BHM SIMULATED DATA 



160 200 



1971— DAYS 



The graph shows the output of blue gramagrass biomass in g/m- as simulated by 

 computer for the growing season of 1971. This simulation is one of the many 

 outputs of the grasslands ecosystem model. Also plotted on the graph are actual 

 field measurements of the blue grama production at the Pawnee site. Although 

 there are differences in the curves, the over-all result indicates that the model is 

 simulating the actual production. The differences can be explained, at least in part, 

 by the fact that the abiotic variables that are an input to the model are not measured 

 at the same location as the sampling point for the blue grama biomass. 



they should be given every encour- 

 agement to develop independently 

 (though with plenty of opportunities 

 for contact) for some years to come. 

 In the meantime, some new approach, 

 not yet conceived, may well show 

 itself superior to any. 



It is clear that modeling of abiotic 

 parts of the ecosystem is considerably 

 in advance of the development of 

 submodels for the living components. 

 More effort needs to be given to de- 

 vising suitable forms for the latter 

 submodels, and this will probably in- 

 volve intensive experimental work 

 on a variety of subjects. 



A good deal more thought needs 

 to be given to the process of validat- 

 ing models and comparing alterna- 

 tives. Statistical considerations will 

 clearly play an important part, but at 

 present most statisticians avoid the 

 subject. 



Techniques To Incorporate Diver- 

 sity — Simplification is essential in 

 the modeling of ecosystems; but the 

 methods of simplification at present 

 in vogue (in terms of compartments, 

 trophic levels, and such) are unlikely 

 to be the most fruitful. There is little 

 doubt that the diversity of an eco- 

 system is an important factor in its 

 dynamics and stability, and means 

 must be found to take this diversity 

 into account in the model. The diver- 

 sity or heterogeneity that is important 

 may take various forms; first and 

 foremost, the division of the biomass 

 into species, each of which has dis- 

 tinctive responses to the environ- 

 ment and, consequently, distinctive 

 niche requirements; second, variation 

 within a species of genetic and ac- 

 quired characteristics, including re- 

 sponses to external factors and the 

 timing of vital processes such as 

 seed germination, metamorphosis, 

 and reproduction; third, spatial dif- 

 ferentiation and patterning, partly 

 dependent on the inanimate substrate, 

 partly developed through the dy- 

 namics of the ecosystem itself. 



Such heterogeneity has mainly 

 been incorporated in models by mul- 



284 



