Sollins, P.; Waring, R. H.; Cole, D. W. 1974. A systematic framework for modeling and studying the physiology of a coniferous forest ecosystem. In: Integrated research in the coniferous forest biome. Coniferous For. Biome Bull. 5. Seattle, WA: University of Washington: 7-20.
A coupled set of models of carbon, water, and mineral element processesis being developed as part of the Coniferous Forest Biome terrestrialresearch program. In this paper we present the rationale and objectives,a summary description of the structure and method of implementation, anda statement of progress as of November 1973. Objectives of the modelinginclude presentation of hypotheses concerning system behavior, researchcoordination, identification of information voids, and study of systemresponse to perturbations. Perturbations of interest include climaticchange, defoliation, fire, thinning, fertilization, and irrigation.Responses of interest include growth of trees, runoff volume and pattern,and nutrient concentrations in the runoff.
Implementation is by means of a coupled set of nonlinear differenceequations, about 80 in all. The equations are divided into six groupsof processes (modules): carbon, water, cationic elements except H+, H+,anionic elements, and HCO3. Documentation accompanies conceptualizationand precedes programing. Both documentation and code use a consistentnotation reflecting what we believe to be structure inherent in thenatural system. The notation permits identification of state variables,and parameters. Mnemonics are not used. Extensive written description
of each variable, function, and parameter is included in the documentation.Only minimal written "comments" appear in the code.
Model parameters for processes that have not been studied extensively
are calculated from annual budgets of transfer and accumulation of carbon,water, and the four "nutrient" element groups. Material balance andelectrical neutrality are principles assumed in calculating these budgets.Function forms for processes that are not well understood are usuallypostulated to be linear and donor-controlled although they often includeeffects of driving variables such as air or litter temperature. Ulti-mately we wish each process to be described by a function of comparablecomplexity and realism. Current information precludes this and we feelthat the most important task at present is identification of processesand construction of an adequate framework for analysis of ecosystemresponse.