WMOD00398

• Project Name: Project Skywater
• Title: Precipitation Management and the Environment - An Overview of the Skywater IX Conference
• Date: 9/1/1977
• Weather Modification - Doc Type: Report

Hammond); natural vegetation; dominant plant taxa; land use; and habitat type. This
stratification leads to partitioning into individualized geomorphic regions with homogeneous
properties which can be generalized throughout the subregion. The generalization may be
extended to other subregions to the extent to which they resemble each other in their important
properties. Methods are available for systematic expression of the degrees of similarity.
Transfer of information.- The theme is generalization from a particular experiment to new
situations. Two methods of generalization are recognized, classification and dynamic models.
Classification goes by important similarities, assuming a corresponding similarity in underlying
processes, qualified by differences in geologic substrate and weather patterns. Dynamic models
are analytic-mathematical expressions of understanding or hypotheses applied to the subject
matter. The models may be validated by experimentation on two levels of generality, that of
total system behavior and that of specific interaction. In ecosystem studies, focus on nutrient
cycling is to be preferred to that on energy flow.
Models for most natural ecosystems cart predict within about 20 percent for most data.
Predictions are limited by errors accumulating from imperfect knowledge of processes and
natural variabilities. A model that verifies over a span of 10 to 25 years can be considered as
validated. One way of , applying the model concept would be to analyze carbon gain in a game of
interspecies competition.
The probl~m of extrapolation may be approached by learning how a set of components critical
to ecosystem functioning changes with precipitation, validating the model in one area, and then
carrying it over to other areas. Critical components include primary above-ground production,
primary production by important species, decomposition, soil moisture and runoff, fixation and
leaching of nitrogen, phytosociology and species diversity, cations, and invertebrates.
Year-to-year and area-to-area variations present a problem.
Ecosystem studies should be repeated in similar but not identical regions and relating the
integr~ty of predictions with similarity between regions.
Process-oriented ecosystem studies should observe the following principles. The degree of
resolution needed depends on the objective, as to whether all processes or only significant ones
are to be included, using standard techniques for interrelating species, decomposition, runoff,
nitrogen fixation/leaching, etc. Complexity and uncertainity will ordinarily limit the number of
components to the most important few. Budget restrictions will also limit the number of
components.
About 20 percent of the effort should be allocated to the setting of objectives. Processes, not
initial conditions, should be the focus of measurement programs. After five iterations of a process
model, initial conditions lose significance. All important analytic levels, from abiotic to
vegetation, animals, etc., should be combined into a single comprehensive systems analysis.
Programs of education and information.-Programs of public education and information
should be continued and expanded.

16

,[

,~

y