SENSITIVITY ANALYSIS: A NECESSITY IN WATER PLANNING1

ABSTRACT In water planning activities, major emphasis has been placed on the development of procedures for devising “optimum plans.” These plans are defined as those which meet prespecified demands for water at “minimum cost.” However, all plans are developed subject to postulated conditions regarding the state of the physical system and of nature. Because planning takes place in a dynamic and uncertain environment in which postulated conditions are known to change, it is imperative that the planner be apprised in the planning phase of the effect of changes which can occur. Using “this information, a planner can temper his judgment with a knowledge of the effect of the uncertainty resulting from changes in the system state variables. This paper presents results of the use of a computer simulation and optimization model to quantify possible variations in system response which could occur as a result of uncertainty in the postulated physical and economic conditions under which the proposed water development system was to perform. The possible effects of these variable responses on planning decision-making is discussed.     

Local and landscape scale factors influencing edge effects on woodland salamanders

We examined local and landscape-scale variable influence on the depth and magnitude of edge effects on woodland salamanders in mature mixed mesophytic and northern hardwood forest adjacent to natural gas well sites maintained as wildlife openings. We surveyed woodland salamander occurrence from June-August 2006 at 33 gas well sites in the Monongahela National Forest, West Virginia. We used an information-theoretic approach to test nine a priori models explaining landscape-scale effects on woodland salamander capture proportion within 20 m of field edge. Salamander capture proportion was greater within 0-60 m than 61-100 m of field edges. Similarly, available coarse woody debris proportion was greater within 0-60 m than 61-100 m of field edge. Our ASPECT model, that incorporated the single variable aspect, received the strongest support for explaining landscape-scale effects on salamander capture proportion within 20 m of opening edge. The ASPECT model indicated that fewer salamanders occurred within 20 m of opening edges on drier, hotter southwestern aspects than in moister, cooler northeastern aspects. Our results suggest that forest habitat adjacent to maintained edges and with sufficient cover still can provide suitable habitat for woodland salamander species in central Appalachian mixed mesophytic and northern hardwood forests. Additionally, our modeling results support the contention that edge effects are more severe on southwesterly aspects. These results underscore the importance of distinguishing among different edge types as well as placing survey locations within a landscape context when investigating edge impacts on woodland salamanders.