Managing resources for multiple purposes: Markov models of southern mixed loblolly pine-hardwood forests

Loblolly pine (Pinus taeda L.) is the most important and widespread forest tree species in the southern United States. To manage naturally regenerated loblolly pine and hardwood stands, computer-based analytical tools are required to deal with uncertainty in growth and market conditions and to achieve multiple goals when making optimal decisions. Such models ought to have the ability to handle landscape analysis and management too.; To this end, first, a stochastic growth model was developed based on a deterministic nonlinear matrix growth model by Lin et al. (1998), to recognize the effects of small-scale, high-frequency disturbances on stand growth.; Second, I developed a space-time method that describes stumpage price dynamics over time and at neighboring locations. The method was applied to quarterly pine sawtimber stumpage prices for 21 geographically contiguous regions in the South.; Third, a Markov Decision Process (MDP) model was formulated to seek management policies satisfying diverse interests under uncertainty. The effect of natural catastrophes was also taken into account. The results show that natural catastrophes create a highly heterogeneous landscape and decrease tree size and species diversities. Current management would in the long run create a diverse landscape with high ecological benefits in terms of tree species and size diversity and basal area, but it produces very low expected timber production and net present value (NPV). The opportunity cost, in terms of NPV, of increasing size and landscape diversity and old growth frequency was much higher than for species diversity. The large difference between the achievable maximum income from timber and that obtained from current management suggested that the forests under examination had substantial non-timber values.