Integration of GIS techniques and heuristic algorithms to address spatial forest planning issues in the Southern U.S.
MetadataShow full item record
GIS techniques have been used in the natural resource field over the past few decades. As spatial analyses in forest planning are increasingly needed by forest managers and forest researchers, a high level of integration of GIS techniques and spatial forest planning is inevitable. In addition, the complexity of spatial forest planning problems requires intelligent use of heuristic methods in order to quickly generate forest plans of high quality. In this research, an extensive literature review in North American forestry journals was performed to identify trends and gaps of GIS applications in the natural resource field. Then, an informed development of meta heuristics based on several standard heuristic algorithms - Monte Carlo integer programming, simulated annealing, threshold accepting, tabu search and the raindrop method - was performed. Meta heuristics were composed by combining different standard heuristics, in an intelligent way without the need for direct human intervention. I composed 24 3-algorithm meta heuristics, and results showed that the meta heuristics presented consistently better solution qualities than standard heuristics in solving typical spatial forest planning problems for the southern U.S. region. Finally, by using both GIS and heuristic techniques, I assessed the effects of a forest planning constraint (maximum clearcut size) on forest fragmentation and found that as maximum clearcut size increased, the effects on forest fragmentation seemed to decrease. This analysis used operational GIS databases and a typical southern U.S. forest management problem formulation, each of which posed an analytical problem. First, roads in the operational GIS database were explicitly recognized, creating an artificial barrier between stands when it came to the fragmentation analysis. When roads were removed, I found that fragmentation indices changed slightly, yet the overall trend was the same. Second, I suspected that woodflow constraints in the problem formulation compounded fragmentation by spreading out harvests relatively evenly over the time horizon. By removing woodflow constraints from the analysis, I found this to be true. This work provided three advances to the forestry sciences: a published literature review illustrating the advances and gaps in the use of GIS in forestry, a novel way to integrate standard heuristics into a meta heuristic in order to develop more efficient forest plans, and an analysis of fragmentation in those forest plans for a southern U.S. landowner.