Homogenization Applied to Integrate Across Spatial and Temporal Scales in Forest/Insect Ecology

点击翻译按钮获取中文摘要

Powell0077663 In one ecological problem of current interest all theelements exist for application and testing of homogenization toresolve the scale issue and simulate climate and human impacts onoutbreak insect species. The investigator and his colleagues havedeveloped a deterministic spatial model of Mountain Pine Beetle(MPB, Dendroctonus ponderosae) dispersal and mass-attack of pinehosts. The model is validated and descriptive on spatial scalesof hectares and time scales of days. In this project they usetechniques of homogenization on the MPB/host system to constructa forest scale ( 100 hectare/years) model reflecting thelarge-scale emergent consequences of current small-scale spatialand temporal models of insect development, dispersal, andresponse to climate. Direct computation on such large scaleswould be prohibitive; similar difficulties face most ecologicalsimulations involving spatio-temporal heterogeneity at realisticscales and life-system/dispersal interactions. The investigatorsdevelop homogenization procedures that make this type ofecological simulation generally tractable, and implement and testthese procedures in the MPB/host system. The homogenizationprocedure generates an integro-difference equation (IDE),operating with year-long clock `ticks' on landscape scales. In avalidation phase, the IDE model is used to classify landscapes.Statistically analyzing misclassifications in comparison toaerial imagery makes the simulation methodology self-correcting.Finally, scenario studies are used to examine consequences ofhypothetical climate and management trajectories on host/insectdynamics. This illustrates the power of homogenizationapproaches in an ecological context, and provides needed examplesto make these techniques accessible to biological andenvironmental researchers. The method of homogenization is a mathematical techniquethat allows deduction of large-scale properties fromdeterministic small-scale models. To the best of our knowledgethe method of homogenization has never been applied to ecologicalproblems, where issues of scale are the focus of much debate. Onepossible angle of attack is joint use of models and simulation,followed by validation with spatially explicit data. Directcomputational approaches to ecological problems with fulldescription of spatial heterogeneity and climatic variation areextremely difficult due to interaction of scales in time andspace. The difficulty is exacerbated on the one hand by thepaucity of validated, deterministic spatial models at anyecological scale, and by the difficulty of extended spatialobservations on the other. In this project the investigators andhis colleagues apply the technique of homogenization to theMountain Pine Beetle/Lodgepole Pine ecosystem to resolve thescale issue and simulate climate and human impacts on populationoutbreaks of insect species.