In central hardwood forests, mean cavity-tree abundance increases with increasing stand-size clas... more In central hardwood forests, mean cavity-tree abundance increases with increasing stand-size class (seedling/sapling, pole, sawtimber, old-growth). However, within a size class, the number of cavity trees is highly variable among 0.1-ha inventory plots. Plots in young stands are most likely to have no cavity trees, but some plots may have more than 50 cavity trees/ha. Plots in old-growth stands often had 25 to 55 cavity trees/ha, but individual plots ranged from 0 to 155/ha. The Weibull probability density function was used to mathematically describe the variation in cavity-tree abundance for plots in stands of differing size (or age) class. A graph of the cumulative probability of cavity-tree abundance is a particularly easy way for managers to estimate the probability that a stand of a given size class will have any specified number of cavity trees per hectare. Results for individual plots or stands can be combined to estimate cavity abundance probabilities for landscapes. Because...
Predicting the effects of global climate change on species interactions has remained difficult be... more Predicting the effects of global climate change on species interactions has remained difficult because there is a spatiotemporal mismatch between regional climate models and microclimates experienced by organisms. We evaluated resource selection in a predominant ectothermic predator using a modeling approach that permitted us to assess the importance of habitat structure and local real-time air temperatures within the same modeling framework. We radio-tracked 53 western ratsnakes () from 2010 to 2013 in central Missouri, USA, at study sites where this species has previously been linked to prey population demographics. We used Bayesian discrete choice models within an information theoretic framework to evaluate the seasonal effects of fine-scale vegetation structure and thermal conditions on ratsnake resource selection. Ratsnake resource selection was influenced most by canopy cover, canopy cover heterogeneity, understory cover, and air temperature heterogeneity. Ratsnakes generally ...
Demographic processes (fecundity, dispersal, colonization, growth, and mortality) and their inter... more Demographic processes (fecundity, dispersal, colonization, growth, and mortality) and their interactions with environmental changes are not well represented in current climate-distribution models (e.g., niche and biophysical process models) and constitute a large uncertainty in projections of future tree species distribution shifts. We investigate how species biological traits and environmental heterogeneity affect species distribution shifts. We used a species-specific, spatially explicit forest dynamic model LANDIS PRO, which incorporates site-scale tree species demography and competition, landscape-scale dispersal and disturbances, and regional-scale abiotic controls, to simulate the distribution shifts of four representative tree species with distinct biological traits in the central hardwood forest region of United States. Our results suggested that biological traits (e.g., dispersal capacity, maturation age) were important for determining tree species distribution shifts. Envi...
Habitat selection is a fundamental component of community ecology, population ecology, and evolut... more Habitat selection is a fundamental component of community ecology, population ecology, and evolutionary biology and can be especially important to species with complex annual habitat requirements, such as migratory birds. Resource preferences on the breeding grounds may change during the postfledging period for migrant songbirds, however, the degree to which selection changes, timing of change, and whether all or only a few species alter their resource use is unclear. We compared resource selection for nest sites and resource selection by postfledging juvenile ovenbirds (Seiurus aurocapilla) and Acadian flycatchers (Empidonax virescens) followed with radio telemetry in Missouri mature forest fragments from 2012-2015. We used Bayesian discrete choice modeling to evaluate support for local vegetation characteristics on the probability of selection for nest sites and locations utilized by different ages of postfledging juveniles. Patterns of resource selection variation were species-sp...
The Central Hardwood Forest (CHF) in the United States is currently a major carbon sink, there ar... more The Central Hardwood Forest (CHF) in the United States is currently a major carbon sink, there are uncertainties in how long the current carbon sink will persist and if the CHF will eventually become a carbon source. We used a multi-model ensemble to investigate aboveground carbon density of the CHF from 2010 to 2300 under current climate. Simulations were done using one representative model for each of the simple, intermediate, and complex demographic approaches (ED2, LANDIS PRO, and LINKAGES, respectively). All approaches agreed that the current carbon sink would persist at least to 2100. However, carbon dynamics after current carbon sink diminishes to zero differ for different demographic modelling approaches. Both the simple and the complex demographic approaches predicted prolonged periods of relatively stable carbon densities after 2100, with minor declines, until the end of simulations in 2300. In contrast, the intermediate demographic approach predicted the CHF would become ...
In central hardwood forests, mean cavity-tree abundance increases with increasing stand-size clas... more In central hardwood forests, mean cavity-tree abundance increases with increasing stand-size class (seedling/sapling, pole, sawtimber, old-growth). However, within a size class, the number of cavity trees is highly variable among 0.1-ha inventory plots. Plots in young stands are most likely to have no cavity trees, but some plots may have more than 50 cavity trees/ha. Plots in old-growth stands often had 25 to 55 cavity trees/ha, but individual plots ranged from 0 to 155/ha. The Weibull probability density function was used to mathematically describe the variation in cavity-tree abundance for plots in stands of differing size (or age) class. A graph of the cumulative probability of cavity-tree abundance is a particularly easy way for managers to estimate the probability that a stand of a given size class will have any specified number of cavity trees per hectare. Results for individual plots or stands can be combined to estimate cavity abundance probabilities for landscapes. Because...
Predicting the effects of global climate change on species interactions has remained difficult be... more Predicting the effects of global climate change on species interactions has remained difficult because there is a spatiotemporal mismatch between regional climate models and microclimates experienced by organisms. We evaluated resource selection in a predominant ectothermic predator using a modeling approach that permitted us to assess the importance of habitat structure and local real-time air temperatures within the same modeling framework. We radio-tracked 53 western ratsnakes () from 2010 to 2013 in central Missouri, USA, at study sites where this species has previously been linked to prey population demographics. We used Bayesian discrete choice models within an information theoretic framework to evaluate the seasonal effects of fine-scale vegetation structure and thermal conditions on ratsnake resource selection. Ratsnake resource selection was influenced most by canopy cover, canopy cover heterogeneity, understory cover, and air temperature heterogeneity. Ratsnakes generally ...
Demographic processes (fecundity, dispersal, colonization, growth, and mortality) and their inter... more Demographic processes (fecundity, dispersal, colonization, growth, and mortality) and their interactions with environmental changes are not well represented in current climate-distribution models (e.g., niche and biophysical process models) and constitute a large uncertainty in projections of future tree species distribution shifts. We investigate how species biological traits and environmental heterogeneity affect species distribution shifts. We used a species-specific, spatially explicit forest dynamic model LANDIS PRO, which incorporates site-scale tree species demography and competition, landscape-scale dispersal and disturbances, and regional-scale abiotic controls, to simulate the distribution shifts of four representative tree species with distinct biological traits in the central hardwood forest region of United States. Our results suggested that biological traits (e.g., dispersal capacity, maturation age) were important for determining tree species distribution shifts. Envi...
Habitat selection is a fundamental component of community ecology, population ecology, and evolut... more Habitat selection is a fundamental component of community ecology, population ecology, and evolutionary biology and can be especially important to species with complex annual habitat requirements, such as migratory birds. Resource preferences on the breeding grounds may change during the postfledging period for migrant songbirds, however, the degree to which selection changes, timing of change, and whether all or only a few species alter their resource use is unclear. We compared resource selection for nest sites and resource selection by postfledging juvenile ovenbirds (Seiurus aurocapilla) and Acadian flycatchers (Empidonax virescens) followed with radio telemetry in Missouri mature forest fragments from 2012-2015. We used Bayesian discrete choice modeling to evaluate support for local vegetation characteristics on the probability of selection for nest sites and locations utilized by different ages of postfledging juveniles. Patterns of resource selection variation were species-sp...
The Central Hardwood Forest (CHF) in the United States is currently a major carbon sink, there ar... more The Central Hardwood Forest (CHF) in the United States is currently a major carbon sink, there are uncertainties in how long the current carbon sink will persist and if the CHF will eventually become a carbon source. We used a multi-model ensemble to investigate aboveground carbon density of the CHF from 2010 to 2300 under current climate. Simulations were done using one representative model for each of the simple, intermediate, and complex demographic approaches (ED2, LANDIS PRO, and LINKAGES, respectively). All approaches agreed that the current carbon sink would persist at least to 2100. However, carbon dynamics after current carbon sink diminishes to zero differ for different demographic modelling approaches. Both the simple and the complex demographic approaches predicted prolonged periods of relatively stable carbon densities after 2100, with minor declines, until the end of simulations in 2300. In contrast, the intermediate demographic approach predicted the CHF would become ...
Uploads
Papers by Frank R. Thompson III