Recipient Organization
OKLAHOMA STATE UNIVERSITY
(N/A)
STILLWATER,OK 74078
Performing Department
Natural Resource Ecology & Management
Non Technical Summary
Forests provide several important ecosystem functions that benefit people, and approximately 20% of Oklahoma's land area is forested. The ecosystem services that are provided by forests include provisioning services such as timber, pulp, and wild game production, supporting services such as net primary production, regulating services including carbon storage and drinking water supply, and cultural services. Therefore, understanding the extent to which forests are resilient to the currently changing environment, and the mechanisms that influence resilience, is important to managing forests to maintain these critical ecosystem functions.The overarching goal of this research is to identify forest management practices that enhance the resilience of forest plant communities to unplanned disturbances. This work will help to bridge the gap between ecological mechanisms that affect resilience, and management practices that are based upon these concepts.
Animal Health Component
60%
Research Effort Categories
Basic
40%
Applied
60%
Developmental
(N/A)
Goals / Objectives
Study 1 - Determine the individual and synergistic effects of disturbances such as prescribed fire, forest thinning, and their interaction on the functional and taxonomic diversity of understory forest plant communities.Study 2 - Develop and apply innovative remote sensing techniques to determine the effects of prescribed fire, thinning, and their interaction on the functional and taxonomic diversity of forest overstory tree communities.
Project Methods
Study 1 - This study will take place at the Pushmataha Wildlife Management Area (PWMA) in southeast Oklahoma. A forest management experiment was initiated at the PWMA in the 1980s. Through this experiment, forest patches have been thinned, burned by prescribed fire, burned and thinned, and left unmanaged to serve as an example of reference conditions with no management. Fire was applied at 1, 2, 3, 4, and never return intervals. These treatments were also applied outside of the experimental area.Understory plant communities and associated environmental variables will be sampled in each treatment area. The percent cover of each species will be determined visually within 1 x 1 m plots. Light conditions will be measured by taking a hemispherical photograph over each plot and then determining the Photosynthetic Photon Flux Density (PPFD) using the software WinSCANOPY (Regent Instruments Inc.). The physiological functional traits will be determined by taking a multispectral or hyperspectral photograph of the plot and analyzing the spectral bands. Morphological traits will be determined by measuring a subsample of plant individuals within each plot. Some traits may be estimated at the species level by looking up average trait values in the literature.Within each site, 1 x 1 m sample plots will be arranged in a cyclically repeating sampling design, which includes two transects intersecting at a 60-degree angle. This sampling design allows the efficient estimation of spatial semivariance, which is a method of characterizing spatial pattern in the data.The functional traits measured in each plot will include morphological and physiological metrics. The morphological traits may include leaf density, node density, and specific leaf area. These traits will be measured on a sample of plants within each sample plot. Physiological traits will be estimated using multispectral or hyperspectral imagery. The imagery will be obtained using a multirotor unmanned aerial vehicle positioned above each sample plot. A benefit of this technique is that it is non-destructive.The physiological traits will include leaf chlorophyll, leaf carotenoids, and leaf water. Each of these traits will be developed from multispectral or hyperspectral imagery. The chosen method of estimating these traits will depend in part on the type of imagery available at an appropriate level of spatial and spectral resolution. Some potential methods are described here. Leaf chlorophyll is the relative content of chlorophyll a and b per unit leaf area.The species diversity of each plot will be determined based on species richness, Shannon's species diversity, and species evenness. The functional diversity of each plot will be based on functional richness (FRic), functional evenness (FEve), and functional divergence (FDiv). To estimate these functional diversity indices, the first step is to develop a multidimensional functional trait space from a reduced set of axes obtained by principal coordinates analysis (PCoA) (Mason et al. 2005; Villéger et al. 2008). The PCoA is based on a species-species Gower dissimilarity matrix computed from the traits (Gower 1971). FRic is the proportion of niche space occupied by the individuals in the plot. It is computed as the convex hull volume of the plot divided by the convex hull volume of all of the plots. FEve is the sum of the minimum spanning tree among the pixels in functional trait space. It is based on the dissimilarity matrix and is analogous to species evenness. FDiv represents the degree of niche differentiation and resource competition and is measured based on the distribution of species in trait space. When these functional diversity metrics are computed using imagery, a pixel or neighborhood of pixels is substituted for species in the formulas described above (Schneider et al. 2017). The species and functional diversity metrics will be computed using the R package FD.Species and functional diversity metrics will be summarized both spatially and non-spatially. For non-spatial comparison, the mean and standard error will be computed for each replicate. The main effects of thinning and prescribed fire, and their interaction, will be assessed using two-factor Analysis of Variance (ANOVA). Regression analysis will be used to examine the shape of the response to increasing fire return interval (1, 2, 3, and 4-year intervals plus no fire). For spatial analysis, semivariograms will be developed for each treatment. Semivariogram analysis will reveal the range of spatial autocorrelation as well as the spatial scale of fine-scale variation and measurement error (the nugget) (Goovaerts 1998). Semivariance analysis will be conducted using the R packages RGDAL and GSTAT. The semivariogram model will be selected using Akaike's Information Criterion (AIC). A model with more parameters, or with a smaller estimated range, indicates a more complex spatial pattern in the data.Study 2 - The taxonomic and functional diversity of the forest overstory will be the focus of this study. Taxonomic diversity will be assessed using 10 BAF prism plots randomly located in each treatment area. The species identity and diameter of trees greater than 10 cm DBH will be recorded and summarized by treatment area.The morphological functional traits will include canopy height (CH), plant area index (PAI), and foliage height diversity (FHD). CH is the distance from the ground to the top of the canopy. PAI is the projected plant area per horizontal ground area. FHD is a measure of the variation and number of canopy layers. These traits will be measured using either laser scanning (LiDAR), structure from motion (SfM) photogrammetry, or a combination of these techniques. The physiological traits will include leaf chlorophyll, leaf carotenoids, and leaf water content, as described for Study 1.The functional diversity indices FRic, FEve, and FDiv will be computed as described above for the understory plant community but will be based on the overstory functional traits obtained from imagery.The hypothesis that functional diversity is increased by forest disturbance will be tested using both spatial and non-spatial analysis. For non-spatial analysis, two-factor ANOVA with interaction will be used to compare the effects of thinning, burning, and their interaction. Each variable will also be analyzed by fire return interval. The spatial analysis will involve the modeling of semivariance as described for Study 1.