Palynologists analyse fossil pollen assemblages to reconstruct temporal changes in plant communities and infer past patterns of environmental change. To quantify the estimation of environmental variables, the relationship between pollen assemblages and environmental conditions must be understood. This research is based on an array of 31 pollen trap sites established by the late Dr. Gurdip Singh in 1970-1975 along a 1400-km transect from Lake Eyre in South Australia to Waldron’s Swamp near the NSW coast. Pollen samples were collected periodically from each trap over a minimum of 3 years (median=6.8 years, maximum=9.9 years). Samples were prepared using standard techniques and pollen types were identified and counted on microscope slides. I have calculated climatic variables were calculated for each trap site using BIOCLIM. Mean annual rainfall ranges from 122 mm in the arid inland to 1073 mm near the coast. I have modelled aggregate influx rates of 35 pollen taxa against mean annual rainfall using generalized linear models with quasi-Poisson errors and log link. Jack-knife estimates of rainfall and 95% confidence limits were calculated for each site using a maximum likelihood approach. Root-mean-squared prediction error was 103 mm. These preliminary results suggest that it will be feasible to refine the modelling approach to provide reasonably accurate estimates of past rainfall from fossilized pollen assemblages in lake sediments and peat cores. Next steps include improving the taxonomic resolution of pollen types, potentially leading to a more powerful estimation model, further testing of prediction accuracy using other modern pollen assemblage data, and application of the model to fossil pollen data to reconstruct palaeorainfall patterns in southeast Australia.
This sabbatical project examines changes in the plant communities and altitudinal distributions of plant species on the Mt. Field Plateau, Tasmania, Australia in response to a warming climate. In January-May, 2019, with the help of John Davies, an old friend and colleague whom I met at the University of Tasmania in the 1970's, a network of 234 vegetation plots, initially surveyed in the summers of 1980-1982 as part of my Ph.D. research, were re-sampled using the same methods. The aim was to examine changes in the plant communities and test for effects of climate change. Results showed shifts in the species composition of plant communities that are consistent with warming and Bayesian models demonstrated upward altitudinal shifts in the distributions of many plant species. A paper reporting initial results was presented at the International Association for Vegetation Science (IAVS) 62nd annual symposium in Bremen, Germany in July 2019. The research has subsequently been reported in several other conference presentations and journal articles are currently in preparation.
Multivariate statistical methods are used to analyze and interpret patterns of variation in the species composition and structure of ecological communities. The development and evaluation of more effective techniques for community analysis has been my major research interest since I was a graduate student. This ongoing project uses a combination of simulated community data with a known structure and real data sets from many kinds of terrestrial and aquatic communities to evaluate and refine novel methods for analyzing community variation in space and time, including ordination, clustering, group discrimination, allocation, and calibration. Dr. Jari Oksanen, at the University of Oulu, Finland has been a major collaborator on some of this work. Current collaborators include Brian Ickes, Upper Midwest Environmental Sciences Center, US Geological Survey and Shane Chalwell, Plantecology Consulting, Perth, Australia. The project aims to develop robust methods that perform well over a wide range of possible variatons in community structure and properties.
One sub-project, which has been an important focus of my research for the past 10 years, is the development of effective techniques to assess the directionality of change in community composition. This work has its origins in contract research that I performed for The Nature Conservancy in 2001-2003, during which I developed a new method, Trajectory Analysis, to assess the success of wetland restoration, using data from The Disney Wilderness Preserve in Florida. The method uses a permutation approach to test whether the time trajectories of sampling units in ordination space are in an hypothesized direction. In the case of a successful restoration, the hypothesized direction would be in the direction of restoration targets or parallel to the community trend correlated with a particular environmental gradient (e.g., for wetland restoration, this may be a gradient of increasing hydroperiod or depth of flooding). My lab at SIUE has continued to develop and refine Trajectory Analysis. The research has been reported in five conference presentations and collaborative research with Mike Ross and Jay Sah at Florida International University, funded by a contract from South Florida Water Management District, led to the publication of a paper describing the technique and applying it to analyze changes in wetland vegetation in the Everglades. Currently, in collaboration with Brian Ickes, USGS, the method is being applied to fish community data collected over 20 years in the Long Term Regional Monitoring Program for the Mississippi River system.
Other sub-projects include the evaluation of methods for estimating the community dissimilarity between sampling units with no species in common using network shortest path techniques, the investigation of the accuracy of axis lengths in detrended correspondence analysis (DCA) in estimating beta diversity, a study of the effects of rare species on the performance of standardization by species in community analyses, and the comparitive evaluation of methods for clustering of community data. Each of these sub-projects has yielded promising results leading to presentations at international conferences and papers are in preparation.
Bottomland forest once covered large areas of the floodplains of the Mississippi River and its major tributaries but since European settlement, this ecosystem has been drastically reduced in area by clearing for agriculture. Among the essential ecosystem services provided by bottomland forest are nutrient cycling, flood mitigation, erosion control, and wildlife habitat. Since 1980, there have been increasing efforts by government agencies and private conservation organizations to restore bottomland forest. This project aims to evaluate restoration success by means of long-term monitoring of restoration sites, chronosequence studies, and experimental studies of factors affecting the growth and survival of species used in restorations.
One sub-project examines bottomland forest restoration in the Lower Mississippi Valley. In collaboration with Dr. Loretta Battaglia, Southern Illinois University Carbondale, we are monitoring passive restoration of bottomland forest on abandoned agricultural land in the Ouachita Wildlife Management Area, near Monroe, Louisiana. Our main study site is a field that was retired from soybean production in 1985. This sub-project examines the processes of vegetation succession, including dispersal and establishment of woody species, and the gradual development of bottomland forest structure and composition. Two journal articles have been published so far from this ongoing work.
Another sub-project is the evaluation of bottomland forest restoration in the Upper Mississippi Valley, using a chronosequence of sites on which restoration began in different years. The study sites were actively restored by government agencies (US Army Corps of Engineers, Missouri Department of Conservation, US Fish and Wildlife Service) by planting of large-seeded tree species (mainly flood-tolerant oaks and pecan) and establishment of grass cover to inhibit the recruitment of small-seeded tree species, followed by periodic mowing. We established five 0.1-ha permanent vegetation plots within each of nine restoration sites, ranging in establishment year from 1990 to 2012. All trees within each plot were tagged for future monitoring of growth and survival. Data on tree diameter and height by species and community properties (e.g., richness, diversity, total density, total dominance) were analyzed using generalized linear modeling to quantify changes with age. Trajectories of change were compared with data from mature bottomland forest reference sites in order to test whether the restorations are on target. The results indicated success in some aspects but not others. Supplemental plantings and some changes in management may be required to attain the structure and composition of mature forest. A Master's thesis has been produced and a journal article is in preparation. Future research will resample the plots to examine patterns of growth and survival.
A third sub-project examines the effects of white-tailed deer on the survival and growth of planted trees in restoration sites. Experiments used metal mesh deer guards to compare levels of deer rubbing, relative growth rates, and mortality between guarded trees and unguarded controls. Results showed that deer rubbing reduces relative growth rates and increases mortality. Successful restoration may require installation of guards to reduce deer damage and/or control of deer populations. A Master's thesis has been produced and a journal article is in preparation. Future research may experimentally examine the interaction between flooding and deer damage (rubbing and browsing) on the growth and survival of planted trees.
In 2010, in response to a proposal developed by a group of faculty from several departments, a 154-ha area on the western edge of campus was set aside as the SIUE Nature Preserve and protected from development for 50 years. The preserve includes Sweet William Woods (SWW), the best example of oak-hickory forest on campus, and a corridor extending northwards along the bluff line that provides potential connectivity to Bohm Woods State Nature Preserve (BWNP), located adjacent to campus and considered by the Illinois Department of Natural Resources to be the only remnant of the original pre-European settlement forest remaining in Madison County. The Minchin lab (in collaboration with the Essner lab) commenced research in SWW in 2006 on a range of research projects aiming to quantify, describe, and monitor the biodiversity of the forest. In 2007 this work was extended to Bluebell Woods (BBW), located in University Park, and BWNP. Following establishment of the SIUE Nature Preserve in 2010, the research was further expanded to include the bluff corridor. The research is based on a network of permanently marked sampling plots first established in 2006 and later upgraded and extended into the bluff corridor region of the SIUE Nature Preserve. The current network consists of 102 0.1-ha circular plots in the SIUE Nature Preserve, 30 in BWNP, and 35 in BBW, with several additional plots in SWW for specific purposes.
Data from the plots have been the basis of a total of 20 undergraduate research projects and four Master's projects, involving students from both the Minchin lab and the Essner lab. Among the topics investigated are patterns of community composition, diversity, and conservation value in the vegetation, including the spring ephemeral community, multivariate habitat modeling for bird species and the southern flying squirrel, seasonal patterns of variation in leaf litter insect communities, the abundance and distribution of woody vines (lianas), variation in conservation value, diversity, and degree of invasion by exotic species with forest age and distance from edge, and the growth and mortality of the dominant tree species. Results have been reported in two journal articles, three reports, and 22 conference presentations. Several additional journal articles are in preparation.
This project, in which we collaborate with Dr. Rick Essner and students in his lab, uses statistical modeling to build multivariate predictors of habitat occupancy and use by bird and mammal species in upland forest, bottomland forest, and prairie ecosystems in southwestern Illinois. The models can be used to predict occurrence and abundance of a species in areas for which survey data and not available and to identify critical habitat requirements of endangered or threatened species, facilitating management of existing habitat to enhance suitability for target species and informing habitat restorations. Results of this research have been reported in nine conference presentations and two journal articles.
Follow us on Facebook
