Research in the GEC lab spans the spectrum from empirical to theoretical approaches in attempting to understand the way ecosystems behave in the face of various stresses (e.g., pollution, invasion, climate change) and their remedies (e.g., restoration). It is now quite clear that the interaction between multiple stressors on ecosystems will interact in ways that that will produce complex synergistic effects. As data sources and theoretical knowledge become increasingly advanced, we will continue to need improved models and measures to both understand and attempt to manage these stressors. Thus, one line of research involves the development of ecological models and quantitative tools to deal with complex eco-informatical data. A consistent theme in the lab is scaling, whether in the form of understanding scaling laws or in the bridging of small-scale/short-term disturbances to large-scale/long-term pattern and process: The need to examine cross-scale phenomena is essential in understanding the ecological architecture of stress-mediated phenomena. Data sources for our research questions include our own field-based studies in forest diversity and dynamics, published databases, paleo-pollen records and recent work in the lab has begun to incorporate tree-ring and carbon isotope analysis to assess historic (past century) stress in forest ecosystems. The main current research areas are described below. Several and diverse side projects have been initiated by the GEC lab over the years in collaboration with others which have involved involve clonal plants, whirligig beetles, heathlands, deserts and make-believe food webs.
Field-ecological studies in the industrially-damaged Sudbury, Ontario region have been undertaking since 2000. We have shown that while restoration efforts can increase the diversity and structure of forests, it can (perhaps alarmingly) also facilitate the colonization of non-native and potentially invasive species and prevent the establishment of native species (Rayfield et al. 2005). This work has already impacted the way forests continue to be restored in the region of Sudbury, since local municipal groups are now considering ways of enhancing restored forests through plantation of native species. We have also shown (Anand et al. 2005), for the first time, that diversity of vascular plants can be used as a surrogate for biodiversity of other taxonomic groups in naturally recovering forests but not in artificially restored ones; this work will impact biodiversity assessment (Desrochers & Anand 2005) and forest management practices at large, since most existing studies on biodiversity surrogacy focus only on pristine ecosystems. This work has attracted international collaborators to the lab (e.g., Dale et al. 2006, Ricotta & Anand 2004, 2006). We have also collaborated with microbiologists and soil scientists on to help assess terrestrial damage on the Sudbury landscape of historic pollution. We found, surprisingly, that there is still a footprint on the landscape of heavy metal presence in the soil as far as 60 km from the historic smelters (Anand et al. 2003) and our model predicts that this effect may be present as far as 100 km from the historic smelter (Tucker & Anand 2003). These results have been confirmed by the recent Sudbury Soils Study (2009). A new project is underway in collaboration with the City of Greater Sudbury (fingers crossed for funding…) to examine the role of forest refugia on landscape recovery. Current PhD student Jennifer Babin-Fenske is determining impacts and relationships between plant and insect diversity and gathering data on Forest Tent Caterpillar population response to historic pollution in the area that will be used to parameterize a model (described below).
We have introduced Hidden Markov models to community ecology and specifically to the study of complex community trajectories traced by ecological recovery and restoration. These models provided a much-needed improvement to the conventional and widely used stationary Markov models for studying ecological success. Stationary Markov models fail when applied to complex (nonlinear) dynamical systems with multiple attractors, however, these dynamics are quite realistic for some ecological recovery trajectories. Hidden Markov models can deal with these more complex dynamics and thus will be much more useful, particularly for disturbed ecosystems and alternative recovery pathways. This work was done with former MSc student Brian Tucker (Tucker & Anand 2005) and has resulted in a menu-driven, UNIX based computer software package (EMAI). We have developed statistical tools that improve information-theoretical definitions of ecological diversity and complexity (e.g., Shannon entropy) by incorporating structural, spatial and taxonomic structure (Anand & Orlóci 1996, 2000, Ricotta & Anand 2004, 2006) and that allow determination of significance levels via randomization testing. These methods have proved to be useful for assessment of short-term ecological disturbance and recovery through the work of an MSc student, Rachelle Desrochers, (Desrochers & Anand 2005) and long-term ecosystem response to climate change (Orlóci, Anand, de Patta Pillar 2002). Dr. de Patta Pillar (Brazil) has incorporated one of the new measures in his excellent ecological analysis software package (MULTIV). Work on spatial techniques with current PhD student (Mark Leithead) and former research assistant (Lorna Deeth) has led to improved spatiotemporal analysis of tree populations to test the Janzen-Connell hypothesis of ecological response to pests (Leithead et al. 2009).
Over the past 6 years we have been working on spatially-explicit computational modeling of forest dynamics to examine questions about stress and disturbance that we cannot answer through data collection due to restrictions of resources or questions that are of theoretical interest. One project was developed as an interdisciplinary collaboration with theoretical physicist and the co-supervision of a physics graduate student (Pagnutti et al. 2005, 7). The project involved studying spatial pattern through small scale-disturbances in a tropical forest. Our results demonstrated both the robustness of forest spatial patterns to changing lattice geometry, as well as the role of small-scale disturbance propagation in maintaining this robust global pattern. This suggests that small changes in local disturbance (or, alternatively, stress through other mechanisms such as selective logging) regimes can have dramatic consequences for global structure of forests. This collaboration is ongoing. Through the development of FORSITE (Forest model with Random, Spatial, Individual and Temporal Effects) with recent postdoctoral fellow (Paul Caplat) and collaborator Chris Bauch, we have shown that cycling in long-term dynamics of tree populations can depend on the way competition is modeled (Caplat et al. 2008a), that species spread across landscapes due to climate change stressors will be much slower than previously predicted due to local interactions with resident species as well as disturbance regimes (Caplat et al., 2008b) and that timing of invasion matters in biological invasions (Caplat et al. 2009a). Very recently we have found profound effects of changing plant traits on successional pathways in forests (Caplat & Anand 2009b). This basic model will be used to test several other hypotheses in forest ecology. With recent MSc student (Aaron Langille) we have discovered a new method based on dynamic modelling to estimate effective dispersal distance of populations from spatial point patterns (Anand & Langille, submitted) and developed a model to determine the relative role of dispersal and habitat in explaining spatial patterns of tropical tree patterns (Langille & Anand, in prep.). PhD student (Jennifer Babin-Fenske) is developing a spatially-explicit model of Forest Tent Caterpillar infestation in Northern Ontario. The model will examine the hypothesis that insect infestations may increase in intensity and return cycle in areas that have been stressed by pollution and other human-induced disturbances.
We are currently involved in a large collaborative international research project funded by the Inter-American Institute for Global Change research (http://www.iai.int/). The GEC lab serves as the Canadian Node for the project. The other nodes are led by labs in Brazil, Chile, Venezuela and Argentina. PhD student (Luca Silva) is examining long-term forest-grassland dynamics though stable carbon isotope studies and has found important results relating climate change to potential forest expansion in the region (Silva et al. 2009). A PhD student in Brazil (Fernando Joner; co-supervised by Prof. Valério de Patta Pillar) is examining trait-based community assembly in Northern Ontario forests and Atlantic rainforest in Brazil. Ongoing collaborations continue with former postdoctoral fellow (Prof. Keming Ma). A former PhD student in China (Prof. Yuxin Zhang; co-supervised with Prof. Ma) demonstrated the global robustness diversity-area relationships in forests (Zhang et al. 2006). A current PhD student (Mark Leithead) is examining disturbance dynamics in old-growth forests in Northern Ontario and Atlantic rainforest in southern Brazil at several scales and the implications of these for biodiversity in the face of global ecological change.