Davies, Chris

Theoretical work on community recovery, development, stability, and resistance to species invasions has outpaced experimental field research. There is also a need for better integration between ecological theory and the practice of ecological restoration. This thesis investigates the dynamics of community assembly following peat mining and subsequent restoration efforts at Canada's most southerly raised bog. It examines mechanisms underlying plant community changes and tests predictions arising from the Dynamic Environmental Filter Model (DEFM) and the Fluctuating Resource Hypothesis (FRH). Abiotic, biotic and dispersal filters were modified to test a conceptual model of assembly for Wainfleet Bog. Hydrology was manipulated at the plot scale across multiple nutrient gradients, and at the whole bog scale using peat dams. Trends in time series of hydrological variables were related to restoration actions and uncontrolled variables including precipitation, evapotranspiration and arrival of beaver. Impacts of a changing hydrology on the developing plant community were compared with those from cutting the invasive Betula pendula. Transplanting experiments were used to examine species interactions within primary and secondary successional communities. Seedlings of B. pendula and the native Betula papyrifera were planted together across a peat volumetric water content (VWC) gradient. Impacts of beaver dams were greater than those of peat dams and their relative importance was greatest during periods of drought. Cutting of B.pendula had little effect on the secondary successional plant community developing parallel to blocked drains. Phosphorus was the main limiting nutrient with optimum levels varying substantially between species. Primary colonisers formed a highly stable, novel plant community. Stability was due to direct and indirect facilitative interactions between all species. Reduction in frost heaving was the major mechanism behind this facilitation. Interactions within the secondary successional community were mostly competitive, driven by light and space availability. However, restricted dispersal rather than competition limited further species recruitment. Predictions based on the DEFM were partially correct. A splitting of this model's biotic filter into competition and facilitation components is proposed. There was little support for the FRH based on nutrient levels and VWC. B. pendula had higher germination and growth rates, tolerance to a wider range of peat VWCs and a greater resistance to deer browsing than native birch. Peat mining, combined with restoration actions and the arrival of beaver has moved much of the bog back to an earlier successional stage circa 350+ years BP. Evidence points to B. pendula being a "back-seat driver" in the ecosystem recovery process. Indirect facilitation of a native by an exotic congener, mediated through herbivory, has not been described previously. Shifts in relative contributions of facilitation, competition and dispersal limitations to community assembly may be useful process-oriented measures for gauging progress in restoration.

Author Keywords: Betula pendula, community assembly, competition, facilitation, peatland, restoration

ABSTRACT

Calving site selection and fidelity in a restored elk (Cervus elaphus) herd in Bancroft

Ontario, Canada.

Michael R. Allan

Parturition site selection by ungulates is believed to be influenced by forage abundance and concealment from predators. In 2011 and 2012, I used vaginal implant transmitters and movements to identify calving sites for 23 GPS collared elk (Cervus elaphus) from a restored herd. I tested the hypothesis that maternal elk used sites with higher forage and denser concealment compared to pre-calving sites at micro and macrohabitat levels. I detected no significant microhabitat differences from direct measurements of vegetation. At the macrohabitat scale, based on proximity of landcover classes, mean distances to hardwood forests was significantly less for calving (153 m) than pre-calving sites (198 m).

Site fidelity is hypothesized to offer security in terms of familiarity to an area. I tested the hypothesis that females demonstrated fidelity to their previous year's location during pre-partum, parturition, post partum, breeding and winter periods. Elk were more philopatric during parturition and post partum than during breeding. Compared to winter elk were more philopatric during pre-partum, parturition and post-partum periods. Expressed as distance between consecutive-year calving locations, site fidelity varied with 27% of females exhibiting high (<1 km), 18% moderate and 55% (>2.9 km) low fidelity.

I measured nearest-neighbour distances at calving time, exploring the hypothesis that females distance themselves from conspecifics. Elk increased the average distances

to collared conspecifics during parturition; however, sample sizes were small. This strategy might influence calving site selection.

Rapid movement prior to parturition, low site fidelity and spacing-out of females during parturition appear to be strategies to minimize predator risk and detection. Little evidence of selection for vegetation structure suggests this may not be limiting to these elk.

Author Keywords: calving, elk, fidelity, movement, parturition, selection