Date of Award

Fall 2013

Degree Type


Degree Name

Doctor of Philosophy (PhD)


Forestry and Natural Resources

First Advisor

Michael Saunders

Second Advisor

Michael Jenkins

Committee Chair

Michael Saunders

Committee Member 1

Michael Jenkins

Committee Member 2

James Speer

Committee Member 3

Hao Zhang


Over the course of the last century, we have observed changes in forest composition and structure related to changes in disturbance regimes. Reduced frequency and severity of disturbance events has allowed for the ingrowth of shade-tolerant, mesophytic species into forest understories; the result is shifting species compositions and changes in forest structure. In some cases, the loss of masting species, such as oak (Quercus spp.) and hickory (Carya spp.), may also accompany a shift in forest function and further accommodate other species shifts among mammal, insect, and bird populations. Management efforts have been suggested as a possible means to imitate natural patterns of disturbance that will sustain historical forest species composition, structure, and function, but studies designed to investigate forest responses have exhibited mixed results. Numerous studies have explored changes in old-growth forest composition and structure over time through long-term or dendrochronological studies, but this research also explores composition and structure of forests over time under natural and management scenarios; to the best of my knowledge, the data used for this work is the oldest spatially explicit inventory of an eastern deciduous mixed-hardwood forest. I used a long-term study of the Purdue Davis Agricultural Center mixed-hardwood deciduous forest stands in east-central Indiana, U.S.A. I examined transitions of species composition and structure in a natural, relict forest as a contrast to stands that were subjected to partial, selective harvesting in the context of reduced disturbance regimes. All stands were grazed by domestic livestock until about the 1920s, at which time the forest stands were fenced to eliminate grazing. Dutch elm disease also impacted these stands around the middle of the 20th century by eliminating many large elms (Ulmus spp.) and reducing elms to a predominantly understory species. The relict stand was generally untouched since 1926, with the exception of the occasional removal of some dead or near-dead trees (removal of 5.8 trees ha-1; 5.3 m3 ha-1). The single-harvest stand was selectively harvested in 1951 (removals of 18.4 trees ha-1; 8.1 m3 ha-1), while the double-harvest stand was harvested in 1951 (removal of 11.6 trees ha-1; 9.8 m3 ha-1) and again in 1964 (removal of 11.3 trees ha-1; 7.8 m3 ha-1). The three stands were initially sampled in 1926 ¡V 1927; all trees ,,d 10 cm at 1.37 m height (dbh) were mapped, identified by species, and measured by dbh, and I followed the same sampling protocol in 2010. Coarse woody debris (CWD), including down dead wood (DDW) and standing dead wood (SDW), abundance and location were estimated using probability proportional to size method based on a 10 m by 10 m sampling grid across each stand. In addition, dendroecological analysis was conducted by coring trees in a subset area of 2.25 ha within the relict and one of the managed stands; I sampled a single tree closest to the sample point based on a 10 m by 10 m sampling grid to estimate disturbance frequency and size over time.

Partial, selective harvests resulted in increased species richness, diversity, and species mingling in different canopy strata relative to changes in the relict stand. All three stands experienced increases in stem density and basal area, particularly in the number of understory shade-tolerant trees, notably sugar maple (Acer saccharum Marsh.), elms, and basswood (Tilia americana L.). Shade-tolerant species became less aggregated over time, and shade-intolerant species, particularly ash (Fraxinus spp.) and hickories, became increasingly aggregated. Spatial arrangement of overstory trees became more random over time, understory trees became less aggregated, and there appeared to be little correlation among strata in either period. At a neighborhood level, neighborhoods became increasingly smaller with a greater number of neighbors over time in all stands, and the relict and single-harvest neighborhood species became more similar, but neighborhoods of the double-harvest stand became less similar. The disturbances of the partial harvest stands resulted in a retrogression of stand development by allowing portions of the stand to persist in an earlier stage of development and maintain more mid-seral, mast- producing species, notably oaks and hickories; this was considered a positive development by diversifying structure and maintaining a degree of functional resilience by ensuring the presence of mast-producing species.

Dendroecological analysis of the relict and single-harvest stands indicated that disturbance was primarily characterized by small-scale (< 200 m2) overstory disturbances, but the relict stand had a significantly higher number of gaps and stand area in gaps annually. It was evident that after 1952 disturbance patterns within the two stands differed; approximately 27% of the sample area of the managed stand had newly-formed gaps because of the Dutch elm disease and harvest, stand structure likely accounted for the observed differences between the two stands. The managed stand had fewer gaps, less area in new gaps, and less variability in gap size because there were fewer large gaps formed after 1952.

Related to differences in disturbance histories, I observed very different temporal patterns of species group establishment and canopy accession between the two stands. With the cessation of grazing around 1920, sugar maple became increasingly established in the understory and recruited into the canopy. Canopy accession patterns indicated virtually all trees rely on disturbance for recruitment, and most were established in gap or high light environments. In many cases differences in mean understory residence time were more related to the timing of establishment and canopy accession of the respective species rather than silvical traits of the individual species.

Natural and anthropogenic disturbance patterns not only influence patterns of stand growth and development, but they also contribute to the structure of the dead wood material that many organisms rely on. The individual piece size, density, and volume of down dead wood (DDW) and standing dead wood (SDW) were all higher in the relict stand compared to the double-harvest stand. In addition, higher densities and volumes of DDW and SDW were observed in the largest size and most highly decayed classes in the relict stand. Higher rates of disturbance and large pieces of DDW and SDW in the relict stand also meant higher connectivity across all size classes, generally fewer patches, and larger average patch size across most size classes. Even after almost 50 years, coarse woody debris in the partially harvested stands has not recovered to natural levels. The differences in abundance and connectivity were attributed to higher disturbance rates in the relict stand and what I termed the storage effect, which suggests that the larger piece sizes and more decay-resistant species will tend to persist on site longer, thus, allowing for greater accumulation and connectivity as trees die.

This long-term study provided insight into forest development during a period of transition in disturbance regimes; less frequent and intensive disturbances have altered stand structures and composition over time. Although the partial harvest stands did successfully recruit mid-successional, shade-intolerant species, such as oak and hickory, into the overstory, without further management, it is likely they too will transition to late-successional species. The fact that these stands were able to recruit oaks and hickories is likely a tribute to past disturbance regimes, which provided a pool of advanced regeneration prior to harvest, and fortuitous timing of the harvest events to release the understory.

This study highlighted both the importance of disturbances on species composition and stand structure, and the fate of ¡§disturbance legacies¡¨ (e.g. CWD) over almost a century. That highlights the importance and value of long-term, detailed planning to ensure sustainability of structure and legacies. Further, efforts to replicate past disturbance regimes as a coarse filter for biodiversity conservation must be extended to match impacts on the understory composition as well as the overstory. Planning must also ensure the provision of suitable coarse woody debris over time to maintain, or even enhance, biodiversity in managed hardwood forests.