Disturbance-based restoration for recovery of pitch pine (Pinus rigida)

in the Thousand Islands Ecosystem: a comparison of prescribed fire and

mechanical disturbance on seedling regeneration.

By

Joshua Frans Van Wieren

A thesis submitted to

the Faculty of Graduate Studies and Research

in partial fulfillment of the requirements for the degree of

Master of Science

Department of Biology

Carleton University

Ottawa, Ontario

January 2014

© 2013, Joshua Frans Van Wieren

ii

Abstract Disturbance plays an important role in maintaining forests worldwide. Many natural

disturbance regimes, especially wildfire, have been disrupted or modified over time,

which can lead to the loss of disturbance adapted forest communities. Determining

historic disturbance patterns and mimicking them can be difficult, and can be further

complicated when disjunct or marginal populations exhibit different patterns than those

at the core. Pinus rigida (pitch pine) is strongly associated with wildfire in the core of its

range, however the association becomes less certain towards the species’ range

margins. I tested the efficiency of two disturbance restoration techniques on increasing

P. rigida seedlings using a Before-After Control-Impact (BACI) design. I compared

prescribed fire with a mechanical treatment and a control at two sites at the northeast

range margin of P. rigida in the Thousand Islands Ecosystem, Canada. I measured

canopy cover, understory cover and depth to mineral soil to control for their effects on

seedling regeneration. Fire had the greatest effect on regenerating P. rigida seedlings

and mechanical treatments were ineffective. My results suggest that the use of

prescribed fire is the best approach to increase P. rigida seedlings in the Thousand

Islands Ecosystem, possibly because these populations are exhibiting phenotypic

plasticity in traits that favour conditions created by fire. Restoration considerations are

discussed including the potential for modified mechanical disturbance treatments where

prescribed fires are not feasible.

iii

Acknowledgements Funding for this project was generously provided by the Parks Canada Agency. I would

like to thank my advisor Andrew Simons for his timely and positive feedback throughout

my thesis and for his patience with me working in spurts as I balanced my career with

my Masters. Thank you to my committee; Naomi Cappuccino and David Currie for their

feedback and positive approach to teaching. A big thank you to Sheldon Lambert, Jeff

Leggo and Harry Szeto at Thousand Islands National Park for providing me with the

opportunity to complete my Masters while continuing my role as a Park Ecologist.

Thank you to Mary Beth Lynch and Pauline Quesnelle, the two best botanists I know,

for their help in identifying the eruption of new species that emerged post treatments!

Thanks to my Thousand Islands National Park fire teammates and Parks Canada Fire

Program staff, especially Derek Bedford, Marie-eve Foisy, Victor Kafka and Raymond

Quenneville, that helped completed these complex fires in settled areas flawlessly.

Thanks to Doug Bickerton for his local expertise and passion for TIE pitch pine

populations. A big thank you to Paul Zorn for always having the time to brainstorm with

me and for his sage guidance throughout my thesis. Thank you to Kristen Zorn and

everybody else that provided assistance and enthusiasm in the field. A special thank

you to my Mother for (as always) expressing interest and her willingness to listen and

provide encouragement. Finally, I owe a big thank you to my wife Pauline for all her

help and encouragement and most of all for all the times we spent talking about

scientific theory, statistical principles and writing techniques, all while working on her

own PhD. Having one of the most well-rounded ecologists I know working back-to-back

with me each weekend was invaluable and enhanced my entire experience!

iv

Table of Contents Abstract ........................................................................................................................................................ ii

Acknowledgements ................................................................................................................................... iii

List of Tables ............................................................................................................................................... v

List of Figures ............................................................................................................................................ vi

1.0 Introduction ........................................................................................................................................... 1

1.1 Disturbance Ecology ............................................................................................................................ 1

1.2 Fire Influences on Forest ..................................................................................................................... 2

1.3 Pinus Fire Adaptations ........................................................................................................................ 2

1.4 Fire Suppression .................................................................................................................................. 4

1.5 Pinus rigida Miller ............................................................................................................................... 5

1.6 Thousand Islands Ecosystem Pitch Pine .............................................................................................. 8

1.7 Restoration Techniques .................................................................................................................... 10

2.0 Methods .............................................................................................................................................. 11

2.1 Study Area ......................................................................................................................................... 11

2.2 Study Design ...................................................................................................................................... 13

2.3 Treatment Types ............................................................................................................................... 14

2.4 Sampling Design ................................................................................................................................ 16

2.5 Variables Measured .......................................................................................................................... 17

2. 6 Statistical Methods .......................................................................................................................... 18

3.0 Results ................................................................................................................................................ 21

3.1 General Responses of Vegetation Community and Soil Depth to Treatments ................................. 21

3.2 Model Selection ................................................................................................................................ 23

4.0 Discussion........................................................................................................................................... 26

4.1 Treatment Effectiveness ................................................................................................................... 26

4.2 Treatment Effects on Competition and Access to Mineral Soil ........................................................ 27

4.3 Possible Explanations of the Positive Effects of fire on Seedling Recruitment ................................ 29

4.4 Possible Phenotypic Responses to Fire ............................................................................................. 30

4.5 Restoration Considerations and Further Studies .............................................................................. 32

5.0 Conclusions ........................................................................................................................................ 37

References ................................................................................................................................................ 39

Appendix .................................................................................................................................................... 46

v

List of Tables 1. Complete list of variables and their measurement units. 2. The model combinations selected against in an MMI framework, analyzed using

a two-factor poisson mixed model ANCOVA. 3. AIC model selection statistics of ANCOVA models. 4. Model-averaged estimates of the variable coefficients and standard error,

calculated using both top models. 5. values for each parameter in the original treatment + depth to mineral soil tests

and post hoc interaction tests, including r squared values for each model.

vi

List of Figures 1. North American distribution of Pinus rigida. 2. Study sites at Georgina Island and Mallorytown Landing. 3. Transect and quadrat placement. 4. Effect of treatments on canopy cover. 5. Effect of treatments on understory cover. 6. Effect of treatments on depth to mineral soil. 7. Change in relative importance of fire compared to depth to mineral soil (soil) for

an ANCOVA model that includes a treatment*soil interaction term versus a model that does not.

8. Pitch pine seedling numbers in 10x10m plots in the three years post treatment. A1. Before and after site photos of the Georgina Island fire. A2. Before and after site photos of the Mallorytown Landing fire.

1

1.0 Introduction

1.1 Disturbance Ecology

Ecological restoration is a key tool used to mitigate the effects of disturbances which

degrade ecosystems (Hobbs and Norton 1996). However, many forest ecosystems are

becoming increasingly imperilled because of a lack of disturbance (Van Sleeuwen

2006). Disturbance is known to influence the development, structure and function of

forests as well as the persistence of many forest-dependent species (Attiwill 1994).

Natural disturbances such as fire, wind, insect attacks, drought and ice storms are

driving forces in shaping many forest ecosystems (Bormann and Likens 1979; Attiwill

1994; Rogers 1996). Maintaining natural patterns and processes, such as disturbance

regimes, is a critical component to preserving the ecological integrity of an ecosystem

(Norton 1992; Grumbine 1994; PCA 2013). Concordantly, it is important to understand

disturbance processes and incorporate them into forest management (Whitmore 1982;

Rogers 1996). However, many disturbance regimes have been suppressed or altered

(Parsons and DeBenedetti 1979; Baker 1992) and in many cases it has become difficult

to determine which disturbance or which regime is required to maintain disturbance

dependent forests (Ne’eman et al. 2004). Hobbs and Norton (1996) suggest that

understanding the processes leading to degradation and decline as well as developing

methods to reverse or ameliorate them are two of the key processes in ecological

restoration. As such, it is imperative that conservation managers understand how to

implement disturbance restoration actions in forests where natural disturbances have

been altered.

2

1.2 Fire Influences on Forest

Many forest ecosystems are wholly or partially dependent on fires (Naveh 1975;

Abrams 1992; Wolf 2004). Wildfire is known to create a patchwork of different

ecological communities across a landscape (Davis et al. 2003). Different ecosystems

are influenced in a variety of ways depending on the local fire regime (Keeley et al.

2011); however, fire generally influences vegetation composition, structure and pattern

by decreasing competition, and preparing seedbeds and soil conditions suitable to

certain species (Van Sleeuwen 2006). Pinus (pine) communities are often more fire

prone than hardwood dominated ecosystems because of low moisture content in their

needles and the high content of flammable organic compounds (Whitney 1986). Fire is

often beneficial for tree regeneration because it provides access to the mineral soil by

removing the duff layer, reduces understory competition by controlling brush and

increases light penetration by killing canopy trees (Van Wagner 1970; Heinselman

1981; Wright and Bailey 1980; Bergeron and Brisson 1990). Pines benefit from these

conditions because many do not grow well in heavy shade and seedlings tend to

survive best on mineral soil and with little or no competition from shrubs and shade-

tolerant trees (Heinselman 1981).

1.3 Pinus Fire Adaptations

Pinus spp. found in fire-dependent ecosystems have developed a wide variety of fire

adaptations. Fire adaptation usually involves either individual resistance (survival of

adult trees) and/or stand resilience (seedling recruitment after the loss of a mature

stand; Keeley and Zedler 1998). Pinus that benefit from lower-severity ground fires

tend to grow taller and shed their lower dead branches thus discouraging ladder fuels

and dangerous crown fires (Keeley et al. 2011). Pinus in ecosystems that experience

3

more intense stand-replacing crown fires often do not shed their lower branches (Keeley

and Zedler 1998) and also retain empty open cones (Ne’eman et al. 2004), both of

which are highly flammable and increase the chances of a high intensity crown fire.

Some have adaptations that help them survive fires such as thick bark and deep rooting

(Heinselman 1981; McRae et al. 1994; Knight 1999). Others have adaptations that

enable adult trees to recover from fire damage, including epicormic sprouting from tree

trunks or branches in damaged areas (Keeley and Zedler 1998). These sprouts can

appear in large densities after intense fires and allow canopies to recover in a matter of

years (Climent et al. 2004). Some Pinus are able to resprout from their root collar as

seedlings (Klaus 1989; Keeley and Zedler 1998) and a few even as adults (Cain and

Shelton 2000; SCBD 2001).

Pinus spp. also exhibit a number of adaptations that improve their chances of

successful regeneration after fire. Cone serotiny is a well-known fire adaptation in

Pinus spp. (Cayford 1970; Vogl 1970; Benzie 1977), whereby cones only open when

they are heated enough to melt the resin on the cone scales allowing seeds to drop

when the conditions are more conducive to regeneration. Some species are known to

increase their reproductive output after fire (Peters and Sala 2008, Haymes and Fox

2012). Thomas and Wein (1985) suggest the delayed seed emergence from the soil

and partial seed retention in tree crowns of jack pine (Pinus banksiana Lamb.)

increases the chances of seedling regeneration if the conditions are poor immediately

after a fire. Larger seed sizes, which are more likely to survive heat shocks, are

exhibited in some Pinus species (Escudero et al. 2000). Furthermore, Escudero et al.

(2000) found that Canary Island pine (Pinus canariensis) seedlings were larger if their

4

cones were heated prior to germination, and Hille and den Ouden (2004) found that

Scots pine (Pinus sylvestris L.) seedlings were significantly taller in recently burnt areas.

Some Pinus seedlings can survive complete defoliation from heat if it occurs before

their terminal buds open (Methven 1971). Moreover, many Pinus species occur in

areas frequented by fires, exhibit fire adaptations and regenerate better after fire,

suggesting fire is a crucial element to their persistence.

1.4 Fire Suppression

Despite the importance of fire to many Pinus species, wildfires have been suppressed in

many forest ecosystems (Thompson, 2000). Human activities and active suppression in

the 19th and 20th centuries have significantly altered many fire regimes with the effects

largely still not understood (Thompson, 2000). Suppression has led to an increase of

shade tolerant species that change light, moisture and soil dynamics (Carleton 2000;

Scheller et al. 2005). Fire suppression also causes fire frequency to decrease until the

fuel loads are well above traditional levels leading to more severe fires than forests are

adapted to, which may cause significant mortality in mature trees (Swezy and Agee

1991; Knight 1999; Van Sleeuwen 2006).

In the continued absence of fire, many pines are negatively affected by increased

canopy and understory competition as well as reduced access to mineral soil.

Competition increases at the understory level as herbaceous and woody species begin

to proliferate and create light and moisture competition for shade-intolerant pines

(Feeney et al. 1998; Parker et al. 2009). Woody competitors such as shrub species or

tree seedlings can reduce light exposure (0’Brien et al. 2007). Canopy competition

creates negative impacts through mortality of shade-intolerant seed trees (Welch et al.

5

2000; Goubitz et al. 2003) and decreased regeneration (O’Brien et al. 2007). Without

fire, forest succession occurs and the litter and duff layers increase quickly (Debano

1991; Arno et al. 1993; Covington and Moore 1994). The tap roots of germinating

seedlings of some pine species are unable to access the mineral soil through thick

layers of litter and duff, which prevents establishment (Little and Moore 1949). As a

result, shallow litter and duff layers are key components to successful germination for

some pine species (Little and Garrett 1990; Waldrop and Brose 1999). Soils with no

leaf litter not only increase seedling establishment, but also provide a more suitable

medium for seedling growth (Hille and den Ouden 2004; Jonsson et al. 2006).

1.5 Pinus rigida Miller

Pinus rigida Miller (pitch pine) is a predominantly fire-adapted coniferous tree found

throughout the eastern United States from Maine to Georgia reaching its northern range

limit in southern Ontario and Quebec in Canada (Gucker 2007; Figure 1). P. rigida is

often a dominant species in pine barrens throughout much of its range, where it can

sometimes be a climax species (Abrams and Orwig 1995). However, it is most often an

early successional species replaced by hardwoods in the absence of fire (Westveld et

al. 1956; Little 1974).

Like many other pines, P. rigida is negatively affected by both canopy and understory

competition and increased depth to mineral soil. Canopy competition can cause

mortality in seed trees and reduce light availability for seedlings (Witzke 1982;

Copenheaver et al. 2000; Welch et al. 2000). Understory competition is also known to

affect P. rigida seedlings by increasing competition, especially light competition (SLINP

1993; Witzke 1996; Gucker 2007). Increased duff and litter layers decrease access to

6

mineral soil and chances for successful regeneration (Bramble and Goddard 1942;

Ledig and Little 1979; Little and Garrett 1990; Srutek et al. 2008). Based on these

preferences it is not surprising that P. rigida has traits characteristic of fire-adapted

species.

Figure 1: North American distribution of Pinus rigida (Little 1971), with northern range including the Thousand Islands Ecosystem inset.

7

P. rigida often occurs in ecosystems with frequent fire (Gucker 2007) and has long been

considered to have strong fire resistance (Starker 1932) and regeneration associated

with fire (Buell and Cantlon 1953; Buchholz and Good 1982). P. rigida develops thick

bark that protects mature trees from fire (Little 1953). Another adaptation is epicormic

sprouting from dormant buds along the main trunk (Ledig and Little 1979). All of the

foliage of a pitch pine may be killed by the heat of a fire and the crown will still produce

new needles (Ledig and Little 1979). Seedlings are also able to recover from fires

through basal buds insulated in or against mineral soil (Little and Somes 1956) or by

initiating a new lead from the bole (Little 1959). Perhaps the most striking fire

adaptation of P. rigida is the presence of serotinous cones throughout the core of the

species’ range (Ledig and Little 1979).

Despite evidence for strong fire dependence in certain parts of P. rigida’s range, in other

parts the species may be less dependent on fire. The serotinous habit in P. rigida is

concentrated towards the center of the species’ range in the New Jersey Pine Plains,

with frequency of serotiny generally decreasing from there in all directions (Ledig and

Fryer 1972). Populations away from the center, including marginal populations such as

those found in Ontario, Canada, have no cone serotiny (Vander Kloet 1981). Fire

appears to be the major selective force determining the patterns of serotiny in P. Rigida

(Frasco and Good 1976; Givnish 1981; but see Ledig and Fryer 1972). If local fire

frequencies drive the evolution of serotiny at a site, then sites with low fire frequencies

will have little to no serotiny.

8

1.6 Thousand Islands Ecosystem Pitch Pine

The Ontario population, found at the northeastern range margin of the species, occurs

within the Thousand Islands Ecosystem (TIE). Most populations found in the TIE occur

on the Frontenac Arch, a narrow portion of Precambrian shield that connects the

Adirondack mountain range in New York State with the Canadian Shield in Canada

(Lynch 2008). Pitch pine in the TIE is found predominantly on rock outcrops and

populations are currently described as over-mature with little to no recruitment

(Ellsworth et al. 2009)

The relationship between TIE pitch pine populations and fire is not currently understood.

Since fire is considered the major selective force on cone serotiny and there is no

serotiny in the TIE, it is possible that fire does not maintain pitch pine in the region.

Although lack of serotiny suggests fire may not strongly influence pitch pine in the TIE,

this is partially contradicted by the presence of other fire adaptations such as thick bark

and epicormic sprouting. These adaptations suggest that fire has played a role in at

least some of the pitch pine stands in the TIE, however it is unclear to what extent fire is

necessary or important to the population. Since it is unknown if TIE pitch pine

populations are the leading edge of expanding populations or relicts from a previous

range contraction (Guries and Ledig 1982), it is possible that phenotypes that respond

favourably to fire have been selected against or that trees will exhibit phenotypic

plasticity after an increase in fire events.

The degree to which fire is important for TIE populations is largely unknown. The strong

fire association with P. rigida in other parts of its range and the fire adaptations

exhibited in the TIE suggest that fire may be important for TIE populations. There is

9

evidence of at least one pitch pine stand in the TIE regenerating after a stand-replacing

fire (Rogeau 2007) and another site demonstrated successful regeneration after a fire

(Srutek et al. 2008). Furthermore, the population has been significantly decreasing

since at least the 1970’s (Lynch 2008), a period in which there was strong fire

suppression in the area (SLINP 1993; Ellsworth et al. 2009). However, while P. rigida

is known to regenerate primarily after fires, it can also be an early successional pioneer

that invades open disturbed sites (SLINP 1993; Jordan et al. 2003). The lack of

serotinous pitch pine in the TIE is a concern when considering the use of fire as a

management tool for pitch pine regeneration (Witzke 1996). Additionally, Vander Kloet

(1981) observed that a fire completely destroyed a pitch pine population in the TIE.

Other authors have found that fire does not appear to play a role in the development of

pitch pine or other plant communities in the TIE (Vander Kloet 1981; Warner and

Marsters1993).

Developing an understanding of which disturbance(s) pitch pine respond to in the TIE is

imperative for conservation managers aiming to maintain the population. Most pitch

pine populations in the TIE are being outcompeted by hardwood species in the canopy

and by shade tolerant seedlings and shrub species in the understory (SLINP 1993;

Donnelly 1996; Lynch 2008). Populations are declining and exhibiting little to no

recruitment (Vander Kloet 1973; Witzke 1982; SLINP 1993; Lynch 2007). These

deteriorating conditions, combined with the fact that most populations in the TIE are too

close to residential areas to allow wildfire to burn free, suggest active management is

required.

10

1.7 Restoration Techniques

Prescribed fire and mechanical restoration approaches, especially canopy thinning and

soil scarification, are widely implemented for pine species (Covington et al. 1997;

Feeney et al. 1998; Karlsson and Orlander 2000; Hille and den Ouden 2004; Béland et

al. 2010). Prescribed fire (Welch et al. 2000) and canopy thinning (Béland et al. 2010)

are tools used to decrease both canopy and understory competition. Prescribed fire

(Heinselman 1981) and soil scarification (Karlsson and Orlander 2000; Béland et al.

2010) are used to reduce or eliminate the litter and duff layers to improve access to

mineral soil. Prescribed fire has been applied extensively to various P. rigida

populations in the United States since the 1930’s (Little and Moore 1949; Little 1964;

Welch et al. 2000). Canopy thinning has also been used, although less frequently than

fire, to both restore P. rigida and reduce fuel loads (Clark and Patterson 2003; APBPC

2010). No mechanical treatments and only one experimental prescribed fire (Srutek et

al. 2008) have been implemented in the TIE. Pitch pine populations in the TIE lack

serotiny and occur in unique forest communities with unique fuel types and micro-

climates (Witzke 1996). Consequently, pitch pine populations in the TIE may respond

differently than core populations do to treatments and warrant regional specific

approaches.

Understanding the most effective restoration technique to promote pitch pine

regeneration is imperative to halt the decline of the species in the TIE. My objective

was to determine what restoration technique will have the greatest effect on increasing

pitch pine seedlings in the TIE; specifically, to test the expectation that prescribed fire

will cause the greatest increase in pitch pine seedlings relative to a mechanical

treatment or control. To this end, I assessed pitch pine seedling recruitment at a

11

prescribed fire plot, a mechanical treatment plot and a control plot at each of two sites in

the TIE, using a Before-After Control-Impact (BACI) design.

2.0 Methods

2.1 Study Area

In order to determine what restoration technique was the most effective at promoting

pitch pine seedling establishment, I compared the results of two different treatments and

a control at two locations (hereafter referred to as “sites”) in the TIE; Georgina Island

(Georgina) and Mallorytown Landing (Mallorytown; Figure 2). Both sites were in similar

ecosites (Lee et al. 1998) to ensure comparable species compositions, tree sizes and

soil conditions. They were characterized by dry rocky ridges interspersed with dips

containing deeper soils. Georgina pitch pine are found in “Dry-Fresh Mixed Woodland

ecosites” and Mallorytown pitch pine stands are found in “Dry-fresh pitch pine

coniferous woodland ecosites” (note: this community is derived from the “Dry-fresh

white pine coniferous woodland ecosite, which was the closest fit in the manual; Lee et

al. 1998). Both sites were conifer dominated, mixed with deciduous species; however,

Mallorytown had a stronger maple (Acer spp.) presence whereas Georgina had a

stronger oak (Quercus spp.) presence. Georgina treatment plot sizes were 80x50m,

50x40m and 60x50m in the fire, mechanical and control plots respectively and

Mallorytown plot sizes were 40x40m, 30x40m and 40x40m in fire, mechanical and

control plots respectively.

Georgina became part of Thousand Islands National Park (TINP) in 1904 and has

remained protected ever since. Pitch pine trees at this site varied in diameter but were

12

all found to be between 88 and 102 years old, which suggests that there was a

disturbance in the early 1900’s that led to the formation of the current pitch pine stand

(Rogeau 2007). Charcoal found on site was determined to have been generated by a

fire in the 1850’s and it is possible that the pitch pines may have originated from a

different fire in the early 1900’s (Rogeau 2007). Mallorytown was used as pasture in the

early 1900’s until as late as the 1940’s and became a part of TINP in 1959 (Leggo,

personal communication) Since then the area has remained undisturbed and

succeeded into a natural state. Although the diameters of the pitch pine at Mallorytown

are similar to those at Georgina, their age is unknown. Although farming was practiced

in the area, the ridge on which the pitch pine occur would have been unsuitable for

crops and used primarily for grazing. As a result, it is likely the pitch pine were present

on the ridge during or prior to farming and they may be as old as the trees on Georgina.

13

Figure 2: Study sites at Georgina Island and Mallorytown Landing, where a prescribed fire, mechanical treatment and control plot were included at both sites.

2.2 Study Design

A Before-After Control-Impact (BACI) design was utilized to test the effect of different

treatments on the number of pitch pine seedlings. BACI designs are used in a wide

variety of impact studies such as mining (Condrey and Gelpi 2010), oil spills (Murphy et

al. 1997), wind farms (Dahl et al. 2012) and urbanization (Price et al. 2012), as well as

in restoration studies (Maccherini and Santi 2012, Boys et al. 2012, Bayne and Nielsen

2011, Dickson et al. 2009). The before- after component of the design involves looking

at environmental data in an area before and after an impact or treatment occurs. These

designs are not controlled experiments and do not control for natural changes in the

environment. However, the inclusion of a non-impacted or untreated control site allows

a test of the impact (in the case of my study; two impacts) while controlling for any

14

natural change that may have occurred (Stewart-Oaten et al. 1986; Smith 2002) and

better control for both spatial and temporal variation (Parker and Wiens 2005). In my

study, one prescribed fire treatment, one mechanical treatment and one control plot

were included at each site. Data were collected in all plots within one year before the

treatment and one year following the treatment.

Many BACI studies only include one or two study sites because impacts are not

planned, or treatments can only be applied in a few locations (due to logistical or

resource constraints; Smith 2002). Similarly, my study only involved two sites because

prescribed fires are expensive and logistically difficult to implement and because P.

rigida is a very rare tree with few stands in Canada (Vander Kloet 1973; Lynch 2008).

The fact that the control plots are in the same ecological communities and located

spatially close to the treatments provides support to that assumption. Additionally, my

use of environmental measures as covariates also lessens the sensitivity to this

assumption (Parker and Wiens 2005).

2.3 Treatment Types

Prescribed fires were hand ignited using drip torches in an “S” pattern across the entire

burn plot. Fires were ignited using prescriptions outlined by a Parks Canada Fire

Behaviour Analyst in part using the Field Guide to the Canadian Forest Fire Behaviour

Prediction System (FBP; Taylor et al. 1997). The prescription for Mallorytown was

slightly more moderate as a precaution for the presence of residential housing within

300m of the prescribed fire. Some canopy thinning took place at Mallorytown to

increase the fuel load to a more comparable level to that found at Georgina. This was

determined using the fuel load classifications outlined in the FBP and by

15

recommendations from a Parks Canada Fire Behaviour Analyst to ensure complete

burn coverage of the treatment plot. Ignition involved lighting the entire treatment area

and allowing the fire to self-extinguish where safely possible. The Georgina fire was

ignited in August of 2009 and Mallorytown was ignited in August of 2011. Although the

treatment areas were quite small, it is likely that a wildfire in the TIE would rarely be

larger than the patch of homogeneous fuel type it is burning and would be limited by the

lack of continuous surface fuel. Thus, it is not unusual for fire to be in the range of only

30x30m to 70x70m in size in the TIE (Rogeau 2007).

Mechanical treatments involved removing 85% of all competing canopy trees including

all size ranges down to trees 1m in height, while leaving all pitch pines standing.

Eighty-five percent was an estimate of the amount of canopy cover that might be

removed in a fire situation and also based on Barden and Woods (1976) who found that

fires that removed 85% of the basal area led to an increase in pitch pine regeneration.

The majority of trees were removed using chainsaws, however some large white pines

(Pinus strobus) were girdled and left standing as snags to provide wildlife habitat.

Shrubs and small trees less than 1m in height were cut as close to the base of the trunk

as possible using a Stihl KM 90R motor with a Stihl FS-KM metal blade attachment. All

woody debris from tree and shrub reductions were removed from the sites or dragged

into large piles to minimise the impact on the study site. The soil was then exposed by

blowing most debris from the surface using a Stihl KW-KM power sweep tool. Finally,

the entire site was scarified using a Stihl BF-KM soil cultivator to increase access to

mineral soil and attempt to remove remaining competing herbaceous and woody

species.

16

2.4 Sampling Design

All variables were measured in 2x2m quadrats placed every 10m along two parallel

transects in both treatments and the control at both sites within one year before and

after the treatments (Figure 3). On Georgina there were 10 2x2m quadrats in each

treatment and the control (transect length of 40m); at Mallorytown there were 8, 8, and 4

2x2m quadrats (transect length 30m, 30m and 20m respectively) in the fire, control and

mechanical plots respectively. There were only eight quadrats in the fire and control

plots at Mallorytown compared to 10 at Georgina due to the slightly smaller size of

Mallorytown. I was only able to collect data at four mechanical quadrats in Mallorytown

because some of the original quadrats were compromised after the mechanical

treatments were completed. As a result, the total sample size was 18 for the fire and

control and 14 for the mechanical. It is important to note that my study design assumes

independence between each of these quadrats. Ideally all the quadrats along a

transect or both transects at a site would have been averaged and used as individual

sample units. However, that would have required having more pitch pine stands than

were available and implementing more fire and mechanical treatments than were

operationally feasible. Quadrats were located along transects rather than being

randomly placed because a) the sites were small enough that these transects traversed

the majority of the site and b) they increased repeatability and decreased error in

subsequent surveys.

17

Figure 3: Transect and quadrat placement, replicated in each plot on Georgina Island (Mallorytown plots were smaller and consequently had shorter transects and fewer quadrats).

2.5 Variables Measured

The response variable and three covariates were measured in each 2x2m quadrat.

The response variable was the number of pitch pine seedlings and the covariates were

canopy cover, understory cover and depth to mineral soil.

To determine understory cover, the percent area covered in the quadrat by all

herbaceous plants and any woody plant less than 4cm in diameter at breast height

(DBH) were included. All plants were identified to species save grasses and sedges,

which were grouped together. The percentage of each species was then combined to

determine an overall percent cover. Percent cover exceeded 100% if plant cover was

thick and layered. For example, a grass species may cover 80% of the quadrat and

then a fern species may cover 30% of the quadrat above the grass, leading to 110%

18

total cover. Quadrats were surveyed in mid-May for spring ephemeral species and from

July – August to capture everything else.

To determine percent canopy cover, I stood at the center of each 2x2m quadrat and

estimated what percentage of the canopy was closed. I considered “closed canopy” to

be any area where the sky was obstructed. A quadrat with 80% canopy cover would

have only 20% of the sky visible from its centroid.

To determine the depth to mineral soil, I measured the distance between the forest floor

and mineral soil. The start of mineral soil was defined as the top of the A horizon (Lee

et al. 1998). Since soil depths can vary even within a site due to rocky outcrops and

dips and rises in the terrain, depth to mineral soil was measured in each of the 2x2m

quadrats. This was done using a soil auger to extract a soil core that was used to

measure the distance between the forest floor and where the mineral soil begins. The

Ecological Land Classification for Southern Ontario field guide (Lee et al. 1998) was

followed to identify different soil horizons.

2. 6 Statistical Methods

To evaluate the relative importance of treatment (fire and mechanical) on pitch pine

seedling establishment, I included three covariates (canopy cover, understory cover and

depth to mineral soil; Table 1) and compared 15 models in a multi-model inference

framework (MMI; Table 2). I analyzed my data using a two-factor poisson mixed model

analysis of covariance (ANCOVA) using R version 3.0.2. An ANCOVA was appropriate

because my response variable was continuous and my explanatory variables were both

continuous and categorical (Crawley 2007). Since each covariate is known to influence

pitch pine regeneration, an ANCOVA was ideal because it allowed me to test the effect

19

of both treatments on the response variable, while controlling for the effect of the

covariates (Quinn and Keough 2002). To facilitate the BACI design, time (before and

after) was included in each model as a categorical factor variable. As a result, each

model assesses the differences among treatments in their effect on the change in the

response variable from the before to after values.

Table 1: Complete list of variables and their measurement units.

Variable Type of variable Measurement

Response Continuous Number of pitch pine seedlings

Fixed factor Categorical Time

Fixed factor Categorical Treatment type

Covariate Continuous % Canopy cover

Covariate Continuous % Understory cover

Covariate Continuous Cm to mineral soil

Random factor Categorical Site

I included site as a random effect in the ANCOVA to control for any differences between

the two sites that might confound my results. This enabled me to test the more general

hypothesis about pitch pine stands in the TIE rather than a hypothesis about the two

sites in my study (Zuur et al. 2009). This is especially important for my BACI design

since I had only two sites to work with. With only two sites in my study, the potential

impact of confounding effects due to unmeasured site differences increased. Including

site as a random variable is further supported because it was a) not something I was

explicitly testing for and b) it was something I expected to influence the variance of the

data not the mean (Quinn and Keough 2002).

I ranked models in the MMI framework based on their Akaike Information Criterion (AIC)

values (i.e. AIC <2.0) and calculated the associated model weights (wi; Burnham and

20

Anderson 2002). To determine which variable(s) were most strongly supported in

determining pitch pine seedling establishment, I calculated model-averaged estimates

of the coefficients for each of the variables in the top models. One of the advantages of

this approach is that it allowed me to measure and reflect model selection uncertainty.

That advantage was particularly important in my study, due to my study’s small sample

size. Using a global model with small sample size may lead to low statistical power

because the model may be too complex for the available degrees of freedom to support

and often results in problems with overinterpreting limited data. The use of AIC allows

assessment of model selection uncertainty and the objective determination of which

model, with which combination of variables, to base conclusions on. A limitation of the

design is that if more than one model is selected on the basis of strong AIC values,

effects exclusive to different models may in fact be driven by correlation with effects not

included in each model. Under these circumstances, an additional post-hoc ANOVA

can partition the competing effects and their interaction.

Table 2: The model combinations selected against in a MMI framework, analyzed using a two-factor poisson mixed model ANCOVA where T=treatment, C=canopy cover, U=understory cover and S=depth to mineral soil.

T+C+U+S T+C+U T+C+S T+U+S T+C T+U T+S C+U+S C+U C+S U+S T C U S

21

3.0 Results

3.1 General Responses of Vegetation Community and Soil Depth to Treatments

There was a variable response by the covariates known to influence P. rigida

regeneration one year after the treatments. Percent canopy cover significantly

decreased following both the fire (df=95, p=<0.0001) and mechanical treatments (df=95,

p=<0.0001; Figure 4). Percent understory cover was not significantly affected by either

the fire (df=95, p=0.362) or mechanical (df=95, p=0.093) treatments (Figure 5). Finally,

depth to mineral soil was significantly decreased in the fire treatment (df=95, p=0.013)

but not the mechanical treatment (df=95, p=0.149; Figure 6).

Figure 4: Canopy cover in A (control), B (fire treatment) and C (mechanical treatment) before and after treatments. ***=p<0.001 using ANCOVA.

22

Figure 5: Understory cover in A (control), B (fire treatment) and C (mechanical treatment) before and after treatments.

Figure 6: Depth to mineral soil in A (control), B (fire treatment) and C (mechanical treatment) before and after treatments. *=p<0.05 using ANCOVA.

23

3.2 Model Selection

The treatment (fire and mechanical) and depth to mineral soil models had the greatest

support in determining an increase in pitch pine regeneration in the initial ANCOVA

analysis (Table 3). Treatment had an Akaike weight (wi) of 0.509 and depth to mineral

soil was 0.331, combined accounting for 84% of the total wi. All other models, including

any with canopy and understory cover included were >4AIC from the top model. The

model-averaged partial regression coefficients were 0.202, 0.065 and -0.012 for the fire

treatment, mechanical treatment and depth to mineral soil respectively (where both

treatments and a decrease in the depth to mineral soil had a positive effect on pitch pine

seedling numbers; Table 4). Fire and depth to mineral soil were the only two variables

whose standard error did not overlap with 0.

Table 3: AIC model selection statistics of ANCOVA models where: T=treatment, S=depth to mineral soil, C=canopy cover and U=understory cover. Response variable is number of pitch pine seedlings.

Model AIC i wi

T 166.769 0.000 0.509

S 167.631 0.861 0.331

C 170.772 4.003 0.069

TS 171.525 4.756 0.047

CS 172.192 5.423 0.034

TC 175.365 8.596 0.007

U 178.768 11.998 0.001

TU 179.178 12.409 0.001

TCS 179.588 12.818 0.001

US 180.315 13.546 0.001

CU 183.569 16.800 0.000

TUS 183.872 17.103 0.000

CUS 184.556 17.787 0.000

TCU 187.329 20.560 0.000

TCUS 191.501 24.731 0.000

24

Table 4: Model-averaged estimates of the variable coefficients and standard error, calculated using both top models.

Parameter SE df p-value

Fire 0.202 0.072 95 0.006

Mechanical 0.065 0.077 95 0.401

Depth to Mineral Soil -0.012 0.004 96 0.006

Since both the treatment and depth to mineral soil models demonstrated good

predictive ability in determining pitch pine seedling regeneration in the initial analysis, a

post-hoc test was completed to test if there was an interaction effect between the two

models. In the post-hoc interaction test I found there was an interaction effect between

treatment and depth to mineral soil. The model containing treatment and depth to

mineral soil as well as the interaction between the two had the highest r2 value (Table

5). Only the fire treatment has a regression coefficient that does not overlap with zero

in this model. The interaction effect between fire and soil seems to potentially confound

the relative importance of both parameters in determining pitch pine seedling

regeneration. When the interaction is not controlled for in the model, the regression

coefficients are fairly close (see hash line in Figure 7). However, when the interaction is

included in the model, the regression coefficients for both fire and soil deviate from each

other and increase from their original values (see solid line in Figure 7).

25

Table 5: values for each parameter in the original treatment + depth to mineral soil tests and post hoc interaction tests, where F=fire treatment, M=mechanical treatment, T= treatment, S = depth to mineral soil and “:” represents an interaction test. r squared values are provided as indicators of the amount of variation explained by each model.

Model Parameter r2 SE df p-value

S+T F 0.269 0.191 0.12 94 0.028

M 0.143 0.126 94 0.262

S -0.024 0.011 94 0.033

S+T+S:T F 1.16 0.318 0.322 92 0.001

M 0.171 0.305 92 0.576

S -0.001 0.026 92 0.98

S:F -0.125 0.037 92 0.001

S:M -0.006 0.03 92 0.848

S:T S:F -0.038 0.174 0.019 94 0.049

S:M -0.026 0.011 94 0.015

Figure 7: Change in relative importance of fire compared to depth to mineral soil (soil) for an ANCOVA model that includes a treatment*soil interaction term versus a model that does not. refers to the partial regression parameter for each ANCOVA model.

26

4.0 Discussion

4.1 Treatment Effectiveness

The results of my study support my prediction that prescribed fire would cause the

greatest increase in pitch pine seedlings compared to the mechanical treatment or

control. In the initial model selection process the treatment model had the greatest

predictive ability on increasing pitch pine seedlings. Within the treatment model, the fire

treatment regression coefficient was three times that of the mechanical treatment.

Since both regression coefficients are measured on the same scale, I can infer that the

fire treatment is expected to produce three times more pitch pine seedlings than the

mechanical treatment. It was further demonstrated that the fire treatment was more

effective than the mechanical treatment, when the interaction between depth to mineral

soil and treatment was assessed in the post-hoc test. In addition to the fire treatment

having the greatest support in the model selection process, it is also important to note

the consistency with which fire had a positive effect on pitch pine regeneration.

Regardless of which variables were included, fire was consistently positively associated

with pitch pine regeneration. Although the efficacy of fire to increase pitch pine

regeneration is well supported in parts of its range (Little and Moore 1949; Buchholz

and Good 1982; Groeschl et al. 1993; Welch et al. 2000), this study represents only the

second and third prescribed fires for P. rigida in Canada and is the first study to

determine that fire is more effective than a mechanical disturbance technique.

Moreover, my results suggest that fire may be more effective than mechanical

disturbance at increasing P. rigida regeneration in ways other than altering habitat

conditions such as canopy / understory cover and depth to mineral soil.

27

The mechanical treatment was ineffective at promoting pitch pine regeneration. There

are few published studies that compare pitch pine regeneration after prescribed fire with

a canopy thinning approach. Furthermore, no mechanical treatments of any kind have

been attempted to increase pitch pine regeneration in the TIE. Thinning approaches

are frequently used to reduce fuel loads in P. rigida stands elsewhere in the species’

range (Clark and Patterson 2003). Additionally, thinning in parts of P. rigida’s range has

led to large, sometimes over-abundant, increases in regeneration (Patterson, personal

communication). However, these mechanical treatments occurred in different

communities than pitch pine in the TIE and the canopy thinning involved removing pitch

pine canopy, as opposed to removing competing canopy trees such as in this study.

Whereas thinning has been effective in increasing P. rigida regeneration in other parts

of the species range, I found it to be ineffective in the TIE.

4.2 Treatment Effects on Competition and Access to Mineral Soil

The fire and mechanical treatments had a similar effect on canopy and understory

cover. As expected, both the mechanical and fire treatment significantly decreased

canopy cover (see fire treatment photos in Figure A1 and A2). Thus there was an

increase in light penetration from the canopy cover reaching the understory. This is

consistent with other studies that found both fire and mechanical thinning decreased

canopy cover (Fulé et al. 2012). What was somewhat unexpected was the lack of effect

both treatments had on decreasing the percent of understory cover. Although non-

significant, the results suggested even a slight increase in understory cover in the fire

treatment (Figure A1 and A2). There was a window after the treatments were

implemented where understory cover was significantly reduced, which could provide an

opportunity for pitch pine seedlings to become established (see fire treatment photos in

28

Figure A1 and A2). However, within one year both herbaceous and shrub species

emerged from existing seedbanks or rhizomes and also likely arrived via seed dispersal

to replace the percentage of understory cover removed by the treatments. Understory

species richness increased remarkably from 28 to 73, 25 to 55 and 35 to 37 species

after the fire treatment, mechanical treatment and control respectively. Moreover, while

the percent understory cover remained relatively constant in both treatments, the

community composition was considerably different with 49 and 35 new species found

after the fire and mechanical treatments that were not present before.

In contrast to the similar effect both treatments had on canopy and understory cover,

depth to mineral soil was effectively decreased by the fire treatment but not the

mechanical treatment. In many locations fire smouldered and burnt through the duff

and litter layers exposing bare mineral soil, ideal for P. rigida seedling germination and

growth (Little and Moore 1952). The mechanical treatment scarified the soil and in

places decreased the depth to mineral soil, but overall was ineffective at increasing

access to mineral soil for pitch pine seedlings. This is possibly because the treatment

did not remove the litter or duff layers, rather mixed them in with some mineral soil tilled

up from underneath. This was the only variable that fire significantly improved that the

mechanical treatment did not. It is possible that this variable may have been

responsible for the fire treatment effectively promoting pitch pine regeneration better

than the mechanical treatment. Considering the test that included the interaction

between treatment and depth to mineral soil had the highest r2 value, fire and depth to

mineral soil may be an important combination to increase pitch pine regeneration.

However, considering the standard error is very large for the depth to mineral soil

29

regression coefficient in this model it is uncertain to what degree this variable is

influencing pitch pine regeneration. As a result, it is also possible that fire could be

modifying the importance of depth to mineral soil to pitch pine regeneration and that

something else about fire is driving improved pitch pine regeneration in the fire

treatments.

4.3 Possible Explanations of the Positive Effects of fire on Seedling Recruitment

The post fire treatment soil chemistry may have improved conditions for pitch pine

generation more effectively than did the mechanical treatment. Fires have been

documented to change a variety of soil properties including physical, physio-chemical,

mineralogical, chemical and biological (Certini 2005). There are also examples of

thinning treatments influencing soil properties in P. rigida stands (Hwang and Son 2006;

Geng et al. 2012). However, there is evidence that fire treatments create favourable

soil conditions for increasing P. rigida recruitment. Groeschl et al. (1993) found that fire

led to significant reductions in nutrient content and an increase in P. rigida recruitment.

They hypothesized that fire degrades site quality and sets back site succession, which

favours P. rigida. Other studies have found that P. rigida seedlings grew significantly

better in the absence of leaf litter (Jonsson et al. 2006) or that abundant leaf litter

reduced seedling growth (Garnett et al. 2004). This may have led to increased seedling

growth in the fire treatments, since fire burns off leaf litter whereas the mechanical

treatments simply tilled existing leaf litter into the soil. Moreover, fire may have created

conditions more conducive to pitch pine seedling establishment and growth and

unfavourable to competing species, than the mechanical treatment.

30

Another possibility is that fire actually promoted a response from mature pitch pine trees

that increased reproductive output and / or efficiency. Some studies have found that

other Pinus species increase cone output after fire (Haymes and Fox 2012), thinning

(González-Ochoa et al. 2004) and after combined fire / thinning treatments (De Las

Heras et al. 2007; Peters and Sala 2008). In the present study on P. rigida, the

maximum number of cones would have been determined prior to the treatments in the

form of fully developed cones (already developed before the treatment) or fertilized

female strobili (which develop into cones the year following the treatment). However,

Peters and Sala (2008) found that there were a higher number of filled seeds per cone

in ponderosa pine after fire compared with after thinning treatments. Other studies have

found that Pinus seedlings had higher initial growth rates after cones were heated prior

to germination (Escudero et al. 2000), or in plots that were burned compared to either a

control site or sites that only had soil scarification (Hille and den Ouden 2004).

Considering understory competition returned to similar pre-treatment levels, it is

possible that smaller seedlings may have been outcompeted and experienced higher

mortality in the mechanical plots. If, in the present study, seed trees in the fire

treatment released a greater number of seeds and / or had seedlings that grew larger,

this could at least partially explain the difference in recruitment between treatments.

4.4 Possible Phenotypic Responses to Fire

My results provide an example of the importance of fire in a population at its range

margin that has experienced lower recent fire frequency and that exhibits fewer fire-

adapted traits than populations in the core. This suggests that it is possible that

populations occurring in areas of lower fire frequency, which originated in areas of high

fire frequency, may respond favourably or even exhibit phenotypic plasticity that

31

provides them with an advantage when a fire does occur. If increased regeneration in

the fire treatment was due to improved reproductive response by pitch pine (increased

filled seeds per cone, increased germination rates, increased seedling growth), TIE

populations may be exhibiting some degree of adaptive phenotypic plasticity.

Cone serotiny provides an opportunity to investigate this possibility. Since fire is

believed to be the major selective force determining the gradient of cone serotiny in P.

rigida and recent fire appears to be less frequent in populations at the range margin, it is

not surprising that populations such as those found in the TIE are not serotinous. It is

not known if the absence of strict serotiny in pitch pine in the TIE is due to selection

against serotiny, or whether this trait persists in the population but the trait is not

expressed in the absence of frequent fire. In fact, even within the high fire frequency

core of the species, there are populations that include many individuals without

serotinous cones. It has been observed that fewer trees exhibit serotiny in non-burnt

clearings, compared to undisturbed adjacent plains which experience fire (Givnish 1981;

Jordan et al. 2003). Indeed, in the absence of frequent fire, many populations in the TIE

may have been colonized by seeds from non-serotinous cones moving into recently

disturbed land. However, it is unknown if trees with non-serotinous cones produce

some serotinous offspring and there is still little known about the genetics of these traits

(Givnish 1981; Jordan et al. 2003). This is not to say that pitch pine in the fire

treatments will now exhibit serotinous cones. Rather I am suggesting that it is possible

that they may have exhibited other fire adapted traits beneficial to establishing seedlings

that were triggered by the fire. If so, it is possible that disjunct populations removed

from regular fire events may retain a level of plasticity that enables them to take

32

advantage of the conditions created when fire returns. These explanations are of

course speculative, and would be worthwhile subjects of future study.

4.5 Restoration Considerations and Further Studies

To further investigate the explanations for the greater effectiveness of fire at increasing

regeneration than mechanical treatments, additional studies are recommended. Soil

chemistry before and after fire and mechanical treatments should be tested to

determine if it is influencing regeneration. If soil chemistry is a key element contributing

to increased regeneration, it would limit the ability of mechanical restoration to increase

natural recruitment.

The possibility that TIE populations of pitch pine increase their reproductive effort and

success after fire but not following mechanical treatments warrants further study. If

confirmed, it would provide a useful insight for conservation managers in the TIE to

improve local pitch pine restoration efforts. It would also be important information for

use in the management of other P. rigida populations outside the high fire frequency

core and for other species found in similar situations worldwide. Studies should focus

on adaptations that promote reproductive effort and success exhibited in other Pinus

species, as discussed (increased serotiny, cone production, seeds per cone,

germination success and seedling growth).

It is important to recognise that my study only included treatment years, and

observations taken during the first year immediately following each treatment. As a

result, the focus is on which treatment created conditions suitable for initial seed

germination and early seedling survival. Although successful regeneration in the first

year after the treatment is likely a good indication that the conditions created are

33

suitable to regenerating the stand, additional years of data collection would be

beneficial. At the Georgina site, seedling numbers have been recorded for an additional

two years beyond the initial year after the treatments in three 10x10m plots per

treatment. Although this represents a small sample size, pitch pine seedlings increased

each year after the fire treatment while not increasing in the mechanical or control

(Figure 8). This implies that, additional years would further support the results of this

study since conditions in the fire treatments clearly created suitable conditions for

recruitment and produced significant regeneration, while mechanical treatments

demonstrated very low recruitment.

Figure 8: Pitch pine seedling numbers in 10x10m plots in the three years post treatment.

Perhaps the strongest indication that fire treatments will continue to be more effective at

regenerating pitch pine compared to the mechanical treatment is that competition may

remain in-check for longer in the fire treatment. Although competing hardwood species

did sprout from root masses of trees in both treatment types, it was considerably more

34

noticeable in the mechanical treatments. I observed limited sprouting in the fire

treatments, where it appears the only hardwoods that were able to sprout were in small

pockets that did not burn very intensely. Overall the majority of hardwoods such as

Quercus rubra, Acer saccharum, Ostrya virginiana, Carya ovata and shrubs such as

Amelanchier arborea and especially Vaccinium spp., exhibited little resprouting in fire

treatments. In contrast, the same species were observed sprouting prolifically the year

after in the mechanical treatments. Rigorous sprouting from hardwoods will increase

understory cover that is significant competition for young shade intolerant seedlings

(Gucker 2007) and eventually, canopy competition for those seedlings that do persist

into trees (Copenheaver et al. 2000). Concordantly, it is likely that the mechanical

treatments will facilitate more hostile conditions for young pitch pines in years following

the treatment than prescribed fires.

Even though mechanical treatments were not effective in promoting pitch pine

regeneration and may facilitate greater competition for young seedlings, it is important

to consider if mechanical treatments could be improved, given that fire treatments are

not always possible. Where the application of prescribed fire is not practical,

mechanical treatments are needed to replicate wild or prescribed fire (Jordan et al.

2003). There are numerous cases where prescribed fire may not be possible, including

proximity to residential areas, private ownership, lack of financial resources or lack of

local fire expertise. In these cases, mechanical approaches may be the only option to

restore pitch pine stands if there is continuing adult pitch pine mortality with no

recruitment. Although mechanical restoration plots were not effective in promoting

regeneration, some seedlings were observed. Considering virtually no regeneration is

35

observed naturally in the TIE, this suggests that mechanical treatments may have some

potential as a restoration approach.

To improve the likelihood of mechanical treatment success, a few options should be

considered. One option that may make mechanical treatments effective is post

treatment seedling plantings. Canopy and understory cover were similarly affected in

both treatments (Figures 4 and 5), yet seedlings germinated and survived in the fire

treatment. It is possible then, that mechanical treatments are unable to provide

conditions suitable to seed germination, young seedling survival or lack seed input.

Consequently, planted seedlings could by-pass any seed germination difficulties or

seed input deficiencies and grow in similar competition conditions to natural recruitment

in the prescribed fire sites. Roughly 100 seedlings planted in two mechanical

treatments in the TIE have demonstrated a high survivorship over a two year post-

treatment period (Van Wieren unpublished). Because thinned hardwoods will likely

necessary. This approach may only work if the limiting factors in the mechanical plot

can be explained by lack seed input, germination or reduced early (<1 year) seedling

growth and not by other factors limiting seedling survival. However, the fact that P.

rigida can be an early pioneer species and have successfully established in non-fire

sites (Jordan et al. 2003), suggests that an investigation of the success of planting pitch

pine seedlings in mechanical restoration plots would be worthwhile. If natural

recruitment is a goal in mechanical restoration sites, experimenting with additional soil

scarification techniques may be beneficial. The approach used to increase access to

mineral soil in this study proved largely ineffective. Mechanical treatments that include

removing the duff layer may improve natural generation in some cases. As discussed in

36

the planting option, continued mechanical thinning would likely be necessary to avoid

seedlings getting outcompeted in subsequent years.

Although determining the fire regime in TIE and other marginal populations of P. rigida

may be difficult (Rogeau 2007), understanding the historical disturbance regime is

important to help guide land management (Cissel et al. 1999). Disturbed ecosystems

are known to have thresholds, which once crossed result in a substantial increase in the

amount of restoration required (Hobbs and Norton 1996; Hobbs and Harris 2001). In

this case it is possible that the natural disturbance cycle was disrupted for so long that it

passed a threshold and fires may be more severe than the forest has historically

experienced. When fire cycles are disrupted through fire suppression, duff layers and

fuel loads increase and fires tend to be more severe and can destroy many adult trees

(Van Sleeuwen 2006). This can be significant for P. rigida, since more intense fires

may cause extreme smoldering which can disrupt basal sprouting (Windisch 1999).

Moreover, since P. rigida is known to lose some of its fire adaptations as it ages

(Windisch 1999), increased periods between fires may reduce the ability of P. rigida

stands to respond favourably to fire. In three 10x10m plots at the Georgina site, 80%

(n=29) of the adult trees (DBH > 4cm) were killed within one year of the fire and 93%

had died two years after the fire. Even though remaining seed trees generated

successful recruitment, it is possible that with lower adult tree mortality rates,

regeneration may have been even better. However, Srutek et al. (2008) found that a

less intense spring fire in the TIE was ineffective at opening up understory and canopy

conditions. Thus, there appears to be a trade-off between treating sites with prescribed

fires severe enough to destroy shade tolerant hardwoods and the root systems of

37

competing shrubs, while retaining some pitch pine seed trees. Conservation managers

should strive to determine the ideal fire frequency to achieve optimum seedling

recruitment.

5.0 Conclusions I found that prescribed fire was more effective than mechanical disturbance at

increasing P. rigida seedling regeneration at the species’ extreme northeast range

margin. I also found that both fire and mechanical disturbance similarly affected canopy

cover and understory cover, and that while only fire decreased depth to mineral soil,

that fire modified the importance of depth to mineral soil on pitch pine regeneration. My

results imply that fire has a significant effect on pitch pine seedling regeneration beyond

a reduction in canopy cover and possibly beyond or in conjunction with depth to mineral

soil. P. rigida at the core of its range exhibits strong fire adaptations that increase

seedling regeneration after fires, whereas TIE populations exhibit fewer adaptations.

Consequently, the effectiveness of fire to increase seedling regeneration compared to

mechanical disturbance demonstrated in my study suggests it is possible that pitch pine

in the TIE have retained fire adapted traits beneficial to establishing seedlings that are

triggered by fire events. This further underscores the importance of researching and

understanding disturbance regimes before conducting wide-scale restoration of a

species or population. My results suggest that prescribed fire should be the primary

conservation management tool to restore declining pitch pine populations in TIE (and

potentially at other locations at the edge of its range distribution). Where prescribed fire

is not feasible, I recommend that additional research to determine why mechanical

38

treatments are failing should be considered and mechanical treatments should be

modified based on the results.

39

References Abrams, M.D., 1992. Fire and the development of oak forests. BioScience 42, 346-

353. Abrams M.D. and Orwig D.A., 1995. Structure, radial growth dynamics and recent

climatic variations of a 320-year-old Pinus rigida rock outcrop community. Oecologia 101, 353-360.

APBPC, 2010. Management Plan and Final Environmental Impact Statement. . Albany Pine Bush Preserve Commission.

Arno, S.F., Reinhardt, E., Scott, J., 1993. Forest structure and landscape patterns in the subalpine lodgepole pine type: a procedure for quantifying past and present conditions. Gen. Tech. Rep. INT-294. Ogden, UT: U.S. Department of Agriculture, Forest Service, Intermountain Research Station. 73pp.

Attiwill, P.M., 1994. The disturbance of forest ecosystems: the ecological basis for conservative management. Forest Ecology and management 63, 247-300.

Baker, W.L., 1992. Effects of settlement and fire suppression on landscape structure. Ecology 73, 1879-1887.

Barden, L.S., Woods, F.W., 1976. Effects of fire on pine and pine-hardwood forests in the southern Appalachians. Forest Science 22, 399-403.

Bayne E., Nielson B., 2011. Temporal trends in bird abundance in response to thinning of lodgepole pine (Pinus contorta). Canadian Journal of Forest Research 41, 1917-1927.

Béland, M., Calmels, S., Bergeron, Y., 2010. Natural regeneration of jack pine on clay soils following harvesting and site preparation: Seventh-year results. The Forestry Chronicle 86, 423-433.

Benzie, J.W., 1977. Managers Handbook for Jack Pine in the North Central States. USDA Forest Service, St. Paul, Minn., General Technical Report. NC-32.

Bergeron, Y., Brisson, J., 1990. Fire regime in red pine stands at the northern limit of the species range. Ecology 71,1352-1364.

Bormann, F.H., Likens G.E., 1979. Catastrophic disturbances and the steady state in northern hardwood forests. American Scientist 67, 660-669.

Boys C.A., Kroon F.J., Glasby T.M., Wilkinson K., 2012. Improved fish and crustacean passage in tidal creeks following floodgate remediation. Journal of Applied Ecology 49, 223-233.

Bramble W.C., Goddard M.K., 1942. Effect of animal coaction and seedbed condition on regeneration of pitch pine in the Barrens of central Pennsylvania. Ecology. 23, 330-335.

Buchholz, K., Good, R.E., 1982. Density, age structure, biomass and net annual aboveground productivity of dwarfed Pinus rigida Mill. from the New Jersey Pine Barren Plains. Bulletin of the Torrey Botanical Club, p. 24-34.

Buell, M.F., Cantlon, J.E., 1953. Effects of prescribed burning on ground cover in the New Jersey pine region. Ecology 34, 520-528.

Burnham, K.P., Anderson, D.R., 2002. Model selection and multi-model inference: a practical information-theoretic approach. Springer.

Cain D., Shelton G., 2000. Survival and growth of Pinus echinata and Quercus

40

seedlings in response to simulated summer and winter prescribed burns. Canadian Journal of Forest Research 30, 1830−1836.

Carleton, T.J., 2000. Vegetation responses to the managed forest landscape of central and northern Ontario. Pp. 178-197 in Perara, A.,D. Euler, D.L., Thompson I., (eds.). Ecology of a managed terrestrial landscape: patterns and processes of forest landscapes in Ontario. The University of British Columbia Press, Vancouver, BC.

Cayford, J.H., 1970. The role of fire in the ecology and silviculture of jack pine. Proceedings of the Annual Tall Timbers Fire Ecology Conference 10, 221-244.

Certini, G., 2005. Effects of fire on properties of forest soils: a review. Oecologia, 143, 1-10.

Cissel, J.H., Swanson, F.J., Weisberg, P.J., 1999. Landscape management using historical fire regimes: Blue River, Oregon. Ecological applications 9, 1217-1231.

Clark, K.H. and Patterson III W.A. 2003. Fire Management Plan for Montague Plain Wildlife Management Area. Department of Natural Resources Conservation, University of Massachusetts, Amherst, Massachusetts. 48pp.

Condrey, R.E., Gelpi, C.G., 2010. Blue crab (Callinectes sapidus) use of the Ship/Trinity/Tiger Shoal Complex as a nationally important spawning/hatching/foraging ground: Discovery, evaluation, and sand mining recommendations based on blue crab, shrimp, and spotted seatrout findings. U.S. Dept. of the Interior, Minerals Management Service, Gulf of Mexico OCS Region, New Orleans, LA. OCS Study MMS 2009-043. 42pp.

Copenheaver, C.A., White, A.S., Patterson, W.A., 2000. Vegetation development in a southern Maine pitch pine-scrub oak barren. Journal of the Torrey Botanical Society 127, 19-32.

Covington, W.W., Fule, P.Z., Moore, M.M., Hart, S.C., Kolb, T.E., Mast, J., Sackett, S., Wagner, M., 1997. Restoring ecosystem health in ponderosa pine forests of the Southwest. Journal of Forestry 95, 23-29.

Covington, W.W., Moore, M.M., 1994. Postsettlement changes in natural fire regimes and forest structure: ecological restoration of old-growth ponderosa pine forests. In: Sampson, R.N., Adams, D.L., (eds.). Assessing the forest ecosystem health in the Inland West. Journal of Sustainable Forestry 2, 153–181.

Dahl, E.L., Bevanger K., Nygard T., Røskaft E., Stokke B.G., 2012. Reduced breeding success in white-tailed eagles at Smøla windfarm, western Norway, is caused by mortality and displacement. Biological Conservation 145, 79-85.

Davis, R.J., Wilkinson, J., Heaman, D., 2003. Fire management efforts in Ontario’s protected areas: a synopsis. Pp. 421-427 in Lemieux, C.J., Nelson, J.G., Parks and Research Form of Ontario.

DeBano, L., 1991. The effect of fire on soil properties. In: Proceedings—management and productivity of western montane forest soils. Gen. Tech. Rep. INT-280. Ogden, UT: U.S. Department of Agriculture, Forest Service, Intermountain Research Station, pp151–156.

De Las Heras, J., Moya, D., López-Serrano, F., Condés, S., 2007. Reproduction of postfire Pinus halepensis Mill. stands six years after silvicultural treatments. Annals of forest Science 64, 59-66.

Dickson, B.G., Noon, B.R., Flather, C.H., Jentsch, S., Block, W.M., 2009. Quantifying

41

the multi-scale response of avifauna to prescribed fire experiments in the Southwest United States. Ecological Applications 19, 608-621.

Ellsworth, K., Bedford, D., Kafka, V., 2009. St. Lawrence Islands National Park of Canada: a framework for the restoration of pitch pine. Unpublished report.

Escudero, A., Nunez ,Y., Perez-Garcia, F., 2000. Is fire a selective force of seed size in pine species? Acta Oecologica 214, 245−256.

Feeney, S.R., Kolb, T.E, Covington, W.W., Wagner, M.R., 1998. Influence of thinning and burning restoration treatments on presettlement ponderosa pines at the Gus Pearson Natural Area. Canadian Journal of Forest Research 28, 1295-1306.

Frasco, B.R., Good, R.E., 1976. Cone, seed, and germination characteristics of pitch pine (Pinus rigida Mill.) from the pine barrens of New Jersey. Bartonia 44, 50-57.

Fulé, P.Z., Crouse, J.E., Roccaforte, J.P., Kalies, E.L., 2012. Do thinning and/or burning treatments in western USA ponderosa or Jeffrey pine-dominated forests help restore natural fire behavior?. Forest Ecology and Management 269, 68-81.

Garnett, E., Jonsson, L.M., Dighton, J., Murnen, K., 2004. Control of pitch pine seed germination and initial growth exerted by leaf litters and polyphenolic compounds. Biology and fertility of soils 40, 421-426.

Geng, Y., Dighton, J., Gray, D,. 2012. The effects of thinning and soil disturbance on enzyme activities under pitch pine soil in New Jersey Pinelands. Applied Soil Ecology 62, 1-7.

Givnish, T.J., 1981. Serotiny, geography, and fire in the Pine Barrens of New Jersey. Evolution 35, 101-123.

González-Ochoa, A.I., López-Serrano, F.R., de las Heras, J., 2004. Does post-fire forest management increase tree growth and cone production in Pinus halepensis? Forest Ecology and Management 188, 235-247.

Goubitz, S., Werger, M.J.A., Ne’eman, G., 2003. Germination response to fire- related factors of seeds from non-serotinous and serotinous cones. Plant Ecology 169, 195-204.

Groeschl, D.A., Johnson J.E., Smith D.W., 1993. Wildfire effects on forest floor and surface soil in a table mountain pine-pitch pine forest. International Journal of Wildland Fire 3, 149-154.

Grumbine, R.E., 1994. What is ecosystem management? Conservation biology 8, 27-38.

Gucker, C.L., 2007. Pinus rigida. In: Fire Effects Information System [Online]. U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fire Sciences Laboratory (Producer). Available: http://www.fs.fed.us/database/feis/

Guries, R.P., Ledig, F.T., 1982. Genetic diversity and population structure in pitch pine (pinus rigida Mill.). Evolution 36, 387-402.

Haymes, K.L., Fox, G.A., 2012. Variation among individuals in cone production in Pinus palustris (Pinaceae). American Journal of Botany 99, 640-645.

Heinselman, M.L. 1981. Fire and succession in the conifer forests of northern North America. Pp. 374-405 in West, D.C., Shugart, H.H., Botkin, D.B., (eds.). Forest Succession: Concepts and Application. Springer Verlag, New York, NY.

Hille, M., den Ouden J., 2004. Improved recruitment and early growth of Scots pine

42

(Pinus sylvestris L.) seedlings after fire and soil scarification. European Journal of Forest Research 123, 213-218.

Hobbs, R.J., Harris, J.A., 2001. Restoration ecology: repairing the earth's ecosystems in the new millennium. Restoration ecology 9, 239-246.

Hobbs, R.J., Norton, D.A., 1996. Towards a conceptual framework for restoration ecology. Restoration ecology 4, 93-110.

Hwang, J., Son, Y., 2006. Short-term effects of thinning and liming on forest soils of pitch pine and Japanese larch plantations in central Korea. Ecological research 21, 671-680.

Jonsson, L.M., Dighton, J., Lussenhop, J., Koide, R.T., 2006. The effect of mixing ground leaf litters to soil on the development of pitch pine ectomycorrhizal and soil arthropod communities in natural soil microcosm systems. Soil Biology and Biochemistry 38, 134-144.

Jordan, M.J., Patterson, W.A., Windisch, A.G., 2003. Conceptual ecological models for the Long Island pitch pine barrens: implications for managing rare plant communities. Forest Ecology and Management 185, 151-168.

Karlsson, C., and Orlander, G., 2000. Soil scarification shortly before a rich seed fall improves seedling establishment in seed tree stands of Pinus sylvestris. Scandinavian Journal of Forest Research. Pp. 256-266.

Keeley, J.E., Pausas, J.G., Rundel, P.W., Bond, W.J., Bradstock, R.A., 2011. Fire as an evolutionary pressure shaping plant traits. Trends in plant science 16, 406-411.

Keeley, J., Zedler, P., 1998. Life history evolution in pines. In: Richardson, D.M., (ed.), Ecology and biogeography of Pinus. Cambridge University Press, pp. 219−242.

Klaus, W., 1989. Mediterranean pines and their history. Plant Systematics and Evolution 162, 133-163.

Knight, D.H., 1999. Ponderosa and limber pine woodlands. In: Anderson R.C, Fralish J.S. and Baskin JM (ed.), Savannas, barrens, and rock outcrop plant communities of North America. Cambridge University Press, pp. 249-261.

Ledig, F.T., Fryer, J.H., 1972. A pocket of variability in Pinus rigida. Evolution 26, 259- 266.

Ledig, F.T. and Little, S., 1979. Pitch pine (Pinus rigida Mill.): ecology, physiology and genetics. In Forman, R.T.T., (ed.) Pine Barrens: Ecosystem and Landscapes. Pp. 347-371. New York: Academic Press.

Lee, H., Bakowsky, W., Riley, J., Bowles, J., Puddister, M., Uhlig, P., McMurray, S., 1998. Ecological land classification for southern Ontario: First Approximation and its Application. Ontario Ministry of Natural Resources, Southcentral Science Section, Science Development and Transfer Branch.

Little, E.L., Jr., 1971. Atlas of United States trees, volume 1, conifers and important hardwoods: U.S. Department of Agriculture Miscellaneous Publication 1146, 9 p., 200 maps.

Little, S., Moore, E.B., 1949. The ecological role of prescribed burns in the pine-oak forests of southern New Jersey. Ecology 30, 223-233.

Little, S., Moore, E.B., 1952. Mechanical preparation of seedbeds for converting oak-pine stands to pine. Journal of Forestry 50, 840-844.

Little, S., 1953. Prescribed burning as a tool of forest management in the northeastern

43

states. Journal of Forestry 51, 496-500. Little, S., 1959. Silvical characteristics of pitch pine. Northeastern Forest Experimental

Station, Upper Darby, Pennsylvania. Station paper 110, pp22. Little, S., 1964. Fire ecology and forest management in the New Jersey Pine Region. In:

Proceedings Third Annual Tall Timbers Fire Ecology Conference. Komarek, E.V. Sr., Chairman. New York: Tall Timbers Research Station. Pp. 34-59.

Little, S., 1974. Effects of fire on temperate forests: northeastern United States. In: Kozlowski, T.T., Ahlgren, C.E., (eds.). Fire and ecosystems. New York: Academic Press, Pp. 225-250.

Little, S., P.W., Garrett., 1990. Pinus rigida Mill. Pitch pine. Pp. 456-462 in R.M. Burns and B.H. Honkala, eds., Silvis of North America. Volume 1. Conifers. Agriculture handbook no. 654, U.S. Department of Agriculture, Forest Service, Washington, D.C.

Little, S., Somes, H.A., 1956. Buds enable pitch and shortleaf pines to recover from injury. U.S. Forestry Service, Northeast Forest Experimental Station. Station paper No. 81.

Lynch, M.B., 2008. Pitch Pine Assessment in St. Lawrence Islands National Park: 2008 Report. Unpublished report. St. Lawrence Islands National Park.

Maccherini, S., Santi, E., 2012. Long-term experimental restoration in a calcareous grassland: Identifying the most effective restoration strategies. Biological Conservation 146: 123-135.

McRae, D.J., Lynham, T.J., Frech, R.J., 1994. Understory prescribed burning in red pine and white pine. The Forestry Chronicle 4, 395-401.

Methven, I.R., 1971. Prescribed fire, crown scorch and mortality: field and laboratory studies on red and white pine. Pettawawa Forest Experiment Station, Chalk River, Ontario. Information Report PS-X-31. Canadian Forestry Service.

Murphy, S.M., Day, R.H., Wiens, J.A., Parker, K.R., 1997. Effects of the Exxon Valdez oil spill on birds: comparisons of pre- and post-spill surveys in Prince William Sound, Alaska. Condor 99:299–313.

Naveh, Z., 1975. The evolutionary significance of fire in the Mediterranean region. Vegetatio 29, 199-208.

Ne'eman, G., Goubitz, S., Nathan, R., 2004. Reproductive traits of Pinus halepensis in the light of fire–a critical review. Plant Ecology 171, 69-79.

Norton, B., 1992. A new paradigm for environmental management. Ecosystem Health. Constanza, R., Norton, B.G., Haskell, B.D., (eds.). Island Press, Washington D.C. Pp 23-41.

O’Brien, M.J., O’Hara, K.L., Erbilgin, N., Wood, D.L., 2007. Overstory and shrub effects on natural regeneration processes in native Pinus radiata stands. Forest Ecology and Management 240, 178-185.

Parker, W.C., Pitt, D.G., Morneault, A.E., 2009. Influence of woody and herbaceous competition on microclimate and growth of eastern white pine (Pinus stobus L.) seedlings planed in a central Ontario clearcut. Forest Ecology and Management 258, 2013-2025.

Parker, K.R., Wiens, J.A., 2005. Assessing Recovery Following Environmental Accidents: Environmental Variation, Ecological Assumption and Strategies. Ecological Applications 15, 2037-2051.

44

Parsons, D.J., DeBenedetti, S.H., 1979. Impact of fire suppression on a mixed- conifer forest. Forest Ecology and Management 2, 21-33.

PCA. 2013. http://www.pc.gc.ca/progs/np-pn/ie-ei.aspx. Viewed on November 4, 2013. Peters, G., Sala, A., 2008. Reproductive output of ponderosa pine in response to

thinning and prescribed burning in western Montana. Canadian Journal of Forest Research 38, 844-850.

Price, S.J., Browne, R.A., Dorcas, M.E., 2012. Evaluating the effects of urbanization on salamander abundances using a before-after control-impact design. Freshwater Biology 57: 193-203.

Quinn, G.G.P., Keough, M.J., 2002. Experimental design and data analysis for biologists. Cambridge University Press.

Rogeau, M.P., 2007. Preliminary Fire Regime Study: St. Lawrence Islands National Park of Canada. Thousand Islands National Park unpublished report.

Rogers, P., 1996. Disturbance ecology and forest management: a review of the literature. Ogden, UT: US Department of Agriculture, Forest Service, Intermountain Research Station.

SCBD, 2001. Impact of human-caused fires on biodiversity and ecosystem functioning and their causes in tropical, temperate and boreal forest biomes. Report no. 5, Secretariat of the Convention on Biologycal Diversity, Montreal.

Scheller, R.M., D.J. Mladenoff, T.R. Crow T.A., Sickley. 2005. Simulating the effects of fire reintroduction versus continued fire absence of forest composition and landscape structure in the Boundary Waters Canoe Area, Northern Minnesota, USA. Ecosystems 8, 396-411.

SLINP, 1993. Pitch pine population study and management plan. Prepared by Ecologistics for St. Lawrence Islands National Park. Unpublished report.

Smith, E.P., 2002. BACI Design. Encyclopedia of Environmentrics 1, 141-148. Srutek, M., Dolezal, J., Findlay, C.S., Andress, R.A., 2008. Regeneration and

seedling-habitat relationships of the marginal population of pitch pine (Pinus rigida) after prescribed burning, Eastern Ontario, Canada. Natural Areas Journal 28, 155-167.

Starker, T.J., 1932. Fire resistance of trees of northeast United States. Forest Worker 8, 8-9.

Stewart-Oaten, A., Murdoch, W.W., Parker, K.R., 1986. Environmental impact assessment: ‘pseudoreplication’ in time? Ecology 67, 929–940.

Swezy, D.M., Agee, J.K., 1991. Prescribed-fire effefects on fine root and tree mortality in old-growth ponderosa pine. Canadian Journal of Forest Research 21, 626-634.

Taylor, S.W., Pike, R.G., Alexander, M.E., 1997. Field Guide to the Canadian Forest Fire Behavior Prediction (FBP) System. Fire Management Network, Canadian Forest Service, Northern Forestry Centre.

Thomas, P.A., Wein, R.W., 1985. Delayed emergence of four conifer species on postfire seedbeds in eastern Canada. Canadian Journal of Forest Research 15, 727-729.

Vander Kloet, S.P., 1973. The biological status of pitch pine, Pinus rigida in Ontario and adjacent New York. Canadian Field Naturalist 87, 249-253.

Vander Kloet, S.P., 1981. Satus of Pinus rigida in Canada (with special reference to St.

45

Lawrence Islands National Park. St. Lawrence Islands National Park. Unpublished report.

Van Sleeuwen, M., 2006. Natural fire regimes in Ontario. Ontario Ministry of Natural Resources, Queen’s Printer for Ontario, Toronto. 143pp.

Van Wagner, C.E., 1970. Fire and red pine. Pp. 211-219 in Proceedings of the 10th Tall Timbers fire ecology conference, 20-21 August 1970. Tall Timbers Research Station, Tallahassee, FL.

Vogl, R.J., 1970. Fire and the Northern Wisconsin pine barrens. Proceedings of the Tall Timbers Fire Ecology Conference 10, 175-209.

Waldrop, T.A., Brose, P.H., 1999. A comparison of fire intensity levels for stand replacement of table mountain pine (Pinus pungens Lamb). Forest Ecology and Management 163, 85-92.

Warner, B.G., Marsters, S.G., 1993. A Paleoenvironmental investigation of the pine ridge wetland, St. Lawrence Islands National Park, Ontario. Wetlands Research Centre, University of Waterloo.

Welch, N.T., Waldrop, T.A., Buckner, E.R., 2000. Response of southern Appalachian table mountain pine (Pinus pungens) and (P. rigida) stands to prescribed burning. Forest Ecology and Management 136, 185-197.

Westveld, M., Ashman, R.I., Baldwin, H.I., Holdsworth, R.P., Johnson, R.S., Lambert, J.H., Lutz, H.J., Swain, L, Standish, M., 1956. Natural forest vegetation zones of New England. Journal of Forestry 54, 332-338.

Whitmore, T.C., 1982. On pattern and process in forests. In: E.I. Newman (Editor), The Plant Community as a Working Mechanism. Blackwell Scientific Publications, Oxford, pp. 45-59.

Whitney, G.G., 1986. Relation of Michigan’s pre-settlement pine forests to substrate and disturbance history. Ecology 67, 1548-1559.

Windisch, A.G., 1999. Fire ecology of the New Jersey pine plains and vicinity. New Brunswick, NJ: Rutgers, The State University of New Jersey. 327pp. Dissertation.

Witzke, C.R., 1982. A preliminary report on the status of Ontario pitch pine (Pinus rigida Miller). Unpublished report, Ontario Ministry of Natural Resources, 60pp.

Witzke, C., 1996. Pitch pine (Pinus rigida Miller) genetics, reproductions and silvicultural management. A literature review and annotated bibliography. Eastern Ontario Model Forest. Information Report No. 20.

Wolf, J., 2004. A 200-year fire history in a remnant oak savanna in southeastern Wisconsin. American Midland Naturalist 152, 201-213.

Wright, H.A., Bailey, A.W., 1980. Fire ecology and prescribed burning in the Great Plains: a research review. U. S. Department of Agriculture, Forest Service, Intermountain Forest and Range Experiment Station, Ogden, Utah. Gen. Tech. Rep. INT-77. 60pp.

Zuur, A.F., Ieno, E.N., Walker, N.J., Saveliev, A.A., Smith, G.M., 2009. Mixed Effects Models and Extensions in Ecology with R. Springer, New York, NY, USA.

46

Appendix

Figure A1: From top to bottom, Georgina Island prescribed fire treatment one year before the fire, one day after the fire and one year after the fire.

47

Figure A2: From top to bottom, Mallorytown Landing prescribed fire treatment one year before the fire, two and half months after the fire and one year after the fire.