CAN SHORT-ROTATION HARVESTS INCREASE MANAGEMENT OPTIONS FOR THE ENDANGERED KIRTLAND’S WARBLER? By Daphna Gadoth-Goodman A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Forestry-Master of Science 2017 ABSTRACT CAN SHORT-ROTATION HARVESTS INCREASE MANAGEMENT OPTIONS FOR THE ENDANGERED KIRTLAND’S WARBLER? By Daphna Gadoth-Goodman Since the early 1980’s, 1550 ha of high-density jack pine (Pinus banksiana) plantations have been established annually in Northern Lower Michigan to serve as habitat for the federallyendangered Kirtland’s warbler (Setophaga kirtlandii). Because these plantations do not produce merchantable sawlogs by their planned 50-year harvest age, I investigated the potential to implement reduced rotation lengths in these stands to produce alternative wood products, namely biomass and pulpwood. I used space-for-time substitution to assess biomass and volume accrual over time, sampling a total of 37 warbler plantations ranging from 7 to 52 years of age. I also destructively sampled 26 living and 8 dead stems to develop allometric equations specific to jack pine grown in these plantations. Potential maximum biomass was estimated to be ~71 Mg ha-1 and potential maximum volume was estimated to be ~71 m3 ha-1. The predicted optimal rotation age for biomass was 20 years and the predicted optimal rotation age for volume was 28 years. I calculated and compared the total land area required for management under these rotation scenarios to continue establishing 1550 ha of habitat annually. Management on the current 50year cycle requires ~77,500 ha. Management for volume would reduce this to ~43,400 ha and management for biomass would require ~31,000 ha. My results suggest that rotation lengths in these plantations could be significantly reduced, allowing for reductions in the total land area dedicated to warbler habitat, allowing for management diversification at the landscape level. ACKNOWLEDGEMENTS I would like to express my deepest appreciation of Dr. David Rothstein for all the mentorship and support he provided me throughout this project. I sincerely thank Dr. Michael Walters and Dr. Kyla Dahlin for their helpful comments on my project proposal and written work. I thank Randy Klevickas, Paul Bloese, Daniel Brown for their time and assistance with data processing. I thank Ashlee Lehner, Robert Froese, Daniel Fell, Kenny Fanelli, Matt Gedritis and Trevor Kubitskey for assistance with field work. I thank Tim Greco, Jason Hartman, and Keith Kintigh of the Michigan DNR for their advice and assistance with all aspects of this project. I gratefully acknowledge USDA National Institute of Food and Agriculture (AFRI Project MICL-08505 and McIntire Stennis Project MICL-06006) for financial support of this project. iii TABLE OF CONTENTS LIST OF TABLES…………………………………………………………………………….......v LIST OF FIGURES………………………………………………………………………………vi CHAPTER 1: Impacts of Rotation Length on Biodiversity and Climate Change Mitigation in Forested Ecosystems………………………………………………………1 1. Introduction……………………………………………………………………………1 CHAPTER 2: Can Short-Rotation Harvests Increase Management Options for the Endangered Kirtland’s Warbler?.....................................................................................4 1. Introduction……………………………………………………………………………4 1.1 Objectives……………………………………………………………………8 2. Methods………………………………………………………………………………..9 2.1 Study area description………………………………………………………..9 2.2 Allometric equation development…………………………………………..10 2.3 Stand inventory……………………………………………………………..13 2.4 Volume estimation………………………………………………………….14 2.5 Production over time………………………………………………………..15 2.6 Covariate analyses…………….……………………………………………16 2.7 MAI and optimal rotation ages……………………………………………..17 2.8 Regional impact assessment………………………………………………..17 3. Results………………………………………………………………………………..19 3.1 Biomass allometrics………………………………………………………...19 3.2 Production over time………………………………………………………..21 3.3 Covariate analyses………………………………………………………….24 3.4 Optimal rotation ages……………………………………………………….24 3.5 Regional impact assessment………………………………………………..25 4. Discussion……………………………………………………………………………28 4.1 Overview……………………………………………………………………28 4.2 Ecological implications……………………………………………………..29 4.3 Economic implications……………………………………………………...31 4.3.1 Overview of economic implications……………………………...31 4.3.2 Local economic opportunity……………………………………...32 4.4 Climate-related implications………………………………………………..34 4.5 Uncertainty and implications for future research…………………………...35 5. Conclusions…………………………………………………………………………..40 APPENDIX………………………………………………………………………………………41 REFERENCES…………………………………………………………………………………..44 iv LIST OF TABLES Table 1. Comparison of plot-level biomass estimate outputs from the Highplains-specific equation, Presque Isle-specific equation, and the combined allometric equation that does not account for effect of region. Differences between region-specific equation outputs and the combined equation outputs are expressed as percentages……………………………………………………………………………...20 Table 2. Statistical outputs of covariate regressions plotted against biomass curve residuals………………………………………………………………………………...24 Table 3. Statistical outputs of covariate regressions plotted against volume curve residuals…...24 Table 4. Potential ecosystem service outputs from 1550 ha of managed KW habitat over 100 years…………………………………………………………………………………….26 Table 5. Plot-level data…………………………………………………………………………..42 v LIST OF FIGURES Figure 1. Fit plots of lnBiomass (kg) against lnDBH (cm) produced by linear regression analyses of (A) live stem data and (B) standing dead stem data. Log-transformed biomass and diameter data show strong positive linear relationships between the two variables. Solid lines represent the linear regression lines described in the text. Shaded areas around these lines represent the 95% confidence limits, and dashed lines represent 95% prediction limits…………………………………………..20 Figure 2. Aboveground biomass content as a function of stand age. Biomass accounts for both living and standing dead trees. Symbols represent stand means (±1 SE), and the solid curve represents the nonlinear regression line described in the text. Dashed lines represent the upper and lower 95% confidence limits across the age range of the dataset…………………………………………………………………….22 Figure 3. Volume content as a function of stand age. Volume is estimated for all living stems with a minimum 2.44 m length to a 10.16 cm top. Symbols represent stand means (±1 SE), and the solid curve represents the nonlinear regression line described in the text. Dashed lines represent the upper and lower 95% confidence limits across the age range of the dataset………………………………….23 Figure 4. Derivatives of (A) biomass and (B) volume growth curves plotted against their respective MAI curves over time. Solid curves represent the growth curve derivatives and dashed curves represent MAI over time. Vertical dotted lines represent the stand age at which culmination of MAI occurs, where MAI equals the derivative of the growth function…………………………………………………..25 Figure 5. Locations of existing wood-based electric power plants and publicly-owned KW plantations in the northern region of Michigan’s Lower Peninsula. Black dots represent power plant locations and shaded gray areas represent areas under KW management. The circles represent 50 km radii around each power plant……………33 vi CHAPTER 1: Impacts of Rotation Length on Biodiversity and Climate Change Mitigation in Forested Ecosystems 1. Introduction Anthropogenic alterations of land use have been the dominant driver of biodiversity loss in terrestrial ecosystems over the last 50 years, and are projected to continue to be major drivers of global biodiversity change over the coming decades (MEA 2005). Rapidly-increasing demands for food, water, timber, fiber and fuel have led to mass land use conversion, largely from natural systems to agriculture or other forms of production management. In forest systems, demands for wood-based commodities and biofuels are addressed in the form of maximization of harvest yields via the implementation of production-based management strategies (Roberge et al. 2016). In the face of global climate change, priorities for forest management have expanded beyond traditional provisioning services to include regulating services related to climate change mitigation and conservation of biological diversity. Simultaneous management for multiple ecosystem services has proven to be a major challenge in forest management, as strong trade-offs exist between strategies for optimization of production, climate change mitigation, and biodiversity (Triviño et al. 2017). A primary silvicultural strategy used to meet these objectives involves the extension or reduction of rotation length, defined as the time elapsed between successive final harvests (Felton et al. 2017). Extending rotation lengths is widely considered to be compatible with promotion of biodiversity conservation (Lindenmayer et al. 2006). Long rotations are associated with increased habitat availability, as they provide structural complexity and microhabitats in the form of large diameter trees, snags, and coarse woody debris in addition to compositional heterogeneity (Lindenmayer et al. 2006; Roberge et al. 2016). Extending rotations can also prove to be a 1 beneficial strategy for climate change mitigation by increasing carbon sequestration (Sohngen and Brown 2008). Alternatively, reducing rotation lengths, a common strategy of commodity-oriented forestry, is typically associated with negative impacts on ecosystem biodiversity (Felton et al. 2015). Strategies that prioritize maximizing economic returns often call for harvesting well before structural and compositional complexity can develop (Franklin et al. 2007). Furthermore, large-scale implementation of commercial forest management can cause a shift to younger landscape-level stand age distributions which favor pioneer shade-intolerant species, and is widely viewed to be inconsistent with aims to emulate historically heterogeneous natural disturbance regimes (Roberge et al. 2016). However, there is a gap in the current literature regarding the potential benefits of reducing rotations in landscapes where biodiversity is limited by a lack of young stands dominated by shade-intolerant species. Jack pine (Pinus banksiana) forests in the northeastern region of Michigan’s Lower Peninsula were historically dominated by frequent, stand-replacing wildfires on a return interval of ca. 60 years (Cleland et al. 2004). Fire suppression efforts in the 20th century led to widespread habitat loss for a diversity of early-successional species adapted to these frequent disturbances, driving one species, the Kirtland’s warbler (Setophaga kirtlandii) to near extinction (MDNR 2014). These birds, which depend on continuous, even-aged jack pine stands of at least 32 ha for breeding habitat, cease to nest in these stands approximately 23 years after establishment (Meyer 2010). In an effort to recover this species, public land agencies have been planting new jack pine stands on an annual basis since 1981 to provide a continuous supply of early-successional habitat on the landscape (MDNR 2014). This strategy has had enormous benefits for the warbler, 2 restoring their population to a size more than double the original recovery goal. These habitat plantations, which currently encompass approximately 77,000 ha, were intended to be managed on a 50-year rotation, a typical rotation length for production of jack pine sawtimber (Byelich et al. 1985). However, it is becoming apparent that these stands will not reach merchantable sawlog size by their planned 50-year rotation, and management agencies are becoming concerned over the long-term economic sustainability of continuing management into the future. Furthermore, implementation of this plan has drastically altered the distribution of mature forest stands on the landscape, resulting in a shift to a younger, more homogenized stand age distribution, which may be detrimental to species dependent on later-successional habitat (Tucker et al. 2016). This study aims to analyze the potential economic and ecological benefits of reducing rotation lengths in a portion of habitat plantations, and reducing the total area under Kirtland’s warbler management, to simultaneously manage for increased wood production, biodiversity, and climate change mitigation services at the landscape level. 3 CHAPTER 2: Can Short-Rotation Harvests Increase Management Options for the Endangered Kirtland’s Warbler? 1. Introduction In the face of global climate change, there is much interest in shifting from traditional forest silvicultural practices to alternative management strategies that utilize forest resources to enhance climate change mitigation services (Chum et al. 2011). Production forests provide climate benefits in the form of carbon storage and sequestration (Zanchi et al. 2014). Implementation of production management for bioenergy can broaden these benefits to include reductions in greenhouse gas emissions via co-generation and provisioning of a sustainable, renewable energy source (Vass 2017). The implementation of climate change adaptation and mitigation strategies (CCAMS) in production forests can directly and indirectly impact the provisioning of other ecosystem services and may involve alterations in rotation length, species composition, and harvest removals (Felton et al. 2015; Immerzeel et al. 2014). Of specific concern is the potential impacts these CCAMS may have on other ecosystem services, such as biodiversity, at both the stand and landscape levels. The extent and nature of these impacts vary system to system and are highly dependent on the climate, natural disturbance dynamics, tree species composition, and historical and current land-use of the ecosystem at hand (Felton et al. 2015). Several studies have identified a direct link between altering rotation lengths and impacts on biodiversity (Felton et al. 2017; Lindenmayer et al. 2006). Extending rotation lengths to increase carbon storage and sequestration is one of the few CCAMS commonly viewed to be compatible with biodiversity goals (Felton et al. 2015). Shifting to longer rotations can increase habitat availability through the provisioning of key habitat structural features such as coarse 4 woody debris and old, large diameter trees (Felton et al. 2017; Chum et al. 2011; Lindenmayer et al. 2006). Alternatively, reducing rotations to produce wood-based bioenergy and mitigate climate associated risks is a strategy commonly considered to be incompatible with biodiversity goals (Felton et al. 2015; Lindenmayer et al. 2006). Forests under short-rotation management are characterized by simplified structural and compositional features, and therefore, widespread implementation of this strategy can lead to increased landscape homogeneity (Spaulding and Rothstein 2009; Berch et al. 2011). However, there is a lack of scientific literature that directly addresses the potential impacts of reducing rotations on biodiversity in landscapes where species of conservational concern are disturbance-adapted and whose populations are limited by a lack of available earlysuccessional habitat. The key to minimizing negative impacts associated with shortened rotation lengths lies in adherence to the principles of ecological forestry, which emphasize the implementation of management strategies that emulate natural disturbance regimes and stand development processes (Franklin et al. 2007). Thus, reducing rotation lengths in forest ecosystems naturally adapted to frequent, stand-replacing disturbances could, in some cases, prove to be both ecologically and economically beneficial (Tarr et al. 2017). In the Lake States region, jack pine (Pinus banksiana) forest systems are adapted to such a disturbance regime. This fast-growing species occurs in even-aged stands historically perpetuated by frequent stand-replacing fires on an average return interval of ca. 60 years (Cleland et al. 2004). The stands provide critical habitat to a variety of species of conservation concern, including the federally endangered Kirtland’s warbler (Setophaga kirtlandii), hereafter referred to as KW (MDNR 2014). This migratory bird is endemic to the region during the summer months and occupies young jack pine stands between the ages of 5 and 23 years that are 5 a minimum 32 ha in size (Meyer 2010). Fire suppression efforts during the early- to mid- 1900’s greatly reduced the amount of early-successional habitat available to KW, and drove its population to near-extinction by the mid-1970’s, with record low population levels of 167 singling males recorded in 1974 and 1987 (MDNR 2014). Efforts to create habitat for KW began as early as 1957 on state lands and 1962 on federal lands (Mayfield 1963; Radtke and Byelich 1963). In 1981, public agencies including the Michigan Department of Natural Resources (MDNR), the United States Forest Service (USFS), and the United States Fish and Wildlife Service (USFWS), established an expanded habitat management program to ensure sufficient breeding habitat for KW population recovery (Kepler et al. 1996). Under this plan, approximately 77,000 ha have been designated as KW management areas, with agencies establishing more than 1500 ha annually to provide a continuous supply of early-successional habitat (MDNR 2014). To mimic the historical structure of jack pine stands maintained by wildfire, which are characterized by a mosaic of dense thickets and scattered openings, managed habitat plantations are planted at high stocking densities (~1.5 m x 1.8 m spacing) in an ‘opposing wave’ pattern that incorporates unplanted gaps to provide structural diversity and foraging opportunities for the bird. These gaps account for approximately 1/5 of total habitat land area. The conservation efforts of this program have been overwhelmingly successful, with the KW population reaching an all-time high in 2015, representing more than a 10-fold increase in population size since its record low levels (MDNR 2014). The current population size is more than double the original goal set out by the recovery plan of 1,000 mating pairs. Under the guidelines of the KW recovery program, these stands are managed on a 45- to 50-year commercial harvest rotation based on the notion that they would provide habitat during 6 earlier stages of stand development, and be of merchantable size for commercial cutting at harvest age (Byelich et al. 1985). However, as those stands established at the onset of recovery efforts begin to reach their 50-year rotation mark, land managers are becoming increasingly concerned over the marketability of stems produced in these habitat plantations. It appears that the extremely dense stocking of these stands is causing growth suppression of individual stems, and in the absence of pre-commercial thinnings they are highly unlikely to produce marketable sawlogs by age 50. Therefore, management for nontraditional wood products, such as pulpwood or biomass, on reduced rotations could contribute to increased financial returns and climate change mitigation benefits, while providing critical endangered species habitat and supporting biodiversity conservation. This system represents a unique opportunity to simultaneously manage for forest products and biodiversity, while imposing minimal negative impacts on conservation efforts that typically coincide with maximizing production yields. For one, implementation of short-rotation production management in this system would not require any land use conversion, a primary driver of global declines in biodiversity (Chum et al. 2011), and would continue to provide breeding habitat for the endangered KW. It is commonly assumed that biomass plantations are established on surplus agricultural land with favorable production conditions (Chum et al. 2011). However, jack pine stands of Northern Lower Michigan occur on acidic, sandy outwash soils of low fertility (Werlein 1998) and thus, KW management for biomass would not conflict with alternative land use interests, such as food production. Additionally, at the stand level, the harvesting methods and planting structure of these habitat plantations largely adhere to the principles of ecological forestry in that they are designed to emulate the disturbance patterns and growth structure of wildfire-originated stands in the region (Franklin et al. 2007; MDNR 2014). 7 The implementation of short-rotation management requires the ability to quantify the amount of aboveground biomass and pulpwood that is produced in these stands over time, such that optimal rotation lengths for maximum yields can be identified. 1.1 Objectives My overall objective was to characterize growth of jack pine in KW plantations in the northeastern region of the Lower Peninsula of Michigan, USA to better predict rates of production over time and to provide agencies with vital information that can be integrated into management decisions, as adaptive management is a primary goal of the KW Breeding Range Conservation Plan (MDNR 2014). My specific objectives were: 1) To estimate harvestable biomass and pulpwood volumes at different stages of stand development following whole-tree harvesting. 2) To determine optimal rotation lengths for biomass and pulpwood production yields and compare them to the current 50-year rotation. 3) To assess the potential impacts of alternate rotation lengths in KW plantations on biomass, volume, and KW habitat provisioning over the coming decades. 8 2. Methods 2.1 Study area description All study sites were located within the KW management areas of northeastern Lower Michigan, USA. The historical disturbance regime of this region was dominated by standreplacing wildfires on a return interval of ca. 60 years, a result of the landscape’s exceedingly dry conditions, relatively level topography, and flammable vegetation (Cleland et al. 2004). KW plantations in the area consist of even-aged, monoculture jack pine plantings interspersed with a minimal component of volunteer hardwood species, primarily Quercus ellipsoidalis and Prunus serotina. Within this area, I designated three geographic regions of study to analyze whether variations in production could be attributed to variations in soils and climate. The regions I selected represent three distinct subsections defined in the Ecosystem Classification of the State of Michigan by Albert (1995), and were labeled accordingly. The Highplains region is characterized by excessively drained sandy soils and a predominantly flat topography, with an extreme frost danger persisting throughout its short growing season (80-120 days). The Arenac region has a growing season ranging from 120-140 days and a flat to gently sloping topography. The third region, Presque Isle, is characterized by drumlins separated by areas of outwash sands and gravels. The growing season for this region ranges from 100-130 days (Albert 1995). In 2015, I established a single chronosequence within each of the three regions to determine whether they were characterized by differences in productivity. Because new KW plantations are established annually, I was able to sample from several plantations (hereafter referred to as ‘stands’) across a spectrum of ages within each region to assess changes in productivity over time. The lack of variation in climate, topography, soil characteristics, species 9 composition, and planting density among stands within a given region provided for an ideal scenario with which to compare differences in growth characteristics across various stages of stand development. The number of stands I selected for each chronosequence was dependent on the overall prevalence of KW plantations existing within the respective region. I selected a total of nine stands from the Highplains region, eight stands from Presque Isle, and seven stands from Arenac. Stands selected from Highplains were aged 8, 11, 15, 18, 22, 28, 35, 41, and 52 years. Stands selected from Presque Isle were aged 7, 10, 13, 16, 20, 23, 30, and 32 years, while those selected from Arenac were aged 10, 13, 19, 22, 28, 32, and 35 years. These initial chronosequences were unreplicated, with the primary intent being to understand the dynamics of stand development over time. In 2016, I sampled an additional 13 stands from the Highplains region to better understand variability in production within two age classes, 17-24 and 31-34 years, and across the two major soil series supporting jack pine forests in this region, Graycalm and Grayling sands (Werlein 1998). 2.2 Allometric equation development To estimate biomass as a function of stand age, I first developed my own local allometric equations predicting biomass of individual stems as a function of diameter at breast height (DBH). Because the geography, climate, and silvicultural practices of the KW management system are distinct from traditional jack pine systems, I opted to develop my own allometric equations, as opposed to adopting pre-existing equations from the literature. For one, these studies were largely conducted in areas geographically distinct from Lower Michigan, such as 10 Canada and Minnesota. Since the Lower Peninsula of Michigan marks the southern limit of jack pine’s natural growth range (Rudolph 1985), I hypothesized that the extreme climate and suboptimal growth conditions of this region would negatively impact the species’ production rates. This, coupled with the fact that KW plantations are planted at a higher density than is practiced in traditional plantations managed for timber, 1.5 x 1.8 m vs. 1.8 x 2.4 m (MDNR 2014; Benzie 1977), could contribute to suppressed growth of individual stems. Therefore, it was imperative that I develop my own allometric equations specific to jack pine grown in this system to achieve reliable growth estimates. To develop an initial equation predicting biomass from stem diameter, in 2015 I destructively sampled a total of 26 living stems from 14 stands within two of the selected regions, Highplains and Presque Isle (7 stands and 13 stems each). The 13 trees sampled from the Highplains region ranged in DBH from 0.7 to 22.9 cm (0.7, 2.5, 2.7, 5.0, 6.5, 7.6, 8.5, 9.5, 10.7, 12.2, 12.8, 17.7, 22.9), whereas the 13 stems harvested from Presque Isle ranged in DBH from 2.0 to 21.5 cm (2.0, 3.3, 4.5, 5.0, 5.8, 7.6, 8.2, 10.1, 11.4, 12.3, 16.5, 17.5, 21.5). Initially, I did not collect destructive samples from the Arenac region because stands in this region fall under USFS jurisdiction, and attaining permission to harvest there was more difficult. Once I observed no significant differences between the Highplains and Presque Isle regions, I decided not to harvest from the Arenac region altogether (see Section 3.1 for details). Within each stand selected for destructive sampling, I harvested a large stem and an average-sized stem. In 2016, I destructively sampled 8 dead stems from 3 stands within the Highplains region, aged 20, 28, and 41 years, to develop an allometric equation specific to standing dead trees, predicting biomass from diameter. The DBH’s for harvested dead stems ranged from 2.4 to 18.4 cm (2.4, 4.3, 5.9, 7.2, 8.1, 11.3, 13.2, 18.4). 11 Each tree was felled and harvested stems were cut into 1.22 m vertical sections, aside from the lowest section, which measured the length between the height of the stump and breast height (1.37 m from the ground). I then separated the branches from each vertical bole section and determined the fresh mass of bole and branches from each section using a portable field scale. I then collected a subsample of representative branches and a 3-6 cm thick stem disc from the bole of each vertical section and recorded their fresh weights in the field. All bole and branch subsamples were returned to the lab and dried in a forced-air oven at 65°C before recording their dry weights. I determined the dry mass proportions of each subsample and applied them to the total branch and bole fresh weights of each respective section to achieve dry mass estimates. I summed the dry mass estimates of the boles and branches of each section to obtain an estimate of total aboveground biomass of each individual stem. Log-transformations were performed on the recorded dry weight and DBH data of the destructively sampled stems and I ran separate linear regressions on the transformed live and dead stem data to determine the parameters of each respective allometric equation. Additionally, ANCOVA was performed on live stem data to test for statistical differences in the biomassdiameter relationship across the two regions sampled (Highplains and Presque Isle). The alpha level of significance set for all statistical analyses was P < 0.05. Log-biomass estimates for live and dead stems were obtained using the following common linear function (Picard et al. 2012): lnB = a + (b * lnD) (1) where ln(B) is the natural log of biomass (kg), a is the y-intercept of the regression line, b is the slope and ln(D) is the natural log of DBH (cm). I later compared biomass estimates produced by 12 my live stem model to those produced by six other allometric equations for jack pine derived from studies conducted throughout the northeastern USA and Canada (reported in Ter-Mikaelian and Korzukhin 1997). 2.3 Stand inventory Within each stand, I established between three and five 6 x 12 m (0.0072 ha) plots for inventory sampling. Each plot was oriented with the long axis running parallel to the planting rows, and each plot contained three rows. For each standing tree within the plot, I recorded species, DBH, and status (live or dead). I estimated the individual log-biomass of each living and dead stem recorded in the inventory data with their respective allometric equation. I then back-transformed these estimates of log-biomass to reflect actual estimates of biomass in kilograms for each tree. Individual stem biomass estimates were then summed for each plot and utilized to estimate total standing biomass in Mg ha-1 for the plot using the following equation: B = S(b1:bn)/A/1000 (2) where B is the biomass estimate for the plot (Mg ha-1), b is the biomass estimate for an individual stem (kg), n is the number of individual stems within a given plot and A is the plot area (ha). Dividing by 1000 converts mass units from kg to Mg. Plot-level estimates were then averaged for each stand to produce mean stand-level biomass estimates in Mg ha-1. It should be noted that these estimates of biomass per unit area only apply to planted zones within KW plantations, and 13 do not account for the unplanted foraging gaps that comprise approximately 20% of the total habitat land area (MDNR 2014). 2.4 Volume estimation To estimate volume production over time, I first calculated the individual volumes of live stems from inventory data using equations and procedures outlined by Hahn (1984). I excluded all dead stems from volume estimates, as they are not considered a source of merchantable timber and thus, their contribution to stand-level volume is of little relevance to land manager decision-making processes. Additionally, volume was estimated only for stems that met the 1stick minimum size requirement for pulpwood production (2.44 m pulp stick below a 10.16 cm top). Sawlog volumes in these stands ranged from negligible to nonexistent and therefore, for the purposes of this study, I only reported estimates of pulpwood volume production. I applied input parameters of field-measured DBH and stand basal area from inventory data to Hahn’s (1984) equations to estimate the merchantable pulpwood volume of each stem. I assumed a species-specific site index of 50 for all stands in the study based on available stand inventory data from the USFS and MDNR. All other equation parameters were derived from Hahn (1984). I then summed the volumes of each stem within a given plot to estimate plot-level volume in cords ac-1. I converted these estimates to metric units and averaged the plot-level estimates within each stand to achieve mean stand-level volume estimates in m3 ha-1. These volume estimates per unit area only apply to planted areas of KW habitat, and do not account for the approximate 1/5 total land area left unplanted as foraging gaps. 14 2.5 Production over time For the purposes of this study, stand age refers to the number of years since plantation establishment, which I acquired from MDNR and USFS year of origin data. To estimate biomass as a function of stand age, a nonlinear relationship was described using a modification of the Richards logistic function (Richards 1959): Bt = a (1-e(-b*t))c (3) where Bt represents aboveground biomass (Mg ha-1) for planted areas at time t, a represents the potential maximum biomass (Mg ha-1), e is the base of a natural logarithm, t is stand age in years, b is a parameter controlling the rate of biomass accumulation, and c is a parameter controlling the inflection point of the curve. Similarly, I described a volume-age relationship using the same modified version of the Richards logistic function (Richards 1959): Vt = a (1-e(-b*t))c (4) where Vt represents aboveground volume (m3 ha-1) for planted areas at time t, a represents the potential maximum volume (m3 ha-1), e is the base of a natural logarithm, t is stand age in years, b is a parameter controlling the rate of volume accumulation, and c is a parameter controlling the inflection point of the curve. Differences in biomass and volume production across regions were estimated by analyzing differences of least squares means. I obtained parameters for the functional forms of 15 each growth curve from iterations produced by the nonlinear regression procedure for ChapmanRichards equations in SAS, using code outlined by Sit and Poulin-Costello (1994). These analyses were performed on stand-level biomass and volume data. 2.6 Covariate analyses To better understand variations in estimated biomass and volume across stands of equal and similar ages, I tested three covariates for statistical significance in the modified Richards models. Covariates I tested included stand density, natural soil drainage index (DI) (Schaetzl et al. 2009), and soil productivity index (PI) (Schaetzl et al. 2012). The Natural Soil Drainage Index (DI) is a general reflection of the amount of water that a soil supplies to plants under natural conditions over long timescales, and is primarily derived from a soil’s taxonomic classification. The DI ranges from 0 for the driest soils (bedrock in a desert) to 99 (open water) (Schaetzl et al. 2009). The soil Productivity Index (PI) ranks soils from 0 (least productive) to 19 (most productive) using interpretations of features or properties of a soil’s family-level taxonomic classification (Schaetzl et al. 2012). I calculated stand-level density estimates from stand inventory data as the mean plot-level density value for all plots within the stand. Stand densities ranged from ~1,917 to ~4,861 trees ha-1. I obtained the DI and PI of each plot from MDNR and USFS GIS data layers. DI and PI values were generally consistent across plots within a given stand. For stands in which index values varied across plots, the mode index value was selected for analyses. DI values ranged from 14 to 35 and PI values ranged from 4 to 9. I regressed each covariate against the residuals 16 of the biomass and volume curves to test for statistical significance and whether there was a need to include any in the final growth models. 2.7 MAI and optimal rotation ages To identify an optimal rotation length for biomass production, I analyzed Mean Annual Increment (MAI) values for biomass estimates produced by the growth model. Cooper (1984) states that maximum sustained yield is attained when a forest is harvested at the age it reaches culmination of MAI. For an S-shaped growth curve, this age can be determined mathematically, and is defined as the age at which MAI equals the derivative of the growth function (Cooper, 1984). I first calculated MAI values for biomass estimates at each year from 0 to 60 years. I then determined the derivative of the growth function for biomass, and calculated estimates for this derivative function at each year from 0 to 60 years. Finally, I plotted the MAI curve against the derivative growth function curve and identified the age at which their response values were equal. This age was then rounded to the nearest whole year, and reported that as the optimal rotation length for biomass production. The same procedure was applied to my volume curve data to identify an optimal rotation age for pulpwood production in KW stands. 2.8 Regional impact assessment Following determination of the optimal rotation ages for biomass and volume production in KW stands, I performed a regional impact assessment to compare 3 rotation lengths: the estimated optimal rotation age for biomass, the estimated optimal rotation age for pulpwood 17 volume, and the current business-as-usual (BAU) rotation age of 50 years. I assessed the potential ecosystem service outputs of each rotation length for a 1550 ha area over the course of 100 years. I selected a land area of 1550 ha for this analysis based on the reported average total land area that is harvested and planted into KW breeding habitat annually (MDNR 2014). For each rotation length, I calculated the number of full rotations that would occur over a 100-year period (assuming establishment at year 0), the potential biomass output per rotation (Gg 1550 ha1 ), the cumulative potential biomass output over a 100-year period (Gg 1550 ha-1 100 yrs-1), the potential volume output per rotation (m3 1550 ha-1), and the cumulative potential volume output over a 100-year period (m3 1550 ha-1 100 yrs-1). For this analysis, all potential biomass and volume yield outputs were calculated to account for the unplanted foraging gaps that are included in KW plantations, assuming these make up 1/5 of the total land area (MDNR 2014). Additionally, I calculated the cumulative number of years that the land would provide suitable breeding habitat for KW over a 100-year period. KW plantations only provide suitable breeding habitat between the ages of 5 and 23 years (Meyer 2010), so I calculated this figure based on the total number of years plantations under each rotation scenario would spend within this age range over a 100-year period, assuming establishment at year 0. Finally, I calculated the total land area that would need to be designated as KW habitat for each rotation scenario to continue to meet the annual habitat development objective of 1550 ha, as outlined in the KW Breeding Range Conservation Plan (MDNR 2014) by multiplying the rotation length (years) by 1550 ha. 18 3. Results 3.1 Biomass allometrics To most accurately estimate tree biomass from stand inventory, I first determined whether it was necessary to utilize separate allometric equations for stems of each region. ANCOVA of the log-transformed biomass-to-diameter relationships of live stems in the Highplains and Presque Isle regions showed an insignificant difference between the two regions (P = 0.097), suggesting the use of one model with a common slope parameter. Although the effect of region was close to statistically significant, parameter differences between the regions were quite small. For example, applying separate, region-specific allometric equations to inventory data resulted in a less than 5% difference from combined equation estimates in stands older than 20 years (Table 1). Therefore, I proceeded with a single generalized model to predict biomass of live stems for all regions in the study. The final combined allometric biomass equation was: lnB (kg) = -0.978 + (1.787 x lnD (cm)); adjusted R2 = 0.929; P < 0.001 (Figure 1A). A comparison of estimates produced by this equation to six other pre-existing allometric equations for jack pine derived from areas throughout the northeastern USA and Canada (reported in Ter-Mikaelian and Korzukhin 1997) resulted in significant differences in stand-level biomass estimates. For example, applying these equations to inventory data from a 20-year old stand resulted in overestimations as high as 25%, or approximately 13 Mg ha-1. Application of these equations to inventory data from a 52-year old stand resulted in overestimations as high as 40%, or approximately 40 Mg ha-1. 19 Stand Age Highplains Eq. -1 Estimate (Mg ha ) Presque Isle Eq. -1 Estimate (Mg ha ) Combined Eq. -1 Estimate (Mg ha ) 23 32 41 63.51 63.79 66.07 58.10 60.55 65.70 60.59 61.62 64.81 % diff. Highplains Eq. 4.60 3.40 1.91 % diff. Presque Isle Eq. 4.28 1.77 1.36 Table 1. Comparison of plot-level biomass estimate outputs from the Highplains-specific equation, Presque Isle-specific equation, and the combined allometric equation that does not account for effect of region. Differences between region-specific equation outputs and the combined equation outputs are expressed as percentages. Wednesday, April 12, 2017 04:36:12 P Monday, May 8, 2017 09:12:48 PM 5 The REG Procedure Model: MODEL1 Dependent Variable: lnB (A) Live Stems (B) Dead Stems The REG Procedure Model: MODEL1 Dependent Variable: lnB Fit Plot for lnB Fit Plot for lnB 6 4 Observations 8 Parameters 2 Error DF 6 MSE 0.0778 R-Square 0.9654 Adj R-Square 0.9596 Observations 26 Parameters 2 2 Error DF 24 MSE 0.1613 R-Square 0.932 Adj R-Square 0.9292 2 lnB lnB 4 0 0 -2 0 1 2 3 1.0 1.5 lnD Fit 95% Confidence Limits 2.0 2.5 3.0 lnD 95% Prediction Limits Fit 95% Confidence Limits 95% Prediction Limits Figure 1. Fit plots of lnBiomass (kg) against lnDBH (cm) produced by linear regression analyses of (A) live stem data and (B) standing dead stem data. Log-transformed biomass and diameter data show strong positive linear relationships between the two variables. Solid lines represent the linear regression lines described in the text. Shaded areas around these lines represent the 95% confidence limits, and dashed lines represent 95% prediction limits. 20 To further strengthen the accuracy of my plot-, stand-, and landscape-level biomass estimates, I developed a separate allometric equation to estimate biomass of dead stems in KW plantations. Dead stems are typically harvested in short-rotation bioenergy management systems, and I anticipated that these stems would contain lower levels of biomass than their livestem counterparts of equal diameter due to mortality-induced damage and decay. The linear regression I performed on log-transformed diameter data confirmed this prediction, and resulted in the following local allometric equation estimating log-biomass of dead stems as a function of log-diameter: lnB (kg) = -2.232 + (2.096 x lnD (cm)); adjusted R2 = 0.960; P <0.001 (Figure 1B). 3.2 Production over time Analysis of differences of least squares means showed no significant effect of region on biomass and volume production in relation to stand age. P-values for regional contrasts of biomass production ranged from 0.164 (Arenac vs Presque Isle) to 0.455 (Highplains vs Presque Isle). P-values for regional contrasts of volume production ranged from 0.742 (Highplains vs Presque Isle) to 0.972 (Highplains vs Arenac). Therefore, I proceeded with a single growth model for biomass and a single growth model for volume for all regions in the study. Stand-level biomass accumulation followed a classic sigmoidal pattern across the chronosequence. This pattern was characterized by a period of rapid accumulation between ca. 10 and 30 years, followed by a decline in the rate of production approaching an asymptote of 71 Mg ha-1 (Figure 2). The pattern of biomass accrual over time conformed well to the modified Richards function for logistic growth (Richards 1959): Bt (Mg ha-1) = 70.856(1-e(-0.118*t))4.201; P < 0.001. These estimates represent biomass production per hectare of planted area and do not account for unplanted areas within KW stands designated as foraging gaps. 21 100 Highplains Arenac Presque Isle 90 80 Biomass (Mg ha-1 ) 70 60 50 40 30 20 10 0 0 10 20 30 40 50 60 Stand Age (years) Figure 2. Aboveground biomass content as a function of stand age. Biomass accounts for both living and standing dead trees. Symbols represent stand means (±1 SE), and the solid curve represents the nonlinear regression line described in the text. Dashed lines represent the upper and lower 95% confidence limits across the age range of the dataset. To compare the optimal rotation lengths for biomass and pulpwood production in KW stands, I developed a similar growth curve to map volume accumulation over time (Figure 3). The shape of the curve for stand-level volume growth differed greatly from that of biomass production, showing no accumulation until ca. 20 years after planting. The minimum size requirements for pulpwood production limit volume accumulation until stems reach a threshold minimum merchantable size (2.44 m pulp stick below a 10.16 cm top), which occurs at approximately 20 years. This lag period is followed by a sharp uptick in the growth curve, reflecting rapid volume accrual between ca. 20 and 30 years, after which there is a decline in the 22 rate of accumulation approaching an asymptote of 71 m3 ha-1 (Figure 3). Stand-level pulpwood volume data was described using a modified Richards function for logistic growth (Richards 1959): Vt (m3 ha-1) = 71.118(1-e(-0.355*t))1932.7; P < 0.001. I found that the upper and lower 95% confidence limits for the volume function did not fit as tightly to the regression curve as those observed in the biomass regression, indicating lower estimate precision (Figs. 2 and 3). 140 Highplains Arenac Presque Isle 120 Volume (m3 ha-1 ) 100 80 60 40 20 0 0 10 20 30 40 50 60 Stand Age (years) Figure 3. Volume content as a function of stand age. Volume is estimated for all living stems with a minimum 2.44 m length to a 10.16 cm top. Symbols represent stand means (±1 SE), and the solid curve represents the nonlinear regression line described in the text. Dashed lines represent the upper and lower 95% confidence limits across the age range of the dataset. 23 3.3 Covariate analyses I regressed three covariates of potential significance against the residuals of both the biomass and volume growth curves. However, none of these regressions yielded statistically significant results for either biomass or volume growth (Tables 2 and 3). Contrary to my initial predictions, stand density and available soil drainage and productivity data did not explain observed variations across stands of equal and similar ages. Therefore, these covariates were not included in the final growth models. Statistical Variable R2 P-value Stand Density 0.0005 0.8933 Drainage Index 0.0003 0.9184 Productivity Index 0.0127 0.5059 Table 2. Statistical outputs of covariate regressions plotted against biomass curve residuals. Statistical Variable R2 P-value Stand Density 0.0337 0.2770 Drainage Index 0.0000 0.9910 Productivity Index 0.0313 0.2945 Table 3. Statistical outputs of covariate regressions plotted against volume curve residuals. 3.4 Optimal rotation ages I determined the optimal rotation age for maximum biomass yields to be 20 years after stand establishment. At this age, MAI was equivalent to the derivative of the biomass growth function, a point also known as the culmination of MAI (Figure 4A). Stands at this age contained an estimated 47 Mg ha-1 of aboveground biomass in planted zones of KW habitat. Additionally, I observed the culmination of MAI for the volume curve to occur at 28 years after 24 establishment, and determined this to be the optimal rotation length to maximize pulpwood yields (Figure 4B). At this age, stands are expected to yield approximately 65 m3 ha-1 of pulpwood volume in planted areas of KW habitat. (A) Biomass (B) Volume 10 10 9 9 8 8 7 7 6 6 5 5 4 4 3 3 2 2 1 1 0 0 0 10 20 30 40 50 60 0 10 Stand Age (years) Derivative MAI 20 30 40 50 60 Stand Age (years) Optimal Rotation Age Derivative MAI Optimal Rotation Age Figure 4. Derivatives of (A) biomass and (B) volume growth curves plotted against their respective MAI curves over time. Solid curves represent the growth curve derivatives and dashed curves represent MAI over time. Vertical dotted lines represent the stand age at which culmination of MAI occurs, where MAI equals the derivative of the growth function. 3.5 Regional impact assessment I compared the potential ecosystem service outputs that would be provided from 1550 ha of KW habitat managed under 3 different rotation scenarios over a 100-year period (Table 4). The three rotation lengths of interest were 20 years (estimated optimal rotation age for maximum biomass yields), 28 years (estimated optimal rotation age for maximum volume yields), and 50 years (current BAU rotation length for KW plantations; MDNR 2014). I selected a land area of 1550 ha to reflect the annual harvest and planting area objective outlined in the Kirtland’s Warbler Breeding Range Conservation Plan (MDNR 2014). All biomass and volume outputs 25 reported in Table 4 reflect forest production on 4/5 of a 1550 ha area (the approximate ratio of planted to unplanted patches within KW stands), equating to a total planted land area of 1240 ha. Rotation Length # Rotations 100 years-1 Biomass Rotation-1 (Gg) Cumulative Biomass (Gg 100 yrs-1) Volume Rotation-1 (103 m3) Cumulative Volume (103 m3 100 yrs-1) Suitable Habitat (yrs 100 yrs-1) 20 28 50 5 3 2 58 75 87 290 225 174 18 80 88 88 241 176 80 69 38 Total Required Habitat Area (ha) 31,000 43,400 77,500 Table 4. Potential ecosystem service outputs from 1550 ha of managed KW habitat over 100 years. Assuming establishment at year zero, management for biomass on a 20-year harvest cycle would undergo 5 full rotations within a 100-year period, followed by management for pulpwood, which would produce 3 harvests in this time, whereas management under the BAU rotation of 50 years would only be harvested twice. A 20-year rotation would yield the most cumulative biomass over 100 years (approximately 290 Gg) and a 28-year rotation would yield the most cumulative pulpwood over this period (approximately 241,000 m3). Although management on a 50-year rotation produces the highest biomass and volume outputs per harvest cycle, cumulative yields over a 100-year period are predicted to be significantly lower than those produced by stands managed on reduced rotations. For each rotation scenario, I calculated the cumulative years that stands would provide early-successional habitat suitable for KW breeding. Lands managed on the current 50-year rotation only fall within the age range of suitability for a total of 38 years per every 100 years of management. Managing for pulpwood production on a 28-year rotation would increase this to a total of 69 years spent as suitable habitat, and managing for biomass on a 20-year rotation would 26 further increase this value to a total of 80 years, more than double that of the BAU 50-year rotation. Additionally, I determined the total land area that would be required under each management scheme to continue harvesting and re-establishing 1550 ha of KW habitat on an annual basis. The 50-year rotation requires a total of 77,500 ha to be dedicated as KW habitat at any given time, whereas a 28-year rotation would reduce this to 43,400 ha. A 20-year rotation would require the lowest total land area to be under KW management at a given time, with a minimum 31,000 ha of designated habitat required to continue to meet current annual development objectives. 27 4. Discussion 4.1 Overview The results of my study indicate that a shift to short-rotation management could maximize pulpwood and biomass outputs in KW stands without affecting their ability to support endangered species conservation. Limiting implementation of this strategy to a portion of current KW stands would also allow for increased management options and ecosystem diversification at the landscape level. Because KW habitat-specificity is restricted to young stands between the ages of 5 and 23 years (Meyer 2010), stands managed under the current plan only provide suitable habitat for a fraction of their 50-year rotation, suggesting that rotation lengths could be substantially reduced without impacting the duration of time spent within the age range of suitability per rotation. Furthermore, reducing rotation lengths would increase the habitat turnover rate, in turn decreasing the total land area required to be under KW management at any given time to maintain adequate levels of available habitat on the landscape. Under this scenario, land managers would gain the opportunity to implement alternative silvicultural strategies in surplus KW stands to simultaneously manage for multiple objectives at the landscape level and increase the long-term economic and ecological sustainability of the system at large. Several studies have reported that optimization of forest multifunctionality cannot be achieved under a single management regime due to strong trade-offs that exist between maximization of provisioning services, regulating services, and biodiversity conservation, recommending management diversification at the landscape level as the only viable method to minimize these trade-offs and simultaneously meet multiple objectives (Triviño et al. 2017; Felton et al. 2015). However, the results of this study indicate that in jack pine stands of 28 Northern Lower Michigan, these often-conflicting goals have the potential to instead complement one another under a short-rotation management regime. At the stand level alone, reducing rotations could simultaneously increase timber revenues through the production of nontraditional wood products, maximize climate regulation benefits through carbon offsets via co-generation of biomass, and contribute to biodiversity conservation efforts by providing critical early-successional habitat, upon which an array of species of conservation concern are dependent (Corace et al. 2010). Supplemental implementation of diversified management at the landscape level would expand both the quantity and variety of benefits and services that may be rendered, allowing for true optimization of ecosystem multifunctionality. The following sections detail the potential ecological, economic, and climate-related implications associated with implementation of limited short-rotation management in KW habitat plantations. 4.2 Ecological implications To effectively maintain the current KW population on fewer lands, it is crucial that land managers be highly selective in determining which stands should remain under KW management, as the birds are not evenly distributed across their breeding range and tend to concentrate in specific geographical areas. Census data recorded since 2000 has shown that more than 86% of all singing males reside within 5 counties in Northern Lower Michigan, 33% in just one of these counties, and 15% in a single township alone (MDNR 2014). Therefore, discontinuing KW management in stands outside these locales should have minimal to no impact on the current bird population. Implementing reduced rotations in core nesting zones may in fact 29 benefit KW population growth as more stands within these zones would be within the age range of suitability at any given time, and distances between suitable habitats would be reduced. Additionally, this strategy could have enormous ecological benefits at the landscape level. In a study conducted by Tucker et al. (2016), current jack pine age distributions in Northern Lower Michigan were compared to estimated historical distributions from preEuropean settlement surveys. The authors compared the distributions of three age classes (<20, 20-50, and >50 years) and found that conversion of older jack pine stands to early-successional KW plantations has caused significant landscape homogenization over time, with a pronounced reduction in the prevalence of mature stands. On the current landscape, they found 31% of jack pine stands fell within the youngest age class (<20 years), 39% in the intermediate class (20-50 years), and 30% in the mature age class of >50 years. In contrast, estimates of pre-European distributions showed much higher levels of landscape variability with a mere 5% of stands belonging to the youngest class, 19% in the intermediate, and 76% in the mature age class. Over time, KW recovery efforts have converted most of these mature stands into habitat plantations, resulting in a major deviation from historical landscape distributions. Furthermore, the study found that KW management has displaced certain major cover types in the region in favor of jack pine, resulting in an estimated 29% decrease in red pine (Pinus resinosa) cover and a 67% reduction in barrens from their pre-European distributions (Tucker et al. 2016). Although the KW recovery plan has proven successful at restoring endangered KW populations in the region, its widespread implementation has come at the expense of landscape diversity, displacing valuable habitat ecosystems once prevalent on the landscape. Reducing the total habitat land area with the implementation of short-rotation management would allow for the restoration of landscape age distributions and cover types that 30 better emulate historical patterns and distributions. A couple viable silvicultural strategies that may be implemented on surplus KW stands include extending rotations to restore historical age class distributions and variability, and replanting harvested stands with cover species that are currently under-represented on the landscape. This ecosystem-based approach, which promotes structural and compositional heterogeneity, would increase biodiversity at the landscape level through provisioning of diverse habitat types, including later-successional forests which are characterized by key habitat features absent in younger stands (Lindenmayer et al. 2006; Franklin et al. 2007). 4.3 Economic implications 4.3.1 Overview of economic implications From an economic standpoint, short-rotation management of KW stands is expected to increase timber revenues at both the stand and landscape levels. At the stand level, this strategy would likely improve land-use efficiency of habitat plantations and boost financial returns via increased harvest frequencies and production of marketable wood products such as biomass and pulpwood. Shifting timber management goals to these nontraditional products should not interfere with current revenues from habitat harvests, as it has become apparent that these highdensity stands do not produce merchantable sawlogs by their planned harvest age. Thus, management for alternative wood products could in fact increase returns from harvest yields. At the landscape level, agencies could manage surplus habitat stands to produce more profitable timber products (such as sawlogs) and further increase financial returns from this system. Managers could extend rotations in surplus KW stands and conduct pre-commercial 31 thinnings to manage for jack pine sawlogs, and replant these stands with under-represented highvalue timber species such as red pine to increase future financial gains. Diversification of timber production on the landscape could help subsidize the high costs of annual habitat harvests and reestablishment to ensure long-term economic sustainability of KW management. 4.3.2 Local economic opportunity IPCC (2014) reports that emissions reduction strategies with low lifecycle emissions, such as fast growing tree species and sustainable use of biomass residues, can be effective in reducing greenhouse gas emissions, but rely on efficient integrated “biomass-to-bioenergy systems.” There are currently five wood-based electric power plants in the northeastern region of Michigan’s Lower Peninsula, which coincide with the core area of KW’s breeding range (Leefers 2011; MDNR 2014). In 2011, managers of these power plants reported that almost 80 percent of the wood fuel they use was sourced from a distance of 97 km or less from the facility, with several managers reporting 100 percent of their wood fuel as being sourced within this distance (Leefers 2011). Nearly all MDNR and USFS KW plantations in this region exist within a 50 km radius of at least one of these plants, and those that do not lie just outside this boundary (Figure 5). Short transport distances, coupled with high processing capacities, gives rise to the potential for an efficient integrated “biomass-to-bioenergy system” with low lifecycle emissions that would simultaneously provide multiple benefits in the form of climate change mitigation, increased options for land managers, biodiversity conservation, and support of the local economy. It should be noted, however, that the viability of implementing bioenergy production 32 management in this system is highly dependent on renewable energy policies and market demand. Figure 5. Locations of existing wood-based electric power plants and publicly-owned KW plantations in the northern region of Michigan’s Lower Peninsula. Black dots represent power plant locations and shaded gray areas represent areas under KW management. The circles represent 50 km radii around each power plant. Additionally, production of conifer pulpwood may soon gain significance and value in this region. The world’s second-largest producer of wood products, ARAUCO, is currently building a state-of-the-art particleboard plant in Grayling, Michigan, situated in the heart of the KW management area, and is expected to begin operation in late 2018. The plant will be North 33 America’s largest single continuous particleboard press with an annual processing capacity of approximately 800,000 m3 (ARAUCO 2017). This presents an outstanding opportunity for land managers to sell low-quality wood produced in KW stands to a high-capacity local manufacturing plant. In addition to the benefits of low transportation costs and distances, and indirect support of the local economy, selling KW harvests to ARAUCO could provide a more stable source of income for land agencies for decades to come, and resolve many of the marketability issues currently experienced with KW jack pine. 4.4 Climate-related implications Limited short-rotation management of KW habitat plantations would diversify and improve climate change mitigation benefits at both the stand and landscape levels. Sohngen and Brown (2008) report that extending rotation lengths, even by just a few years, is the quickest way to increase carbon stock on a landscape. Therefore, extending rotations in surplus KW stands would diversify climate change mitigation services at the landscape level to include enhanced carbon sequestration benefits alongside potential fossil-fuel reductions rendered from KW biomass harvests for bioenergy production. Furthermore, this mixed-management strategy could have significant benefits for the ecosystem in terms of risk mitigation. At the stand level, shortrotation KW habitats reduce risks associated with natural disturbances such as wind storms, as well as insect pests and diseases that target mature jack pine (Felton et al. 2015; Carey 1993). Whereas, at the landscape level, heterogeneity of species compositions, age distributions, and cover types increase landscape resilience and mitigate risks associated with species-specific pests 34 and diseases, which can be devastating in regions with widespread monoculture plantings (Felton et al. 2015; Ennos 2014). 4.5 Uncertainty and implications for future research Although this study identified optimal rotation lengths for biomass and volume production (20 and 28 years, respectively), these conclusions are highly dependent on the accuracy of my growth curves. There are a few key limitations that should be considered before these rotations are implemented on a large scale. The primary source of uncertainty related to my recommended harvest ages pertains to the lack of data I had for older KW stands. Because large-scale KW habitat establishment began in 1981, few mature plantations were available for sampling. Therefore, I was only able to collect data in two stands >35 years, aged 41 and 52. It is possible that parameter values for the biomass and volume growth curves could change as more data from mature KW stands becomes available. Alterations to model parameters would affect predictions of optimal rotation lengths for biomass and pulpwood production. In the absence of sufficient data from older KW plantations, a comparison of my estimates of biomass and volume production to those from the literature will be used to determine whether reported rotation ages from this study seem plausible, or whether they appear to be skewed by the limited age range of the dataset. To address uncertainties related to biomass production over time, I compared my results to two studies conducted on jack pine grown in the northern region of Michigan’s Lower Peninsula, the same geographic location in which this study is based. Rothstein et al. (2004) studied the loss and recovery of carbon pools following stand-replacing fire in jack pine stands in 35 this region. They estimated overstory biomass in a chronosequence ranging in age from 1-72 years and found that growth over time in this system followed an S-shaped pattern with biomass production peaking at 16 years and approaching an asymptotic value of 106 Mg ha-1 by age 40. In a separate study, Spaulding and Rothstein (2009) evaluated stand structural differences between fire-origin jack pine stands and jack pine plantations. In this study, they collected data from two jack pine plantations aged 65 and 69 which averaged 100 and 105 Mg ha-1 of biomass, respectively. Although both studies’ estimates of maximal biomass were higher than my estimate of 71 Mg ha-1, both studies utilized Perala and Alban’s (1994) allometric equation to estimate jack pine biomass, which I found could lead to a nearly 40% overestimation of standlevel biomass in older KW stands. Overall, results from these two studies support the conclusions of this study on both the timing and magnitude of biomass production for jack pine in this area. To my knowledge, no studies examining volume accumulation by jack pine in Northern Lower Michigan exist in the current literature. However, studies of jack pine and closely related species from other regions are available for comparison. Hébert et al. (2016) compared volume increment rates of individual stems in jack pine plantations of varying densities (1111 trees ha-1 to 4444 trees ha-1) in Quebec, Canada. Although values for volume increment varied by stand density, all stem increment rates followed the same general pattern across the 25-year study period, peaking at an age of approximately 15 years. Long and Smith (1992) measured stem volume increment as a function of age and relative density for the closely-related lodgepole pine (Pinus contorta) in south-central Wyoming, between the ages of 10 and 117 years. They found that at a stand density of 1200 trees ha-1 volume increment peaked approximately 40 years after establishment. Considering the high density and harsh environmental conditions under which 36 jack pine grow in KW plantations, these results suggest that a culmination of MAI for volume occurring at 28 years appears to be a valid assessment. I also compared my results for volume production to two studies which compared growth and development of jack pine grown in extremely dense fire-originated stands to less dense plantations. Morris et al. (2014) found that 5-year periodic increments for stand-level volume peaked at 20 years following establishment in both planted and naturally-regenerated stands in Ontario, Canada. The authors associated the age of peak volume increment to the age at which crown closure occurs, and photosynthetic capacity is reduced in the stand. In another study in Ontario, Canada, Janas and Brand (1988) found that volume increment for high-density natural jack pine stands peaked at 18 years, and plantations of 2.13 m spacing peaked at 15 years. They concluded that an optimal biological rotation for volume production would be shortest for the densest stands and longer for stands of lower density, unless those stands were to be managed for sawlog production, in which case the reverse would be true. Overall, these studies provide strong support for my conclusion that rotation lengths in high-density KW plantations could be substantially reduced to maximize volume yields prior to growth stagnation. Despite supportive data from other studies of jack pine and related species, it is important that we continue to monitor trends in biomass and volume accrual as younger KW habitats mature and enter older age classes. Future research should focus on KW stands grown beyond 50 years to determine if and when they reach merchantable sawlog size. Due to the high planting density of these stands, it may be necessary to conduct one or more pre-harvest thinnings to release residual stems and allow them to grow to merchantable size (Morris et al. 2014; Janas and Brand 1988). Continued monitoring and experimentation is needed to adequately assess 37 which management strategies, or combination of management strategies, prove to be the most feasible and beneficial to the system. It is also important to note that my 100-year regional impact assessment is based on current growth patterns, which are likely to change over the next century as the effects of global climate change continue to intensify (Chum et al. 2011). Increased levels of atmospheric carbon dioxide could contribute to increased production rates, whereas rapidly warming temperatures and drought could have the opposite effect on jack pine growth in KW stands. Geographically, KW plantations exist at the southern limit of jack pine’s current natural growth range (Rudolph 1985). However, it has been projected that in the northern hemisphere, tree species will begin to migrate northward as changes in temperature and precipitation patterns at the southern margins of their current distributions become unsuitable for growth (Case and Lawler 2017). Furthermore, it is predicted that over the next 100 years, the frequency and severity of natural disturbances will increase, raising concerns that a suite of threats could impact growth patterns over the coming decades (MEA 2005; Felton et al. 2015). Because KW plantations are a monoculture system, they are prone to species-specific pest outbreaks (Thompson et al. 2009; Ennos 2014). Additionally, pathogen pressures tend to be consistently high in stands with high host densities (Ennos 2014). Implementing short rotations could potentially mitigate these risks by harvesting stems before they reach the peak age of susceptibility (Roberge et al. 2016). Another important caveat to consider is the potential long-term consequences of repeated whole-tree harvests on site productivity and nutrient availability in forest ecosystems. Several studies have shown that low-fertility sites, such as those utilized in KW management, are most likely to be negatively impacted by repeated intensive management practices, such as whole-tree harvests, and associated nutrient removals over the long-term (Blanco et al. 2005; Kaarakka et al. 38 2014). There is also evidence that pioneer shade-intolerant tree species such as jack pine, which exhibit rapid rates of resource acquisition, are most strongly impacted by changes in soil nutrient status associated with increased harvest intensities (Thiffault et al. 2006). Therefore, it is important to monitor for changes in soil quality and productivity over time in intensivelymanaged KW stands and adjust management accordingly; this may involve alternating harvesting methods and rotation lengths on a given site to balance production with long-term soil sustainability. 39 5. Conclusions Implementation of short-rotation management in jack pine habitat plantations in Michigan’s Lower Peninsula has the potential to benefit the system on multiple levels. Reducing rotation lengths in a portion of these stands would expand the quantity of ecosystem services that can be rendered to include provisioning of forest products and climate change mitigation benefits, without negatively impacting their ability to provide critical early-successional habitat to endangered species and species of conservation concern. Additionally, shifting to shorter rotations would reduce the total land area required to be under KW management at any given time, allowing for significant ecological and economic benefits while continuing to meet annual habitat development objectives. For one, land agencies would gain the opportunity to diversify management goals at the landscape level and produce more valuable timber species on extended rotations to subsidize costs associated with annual KW habitat harvests and development projects. Diversification would also increase habitat variation and availability, improving the landscape’s ability to support a variety of species adapted to various disturbance regimes and forest cover types. Limiting KW management to a portion of its current domain would improve landscape resilience to disturbance events, while better emulating the region’s historical forest distributions. 40 APPENDIX 41 Table 5. Plot-level data Stand ID Age 1 1 1 2 2 2 2 2 3 3 3 3 3 4 4 4 5 5 5 5 5 6 6 6 6 6 7 7 7 8 8 8 8 8 9 9 9 9 9 10 10 10 10 10 11 11 11 11 11 12 12 12 12 12 13 13 13 13 13 14 14 14 14 14 15 15 15 16 16 16 16 16 17 17 17 17 17 18 18 18 18 18 52 52 52 41 41 41 41 41 35 35 35 35 35 35 35 35 34 34 34 34 34 33 33 33 33 33 32 32 32 32 32 32 32 32 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 30 30 30 30 30 28 28 28 28 28 28 28 28 24 24 24 24 24 23 23 23 23 23 22 22 22 22 22 Region Ownership Plot Latitude Longitude Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Arenac Arenac Arenac Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Arenac Arenac Arenac Presque Isle Presque Isle Presque Isle Presque Isle Presque Isle Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Presque Isle Presque Isle Presque Isle Presque Isle Presque Isle Highplains Highplains Highplains Highplains Highplains Arenac Arenac Arenac Highplains Highplains Highplains Highplains Highplains Presque Isle Presque Isle Presque Isle Presque Isle Presque Isle Highplains Highplains Highplains Highplains Highplains MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR USFS USFS USFS MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR USFS USFS USFS MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR USFS USFS USFS MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR 1 2 3 1 2 3 4 5 1 2 3 4 5 1 2 3 1 2 3 4 5 1 2 3 4 5 1 2 3 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 44.8422 44.8453 44.8427 44.4403 44.4383 44.4356 44.4393 44.4389 44.4325 44.4347 44.4343 44.4338 44.4340 44.5130 44.5120 44.5113 44.5363 44.5360 44.5358 44.5352 44.5361 44.5966 44.5962 44.5964 44.5965 44.5964 44.4844 44.4838 44.4839 45.1962 45.1953 45.1964 45.1971 45.1971 44.4323 44.4332 44.4317 44.4331 44.4312 44.7470 44.7461 44.7456 44.7459 44.7461 44.5733 44.5729 44.5737 44.5737 44.5725 44.8379 44.8384 44.8389 44.8397 44.8377 45.1956 45.1949 45.1955 45.1956 45.1954 44.4776 44.4771 44.4771 44.4771 44.4772 44.5363 44.5345 44.5401 44.5434 44.5438 44.5433 44.5438 44.5431 45.1269 45.1253 45.1271 45.1270 45.1267 44.4329 44.4341 44.4336 44.4334 44.4338 -84.4624 -84.4654 -84.4700 -84.2972 -84.2973 -84.3016 -84.2943 -84.2930 -84.2673 -84.2602 -84.2505 -84.2557 -84.2613 -83.6298 -83.6328 -83.6426 -84.8270 -84.8273 -84.8269 -84.8267 -84.8278 -84.6245 -84.6236 -84.6223 -84.6213 -84.6248 -83.5455 -83.5457 -83.5445 -84.1577 -84.1566 -84.1514 -84.1527 -84.1506 -84.2475 -84.2482 -84.2480 -84.2480 -84.2481 -84.4324 -84.4320 -84.4312 -84.4300 -84.4314 -84.5569 -84.5571 -84.5571 -84.5579 -84.5579 -84.4901 -84.4900 -84.4896 -84.4896 -84.4903 -84.1440 -84.1439 -84.1419 -84.1430 -84.1428 -84.3174 -84.3123 -84.3080 -84.3140 -84.3133 -83.6261 -83.6273 -83.6206 -84.8321 -84.8333 -84.8338 -84.8338 -84.8325 -84.1947 -84.1947 -84.1919 -84.1924 -84.1916 -84.2732 -84.2741 -84.2745 -84.2754 -84.2722 Total Area Planted Area Total Area Planted Area Live Total Area Live Planted Area Total Area Planted Area 3 Density (stems Density (stems Stem Density Stem Density Volume (m ha -1 -1 3 -1 Biomass (Mg ha ) Biomass (Mg ha ) Volume (m ha ) -1 -1 -1 -1 1 ha ) ha ) (stems ha ) (stems ha ) ) 2917 1389 1944 3611 2778 3194 3194 1806 3194 2917 2500 4167 3194 2500 4167 2361 2639 2778 1806 3611 1944 3333 2222 2639 2222 2917 2222 4167 2500 3194 3472 2639 2361 2500 2361 2500 1944 1667 1111 2639 2639 1806 2500 1667 2778 3472 3333 2639 3472 1806 2500 1944 2222 2917 2778 2778 2917 2361 2917 2778 3611 2778 2917 2361 3056 2778 3611 3194 3611 4028 2639 4028 4028 2639 3056 2639 3611 3333 2639 2500 2083 3611 2333 1111 1556 2889 2222 2556 2556 1444 2556 2333 2000 3333 2556 2000 3333 1889 2111 2222 1444 2889 1556 2667 1778 2111 1778 2333 1778 3333 2000 2556 2778 2111 1889 2000 1889 2000 1556 1333 889 2111 2111 1444 2000 1333 2222 2778 2667 2111 2778 1444 2000 1556 1778 2333 2222 2222 2333 1889 2333 2222 2889 2222 2333 1889 2444 2222 2889 2556 2889 3222 2111 3222 3222 2111 2444 2111 2889 2667 2111 2000 1667 2889 42 83 55 73 61 45 75 76 52 56 50 64 69 62 61 60 67 75 104 60 66 90 62 52 67 60 58 55 67 53 63 66 53 49 53 69 59 59 53 37 74 76 62 72 47 69 69 75 45 64 66 91 66 72 91 62 54 68 38 63 48 62 71 74 59 69 74 76 74 81 62 70 96 61 71 55 65 75 43 49 44 51 46 66 44 59 49 36 60 61 42 45 40 51 55 49 49 48 53 60 83 48 52 72 49 42 54 48 46 44 54 42 50 53 42 39 42 55 47 47 42 30 59 61 50 57 38 55 55 60 36 52 53 72 53 57 73 50 43 55 31 50 39 49 57 59 47 55 59 61 59 65 50 56 77 48 57 44 52 60 34 39 35 40 36 1667 833 1250 1806 1944 2222 2361 1389 2222 1250 1250 2222 1250 1528 2361 1667 1806 1944 1389 2222 1806 1250 1389 1528 1111 1667 1806 2778 1806 2778 2500 1944 1528 1667 1806 1111 1250 1250 972 2083 1944 1528 2083 1250 1806 3194 2639 1806 2083 1528 2222 1528 2083 2083 2361 2361 1944 1111 1806 2083 2083 2361 2083 1806 3056 2778 3472 2500 2917 2500 2222 3472 4028 2222 2917 2639 2917 2500 2500 2222 2083 3333 1333 667 1000 1444 1556 1778 1889 1111 1778 1000 1000 1778 1000 1222 1889 1333 1444 1556 1111 1778 1444 1000 1111 1222 889 1333 1444 2222 1444 2222 2000 1556 1222 1333 1444 889 1000 1000 778 1667 1556 1222 1667 1000 1444 2556 2111 1444 1667 1222 1778 1222 1667 1667 1889 1889 1556 889 1444 1667 1667 1889 1667 1444 2444 2222 2778 2000 2333 2000 1778 2778 3222 1778 2333 2111 2333 2000 2000 1778 1667 2667 107 72 113 55 21 65 81 63 42 25 73 46 65 70 57 90 84 156 87 41 145 43 54 66 75 56 76 57 55 40 57 48 52 58 82 41 64 70 46 91 92 82 91 56 84 32 61 38 52 96 122 92 93 127 49 48 58 32 63 21 55 72 69 62 48 75 66 72 93 49 83 98 34 93 32 52 45 22 16 28 51 14 85 57 90 44 17 52 65 50 34 20 59 37 52 56 45 72 67 125 69 33 116 34 44 53 60 45 61 46 44 32 46 38 41 47 65 33 51 56 37 72 74 66 73 45 67 26 49 30 42 77 98 73 74 101 39 38 46 25 50 16 44 58 55 50 38 60 53 57 74 39 66 78 27 75 26 42 36 18 13 23 41 11 Table 5 (cont’d) Stand ID Age Region Ownership Plot Latitude Longitude Planted Area Total Area Density (stems Density (stems -1 -1 ha ) ha ) 19 22 Highplains MDNR 1 44.4625 -84.2931 3594 2875 19 19 19 19 20 20 20 21 21 21 21 21 22 22 22 22 22 23 23 23 23 23 24 24 24 24 24 25 25 25 25 25 26 26 26 27 27 27 27 27 28 28 28 28 28 29 29 29 29 29 30 30 30 30 30 31 31 31 32 32 32 32 32 33 33 33 33 33 34 34 34 35 35 35 35 35 36 36 36 36 36 37 37 22 22 22 22 22 22 22 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 19 19 19 19 19 19 19 19 19 19 19 19 19 18 18 18 18 18 17 17 17 17 17 16 16 16 16 16 15 15 15 15 15 13 13 13 13 13 13 13 13 11 11 11 11 11 10 10 10 10 10 10 10 10 8 8 8 8 8 7 7 Highplains Highplains Highplains Highplains Arenac Arenac Arenac Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Presque Isle Presque Isle Presque Isle Presque Isle Presque Isle Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Arenac Arenac Arenac Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Highplains Presque Isle Presque Isle Presque Isle Presque Isle Presque Isle Highplains Highplains Highplains Highplains Highplains Arenac Arenac Arenac Presque Isle Presque Isle Presque Isle Presque Isle Presque Isle Highplains Highplains Highplains Highplains Highplains Arenac Arenac Arenac Presque Isle Presque Isle Presque Isle Presque Isle Presque Isle Highplains Highplains Highplains Highplains Highplains Presque Isle Presque Isle MDNR MDNR MDNR MDNR USFS USFS USFS MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR USFS USFS USFS MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR USFS USFS USFS MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR USFS USFS USFS MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR MDNR 2 3 4 5 1 2 3 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 1 2 3 4 5 1 2 3 4 5 1 2 3 1 2 3 4 5 1 2 3 4 5 4 5 44.4559 44.4510 44.4546 44.4551 44.5175 44.5210 44.5172 44.4685 44.4707 44.4714 44.4723 44.4679 44.6247 44.6252 44.6257 44.6255 44.6248 45.1458 45.1459 45.1486 45.1468 45.1469 44.5981 44.5981 44.5979 44.5979 44.5980 44.7579 44.7580 44.7617 44.7626 44.7608 44.5262 44.5262 44.5301 44.4731 44.4756 44.4734 44.4766 44.4765 44.5838 44.5834 44.5837 44.5833 44.5835 45.2018 45.2034 45.2057 45.2069 45.2039 44.4796 44.4829 44.4790 44.4789 44.4799 44.5210 44.5253 44.5253 45.1487 45.1487 45.1457 45.1425 45.1435 44.4051 44.3976 44.4013 44.4008 44.4088 44.5411 44.5406 44.5370 45.1497 45.1514 45.1492 45.1496 45.1504 44.3570 44.3481 44.3453 44.3550 44.3529 45.2190 45.2170 -84.2913 -84.2935 -84.2916 -84.2921 -83.6221 -83.6233 -83.6216 -84.3514 -84.3517 -84.3516 -84.3520 -84.3515 -84.6266 -84.6263 -84.6272 -84.6280 -84.6269 -84.1741 -84.1708 -84.1736 -84.1736 -84.1742 -84.5836 -84.5827 -84.5820 -84.5815 -84.5835 -84.3545 -84.3542 -84.3541 -84.3542 -84.3546 -83.6114 -83.6029 -83.6031 -84.2791 -84.2786 -84.2751 -84.2619 -84.2616 -84.6198 -84.6193 -84.6201 -84.6203 -84.6199 -84.1653 -84.1642 -84.1610 -84.1627 -84.1627 -84.3463 -84.3451 -84.3416 -84.3451 -84.3443 -83.5290 -83.5603 -83.5581 -84.1980 -84.1957 -84.1968 -84.1909 -84.1933 -84.3876 -84.3872 -84.3806 -84.3898 -84.3764 -83.5867 -83.5846 -83.6031 -84.1941 -84.1996 -84.1959 -84.1978 -84.1999 -84.3669 -84.3783 -84.3366 -84.3579 -84.3525 -84.1692 -84.1702 2778 3056 3194 3194 3472 3056 3333 3056 3056 2361 1389 2500 3333 3056 4028 3472 3194 3472 3056 3750 3333 3472 3194 3750 3611 3056 4583 2778 2639 3194 2083 3611 3333 3333 2500 2639 4861 2222 4306 4167 4861 4444 2917 2500 1667 3889 3472 3750 3472 4583 6944 5139 3333 4444 4444 3333 3750 2778 6528 2500 3611 4444 4861 2778 2639 3611 3750 4028 3611 3611 3056 3889 3750 3889 4444 3889 3750 3472 3056 4167 3056 2917 3472 2222 2444 2556 2556 2778 2444 2667 2444 2444 1889 1111 2000 2667 2444 3222 2778 2556 2778 2444 3000 2667 2778 2556 3000 2889 2444 3667 2222 2111 2556 1667 2889 2667 2667 2000 2111 3889 1778 3444 3333 3889 3556 2333 2000 1333 3111 2778 3000 2778 3667 5556 4111 2667 3556 3556 2667 3000 2222 5222 2000 2889 3556 3889 2222 2111 2889 3000 3222 2889 2889 2444 3111 3000 3111 3556 3111 3000 2778 2444 3333 2444 2333 2778 43 Planted Area Total Area -1 -1 Biomass (Mg ha ) Biomass (Mg ha ) 47 40 36 42 45 61 53 58 48 42 42 27 47 43 53 73 44 65 42 36 37 40 44 23 22 30 23 35 51 45 58 37 45 55 46 40 32 40 21 42 47 46 41 36 34 17 20 26 30 21 31 29 26 20 34 45 30 34 22 28 9 26 36 32 8 11 13 22 22 29 24 36 23 13 19 11 17 14 12 7 13 6 3 4 Planted Area Live Total Area Live Stem Density Stem Density -1 -1 Planted Area 3 -1 Volume (m ha ) Total Area Volume (m3 ha1 (stems ha ) (stems ha ) 38 3047 2438 10 8 ) 32 29 34 36 49 42 46 38 34 34 22 37 34 42 58 35 52 34 29 30 32 35 18 18 24 18 28 41 36 47 29 36 44 37 32 26 32 17 34 38 37 33 28 27 13 16 21 24 17 25 23 21 16 27 36 24 27 17 22 7 20 29 26 6 9 10 17 18 23 19 29 19 10 15 9 14 11 9 6 10 5 3 3 2778 2778 2639 2639 2917 3056 3056 2917 2500 2222 1389 2361 2639 2639 2361 2222 2778 3333 2917 3611 3194 3472 3194 3750 3611 3056 4583 2778 2639 3194 2083 3472 3333 3333 2361 2639 4583 2222 4028 3750 4722 4444 2917 2500 1667 3750 3333 3611 3472 4583 6944 5139 3056 4444 4444 3333 3750 2778 6528 2361 3611 4444 4861 2778 2639 3611 3750 3750 3333 3611 3056 3889 3750 3889 4306 3611 3750 3472 2917 4167 3056 2778 3472 2222 2222 2111 2111 2333 2444 2444 2333 2000 1778 1111 1889 2111 2111 1889 1778 2222 2667 2333 2889 2556 2778 2556 3000 2889 2444 3667 2222 2111 2556 1667 2778 2667 2667 1889 2111 3667 1778 3222 3000 3778 3556 2333 2000 1333 3000 2667 2889 2778 3667 5556 4111 2444 3556 3556 2667 3000 2222 5222 1889 2889 3556 3889 2222 2111 2889 3000 3000 2667 2889 2444 3111 3000 3111 3444 2889 3000 2778 2333 3333 2444 2222 2778 6 0 6 6 29 6 37 20 15 36 21 21 0 37 51 12 45 0 0 0 0 0 0 0 0 0 0 15 30 29 19 0 33 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 14 0 0 0 0 0 0 0 0 0 0 0 5 0 5 5 23 5 30 16 12 29 17 17 0 29 41 10 36 0 0 0 0 0 0 0 0 0 0 12 24 23 15 0 26 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11 0 0 0 0 0 0 0 0 0 0 0 REFERENCES 44 REFERENCES Albert, D. A. (1995). Regional landscape ecosystems of Michigan, Minnesota, and Wisconsin: a working map and classification. General technical report NC (USA). ARAUCO. (2017). ARAUCO Conducts Groundbreaking Ceremony for New Particleboard Plant in Michigan. ARAUCO North America. http://www.arauco-na.com/news-and-pressreleases/arauco-conducts-groundbreaking-ceremony-for-new-particleboard-plant-inmichigan Benzie, J. W. (1977). Manager's handbook for jack pine in the north central states. North Central Forest Experiment Station, Forest Service, U.S. Dept. of Agriculture, Saint Paul, MN. Berch, S., Morris, D., & Malcolm, J. (2011). Intensive forest biomass harvesting and biodiversity in Canada: a summary of relevant issues 1. The Forestry Chronicle, 87(4), 478-487. Blanco, J. A., Zavala, M. A., Imbert, J. B., & Castillo, F. J. (2005). Sustainability of forest management practices: Evaluation through a simulation model of nutrient cycling. Forest Ecology and Management, 213(1), 209-228. Byelich, J., M. E. DeCapita, G. W. Irvine, R. E. Radtke, N. I. Johnson, W. R. Jones, H. Mayfield, and W. J. Mahalak. 1976 (revised 1985). Kirtland’s Warbler Recovery Plan. U. S. Fish and Wildlife Service, Twin Cities, MN. Carey, J. H. (1993). Pinus banksiana. 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/ Case, M. J., & Lawler, J. J. (2017). Integrating mechanistic and empirical model projections to assess climate impacts on tree species distributions in northwestern North America. Global change biology, 23(5), 2005-2015. Chum, H., A. Faaij, J. Moreira, G. Berndes, P. Dhamija, H. Dong, B. Gabrielle, A. Goss Eng, W. Lucht, M. Mapako, O. Masera Cerutti, T. McIntyre, T. Minowa, K. Pingoud, (2011): Bioenergy. In IPCC Special Report on Renewable Energy Sources and Climate Change Mitigation [O. Edenhofer, R. Pichs-Madruga, Y. Sokona, K. Seyboth, P. Matschoss, S. Kadner, T. Zwickel, P. Eickemeier, G. Hansen, S. Schlömer, C. von Stechow (eds)], Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Cleland, D.T., T.R. Crow, S.C. Saunders, D.I. Dickmann, A.L. Maclean, J.K. Jordan, R.L. Watson, A.M. Sloan, and K.D. Brosofske. (2004). Characterizing historical and modern fire regimes in Michigan (USA): a landscape ecosystem approach. Landscape Ecology, 19(3), 311-325. 45 Cooper, C. (1984). Carbon storage in managed forests. Canadian Journal of Forest Research, 13(1), 155-166. Corace III, R. G., Goebel, P. C., & McCormick, D. L. (2010). Kirtland's warbler habitat management and multi-species bird conservation: considerations for planning and management across jack pine (Pinus banksiana Lamb.) habitat types. Natural Areas Journal, 30(2), 174-190. Ennos, R. A. (2014). Resilience of forests to pathogens: an evolutionary ecology perspective. Forestry: An International Journal of Forest Research, 88(1), 41-52. Felton, Adam, Johan Sonesson, Urban Nilsson, Tomas Lämås, Tomas Lundmark, Annika Nordin, Thomas Ranius, and Jean-Michel Roberge. (2017). Varying rotation lengths in northern production forests: Implications for habitats provided by retention and production trees. Ambio 46(3), 324. Felton, A., L. Gustafsson, J-M. Roberge, T. Ranius, J. Hjältén, J. Rudolphi, M. Lindbladh et al. (2015). How climate change adaptation and mitigation strategies can threaten or enhance the biodiversity of production forests: Insights from Sweden. Biological Conservation 194, 11-20. Franklin, J. F., R.J. Mitchell, and B. Palik. (2007). Natural disturbance and stand development principles for ecological forestry. Gen. Tech. Rep. NRS-19. Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 44 p. Hahn, Jerold T. (1984). Tree volume and biomass equations for the Lake States. Research Paper NC-250. St. Paul, MN: U.S. Dept. of Agriculture, Forest Service, North Central Forest Experiment Station. Hébert, F., Krause, C., Plourde, P. Y., Achim, A., Prégent, G., & Ménétrier, J. (2016). Effect of Tree Spacing on Tree Level Volume Growth, Morphology, and Wood Properties in a 25Year-Old Pinus banksiana Plantation in the Boreal Forest of Quebec. Forests, 7(11), 276. Immerzeel D., Verweij P., Hilst F., Faaij A. P. (2014). Biodiversity impacts of bioenergy crop production: a state-of-the-art review. Global Change Biology Bioenergy, 6(3), 183–209. IPCC, (2014). Summary for Policymakers, In: Climate Change 2014, Mitigation of Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Edenhofer, O., R. Pichs-Madruga, Y. Sokona, E. Farahani, S. Kadner, K. Seyboth, A. Adler, I. Baum, S. Brunner, P. Eickemeier, Kriemann, J. Savolainen, S. Schlömer, C. von Stechow, T. Zwickel and J.C. Minx (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Janas, P. S., & Brand, D. G. (1988). Comparative growth and development of planted and natural stands of jack pine. The Forestry Chronicle, 64(4), 320-328. 46 Kaarakka, L., Tamminen, P., Saarsalmi, A., Kukkola, M., Helmisaari, H. S., & Burton, A. J. (2014). Effects of repeated whole-tree harvesting on soil properties and tree growth in a Norway spruce (Picea abies (L.) Karst.) stand. Forest Ecology and Management, 313, 180-187. Kepler, C. B., Irvine, G. W., DeCapita, M. E., & Weinrich, J. (1996). The conservation management of Kirtland's Warbler Dendroica kirtlandii. Bird Conservation International, 6(1), 11-22. Leefers, L. (2011). Wood-based electric power generation in Michigan: wood use and policies. Forest products journal, 61(7), 586-591. Lindenmayer, D. B., Franklin, J. F., & Fischer, J. (2006). General management principles and a checklist of strategies to guide forest biodiversity conservation. Biological conservation, 131(3), 433-445. Long, J. N., & Smith, F. W. (1992). Volume increment in Pinus contorta var. latifolia: the influence of stand development and crown dynamics. Forest Ecology and Management, 53(1-4), 53-64. Mayfield, H. F. (1963). Establishment of preserves for the Kirtland's Warbler in the state and national forests of Michigan. The Wilson Bulletin, 75(2), 216-220. MDNR (2014). Kirtland’s Warbler Breeding Range Conservation Plan. https://www.michigan.gov/documents/dnr/Kirtlands_Warbler_CP_457727_7.pdf MEA (Millennium Ecosystem Assessment). (2005). Ecosystems and human well-being: Synthesis. Washington: Island Press. Meyer, Rachelle. (2010). Dendroica kirtlandii. 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/ [2013, June 17]). Morris, D. M., Reid, D. E., Kwiaton, M., Hunt, S. L., & Gordon, A. M. (2014). Comparing growth patterns of jack pine and black spruce in mixed natural stands and plantations. Perala, D.A., Alban, D.H., (1994). Allometric biomass estimators for aspen-dominated ecosystems in the upper Great Lakes. In: Research Paper NC-314, Forest Service, U.S. Department of Agriculture, North Central Forest Experiment Station, St. Paul, MN. Picard, N., Saint-André, L., Henry, M. (2012). Manual for building tree volume and biomass allometric equations: from field measurement to prediction. Food and Agricultural Organization of the United Nations, Rome, and Centre de Coopération Internationale en Recherche Agronomique pour le Développement, Montpellier, 215 pp. 47 Radtke, R., & Byelich, J. (1963). Kirtland's warbler management. The Wilson Bulletin, 208-215. Richards, F. J. (1959). A flexible growth function for empirical use. Journal of experimental Botany, 10(2), 290-301. Roberge, J. M., Laudon, H., Björkman, C., Ranius, T., Sandström, C., Felton, A., ... & Bergh, J. (2016). Socio-ecological implications of modifying rotation lengths in forestry. Ambio, 45(2), 109-123.
 Rothstein, D. E., Yermakov, Z., & Buell, A. L. (2004). Loss and recovery of ecosystem carbon pools following stand-replacing wildfire in Michigan jack pine forests. Canadian Journal of Forest Research, 34(9), 1908-1918. Rudolph, T. D. (1985). Jack pine, an American wood. FS–252. Washington, DC: US Department of Agriculture, Forest Service. Schaetzl, R. J., Krist Jr, F. J., & Miller, B. A. (2012). A taxonomically based ordinal estimate of soil productivity for landscape-scale analyses. Soil Science, 177(4), 288-299. Schaetzl, R. J., Krist, F. J., Stanley, K., & Hupy, C. M. (2009). The natural soil drainage index: an ordinal estimate of long-term soil wetness. Physical Geography, 30(5), 383-409. Sit, V., & Poulin-Costello, M. (1994). Catalog of curves for curve fitting (Biometrics Information handbook series, No. 4). British Columbia Ministry of Forests, Victoria, BC. Sohngen, B., & Brown, S. (2008). Extending timber rotations: carbon and cost implications. Climate Policy, 8(5), 435-451. Spaulding, S. E., & Rothstein, D. E. (2009). How well does Kirtland's warbler management emulate the effects of natural disturbance on stand structure in Michigan jack pine forests?. Forest ecology and management, 258(11), 2609-2618. Tarr, N. M., Rubino, M. J., Costanza, J. K., McKerrow, A. J., Collazo, J. A., & Abt, R. C. (2017). Projected gains and losses of wildlife habitat from bioenergy-induced landscape change. GCB Bioenergy, 9(5), 909-923. Ter-Mikaelian, M. T., & Korzukhin, M. D. (1997). Biomass equations for sixty-five North American tree species. Forest Ecology and Management, 97(1), 1-24. Thiffault, E., Paré, D., Bélanger, N., Munson, A., & Marquis, F. (2006). Harvesting intensity at clear-felling in the boreal forest. Soil Science Society of America Journal, 70(2), 691-701. Thompson I., Mackey, B., McNulty, S., Mosseler, A. (2009). Forest Resilience, Biodiversity, and Climate Change. A synthesis of the biodiversity/resilience/stability relationship in forest ecosystems. Secretariat of the Convention on Biological Diversity, Montreal. Technical Series no. 43, 67 pages. 48 Triviño, M., Pohjanmies, T., Mazziotta, A., Juutinen, A., Podkopaev, D., Le Tortorec, E., & Mönkkönen, M. (2017). Optimizing management to enhance multifunctionality in a boreal forest landscape. Journal of Applied Ecology, 54(1), 61-70. Tucker, M. M., Corace, R. G., Cleland, D. T., & Kashian, D. M. (2016). Long-term effects of managing for an endangered songbird on the heterogeneity of a fire-prone landscape. Landscape ecology, 31(10), 2445-2458. Vass, M. M. (2017). Renewable energies cannot compete with forest carbon sequestration to cost-efficiently meet the EU carbon target for 2050. Renewable Energy, 107, 164-180. Werlein, J.O. (1998). Soil survey of Crawford County. Department of Agriculture, Nat- ural Resources Conservation Service and U.S. Forest Service, MI, US, 274 pp. and 96 sheets. Zanchi, G., Belyazid, S., Akselsson, C., & Yu, L. (2014). Modelling the effects of management intensification on multiple forest services: a Swedish case study. Ecological Modelling, 284, 48-59. 49