Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity

Executive Summary November 2013 Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity

ACKNOWLEDGMENTS

This study was commissioned by the National Wildlife Federation and Southern Environmental Law Center with funds provided by Doris Duke Charitable Foundation.

Patient project management was provided by F.G. Courtney-Beauregard, Julie Sibbing, Ben Larson, and Aviva Glaser of the National Wildlife Federation, as well as David Carr and Derb Carter from Southern Environmental Law Cen- ter. Bruce Stein and Barbara Bramble from National Wildlife Federation provided important suggestions and contri- butions in technical review that greatly improved the fi nal report. Jovian Sackett from Southern Environmental Law Center provided key GIS datasets and other insights that also were critical to project development and completion.

We greatly thank Jacquie Bow, Kristin Snow, Jason McNees, and Leslie Honey of NatureServe for assistance in con- ducting and interpreting overlay analyses of at-risk (G1-G3) ecological associations. Additional assistance in develop- ing G1-G3 analyses was provided by Matt Elliott, Anna Yellin, and John Ambrose at the Georgia Natural Heritage Program; John Finnegan from the North Carolina Natural Heritage Program; and Kirsten Hazler and Karen Patter- son of the Virginia Natural Heritage Program.

Research for this project was conducted through a collaborative effort between faculty and graduate student research- ers at the University of Georgia, University of Florida, and Virginia Polytechnic Institute and State University (a.k.a., Virginia Tech University). Chapter 2, authored by Daniel Geller (University of Georgia, College of Engineering) and Jason M. Evans (University of Georgia, Carl Vinson Institute of Government), provides an overview of facilities cho- sen for the study’s focus. Chapter 3, authored by Divya Vasudev, Miguel Acevedo, and Robert J. Fletcher, Jr. (all from University of Florida, Department of Wildlife Ecology and Conservation), provides a presentation of conservation analysis methods and identifi cation of indicator species. Chapter 4, authored by Jason M. Evans, provides a techni- cal explanation of spatial modeling methods employed for the facility case studies. Chapters 5-10, authored by Jason M. Evans, Alison L. Smith (University of Georgia, College of Environment and Design), Daniel Geller, Jon Calabria (University of Georgia, College of Environment and Design), Robert J. Fletcher, Jr., and Janaki Alavalapati (Virginia Tech University, Department of Forest Resources and Environmental Conservation) provide the results and discus- sion of facility case study analyses. Chapter 11, authored by Pankaj Lal (Montclair State University, Department of Earth and Environmental Studies), Thakur Upadhyay (Virginia Tech University, Department of Forest Resources and Environmental Conservation) and Janaki Alavalapati, provides an overview of forestry biomass energy policies within state, federal, and international contexts, as well as the increasing policy attention to biodiversity concerns. Chapter 12, co-authored by all investigators, synthesizes the results of the report into a series of suggestions for policy consider- ation and future research studies. Executive Summary and Final Report layout completed by Alison L. Smith.

Christopher Stebbins, a graduate student at the University of Georgia’s College of Environment and Design, provided key technical support for developing a number of GIS analyses and map designs. Robinson Schelhas, an undergradu- ate intern at the University of Georgia, provided tireless assistance with developing maps, assembling literature data- bases, formatting tables, and taking numerous photographs. Sumner Gann, a graduate student at the University of Georgia’s College of Environment and Design, provided document layout and formatting assistance. Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity Page 1

EXECUTIVE SUMMARY

Introduction and Background construction of woody biomass electrical The southeastern United States (SE U.S.) generating facilities has been prompted by is currently experiencing what is likely the various other local, state, and federal incen- world’s most rapid growth in the develop- tives that promote development of biomass ment of woody biomass energy facilities. energy projects. Expansion of this new industry is prompt- ing wide-ranging discussion about op- The research in this report represents one portunities and risks that biomass energy of the fi rst efforts to develop spatially demands may pose for SE forest lands. This explicit analyses of long-term woodshed study, commissioned by the National Wild- sourcing for biomass energy facilities in the life Federation and Southern Environmental SE U.S. Sourcing models were developed Law Center with funds provided by Doris using advanced geographic information Duke Charitable Foundation, was developed systems (GIS) methodologies that incor- to help inform and guide this emerging porate biomass travel distances, woodshed body of basic and applied forest research. competition, and various environmental factors known to infl uence forestry biomass Growth in the woody biomass energy sec- sourcing practices. Sourcing models were tor has emerged due to a variety of state, then coupled with information on existing national, and international policy initiatives land cover types, protected lands, imper- designed to encourage renewable energy iled forest community types, and species generation. A large percentage of the new habitat information to develop comparative demands for woody biomass in the SE U.S. assessments of potential impacts that dif- is associated with the manufacturing of ferent biomass sourcing practices may have wood pellets for export to European Union across a range of SE forest ecosystems and (EU) nations. This industry is being driven wildlife species. These assessments provide by EU directives and subsidy programs that information that is highly relevant for ongo- promote biomass energy as a strategy for ing development of biodiversity protec- meeting greenhouse gas reductions, as man- tion criteria in sustainable forestry biomass dated for signatories of the Kyoto Protocol certifi cations, as well as highlighting areas to the United National Framework Conven- of future scientifi c research and monitoring tion on Climate Change. needs.

In the U.S., Renewable Portfolio Stan- Goals and Methodology dards (RPS) that phase in mandated utility The goals of this study were fourfold: purchase of renewable energy sources have served as an important driver for develop- 1. To develop spatial analyses that provide ment of woody biomass demand in some specifi c information about the likely states, including the SE states of North land cover base for long-term feedstock Carolina and Virginia. In other SE states sourcing for six woody biomass facili- that do not have RPS programs, recent ties. Page 2 Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity

2. To analyze potential effects of biomass Facility Selections sourcing scenarios on a selection of na- A total of six woody biomass facilities dis- tive wildlife species identifi ed as having tributed across the Coastal Plain, Piedmont, high conservation concern. and Mountain provinces of the SE U.S. 3. To review state, national, and interna- were selected for this analysis: tional policies related to deployment of biomass-based energy, with specifi c 1. Georgia Biomass, LLC, a wood pellet focus on sustainable sourcing criteria manufacturing facility located near Way- that pertain to wildlife habitat and bio- cross, GA in the lower Atlantic Coastal diversity maintenance. Plain. 4. To synthesize the land cover analyses, 2. Enviva Pellets Ahoskie, a wood pel- wildlife assessments, and policy review let manufacturing facility located in as a guide for future research focus and Ahoskie, NC in the upper Atlantic associated policy development. Coastal Plain. 3. Piedmont Green Power, a biomass fi red electrical generating unit located near

Wood to Bioenergy Facilities and 75-mile Woodsheds Proposed for Wildlife Impact Analyses Facility Type Cellulosic Ethanol (CE) Electric Generating Unit (EGU) Coal-Fired EGU Retrofitted for Biomass Virginia Wood Pellets (WP) All other facilities

75-mile Woodshed Buffer Virginia City Hybrid Energy Center South Boston (! Energy (! Enviva Pellets Ahoskie(!

Tennessee NorthN o r t h CCarolinao l i n a Carolina Wood Pellets,(! LLC

GeorgiaG e o r g i a SouthS o u t h Carolina Piedmont Green(! Power

Mississippi Alabama

Physiographic Province Appalachian Plateaus Georgia Blue Ridge Biomass, LLC Coastal Plain (! Interior Low Plateaus Piedmont Valley and Ridge

Florida Miles 050 100 200 Data Sources: Southern Environmental Law Center, Esri, USGS Figure 1. The six facilities chosen to model land use change and habitat impact risks Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity Page 3

Barnesville, GA in the southern reaches Land cover information for all sourcing of the Piedmont province. models was based on the United States 4. South Boston Energy, a biomass fi red Geological Survey’s 2011 Gap Analysis Pro- electrical generating unit located in gram (GAP) National Land Cover dataset. South Boston, VA and in the northern This dataset is designed for use in conserva- reaches of the Piedmont province. tion planning and assessments, which can 5. Carolina Wood Pellets, a wood pellet include large-scale evaluations of biomass manufacturing facility located in Otto, and renewable energy sourcing from forest NC and in the southern Appalachian ecosystems. mountains. 6. Virginia Hybrid Energy Center, a co- A series of customized “scenario screens” fi red coal and biomass electrical gener- were run for each facility to simulate ating unit located in St. Paul, VA and in sourcing under different sets of sourcing the southern Appalachian mountains. constraints that refl ect various protocols for sustainable forest management criteria. These selections were intended to represent Woodshed areas with public ownership sta- a wide cross-section of the different types tus or conservation easements that exclude of biomass sourcing practices and diverse extractive timber harvests were removed biodiversity conservation concerns associ- from consideration for all sourcing model ated with these different practices across the scenarios. SE U.S. Softwood sourcing Biomass Sourcing Models For plantation pine-based biomass, a series Using a comprehensive geographic infor- of fi ve scenario screens were applied for mation systems (GIS) work fl ow, we de- softwood sourcing on private lands. These veloped biomass sourcing models for each ranged from the most permissive criterion facility based that take into account two of allowing conversion of any upland land primary spatio-economic factors: 1) Road cover with the exception of row crops and transport distance of biomass material from developed areas, to the most restrictive of the forest to the facility; and 2) Competition only sourcing biomass from existing planta- with other woody biomass consumers in the tion pine forestry land covers. woodshed sourcing area. Sourcing models assumed that facilities will preferentially Ecosystem and wildlife habitat overlap source from woodshed areas that minimize assessments for softwood sourcing were costs through less road transport distance, performed on a subset of two intermediate while also minimizing bid pressure from scenario screens: 1) a permissive scenario competing biomass facilities. For softwood that allowed for conversion of natural sourcing, additional modeling consideration upland forest stands into plantation pine was given to soil type, elevation, slope, and based on landscape factors, while assuming distance to road factors that infl uence land no conversion of agricultural (i.e., row crop owner decisions for establishing plantation and pasture), developed lands, or wetland pine across the landscape. The spatially ex- areas into plantation pine; and 2) a restric- plicit integration of these disparate factors tive scenario that limited the resource base into biomass sourcing models is a novel re- of softwood sourcing to existing plantation search contribution provided by this study. pine and other disturbed lands (i.e., har- Page 4 Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity

vested, cleared, and ruderal succession) that Five landscape risk classes were defi ned are presumed to form the existing resource through this approach: 1) High; 2) Moder- base for extractive softwood forestry. ately high; 3) Moderate; 4) Moderately low; and 5) Low. Areas that were predicted to Hardwood sourcing contribute to biomass sourcing under the Two scenario screens were applied for lowest landscape area requirement (i.e., 50% hardwood sourcing on private lands. The biomass allocation) were classifi ed as High permissive screen for hardwood forestry risk, while those only predicted to contrib- assumed no restriction against sourcing ute to biomass sourcing under a residuals- from wetland and riparian forests. A more only landscape area requirement (i.e., 10% restrictive screen limited all sourcing to biomass allocation) were classifi ed as Low upland hardwood forests, and thus allowed risk. Higher risk in this context is technically no sourcing from forested wetlands. All defi ned as having a higher relative suitability hardwood sourcing screens excluded agri- for biomass sourcing based on model fac- cultural (including pasture and row crop) tors, and does not necessarily imply vulner- and developed land covers from the forestry ability to an adverse biodiversity impact biomass resource base. In two woodsheds from this sourcing. with large areas of land held publicly by the U.S. Forest Service, an additional screen that Sourcing models for each facility were inter- allowed for sourcing from all non-Wilder- sected with the 2011 GAP National Land ness National Forest lands was compared to Cover dataset to develop area-based calcula- a scenario screen that prohibited all sourc- tions of relative conversion or harvest risk ing from National Forests. for forest ecosystems assumed to serve as potential supply sources for biomass energy. Risk Assessments by Ecosystem Class These area calculations are spatially explicit Facility biomass demands and literature- by risk class, and are comparable across based estimates of local forestry productiv- woodsheds due to the standardized land ity provided the basis for calculations of units defi ned by the 2011 GAP National minimum landscape area sourcing require- Land Cover dataset. ment, which corresponds to 100% biomass allocation to the facility at the time of for- Identifi cation of At-Risk (G1-G3) Eco- estry harvest. From this minimum sourcing systems area, actual sourcing scenarios were distrib- Through a partnership with NatureServe, an uted across a consistent set of harvest in- additional intersection procedure was per- tensity and biomass allocation scenarios for formed to identify lists of forest vegetation each facility. These scenarios ranged from types (ecological associations) classifi ed as intensive sourcing of pulpwood quality critically imperiled (G1), imperiled (G2), or forestry material (50% biomass allocation) vulnerable (G3). These intersection analyses to a residuals-only sourcing practice (10% were performed for each facility woodshed biomass allocation). Sourcing model results with known conservation areas excluded based on each of these biomass allocation from consideration. assumptions are translated into maps of relative landscape risk for biomass harvest. Identifi cation of such at risk (G1-G3) asso- ciations for the purpose of avoiding adverse impacts on forests of high conservation Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity Page 5

value is a component of most sustainable Birds forest management certifi cations. However, 3. Brown-headed nuthatch (Sitta spatial resolution of G1-G3 ecological pusilla) community data is currently not suffi cient 4. Northern bobwhite (Colinus virgin- to support landscape-level analyses explicit ianus) to the risk classes identifi ed by the biomass 5. Prothonotary warbler (Protonotaria sourcing models. Lists of at risk associa- citrea) tions in each woodshed are also not directly 6. Swainson’s warbler (Limnothlypis comparable across different facility wood- swainsonii) sheds due to highly variable data availability and quality within and among the SE states.

Indicator Species Through literature reviews and dialogue with regional wildlife experts, a list of six- teen wildlife species of known conservation concern and with forest habitat dependence was developed for the SE region. Based on data availability, formal geo-spatial analyses using species distribution models available through the National Gap Analysis Program (GAP) were developed for a sub-selection of nine species. The list of indicator species analyzed through formal geo-spatial over- lays includes:

Mammals 1. Eastern spotted skunk (Spilogale putorius) 2. Long-tailed weasel (Mustela frenata) Figure 3. Brown-headed nuthatch Sitta pusilla. Photo credit: http://www.fl ickr.com/ photos/vickisnature/3297971410/

Figure 2. Eastern spotted skunk Figure 4. Prothonotary warbler Spilogale putorius Photo credit: NPS Colinus virginianus. Photo credit: Jeff Lewis Page 6 Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity

Amphibians Modeling Limitations 7. Gopher frog (Lithobates capito) It is important to note that the biomass 8. Northern cricket frog (Acris creitans) sourcing models and scenario screens in this study were developed using static assump- Reptile tions about landscape risk, the local forestry 9. Timber rattlesnake (Crotalus hor- base, and facility sourcing practices. First, ridus) models do not account for other private landowner values (e.g., hunting leases, or aesthetic enjoyment) that may be associ- ated with holding natural stands of forest in uses not associated with bioenergy produc- tion and/or competing extractive practices (e.g., pulp and paper production). Second, comparative risk assessments of ecosystems and species utilize conservative assumptions regarding complete maintenance of agricul- ture and pasture lands in existing uses (i.e., no demand-driven conversion into forestry). Third, because models are strictly based on Figure 5. Swanson’s warbler Limnothlypis existing forestry supply chains and known swainsonii. Photo credit: http://www.fl ickr.com/ photos/juliom/7158750123/ sourcing practices, they do not consider the potential for novel deployment of short ro- tation woody crops (SRWC) such as hybrid poplar (Populus sp.) as a supply response for facilities with large hardwood demands. While the results in this study do provide an objective basis for relative landscape risk comparison under the given scenario assumptions, more complex landscape deci- sion and stochastic options modeling stud- Figure 6. Northern cricket frog Acris crepitans. ies are warranted as additional data become Photo credit: http://www.fl ickr.com/photos/ available. pcoin/369987905/ Results by Modeled Facility The GAP species distribution models Georgia Biomass, LLC, located in the for these species provide spatially explicit Coastal Plain province near Waycross, GA, predictions of species habitat suitability, and is a wood pellet facility with an estimated correspond directly to the National GAP output of 750,000 Mg/yr. This facility cur- Land Cover dataset used as the basis for all rently demands 100% softwood material, biomass sourcing models. Thus, risk-based which is currently sourced from local plan- overlay area calculations for each species tation pine forestry operations. Most pellets provide a consistent basis for comparing the from this facility are exported from the Port relative potential for habitat impact based of Savannah, GA to markets in Europe for on different biomass sourcing scenarios. power generation. Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity Page 7

The sourcing model for Georgia Biomass northern bobwhite and long-tailed weasel suggests that approximately 43,500 hect- were more ambiguous, which may refl ect ares of extant native forests, including over the higher adaptability of these species to 34,000 hectares of native longleaf pine and plantation pine forestry landscapes. other pine savannah forest types, may be at high risk of conversion into plantation A restrictive scenario screen that limits forestry over the life time of the facility. The biomass sourcing to the existing base of current land cover base for plantation pine plantation pine, disturbed, and ruderal forestry in this facility’s woodshed appears forestry lands strongly suggests that the suffi cient for meeting long-term softwood facility’s long-term biomass demands can demands if biomass sustainability criteria be met without primary sourcing or future that prohibit natural forest stand conversion conversion of extant natural forest stands. are adopted. Due to the large existing base of plantation forestry lands in this woodshed, the full suite of sourcing model runs suggests that The softwood sourcing model for Georgia between 73%-76% of the facility’s softwood Biomass suggests that approximately 43,500 biomass demand would be met by current hectares of native forests could be at high plantation pine and disturbed forestry lands risk of conversion into plantation pine for- even when assuming no restriction against estry when assuming the permissive scenar- natural stand conversion. io screen that provides no restriction against natural stand conversion. Over 34,000 Enviva Pellets Ahoskie, located in the hectares of this high risk area is classifi ed as Coastal Plain province near Ahoskie, NC, longleaf pine and other upland pine for- is a wood pellet facility with an estimated est ecosystems. Due to large-scale historic output of 350,000 Mg/yr. This facility conversion of longleaf and other upland reports a mixed sourcing of 80% hardwood pine forests into agricultural, forestry, and and 20% softwood. Materials include lower developed land covers, remaining examples quality stemwood and residual material of these ecosystem types in the SE region associated with local logging operations. are widely recognized for the high conserva- Based on the existing resource base, low tion value provided to many native animal quality stemwood and residuals from natu- and plant species. rally regenerating forest stands are expected to provide the supply base for hardwood Most indicator species analyses showed material. Softwood material will likely be higher areas of impacted habitat under scenarios where natural stand conversion was allowed as compared to scenarios where The land cover base and sourcing model in sourcing was constrained to the existing the Enviva Pellets Ahoskie woodshed suggest forestry land cover base. Particularly high that between 46% - 63% of the long-term amounts of relative habitat risk were shown sourcing area for hardwood biomass will for the eastern spotted skunk, gopher frog, be composed of forested wetlands. Over timber rattlesnake, brown-headed nuthatch, 68,000 hectares of wetland forest in this and Swainson’s warbler, all of which may be woodshed may be at high risk of biomass generally expected to show negative re- sourcing and may be at risk over the facility sponses when natural upland forest habitat life time. is converted into plantation pine. Results for Page 8 Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity

composed of low quality stemwood from which is a direct function of these taxa local pine plantation forestry operations. showing strong fi delity to bottomland for- Most pellets from this facility are exported est habitat in the Enviva Ahoskie wood- from the Port of Chesapeake, VA to mar- shed area. Similarly, higher habitat overlays kets in Europe for power generation. under no wetland sourcing scenarios for the brown-headed nuthatch, northern A sourcing model based on 80% hardwood bobwhite, and long-tailed weasel refl ect the and 20% softwood was developed for the upland habitat preference of these species. Enviva Pellets Ahoskie facility. Results for the hardwood component of the model Literature for most indicator species sug- suggest that between 46%-63% of the gests a number of habitat degradation biomass sourcing area would be composed concerns related to increased sourcing of of forested wetlands under the scenario primary and residual biomass from hard- screen where wetland forests are assumed wood forests, particularly on wetland sites. as available for bioenergy supply. Forested However, selective thinning of overstory wetland ecosystems predicted to have the and understory hardwoods on upland sites highest risk for bioenergy sourcing include was identifi ed as a potential strategy for over 48,000 hectares of riverine fl oodplain enhancing habitats for both the northern forests and almost 21,000 hectares of iso- bobwhite and brown-headed nuthatch. lated basin swamps not directly associated Direct and secondary impacts to aquatic with river systems. ecosystems, as well as unpredictable climate interactions with forest regeneration, are Hardwood models that assumed no wetland additional concerns associated with wetland sourcing for the Enviva Pellets Ahoskie forest harvesting. Specifi c habitat changes facility suggest that long-term biomass in wetland and upland harvest practices demands likely cannot be met through a specifi cally associated with biomass energy residuals-only sourcing strategy (i.e., 10% - sourcing in the Enviva Pellets Ahoskie 12.5% biomass harvest allocation) that relies woodshed will require additional research to solely upon existing stands of natural up- resolve more fully. land hardwood forests. Afforestation with fast growing hardwood biomass crops, such The softwood component of the model as hybrid poplar (Populus sp.), on marginal suggests that a little over 1,600 hectares pasture and crop lands may be a potential of native forests could be at high risk of strategy for long-term supply of this facility conversion into plantation pine forestry if a no wetlands sourcing policy is required. when assuming a permissive scenario screen that provides no restriction against natural Although comparison of indicator species stand conversion. Almost 1,500 hectares overlay areas under the “wetlands allowed” of this high risk area is classifi ed as mesic and “no wetlands” sourcing screens showed hardwood forests, with less than 30 hectares a large amount of differences, these area classifi ed as longleaf pine. Based on these differences are generally related to wetland results, it was determined that the softwood vs. upland habitat preferences. For example, demands of Enviva Pellets Ahoskie likely the prothonotary warbler and Swainson’s do not pose a major conservation concern warbler showed much higher habitat area for natural upland forest stands within the overlays under wetland sourcing scenarios, local woodshed area. Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity Page 9

Piedmont Green Power, located in the of impacted habitat under scenarios where Piedmont province near Barnesville, GA, is natural stand conversion was allowed to a 60.5 MW electric generating unit. Electri- occur. All of these species can be expected cal production by this facility is entirely for to show negative responses when natural use in the local and regional utility market. upland forest habitat is converted into plan- This facility sources low quality stemwood tation pine. By contrast, the northern bob- and residual material from local logging white consistently showed relatively higher operations. Based on local forestry practices areas of impact under scenarios where and the existing resource base, plantation sourcing was restricted to plantation pine pine-based softwood material is expected to and disturbed forestry lands. This result is provide the long-term biomass supply for likely refl ective of the northern bobwhite’s this facility. preference for relatively open overstory and understory conditions. Results for the long- The sourcing model suggests that approxi- tailed weasel and northern cricket frog were mately 48,500 hectares of extant native ambiguous, perhaps refl ecting the adaptabil- forests may be at high risk of conversion ity of these species to plantation pine forest into plantation forestry over the life time landscapes. of the Piedmont Green Power facility. The current land cover base for plantation pine A restrictive softwood scenario screen that forestry in this facility’s woodshed appears limits biomass sourcing to the existing base suffi cient for meeting long-term softwood of plantation pine, disturbed, and ruderal demands if biomass sustainability criteria forestry lands indicates that the facility’s that prohibit natural forest stand conversion biomass demands can be met without are adopted. conversion of natural forest stands. How- ever, the full suite of sourcing model runs The softwood sourcing model for Piedmont suggests that a little less than half (45%- Green Power suggests that approximately 49%) of softwood biomass would be met 48,500 hectares of native forests could be by the existing base of plantation pine and at high risk of conversion into plantation disturbed forestry lands when assuming pine forestry when assuming a permissive no restriction against natural forest stand scenario screen that provides no restriction conversion. While this conversion model against natural stand conversion. Almost likely overstates absolute risk due to other 41,000 hectares of this high risk area is clas- landowner values associated with holding sifi ed as southern piedmont dry oak forest forest land in hardwood forest, this result is types, with the remainder composed of nevertheless suggestive of potential long- southern piedmont mesic forest. Historic term land cover change without adoption and ongoing conversion of piedmont hard- of sourcing criteria that are protective wood forests into plantation pine forestry is against the conversion of existing natural a conservation concern in this woodshed. stands of piedmont forest.

The brown-headed nuthatch, Swainson’s warbler, eastern spotted skunk, and timber rattlesnake showed substantially higher areas Page 10 Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity

South Boston Energy, located in the can be readily met through a residuals-only Piedmont province near South Boston, sourcing strategy (i.e., 10% - 12.5% biomass VA, is a proposed 49.95 MW power facil- harvest allocation). ity. Feedstocks are expected to include low quality stemwood, slash, wood wastes, and Comparison of indicator species overlay re- wood chips and slash. Based on the exist- sults showed the potential for substantially ing forestry resource base in the woodshed, higher habitat impacts for the Swainson’s a mixture of 50% hardwood from natural warbler, northern cricket frog, and timber stands and 50% softwood from pine planta- rattlesnake under scenarios where wetland tions is assumed as the long-term biomass sourcing was allowed. Each of these spe- sourcing strategy for this facility. Therefore, cies, and various other riparian-dependent a sourcing model based on 50% hardwood species, would generally be expected to re- and 50% softwood was developed for the act negatively to intensive biomass sourcing South Boston Energy facility. from wetland forest habitats in this region of the piedmont province. Other taxa with The land cover base for the South Boston upland preferences showed little difference Energy facility suggests a 50% hardwood to in overall habitat area overlays between 50% softwood sourcing strategy. The sourc- “wetland allowed” and “no wetlands” ing model suggests that approximately 5%- sourcing scenarios. While biomass thin- 6% of the hardwood resource base could ning practices on upland hardwood forests be composed of wetland forests, but that may potentially have habitat enhancement suffi cient upland hardwood area is available benefi ts for northern bobwhite and brown- if sustainability criteria that exclude wetland headed nuthatch, species-level responses sourcing are adopted. The model suggests to upland harvest practices associated that approximately 6,700 hectares of with biomass energy sourcing in the South extant native forests may be at high risk of Boston Energy woodshed will require ad- conversion into plantation forestry to meet ditional research to resolve more fully. softwood demand. However, suffi cient areas of extant plantation pine are available Results for the softwood component of the in the woodshed to source this facility if model results suggest that approximately biomass sustainability criteria that prohibit 6,700 hectares of native forests could be natural forest stand conversion are adopted. at high risk of conversion into plantation pine forestry when assuming a permissive Results for the hardwood component of scenario screen that provides no restriction the model suggest that between 5%-6% of against natural stand conversion. Almost the biomass sourcing area would be com- 6,400 hectares of this high risk area is clas- posed of forested wetlands under a scenario sifi ed as southern piedmont dry oak forest screen where wetland forests are assumed as types, with the remaining 300 hectares clas- available for bioenergy supply. The highest sifi ed as southern piedmont mesic forest. risk areas for bioenergy sourcing include However, the restrictive softwood scenario approximately 4,900 hectares of piedmont screen that limits biomass sourcing to the fl oodplain forests. Hardwood sourcing existing base of plantation pine, disturbed, models that assumed no wetland sourcing and ruderal forestry lands indicates that the for the South Boston Energy facility suggest facility’s biomass demands can very likely be that long-term hardwood biomass demands met without further conversion of natural Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity Page 11

forest stands. Sourcing models suggest that production of 68,000 Mg/yr. This facility between 78%-84% of the softwood bio- sources 100% hardwood material from low mass would be met by the existing base of quality stemwood, logging residual materi- plantation pine and disturbed forestry lands als, and waste material from construction even when assuming no restriction against sources. All stemwood and residual ma- natural stand conversion. terials are assumed to come from natural hardwood stands. Most hardwood pellets The brown-headed nuthatch, Swainson’s from this facility are bagged and sold in the warbler, and timber rattlesnake showed sub- domestic U.S. market for use in home heat- stantially higher areas of impacted habitat ing stoves. under scenarios where natural stand con- version was allowed to occur. All of these The hardwood sourcing model for the species can be expected to show negative Carolina Wood Pellets facility suggests that responses when natural upland forest habi- between 1%-3% of the biomass sourc- tat is converted into plantation pine. Similar ing area would be composed of forested to the results for Piedmont Green Power, wetlands under a scenario screen where overlays for the long-tailed weasel and wetland forests are assumed as available for northern cricket frog showed ambiguous bioenergy supply. Somewhat larger areas differences between the conversion allowed of potential riparian wetland forest impact and no conversion scenarios. are predicted for scenarios where sourc- ing is excluded from all Natural Forest Carolina Wood Pellets, located in the lands, with approximately 780 hectares of Mountain province near Otto, NC, is a Mountain riparian wetlands and 20 hectares wood pellet facility with an estimated of Piedmont riparian wetlands identifi ed as having high risk. Because of this facil- The sourcing model for the Carolina Wood ity’s comparatively low biomass demand, Pellets facility suggests that this facility’s hardwood sourcing models that assumed demands can be met with residual hard- no wetland sourcing for the Carolina Wood wood biomass from the local woodshed, Pellets facility indicate that long-term hard- including in scenarios where all wetlands wood biomass demands can be readily met and U.S. Forest Service lands are assumed through a residuals-only sourcing strategy as unavailable. Avoidance of riparian areas (i.e., 10% - 12.5% biomass harvest alloca- appears particularly important for biodi- tion) on upland forests only. versity protection, as it is expected that the northern cricket frog and other amphibian The northern cricket frog is the only indica- species in this woodshed would show signifi - tor species that shows a consistent effect of cant negative responses to biomass sourcing higher overlay areas under scenarios where from riparian forest habitats along moun- wetland sourcing is allowed. It is expected tain streams. Other wildlife and biodiversity that the northern cricket frog and other am- concerns for this facility relate to uncertain phibian species in the Carolina Wood Pellets taxa and landscape responses to increased woodshed would show signifi cant negative residual harvest pressure for biomass sourc- responses to biomass sourcing from riparian ing from Appalachian forests. These issues forest habitats along mountain streams. require additional fi eld research to resolve more fully. Page 12 Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity

As with other facilities, forestry thinning The sourcing model for the Virginia City practices associated with biomass extraction Hybrid Energy facility suggests that biomass on upland hardwood forests may poten- demands can likely be met with residual for- tially have habitat enhancement benefi ts estry from the local woodshed, including in for northern bobwhite and brown-headed scenarios where all wetlands and U.S. Forest nuthatch. However, species-level responses Service lands are assumed as unavailable. to increased removal of residual material, Avoidance of riparian areas for sourcing particularly downed woody matter (DWM), this facility appears particularly important from Appalachian forests is a noted habitat for biodiversity protection due to relative concern that will require additional research rarity of these habitats in the local land- to resolve more fully. scape, as well as the potential for severe stream erosion and habitat degradation The sourcing models suggest that exclusion effects with riparian logging practices in the of non-Wilderness National Forest stands Appalachian Mountains. Other wildlife and from the biomass sourcing land base for biodiversity concerns for this facility relate the Carolina Wood Pellets could have the to uncertain taxa and landscape responses effect of transferring much of the facil- to increased residual harvest pressure for ity’s hardwood supply to piedmont forests. biomass sourcing from Appalachian forests. This result was surprising, and may be an The relatively high biomass demands of artifact of the transport model not includ- this facility compared to other Appalachian ing increased fuel costs associated with biomass energy sites make such concerns up-gradient elevation. especially notable.

Differences in areal overlays for ecosystem MW). Electrical production by this facility and indicator species in scenarios that al- is entirely for use in the local and regional lowed for non-Wilderness National Forests utility market. This facility sources low qual- versus those that excluded all National ity stemwood and logging residues from the Forests are largely a function of ecological local area. Based on the existing forestry differences between piedmont and moun- resource base, it is likely that most sourcing tain ecosystems. Further research will be will be composed of low quality stemwood needed to provide additional insights into and residuals from naturally regenerating the stand-level and landscape tradeoffs be- hardwood stands. tween sourcing strategies that may include National Forests as compared to those that The sourcing model for the Virginia Hybrid are restricted to private landholdings. Energy Center facility suggests that less than 1% of the biomass sourcing area Virginia City Hybrid Energy Center, would be composed of forested wetlands located in the Mountain province near St. under scenarios where wetland forests are Paul, VA, is a 585 MW electrical genera- assumed as available for bioenergy sup- tion unit operated by Dominion Virginia ply. These results are generally consistent Power. This facility is designed to co-fi re up among scenarios where non-Wilderness to 20% biomass in its coal-fuelled electric Natural Forest lands are assumed as avail- production facility, although is operationally able for harvest, as well as scenarios where running on a 10% biomass capacity (~59 all National Forest lands are excluded. Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity Page 13

Likely due to this very low area of riparian particularly downed woody matter, from habitat overlay, wildlife indicator species Appalachian forests is a noted concern that showed little difference in scenarios where will require additional research to resolve wetland harvest was allowed as compared more fully. to those restricted to upland forestry areas. However, it is generally expected that high Policy Summary intensity biomass harvest from riparian The European Commission is currently de- areas in this woodshed could have high bating sustainability certifi cation directives habitat and biodiversity impacts, particularly for all bioenergy pathways being imple- due to the relative rarity of these habitats mented under the Kyoto Accord. Draft within the local landscape. proposals and current policy discussions suggest that Sustainable Forestry Manage- Inclusion of non-Wilderness National For- ment (SFM) standards may be required for est in the biomass sourcing area resulted in all imported forestry biomass utilized in EU somewhat higher areas of predicted overlay electrical generation. impact for the Swainson’s warbler. While this species can be attracted to small-scale Maintenance and enhancement of biodiver- disturbances, it generally prefers disturbanc- sity through forestry sourcing is a funda- es found within the context of otherwise mental objective in SFM defi nitions applied unfragmented hardwood mountain forest in EU nations. Because of the increasing habitats. Residuals-only sourcing demands importance of EU demands on the biomass for the Virginia Hybrid Energy Center energy market in the SE U.S., European suggest potential long-term impact on over SFM standards may have signifi cant infl u- 50% of the Swainson’s warbler habitat in ence on the development of SFM for other this woodshed. Monitoring of potential biomass energy procurement in this region. impacts from different biomass harvest practices may be particularly warranted for The EU broadly defi nes SFM as use of the Swainson’s warbler in this woodshed. forest lands in a way that “maintains their biodiversity, productivity, regenerative capac- Exclusion of non-Wilderness National For- ity, vitality and their potential to fi ll, now and ests from the biomass sourcing area consis- in the future, relevant ecological, economic tently resulted in somewhat higher areas of and social functions, at local, national, and predicted overlay impact for the northern global levels” (European Commission 2013, bobwhite and long-tailed weasel. This result pg. 11). is most directly related to increased per- centages of piedmont forests serving as a The U.S. forestry sector has a number of supply source for this facility when National existing sustainable forest management Forests are excluded. While increased bio- (SFM) certifi cation programs, including the mass harvest on private forestry lands may Forest Stewardship Council (FSC), Sustain- potentially have habitat enhancement values able Forestry Initiative (SFI), American Tree for northern bobwhite, potential effects on Farm System (ATFS), and the Program on long-tailed weasel populations are less clear the Endorsement of Forest Certifi cation and would require additional research. As (PEFC). These programs are voluntary for noted previously, species-level responses to forest landowners, and only 17% of SE increased removal of residual material, and forestry lands are presently certifi ed through Page 14 Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity

one of the major sustainability certifi cation the regional nature of this study and the programs. FSC certifi cations are generally idiosyncratic or even unknown responses of regarded as the most protective of remain- wildlife taxa to variable sourcing practices, ing natural forest stands and associated it is important to note that it not possible wildlife habitat in the working forestry to make fi rm conclusions of impact that landscape, although all of the major SFM would apply to all sites, species, and harvest certifi cation programs in the U.S. do contain practices. However, the results of this study specifi c biodiversity protection criteria. do support several generalizations that can There currently is no SFM certifi cation in inform policy discussions and research the U.S. that specifi cally applies to bioenergy priorities for promoting increased sustain- production. However, recommendations for ability of forestry biomass energy moving minimizing biomass energy sourcing im- forward. pacts on wildlife habitat have recently been developed by the Forest Guild. 1. The primary conservation concern for softwood biomass sourcing in the SE State level best management practices Coastal Plain and Piedmont provinces (BMPs) that provide specifi c guidance for is land cover change away from exist- federal regulatory compliance are widely ing natural forest stands into planta- applied throughout the SE forestry industry. tion pine forestry. Such conversion is While such BMPs are implemented on a historically known as a primary factor in voluntary basis, they are specifi cally de- the loss or degradation of much wildlife signed to prevent erosion and sedimentation habitat in the SE U.S., including for violations of the Clean Water Act that are many species of conservation concern. more likely to occur if BMPs are not fol- Using existing land cover as a base, lowed in forestry operations. South Carolina woodsheds with relatively large extant is the fi rst SE state to develop specifi c BMP areas of plantation pine and ruderal criteria for forestry bioenergy sourcing, and forestry lands generally pose less con- forestry commissions in other SE states are cern for future biodiversity impacts as considering similar BMP approaches for compared to those with relatively large biomass energy. The South Carolina BMPs areas of natural forest stands. provide general assurances against Clean Water Act violations, but are not designed 2. Biomass thinning for energy produc- or intended to maintain overall forest bio- tion may in many cases provide wildlife diversity and wildlife habitat at either the habitat enhancement within the extant stand or landscape level. Forest Guild guid- plantation pine forestry landscape in the ance for minimizing the impacts of forest Coastal Plain and Piedmont, particularly biomass sourcing on biodiversity is included in cases where lack of other markets as a voluntary set of recommendations in has resulted in landowner neglect of the South Carolina BMP manual. planted forests. Such thinning prac- tices can provide some structural and Synthesis and Conclusions functional simulation of longleaf pine This study represents one of the fi rst ecosystems that were historically a detailed analyses of biodiversity and habitat dominant upland habitat throughout concerns associated with forestry biomass the SE. energy in the SE U.S. region. Because of Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity Page 15

3. In the Coastal Plain, the most bio- may in some cases have habitat en- logically productive hardwood forests hancement effects, particularly when are generally located in bottomland coupled with understory thinning. areas, including riparian fl oodplain However, biomass harvest practices that and isolated basin wetland systems. result in large-scale reductions of cavity Large-scale hardwood biomass sourcing trees, snags, and downed woody matter within the Coastal Plain sites therefore are a habitat concern for a wide variety will generally imply substantial logging of wildlife taxa found in Piedmont and pressure on natural wetland forests. Mountain hardwood forests. While wetland forestry BMPs such as streamside riparian buffers and erosion 6. Specifi c fi eld research is required to control measures are available in the SE better understand long-term habitat U.S., unique challenges associated with responses associated with different bio- degradation of wildlife habitat, local mass management regimes, and to com- and downstream water quality impacts, pare these responses to control regimes and uncertainties of natural stand that are not sourced for biomass. Such regeneration after intensive harvests of research will be critical for the long- SE bottomland and fl oodplain forests term co-management of SE forestry are all documented within the ecological ecosystems for both wildlife habitat literature. However, sourcing of upland maintenance and sustainable bioenergy hardwood biomass in the Coastal production. Plain may in some cases have habitat enhancement effects, particularly on 7. State-level BMP guidelines for forestry sites that have experienced hardwood operations are not designed for the en- succession away from pines due to fi re hancement or maintenance of biodiver- exclusion. sity at stand or landscape levels, but in- stead to ensure compliance with federal 4. In the Piedmont and Mountain prov- regulations related to the Clean Water inces, relatively low amounts of hard- Act. Biomass sustainability policies that wood forest land are contained within aspire to be protective or restorative of wetland forests. There are substantial landscape biodiversity and native forest concerns regarding erosion, sedimen- vegetation types of high conservation tation, and wildlife habitat impacts value are likely to require additional associated with riparian and wetland certifi cation regimes and compliance forest harvesting in both the Piedmont procedures beyond those provided by and Mountain provinces. Avoidance of BMP guidelines. such wetland areas for biomass sourc- ing through riparian buffer strips can be 8. The U.S. forestry sector has a number recommended with minimal effects on of existing sustainable forest manage- overall hardwood biomass supply in the ment (SFM) certifi cation programs, in- Piedmont and Mountain provinces. cluding the Forest Stewardship Council (FSC), Sustainable Forestry Initiative 5. Moderate biomass sourcing of upland (SFI), American Tree Farm System hardwood biomass from the Piedmont (ATFS), the Program on the Endorse- and, to a lesser extent, the Mountains ment of Forest Certifi cation (PEFC), Page 16 Forestry Bioenergy in the Southeast United States: Implications for Wildlife Habitat and Biodiversity

and more recently the Roundtable on 10. Biomass energy sourcing from natural Sustainable Biomaterials. These pro- stands of hardwood forests poses a grams are voluntary for forest landown- more complex set of potential impacts ers, and only 17% of SE forestry lands for wildlife habitat and associated for- are presently certifi ed through one of est sustainability certifi cation regimes. the major sustainability certifi cation Specifi c recommendations for reducing programs. There currently is no SFM habitat impacts from biomass sourc- certifi cation in the U.S. that specifi cally ing in natural hardwood stands have applies to bioenergy production. recently been developed by the Forest Guild (2012). These recommendations 9. While all of the major SFM certifi cation generally focus on retaining suffi cient programs in the U.S. contain biodiver- snags, cavity trees, and downed woody sity protection criteria, FSC certifi ca- matter to maintain opportunities for tions are generally regarded as the most wildlife habitat regeneration. However, protective of remaining natural forest rates of voluntary compliance with stands and associated wildlife habitat in these practices are currently unknown, the working forestry landscape. In par- as are specifi c wildlife responses to rec- ticular, the FSC Controlled Wood and ommended and actual biomass harvest Forest Management Standards provide practices across different habitat and restrictions against conversion of extant ecosystem types. Additional research natural stand forests into plantation for- will be required to more fully resolve estry. If adopted for softwood sourc- these questions, and thus provide ing biomass energy facilities that rely adaptive guidance for the long-term upon plantation pine feedstocks, these protection of biodiversity and wildlife FSC standards would be expected to resources under sustained forestry bio- offer a high level of protection against mass sourcing. biodiversity degradation associated with natural stand conversions to plantation forestry.