Potential Effects of a Forest Management Plan on Bachman's by gqt76194


									Potential Effects of a Forest Management Plan on
Bachman's Sparrows (Aimophila aestivalis): Linking
a Spatially Explicit Model with GIS
Institute of Ecology, University of Georgia, Athens, GA, 30602, U.S.A.

Abstract: By combining a spatially explicig individual-based population simulation model with a geo-
graphic information systerrg we have simulated the potential effects o f a US. Forest Service management p l a n
on the population dynamics o f Bachman's Sparrow (Aimophila aestivalis) at the Savannah River Site, a U.S.
Department o f Energy facility in South Carolir~ Although the Forest Service's management p l a n explicitly
sets management goals f o r many specie& most o f the prescribed management strategy deals with the endan-
gered Red-cockaded Woodpecker (Picoides borealis) because o f legal requirements. We explored h o w a species
(the sparrow) that is not the target o f specific management strategies but that shares some habitat require-
ments with the woodpecker, would fare under the management plar~ We f o u n d that the major components
o f the proposed management p l a n may allow the sparrow population to reach and exceed the m i n i m u m
management goal set f o r this specie& but only after a substantial initial decline in sparrow numbers and a
prolonged transition period, In the model, the sparrow population dynamics were most sensitive to demo-
graphic variables such as adult and juvenile survivorship and to landscape variables such as the suitability
o f y o u n g clearcuts and mature pine stands. Using various assumptions about habitat suitability, we esti-
mated that the 50.year probability o f population extinction is at least5% or may be much higher if juvenile
survivorship is low. We believe, however, that modest changes in the management plan might greatly increase
the sparrow population and presumably decrease the probability o f extinctior~ Our results suggest that
management plans focusing on one or a f e w endangered species may potentially threaten other species o f
management concerr~ Spatially explicit population models are a useful tool in designing modifications o f
management plans that can reduce the impact on nontarget species o f management concer~

Impactos potenciales de u n plan de manejo del bosque sobre los gorriones de Bachman (Aimophila
aestivalis): vinculando un modelo espacialmente explicito con SIG
R e s u m e n : Hemos simulado los impactos potenciales de un plan de manejo del Servicio Forestal de los
Estados Unidos, sobre la dindmica poblacional del gorri6n de Bachman (Aimophila aestivalis) en un sitio de
Savannah River, que es propiedad del Departamento de Energia de los Estados Unido& en Carolina del Sur
La estrategia utilizada combtn6 un modelo de simulaci6n espacialmento explicito de las individuos miem.
bros de la poblaci6rg con un Sistema de Informaci6n Geogrdfico. Si bien el plan de manejo del Servicio
Foresta~ explicitamente establece medidas de m a n g o para varias especie& debtdo a requerimientos legales
la mayoria de las estrategias de manejo prescripta~ consideran al pdjaro carpintero de cucarda coloradd
(Picoides borealis) en peligro de extinci6r~ Exploramos como se desenvolveria dentro del plan de manejo,
una especie (el gorri6n), que no estd en la mira de ninguna estrategia especifica de manc~o, pero que
comparte algunos de los requerimientos de habitat con el pdjaro carpintero. Encontramos que los compo-
nentes mds importantes del p l a n de rnanejo le permiten a la poblaci6n de gorriones llegar y exceder la metal
m i n i m a trazada para esta especie, pero s6lamente despuds de una disminacion sustancial en el nt~mero de
gorriones y de u n p r o l o n g a d o periodo de transici6n. En el modelo, la d i n d m i c a de p o b l a c i 6 n

* Current address: Department of Fisheries and Wildlife, Michigan State University, East Lansing 311 48824.
~Current address: The National Biological Survey, Department of the Interior, Matlstop 3070-MIB, 1849 C Street N.W., Washington, DC 20240,
Paper submitted February 24, 1993; revised manuscript accepted March 1, 1994.

ConservationBiology,Pages 62-75
Volume9, No. 1, February 1995
£iu et al.                                                                            Spatially l!gplicttModels and Forest Management      63

del gorrl6n f u e mds sensible a variables den~grdflcag tales como supervit~ncta de adultos y jut~nileg y
variables del paisaj~ tales como la aptitud de los clareos recientes del bosque y los rodales de p i n o s maduro~
Utilizando varlas suposiciones sobre la aptitud del hdbita~ estimamos que la probabilidad de extinci6n al
cabo de 5 0 afzos es de p o r 1o menos un 5% pudiendo set rads aIta si la supervivencla de juveniles baja Una
expansi6n de un programa del Servicio Forestal podrla incrementar ampllamente la poblaci6n de gorriones
y presumiblemente disminuir la probabilidad de extinci6n` Nuestros resultados sugieren que los planes de
manejo que se centran en una o en unas pocas especies en peligro de extincibrg pueden potenclalmente p o n e r
en riesgo otras especies cuyo manejo es de importancia Los modelos poblacionales espaclalmente explicito~
constituyen una herraraienta (~til en el disetio de modificaciones de planes de mane]o, pudiendo reducir el
impacto en especies cuyo manejo es de imporlanci~ pero que no constituyen el objetivo principal del plan.

Introduction                                                         changes and forest management practices on animal
                                                                     populations that inhabit changing landscapes (Pulliam
l a n d mangers face a major challenge in understanding
                                                                     et al. 1992; Liu 1992; 1993a~ 1993b; Liu et al. 1994). In
the long-term effects of management strategies on the
                                                                     this paper w e link the latest version of this model with
population dynamics of animal and plant species. Many
                                                                     a geographic information system (GIS) database of the
aspects of m a n a g e m e n t strategies will have conse-
                                                                     Savannah River Site, a Department of Energy facility in
quences that are difficult to predict prior to the imple-
                                                                     Aiken and Barnwell counties, South Carolina. The U.S.
mentation of a management plan. This is especially true
                                                                     Forest Service manages the majority of the 770-kin 2 site
of management options that affect habitat distribution
                                                                     for timber production and biodiversity conservation.
and quality o v e r large spatial scales or that cause
                                                                     The Forest Service's Savannah River Forest Station
changes that unfold over long periods of time. For ex-
                                                                     (SRFS) recently developed a long-term management
ample, increases in timber-harvest rotations (the age at
                                                                     plan called the "Savannah River Site Wildlife, Fisheries,
which trees are harvested) in a forest district managed
                                                                     and Botany Operation Plan" (SRFS 1992), hereafter
for timber production will change the age distribution
                                                                     called the "operation plan". The operation plan gives
of the f o r e s t and the a m o u n t and distribution of
                                                                     general and specific management goals in quantifiable
clearcuts p r o d u c e d t h r o u g h t i m b e r harvest. These
                                                                     terms, such as the number of animals desired of each
changes will affect both the organisms found in forest
                                                                     endangered, threatened, or key species. The operation
patches of different ages as well as early successional
                                                                     plan also proposes a series of manipulations, such as
species found in the clearcuts, but the impact on the
                                                                     changing the composition of forest stands over a 50-year
species may not be identifiable for years. It is difficult
                                                                     period, that are designed to achieve specific manage-
prior to the implementation of management changes to
                                                                     ment goals. We incorporated the basic components of
predict what the effects will be and h o w to mitigate
                                                                     the forest management plan in a series of simulations
severe effects before they occur. This is especially true
                                                                     designed to test their effects on the population dynam-
w h e n field experiments are hard to design and imple-
                                                                     ics of Bachman's Sparrow ( A i m o p h i l a a e s t i v a l i s ) , a po-
ment at the p r o p e r spatial and temporal scales.
                                                                     tentially threatened species found in pine woodlands.
    Spatially explicit population models provide a tool for
                                                                     We w e r e able to simulate the primary c o m p o n e n t s
exploring the possible effects of management plans on
                                                                     (such as harvesting, burning, and thinning) of the oper-
selected species prior to implementation (Dunning et
                                                                     ation plan but not some secondary components (such as
al. 1995). Population models can be designed to simu-
                                                                     retention of standing snags or cavity trees and applica-
late the life history of a species on a specific landscape
                                                                     tion of herbicides). Thus, our simulations should be
and to measure population change resulting from spe-
                                                                     considered suggestive of the effects of the overall man-
cific changes to the landscape. Models can be used t o
                                                                     agement strategy but not precise predictions of h o w the
study landscape changes that cover large spatial areas
                                                                     actual sparrow populations will change. We also discuss
and occur over long periods. Although the predictive
                                                                     the implications of our results for management of bio-
ability of such models is limited by their assumptions
                                                                     diversity and address the potential danger of using
and the researchers' ability to validate and verify the
                                                                     knowledge of habitat requirements of the Red-cockaded
model (Conroy et al. 1995), spatially explicit models
                                                                     Woodpecker to infer the needs of other species such as
can be used to suggest possible population responses to
                                                                     Bachman's Sparrows.
landscape change and therefore may be especially useful
in examining w h e t h e r management strategies designed
to meet particular goals have detrimental effects on non-
target wildlife species.                                             The Simulation Model and Geographic Information System
    As an example of this approach, w e have developed a             ECOLECON is an ecological-economic model that is ca-
class of spatially explicit, individual-based simulation             pable of simulating animal population dynamics and
models to e x p l o r e the possible effects of land-use             economic revenues in response to different forest land-

                                                                                                                Conservation Biology
                                                                                                                Volume 9, No. 1, February 1995
64              Explicit Models and Forest Management
        Spatially                                                                                                                           ldu et al.

scape structures and various timber management sce-                          cility of the U.S. Department of Energy and an "experi-
narios (Liu 1992, 1993b). ECOLECON is spatially ex-                          mental forest" managed by the U.S. Forest Service (see
plicit and o b j e c t - o r i e n t e d . It w a s p r o g r a m m e d in   Dunning & Watts 1990). Past silvicultural practices
Borland C + + 2.0 (Borland International 1991 ) and can                      have p r o d u c e d many even-aged forest stands ranging
be i m p l e m e n t e d on IBM-compatible computers.                        from 5 to 100 ha in size. A stand is a patch of forest in
   ECOLECON's e c o n o m i c information includes subrou-                   which the trees are of the same age and that has b e e n
tines for forest growth and yield, income from timber                        subjected to uniform management. The largely forested
harvest, cost, and net income. Its ecological compo-                         landscape consists of uplands dominated b y loblolly
nents integrate information about animal habitat use,                        pine (Pinus taeda) and longleaf pine (P. palustris), bot-
demography, and dispersal. All of these subroutines are                      t o m l a n d w e t l a n d s d o m i n a t e d b y s w e e t g u m (Liq-
hierarchically structured (Liu 1992, 1993b). The model                       uidambar sOwaciflua) and yellow poplar (Lirioden-
predicts animal population dynamics, spatial distribu-                       dron tulipifera), and s w a m p s d o m i n a t e d by b a l d
tions, and extinction probabilities, as well as forest                       c y p r e s s (Taxodium distichum) and w a t e r t u p e l o
growth and yield and e c o n o m i c income from timber                      (Nyssa aquatica) (Dukes 1984). The region was largely
harvest. The outputs of the model provide valuable in-                       agricultural w h e n the Federal g o v e r n m e n t assumed con-
formation for balancing the conflicts b e t w e e n the gen-                 trol of it in the early 1950s. The Forest Service planted
eration of e c o n o m i c revenues and the conservation of                  the former agriculture fields with pines that are n o w
endangered species.                                                          mostly 3 0 - 5 0 years old and that w e r e being harvested
    ECOLECON is a "second generation" model that is                          under the previous m a n a g e m e n t plan. Because of the
built on BACHMAP, a spatially explicit model of the                          historical use of the land, relatively few mature pine
population dynamics of Bachman's Sparrows. BACHMAP                           stands (older than 80 years) are currently found on the
is one of a class of n e w models, referred to as MAP                        Savannah River Site. Mature longleaf pine stands cover
models, an a c r o n y m for Mobile Animal Population                        only 0.2% of the total forested area on the site, while
(Pulliam et al. 1992). ECOLECON adds m o r e features                        intermediate-aged ( 3 0 - 8 0 years old) stands of loblolly
(such as dispersal rules) and flexibility to the ecological                  and longleafpines cover 47.7% of the entire site. About
c o m p o n e n t s of BACHMAP.. This paper reports simula-                   14% of the Savannah River Site is in early successional
tion studies done with ECOLECON but does not include                         habitats (Savannah River Forest Station 1992).
the e c o n o m i c results because the Savannah River Site is                   Our study area was a 5924-ha region in the southeast-
not managed solely for timber revenues and the eco-                          ern corner of the Savannah River Site. In 1989, this area
nomic parameters in ECOLECON do not apply directly                           included 4725 ha of pines and 1199 ha of hardwoods
to the Savannah River Site. For additional information                       (Fig. 1 ). ECOLECON represents a landscape as a grid of
about the e c o n o m i c aspects of ECOLECON, see Liu                       hexagons in which each hexagon approximates the size
(1992, 1993b) and Liu et al. (1994).                                         of a typical breeding territory of the species being mod-
    All MAP models, including ECOLECON and BACH-                             eled. Bachman's Sparrow territories average 2.5 ha
MAP, are individual-based grid models (Fahrig 1988).                         (Haggerty 1986), so w e overlaid the landscape map
Each individual in the population is followed through its                    (Fig. l a ) with a grid of approximately 2370 2.5-ha hex-
annual cycle of birth, dispersal, reproduction, and death                    agonal cells (Fig. lb). If a hexagonal cell was completely
(Pulliam 1988). Dispersal strategies and habitat-specific                    within a forest stand, w e assigned the cell the same
fecundity are incorporated into MAP models. Forest suc-                      characteristics (such as forest type and age) as found in
cession and m a n a g e m e n t schemes are also mimicked by                 that stand. When a hexagonal cell contained portions of
annually increasing the age of the habitat in each grid cell                 two or m o r e stands, w e assigned the cell to the stand
and harvesting forest stands w h e n they reach an appro-                    type that covered the greatest portion of the cell. Thus,
priate age as specified by the timber management regime.                     there existed s o m e discrepancies b e t w e e n the real
    Forest c o m p a r t m e n t maps supplied by the Savannah               stand and its hexagonal representation. The average dis-
River Forest Station w e r e used to create several data-                    crepancies w e r e less than 10% ( c o m p a r e Fig. l c with
bases, including the age, size, and spatial distribution of                  Fig. l a ) ; that is, the average accuracy of hexagonal rep-
forest stands. This information was represented in ARC/                      resentation was greater than 90%. The resulting hexa-
 INFO, a GIS operated on a Sun Sparc workstation (Sun                        gon-based map (Fig. l b ) was used for simulations.
 Microsystem 1991). The data generated from the GIS
was transferred to IBM-compatible c o m p u t e r s as simu-                 Savannah River Site Wildlife, Fisheries, and Botany
 lation inputs. The operation plan (Savannah River Forest                    Operation Plan
 Station 1992) was used to project land use patterns up
 to 50 years into the future.                                                The operation plan was p r e p a r e d to provide b o t h short-
                                                                             term (10 years) and long-term ( 5 0 years) objectives for
study Area                                                                   managing endangered, threatened, and key wildlife spe-
The Savannah River Site is located in western South                          cies and communities. Specific m a n a g e m e n t strategies
Carolina along the Savannah River. It is a 77,000-ha fa-                     w e r e p r o p o s e d in the operation plan to reach these oh-

Volume9, No. 1, February1995
lJu et al.                                                                SpstiJllyE~plicitModels sad ForestMsaagem~t      65

Figure 1. (a) A GIS m a p o f forest stands in our study area at the Savannah River Site. The black lines are stand
boundarie~ Shaded stands are suitable habitat patches f o r Bachman's Sparrows (1- to 5-year a n d mature p i n e
stands); unshaded stands are unsuitable habitat (hardwoods, intermediate.aged p i n e stands (6- to 79-year)
a n d stands that have j u s t been harvested (0 year age class). Co) Study area as represented by ECOLECON show-
ing the m a p overlaid with a grid o f hexagonal cell~ (c) A m a p o f forest stands with each cell assigned to one
habitat type This m a p was used to run ECOLECON f o r s i m u l a t i n g the sparrow p o p u l a t i o n dynamtc~

                                                                                                Conservation Biology
                                                                                                Volume 9, No. 1, February 1995
 66      S~s.lly JF~plidtModelsand ForestMan~sement                                                                l,iu et ~.

jectives. For instance, an increase in harvest rotation          density of five breeding pairs per 640 acres (about 259
length was proposed to create more mature pine stands            ha) in the region where our study area is located. Ex-
of the sort favored by the Red-cockaded Woodpecker              trapolation from the operation plan to our 5924-ha
(Picoides borealis), an endangered species whose hab-            study site yields a goal of about 115 breeding pairs of
itat needs are a management priority of the Forest Ser-          Bachman's Sparrow for the study area.
vice as mandated by the listing of woodpecker under
the Endangered Species Act.
                                                                 mology of aachman's Sl..'row
   The forest composition of our study area was fairly
representative of the entire Savannah River Site (Liu,          Bachman's Sparrow is a potentially threatened species of
unpublished data), so we used the operation plan to             the southeastern United States (Dunning & Watts 1990;
project the future composition of the pine stands in our        Pulliam et al. 1992). Its range has declined significantly
study area. In 1991 the pine regions of the Savannah            in the past several decades (Haggerty 1986; Dunning
River Site were dominated by a few age classes of pines.        1993). The species breeds in pine woodlands and feeds
Over 50% of the pines were 31 t o 40 years old and              on insects and seeds on the ground (Dunning & Watts
more than 20% were 0 to 10 years old (Fig. 2). There            1990). Dunning and Watts (1990) reported that the
were no pine stands older than 90 years. The operation          sparrows' favored habitats are those with substantial
plan calls for harvest of the most common pine age              amounts of herbaceous vegetation (grasses and forbs)
classes to produce a forest with a more even composi-           within the first meter above the ground and with an
tion (Fig. 2). The Operation Plan establishes minimum           open understory (the second through fourth meters
objectives for maintaining 42 species of management             above ground) with few tall shrubs. The sparrows usu-
interest, including the Red-cockaded Woodpecker and             ally shun sites with dense understory vegetation. On the
Bachman's Sparrow. For Bachman's Sparrow, the mini-             Savannah River Site, they are found breeding in 1-to-5-
mum objective is to maintain 1100 breeding pairs on             year-old pine stands (clearcuts) and powerline rights-
the entire Savannah River Site and to attain a sparrow          of-way as well as in mature pine stands (older than 80


                                                                                                []   1991

 #.                                                                                             [ ] 2000
                                                                                                m 2040



                0-10        11-20   21-30      31-40   41-50   51-60     61-70    71-80     81-90     91-100    100+

                                                       Age Classes   (year)
Figure 2. Frequency distribution of forest stands in the field study area The data for 1991 were derived from a
GIS database o f 1989, and the distributions o f age classes for 2000 and 2040 were projected according to the
proposed operation plarL

Conservation Biology
Volume 9, No. I, February 1995
l,iu et sl.                                                                               S p s t ~ Explicit Models and Forest Msnsgement     67

years). The mature stands are suitable habitat for the                    Simulated Management Options
sparrows because mature stands also have dense ground
vegetation and o p e n understories w h e n burned on a
                                                                          BAIt-YK~ 0FII01¢3
frequent basis (Dunning & Watts 1990; Gobris 1992).
   The survivorship, r e p r o d u c t i o n , and dispersal of           The operation plan describes h o w m u c h forest should
Bachman's Sparrow are currently under intensive inves-                    be harvested over the 50-year period but does not spec-
tigation at the Savannah River Site, but there remains                    ify w h i c h forest stands should b e harvested or h o w
uncertainty concerning s o m e parameters needed for the                  m u c h forest should b e cut each year. We assumed that
model. Accordingly, w e have made some assumptions                        an equal area of forest stands w o u l d be harvested each
for the simulations based on the current field investiga-                 year until the desired forest structure is reached. We
tions and a literature survey. The assumptions are dis-                   considered three options for selecting the eligible forest
cussed in detail by Liu ( 1 9 9 2 ) and Pulliam et al. (1992).            stands for harvest each year: ( 1 ) harvesting the forest
For m o s t simulations, r e p r o d u c t i v e success was as-          stands randomly, ( 2 ) harvesting the oldest stands first,
sumed to be 3.0 offspring p e r pair per year in 1-to-2-year              and ( 3 ) harvesting clusters of adjacent stands. Under the
clearcuts and in m a t u r e p i n e stands (/>80 years) and 1.0          random harvest option, w e selected stands without re-
offspring in 3-to-5-year pine stands (Haggerty 1988; Liu                  gard to their proximity to other selected stands. The
1992). Adult survivorship of Bachman's Sparrow was set                    option of harvesting the oldest eligible stands first rec-
at 60% and juvenile survivorship at 40%. Survivorship                     ognizes that older timber often p r o d u c e s greater eco-
was annual and it w a s assumed constant in every year.                   nomic return. Under this option, the stands to be cut
For each individual, however, survival or death in a par-                 w e r e grouped according to age, and the oldest eligible
ticular year was stochastic and was randomly assigned.                    stands w e r e selected until the required acreage for that
This p a p e r presents the results of varying the values of              year was attained. We assumed that .this option would
selected demographic and behavioral parameters. For a                     approximate a management strategy of maximizing in-
m o r e detailed sensitivity analysis, see Liu ( 1 9 9 2 ) and            c o m e within the constraints of the operation plan. The
Pulliam et al. (1992).                                                    final option, harvesting by clusters, recognized that the
   Juveniles are assumed to disperse from their natal                     Forest Service classifies stands into c o m p a r t m e n t s of
sites in the spring, just prior to territory acquisition. The             3 0 - 1 0 0 stands for administrative purposes. Timber sales
model assumes an even sex ratio but traces the dispersal                  are organized on a c o m p a r t m e n t level, with the result
m o v e m e n t s of females only. A juvenile female is as-               that the stands harvested each year are often c l u m p e d
sumed to inherit her natal territory only if the female                   spatially. The option of harvesting in clusters was there-
parent on that territory has died. Juvenile females not                   fore designed to be similar to the actual harvest strategy
inheriting a natal territory must disperse in search of an                used at the Savannah River Site. The algorithm to cluster
unoccupied, suitable breeding site. In the model, this                    the cuts was that the stands in one c o m p a r t m e n t w e r e
means finding an u n o c c u p i e d cell ( h e x a g o n ) of the cor-   cut first, then stands in a neighboring c o m p a r t m e n t
rect vegetation type and age ( 1-5-year or --->80-year pine               w e r e harvested until the required acreage for a specific
stands). Dispersing juveniles can m o v e to any adjacent                 year was obtained. A stand in our m o d e l was defined as
cell, but w e assumed that they m o v e preferentially to                 a group of cells that w e r e uniform in species composi-
adjacent cells of suitable vegetation type and age. Thus,                 tion and age. For example, if a single hexagonal ceil was
in the m o d e l a dispersing female m o v e s with equal prob-           surrounded by cells of different ages, then the single cell
ability into any adjacent suitable cells. If none of the                  was a stand. If adjacent cells w e r e of the same age and
adjacent cells are suitable for breeding, the female                      species composition, then all similar adjacent cells be-
m o v e s at r a n d o m to an adjacent unsuitable cell. The              longed to the same stand.
search continues until an unoccupied, suitable cell is                        Our GIS database included the age, size, and bound-
found or the disperser dies. Dispersal involves extra                     aries of forest stands in the study region as of 1989.
mortality risks, and w e assume that with each m o v e to a               Since the operation plan had a 1992 starting date, w e
n e w cell a dispersing female has a 2% chance of dying.                  wished to begin our simulations with the 1991 land-
Further details about the dispersal rules can be found in                 scape. To generate this landscape, w e increased the age
Pulliam et al. ( 1 9 9 2 ) and Liu (1992, 1993a~ 1993b).                  of each stand in the 1989 database by two years. We
    The field data presented on the population abundance                  therefore assumed that there was no harvest in 1989 or
and distribution of Bachman's Sparrow in the study re-                     1990. In fact, some harvest did o c c u r during these years,
gion was collected in the s u m m e r of 1989 (J. B. Dun-                 especially in an area damaged by a tornado in the fall of
ning, Jr., B.J. Danielson, and B. D. Watts, unpublished                    1989 (Dunning et al., in press). The increased harvest
data). There w e r e 64 pairs of Bachman's Sparrow on the                 due to salvage logging made the region affected by the
study area, of which 12 pairs w e r e in mature stands, 40                tornado less typical of the Savannah River Site as a
pairs in 1-to-2-year stands, and 12 pairs in 3-to-5-year                  whole, so w e preferred to use the original distribution
stands.                                                                   of habitat patches to explore the operation plan. Thus,

                                                                                                                   Conservation Biology
                                                                                                                   Volume 9, No. I, February 1995
68              ~
        Spstislly l i a t Modelssad ForestManagement                                                                IJu et sl.

our landscape maps do not portray the exact stand dis-          plete as those of lobloUy pines, w e used the models of
tributions present in 1991. Because w e are attempting          loblolly pine growth for the entire pine area. This sim-
to explore possible responses of the sparrow population         plification certainly caused some discrepancies, espe-
to landscape change and not to predict precise popula-          cially for longleaf pines. While the growth rate of slash
tion change, the projected 1991 maps are adequate for           pines is about the same as that of loblolly pines, longleaf
this study.                                                     pines grow m u c h slower than loblolly pines, at least in
                                                                their early years (Clutter et al. 1983). This suggests that
OPTIONS ON THINNING AND BURNING MIDDLE-AGED PINg STANDS         the period during which clearcuts planted in longleaf
                                                                pine are suitable for Bachman's Sparrows might b e
Thomas et al. ( 1 9 9 0 ) suggested that to create more suit-
                                                                longer than that of young loblolly pine stands (Hunter
able habitat for the Northern Spotted Owl (Strix occi-
                                                                1990). In fact, field surveys on the Savannah River Site
dentalis caurina), younger stands may be modified to
                                                                show that the sparrows occupy longleaf pine stands 1-8
include structural characteristics of old-growth forests
                                                                years after planting, while the birds o c c u p y loblolly
(Murphy & Noon 1992). This option may also be a vi-
                                                                stands only from 1-5 years after planting (Dunning, un-
able management strategy for the Bachman's Sparrow
                                                                published data).
because some individuals have been found to breed in
                                                                   We simulated the effects of planting different pine
pine stands b e t w e e n 50 and 80 years old where the
                                                                species by varying the maximum age at which y o u n g
structure of these stands appeared similar to that of ma-
                                                                stands are suitable. We ran a series of simulations in
ture forest (Haggerty 1986; Dunning & Watts 1990; Go-
                                                                which the maximum suitable age of young stands varied
bris 1992). Dunning and Watts ( 1 9 9 0 ) stated that forest
                                                                from 2 to 10 years. Reproductive success (1.5 female
management practices, particularly burning and thin-
                                                                offspring per season) was assumed to be equal in all
ning, can reduce the density of understory vegetation
                                                                clearcuts of suitable age. We assumed that a smaller
and thus improve habitat suitability for the sparrows.
                                                                maximum suitable age reflects the effects of planting
Also, Gobris ( 1 9 9 2 ) found that middle-aged pine stands
                                                                loblolly and slash pine, while a larger maximum suitable
on the Piedmont National Wildlife Refuge in central
                                                                age mimics the effects of planting longleaf pines.
Georgia regularly support Bachman's Sparrows w h e n
maintained with the same burning regimes as used there
in mature pine forest. This suggests that thinning and          s~ittvity ~ r s l s
burning 50-to-80-year stands o n the Savannah River Site
to generate ground-layer and understory conditions              Pulliam et al. ( 1 9 9 2 ) performed extensive sensitivity
similar to older forest may render these stands more            analyses on the BACHMAP model and found that the
suitable for Bachman's Sparrow, increasing the total            model results were, in general, much m o r e sensitive to
amount of suitable habitat.                                     demographic parameters, particularly survivorship, than
   A program of thinning and burning middle-aged pine           to model parameters associated with dispersal behavior.
stands has been initiated by the Forest Service on the             For the current study, we examined the sensitivity of
Savannah River Site. The purpose of this program is to          estimated total population size to variation in five vari-
generate suitable c o n d i t i o n s for the Red-cockaded      ables. These variables w e r e dispersal mortality, adult
Woodpecker and other species found in open, mature              survivorship, juvenile survivorship, and two landscape
pine forest. The operation plan did not contain enough          variables---the number of years that clearcuts w e r e suit-
details o n this program to include these management            able (young stands) and the minimum age at which
options in our main simulations. To explore the possible        older stands became suitable (older stands).
effects of such management, we assumed in a series of              The sensitivity index (Sx) was calculated as
simulations that stands of 50 +, 60 + , or 70 + years old
w e r e thinned and then burned at rotations of 3-5 years
                                                                                      Sx = ( Ax/x y ( Ae/P ),
each, and that sparrows in the modified stands had the
                                                                where Ax/x is the observed change in the total popula-
same reproductive success as those in mature stands of
                                                                tion size due to a change (AP/P) in the parameter P
80 years or older. We assumed that the eligible forest
                                                                Oorgensen 1986; Puliiam et al. 1992). A larger Sx indi-
stands in these simulations w e r e harvested in clusters.
                                                                cates a higher sensitivity of population size to change in
                                                                a particular parameter.
                                                                   To see h o w the sparrow population size changed as
In the study area, there were three primary species of          several factors varied simultaneously, w e used a 23 fac-
pine trees: loblolly, longleaf, and slash pines (Pinus el-      torial design to determine the interliction effects among
liotti). These three species accounted for 41%, 48%,            three variables each at two levels. The three variables
and 11%, respectively, of the area covered with pine            included adult and juvenile survivorship and reproduc-
forests o n the Savannah River Site. Because the growth         tive success of 3-to-5-year stands because these vari-
functions of longleaf and slash pines w e r e not as com-       ables w e r e among the most important factors that influ-

Conservation Biology
Volume 9, No. 1, February 1995
IJu et al.                                                                               SpatiallyExplicitModelsand ForestManagement                69

enced the sparrow's population size and dynamics in                       a)
previous analyses (Pulliam et al. 1992).
Simulation and Statistical Methods                                  150

All simulations w e r e run for 50 years, the length of the
long-term objectives set by the operation plan (Savan-              190"   ~
nah River Forest Station 1992). Each simulation run had
100 replicates. Extinction probability was calculated as       .=   50"
the n u m b e r of simulations in which population extinc-
tion took place divided by the total number of repli-                 0                    _~^             '                '                      2--004
                                                                      1990                2000           2010             2020            2030
cates. Population sizes w e r e averaged over the entire
simulation period for runs in which populations did not                                                    Time (year)
go extinct. To test for significance of differences in pop-
ulation sizes, w e used two-tailed t-tests and F tests. Dif-
ferences in extinction probabilities w e r e tested with a
test for binominal proportions (Ott 1988).

                                                               g       190"
Effects of Management Options
                                                               eL         50'

The general pattern of sparrow population dynamics
                                                                           1990              2000         2010             2020           2030     2040
was very similar under the three harvesting options con-
sidered. In all three cases the population first decreased                                                       Time (year)
to a "valley" and then increased gradually (Fig. 3). There                     c)
were, however, three major differences between the
harvest options regarding management goals. First, har-
vesting the eligible stands randomly (Fig. 3a) never al-
lowed the sparrow population to reach the minimum
                                                                      ,5o.                                           b7, 7
objective set by the operation plan. Population sizes                           ......................    M':7: °
reached and e x c e e d e d the minimum objective (115                190"
breeding pair) before the year 2040, w h e n the stands
w e r e cut in clusters (Fig. 3b) or the oldest stands were    .=
harvested first (Fig. 3c). Second, clustered harvesting
had the highest average population size over the 50
years of simulations (Table 1 ), while random harvesting                   0                  -             .        .        .       .
                                                                           1990              20(10        2010             2020            2030     204C
had the lowest. Finally, the average yearly standard error
under random harvesting (11.14) was about twice as                                                              Time (year)

large as that under harvesting in clusters (5.60) or har-      Figure 3. The sparrow population dynamics result-
vesting the oldest stands first (5.28).                        ing from harvesting the stands randomly (a), har-
                                                               vesting the stands in clusters (b), and harvesting the
EFFECF OF THINNING AND BURNINGMIDDLE-AGEDSTAND~                oldest stands first (c). M i n i m u m objective refers to
When w e varied the minimum age at which stands be-            the m i n i m u m management goal proposed by the
came suitable for Badlman's Sparrow, the sparrow pop-          operation plat6 extrapolated to our study area Verti.
ulation became much larger as the minimum age was              cal bars indicate 1 ~ e
reduced (Fig, 4a). The average population size over the
50 years of simulations (POP) decreased linearly as the        jective set by the operation plan. When 5t-year-old pine
minimum suitable age of 50-80-year stands (OS) in-             stands were maintained in a condition suitable for the
creased according to the relationship POP = 648.51 -           sparrow, the minimum population objective of the op-
7.534 OS (R 2 = 0.969).                                        eration plan was reached in 2007, 17 years into the
  The simulation results (Fig, 4b) also indicated that the     simulation period. The relationship between the mini-
younger the suitable age of forest stands the earlier the      mum suitable age of 50--8t-year stands (OS) and the
population size reached and e x c e e d e d the minimum ob-    first year by which the minimum objective was reached

                                                                                                                         Conservation Biology
                                                                                                                         Volume 9, No. I, February 1995
70               b:X'plicitModels and ForestManagement
         Spatially                                                                                                                             Liu et al.

Table 1. g~ects of tJm~ hat~estieg methods oQ the ~ e r a ~                            ship, the population size (POP) w h e n j u v e t ~ ¢ survivor-
populationsize (±SE).                                                                  ship was low decreased with an increase in m i n i m u m
                                                             Population Size           suitable age of older stands (OS) (POP = 61.97 - 0.50
Harvesting option                                                (no.)*                OS, R 2 = 0.943, Fig. 6). The population sizes r e c o r d e d
Harvest oldest stands first                                    54.81 +- 0.79 ~         in the simulations with juvenile survivorship at 0.3,
Harvest stands randomly                                        50.08 - 2.74 ~          however, w e r e an o r d e r of magnitude l o w e r than pop-
Harvest stands in clusters                                     59.69 -+ 0.80 ~         ulation sizes in previous simulations (Fig. 4 a ) and never
* Letters a, b, a n d c indicate w h e t h e r t w o n u m b e r s are significantly
                                                                                       reached the minimum objective.
different a t the 5% level I f t w o n u m b e r s h a v e different letter~ they         The sparrow population frequently w e n t extinct in
are Mgntficantly dtffere~Z                                                             the simul~ti~ts with low juvenile survivorship (Table
                                                                                       3). Under the most restrictive conditions, the frequency
(Y) is described by the linear function Y = 1957.6 +                                   of extinction was extremely high. W h e n b o t h adult and
0.96 OS (R 2 = 0.977).                                                                 juvenile survivorships w e r e low and the suitable age of
                                                                                       older stands set at the m a x i m u m of 80 years, the extinc-
         OF PLANTING D        ~       PII~ SPS¢ISS                                     tion probability was 0.85 or higher. With adult survivor-
                                                                                       ship at 0.6, extinction probability was less than O. 1. Even
As the m a x i m u m suitable age of young stands (YS) in-
                                                                                       with the minimum suitable age of older stands r e d u c e d
                                                                                       to 70 years, there was still a 5% chance for the popula-
linearly (Fig. 5a~ POP = 25.42 + 10.83 YS, R 2 =
                                                                                       tion to go extinct before the year 2040. The level of
0.942). Also, a larger YS enabled the sparrow population
                                                                                       reproductive success in young stands also influenced
to reach the m i n i m u m objective sooner (Fig. 5b), this
                                                                                       the total population size. For example, w h e n reproduc-
relationship was a negative linear function (Y = 2042.8
                                                                                       tive success was low in 3- to 5-year clearcuts (0.5 female
 - 2.2941 YS, R 2 = 0.913).
                                                                                       offspring/year ), average population size was significantly
   The time w i n d o w during which clearcuts w e r e suit-
                                                                                       higher than w h e n reproductive success was zero in
                                                                                       these stands (0.0 offspring/year, Fig. 7).
When the m a x i m u m suitable age of young stands was 2,
                                                                                          There w e r e very strong interactions a m o n g adult sur-
the population size decreased to a valley before even-
                                                                                       vivorship, ~uvenile survivorshi~, and the reproductive
tualfltv increas'mg ~o e x c e e d r3ae management goa~ ofpop-
                                                                                       success in 3- to 5-year stands (.Table 4). Even the sec-
u~ar~on s~ze. Ti-hs Xs fiae pattern w e o~rservefi w i ~ rmr
                                                                                       ond-order interaction a m o n g the three variables was sig-
first set of simulations (Fig. 3). W h e n the time w i n d o w
                                                                                       nificant at the 5% level (Table 5).
of suitability increased to at least 5 years after planting,
population size first increased for a few years and then
decreased before increasing again. Finally, w h e n the
                                                                                       Discussion and Conclusions
time w i n d o w of suitability was at least 7 years after                             Our simulations suggest that m a n a g e m e n t strategies that
planting, the peak population that o c c u r r e d prior to                            combine harvesting, thinning, and burning of selected
population decline reached o r e x c e e d e d the minimum                             forest stands could achieve the m i n i m u m objectives set
objective goal before declining (Fig. 5c).                                             by the operation plan of 115 breeding pairs of the spar-
                                                                                       rows in the study area. This assumes that the p a r a m e t e r
Sensitivity Analysis                                                                   estimates used in the majority of our simulations are
                                                                                       correct. W h e n juvenile survivorship was low in our sim-
Similar to the findings by Pulliam et al. (1992), the                                  ulations, however, the sparrow population size was al-
model was m u c h m o r e sensitive to survivorship than to
                                                                                       ways lower than the initial population size and never
dispersal mortality (Table 2). The minimum suitable age
                                                                                       reached the minimum objective. Although w e have little
of older stands was the most sensitive parameter deter-
                                                                                       direct data on juvenile survivorship from the field stud-
mining the average population size. The sensitivity of                                 ies at the Savannah River Site, a comparison of mortality
population size to adult and juvenile survivorship was
                                                                                       estimates for other sparrow species suggests that 30%
intermediate, with similar Sx indices for these two pa-                                survivorship would not be unusually low (see Puliiam et
r a m e t e r s . Although t h e t i m e w i n d o w for c l e a r c u t               al. 1992). This result underscores the need for caution
suitability (YS) had a lower sensitivity than the two sur-                             w h e n applying results of the model and for the n e e d for
vivorship parameters, and the other landscape parame-                                  further field investigations to reduce p a r a m e t e r uncer-
ters. YS had a sensitivity value that was three times as
high as the Sx value for dispersal mortality.
    The effects of the m i n i m u m suitable age of older                             Management Simulations
stands (OS) w e r e dramatically offset w h e n juvenile sur-
vivorship was low. We ran a set of simulations with                                    HARY]~T oPrIONS

juvenile survivorship at 0.3 instead of 0.4. As in previous                            The reason clustered harvesting resulted in the highest
simulations with the higher value of juvenile survivor-                                population size in the simulations ks probably that clus-

Conservation Biology
Volume 9, No. 1, February 1995
Uu et al.                                                                                    Spgti~i~,B~plicitModelsand ForestMmmgement           71

              n)                                                                        b)

 .-¢                                                                             2°3°


                                                                          ~O     2010]

        0                                                                        2000 '
         40        50           60           70           80         90                 ~o        5'0          ~           7o           so        9~

                   Minimum Suitable Age of Older Stands (year)                                   Minimum Suitable Age of Older Sta.ds (year)

Figure 4. Changes in total population size caused by manipulating the age at which 50- to 80-year stands be-
come suitable for breeding (a), and effects of manipulating the age at which 50- to 80.year stands become suit-
able for breeding on the year when the sparrows first reach the objective population size (115 breeding pairs)
tered harvesting p r o d u c e d clustered patches of newly               suitable stands in the landscape. As shown in Fig, 4a; the
suitable habitat, making it easier for a dispersing juvenile              sparrow population increased with a decrease in the age
to locate a suitable patch. Random harvesting created a                   at which stands w e r e first suitable as breeding habitat.
landscape that was probably presented the most diffi-
culties for a dispersing juvenile seeking a suitable patch,               EFFECT OF PLANTINGDIFFgRENT PiNE SPECIES
and the population size was the lowest under this op-
tion.                                                                     According to the operation plan, while the total area of
    The p r o j e c t e d sparrow population sizes under all              pines at the Savannah River Site will remain the same,
three harvesting options w e r e closely related to the                   slash pine stands will be converted into longleaf pine
amount of suitable habitat. For example, under clus-                      stands. By the year 2040, the proportion of pine species
tered harvesting the mature stands increased over time,                   on the Savannah River Site will be about 61% longleaf
while clearcut habitat decreased during the first few                     and 39% loblolly pines. Because longleaf pines grow
years and then later increased to a stable level (Fig, 8a).               slower than loblolly pines, the young longleaf stands are
The relationship between the population size (POP) and                    suitable for sparrows for longer periods of time. In our
the amount of suitable habitat, including young and ma-                   main set of simulations, w e assumed that no difference
ture habitat (X), under clustered harvesting is shown in                  existed for longleaf and loblolly pines in terms of max-
Fig. 8b and can be represented as POP = - 2 0 . 6 2 5 +                   imum suitable ages. In our later simulations w e varied
0 . 1 2 4 1 X (R 2 = 0.934).                                              the time window during which stands w e r e suitable for
                                                                          the sparrow and found that the sparrow population in-
                                                                          creased with an increase in the suitability window. The
                                                                          model results will be more accurate w h e n the maxi-
Two harvest options---harvesting the oldest stands first                  mum suitable ages of both longleaf and loblolly pine
and harvesting the stands in clusters---allowed the sim-                  stands are treated differently, but this will require addi-
ulated sparrow population sizes to achieve the mini-                      tional local growth information on longleaf pines.
mum objective proposed by the operation plan (Savan-
nah River F o r e s t S t a t i o n 1992), b u t the s p a r r o w
                                                                          Implications for Land Management of Biodiversity
populations w e n t through a long transition period with
a very low abundance (Fig. 3b & c). One silvicultural                     Land managers are focusing more attention on conserv-
practice incorporated in the operation plan that would                    ing and managing for biodiversity. In many cases, this
make stands suitable at earlier ages is to thin and then                  new emphasis results in management for particular spe-
periodically burn 50- to 80-year stands so that the un-                   cies of management concern, especially endangered
derstory becomes less dense and ground vegetation is                      species. Single-species management is often emphasized
encouraged (Dunning & Watts 1990). This practice                          for two reasons. First, the selected species are often
would make forested stands suitable for the spatax>w at                   habitat specialists, and it is assumed that habitat patches
an age earlier than the mature stands currently found on                  that can support specialists will also be of sufficient
the Savannah River Site. Making 50- to 80-year stands                     quality to support a wide range of other species. For
suitable for the sparrows increased the total amount of                   example, the forest stands on the Savannah River Site

                                                                                                                       Conservation Biology
                                                                                                                       Volume 9, No. 1, February 1995
72                  EgplicttModelsand ForestManagement
            Spatially                                                                                                                                                                                                                                 IAu et al.

                  a)                                                                                                                                                 b)
           160'                                                                                                                                             2040

           120"                                                                                                                                             2030

                                                                                                                                              @ ¢o
           40                                                                                                                i                                                         2                4                      6        8        10          12
                  0                          4                    6                8                   10                   12
                                                                                                                                                                                       Maximum Suitable Age of Young Stands (year)
                        Maximum Suitable Age of Young Stands (year)


                                             3°°l                                      .... --o-.... Maximum Suitable Age = 10
                                             250                                       --!--              Maximum Suitable Age = 7                                                                               d ~ "°
                                  @                                                    ------o---         Maximum Suitable Age = 5                                                                        o.d"
                                                                                       ----e--.           Maximum Suitable Age = 2                                                             a,~.o ~r              ./m
                                  ¢0         200-                                                                                                                                      ,4-                           .ll

                                                                                                                                                                                   /                        /j.
                                              150"                          /'~                     Minimum Objective                                       •,r,=*J                              ._.'l~,o             '?
                                                        ......        -#,    .......
                                                                                               ..........................                                    ..........           =,----:--7-..,-'----
                                                                                                                                                                                           •If     ir        ,e,.O

                                                                              ~,           "o..o.                  ..cr"                                                  t ~ ~                --u"a

                                                                 •liL,,LO.0 .o .4i,,,0.4D.                                         ..q ..iP      o..0

                                                  0                                    i                                    i                           i                                  i                               |
                                                  1990                             2000                               2010                           2020                           2030                            2040

                                                                                                                                Time (year)

Figure 5. Effects of maximum suitable age of young stands on average population size (a), the first year to per-
manently reach the minimum objective set by the operation plan (b), and population dynamics (c).

Table 2. Sensitivity of estimated total population to five
vambl~ (Sx = (& ~ ) / ( A P/~).
Adult Survivorship
Juvenile Survivorship
                                                   A x/x

                                                                                       A WP


                                                                                                                                      30]               25

  (0.40)                                         49.40/72.71                       0.10/0.40                          2.72                              20
Dispersal Mortality                                                                                                                                          40                     50                      60                     ;0       80               90
  (0.01)                                      10.38/72.71                          0.01/0.01                        0.14
                                                                                                                                                                                  Minimum Suitable Age of Older Stands (year)
Young Stands** (5)                            18.08/72.71                             3/5                           0.41
Older Stands*** (80)                        - 37.31/72.71                            10/80                        - 4.10
                                                                                                                                          Figure 6. Effects of manipulating the age at which
*D e f a u l t v a l u e s f o r e a c h v a r i a b l e g i v e n i n p a r e n t h e s e ~ A n a l y s e s                              50- to 80-year stands become suitable for breeding
varied one parameter at a time in a simulation while using default                                                                        on average population size when juvenile survivor.
values as constants for all other parameter~
** N u m b e r o f y e a r s c l e a r c u t s w e r e s u i t a b l ~                                                                    ship = 0.3. (Compare with Fig. 4a in which juvenile
*** M i n i m u m a g e a t w h i c h o l d e r s t a n d s b e c a m e s u i t a b l e                                                   survivorship was 0. 4.)

Conservation Biology
Volume 9, No. 1, February 1995
Liu et ~.                                                                                                     Spatially £xplicit Models and Forest Management           73

Table 3.      Extinction probability under low juvenile survivorship                       Table 4. Interaction effects among three variables (PA, P~ and
(0.3).                                                                                     I~_,)* on population size (mean ± standard error).
Adult                       Young                Older                Extinction                              ~3-5 = 0.0                         ~3-~ = 1.5
Survivorshtp               Stands*             Stands**             Probability***                                 Pj                                  Pj
       0.6                      2                   80                      O.08a          PA           O.3                 0.4            0.3                  0.4
       0.6                      2                   70                      0.05 a
                                                                                           0.5      8.90 +- 0.25        33.10• 0.74    10.26-+0.37          38.83-+ 0.76
       0.6                      5                   80                      0.06 a         0.6     16.76--+0.66         54.63--+0.56   23.31 -+ 0.62        72.71 --+0.67
       0.5                      2                   80                      0.89 b
       0.5                      5                   80                      0.85 b         *PAiS the adult Survivorsbip, Pj is thejuvenile survitmmhip, and ~3-5
                                                                                           is the reproductive success (female offsprtng per pair) in pine stands
*Number of years clearcuts were suitable                                                   of 3-5 year old
** Minimum age at which older stands became suitable
***Letters a and b indicate whether two numbers are significantly
different at the 5% level I f two numbers have different letter~ they
a r e significantly different
                                                                                           Woodpecker. Eighteen of the 26 specific standards and
                                                                                           guidelines for Management Area II, w h e r e our study
                                                                                           area is located, on the Savannah River Site contain a
that contain suitable habitat for the endangered Red-
                                                                                           reference to the Red-cockaded Woodpecker, while an
cockaded W o o d p e c k e r also supply habitat for Bach-
                                                                                           additional three guidelines refer indirectly to habitat im-
man's Sparrows and 16 of the 42 other vertebrates on
                                                                                           p r o v e m e n t for this species (Savannah River Forest Sta-
the Forest Service's list of m a n a g e m e n t indicators for
                                                                                           tion 1992). The dominance of woodpecker-related stan-
this region (Savannah River Forest Station 1992).
                                                                                           dards and guidelines leaves the impression of single-
   The second reason m a n a g e m e n t plans often empha-
                                                                                           species management, even though the operation plan
size single species is that the habitat requirements for
                                                                                           explicitly states that "The need to use [Red-cockaded
endangered species or other "charismatic megafauna"
                                                                                           Woodpecker] terminology to describe habitat parame-
are relatively well k n o w n or at least under intensive
                                                                                           ter objectives should in no way detract from the com-
study (see Ligon et al. 1986; Thomas et al. 1990; Murphy
                                                                                           prehensive wildlife and botanical objectives that have
& N o o n 1992). Research can therefore give managers a
                                                                                           b e e n developed for this area" (Savannah River Forest
specific set of guidelines for h o w to manage for these
                                                                                           Station 1992:15).
species. It is less clear h o w to manage for a variety of
                                                                                                These standards and guidelines might be thought to
species simultaneously or for "biodiversity," with all its
                                                                                           provide adequate protection for Bachman's Sparrow be-
                                                                                           cause sparrow and the w o o d p e c k e r share a critical hab-
   Our study provides an example of the problems faced
                                                                                           i t a t - m a t u r e pine forest. But our study indicates that
by managers c o n c e r n e d with multiple species manage-
                                                                                           because the sparrow also inhabits clearcuts and other
ment. The operation plan p r o p o s e d b y the Forest Ser-
                                                                                           early successional habitats, it does not necessarily ben-
vice at the Savannah River Site is a m o d e l plan of mul-
                                                                                           efit from woodpecker-based management, at least in the
tiple s p e c i e s m a n a g e m e n t , in that it sets specific
                                                                                           short term. In effect, the two species o c c u p y different
m a n a g e m e n t goals for m a n y plants, fish, terrestrial ani-
                                                                                           portions of the landscape, although they share one land-
mals, and e v e n w h o l e communities. Most of the man-
                                                                                           scape component. Thus, implementation of the opera-
agement strategies outlined in the Operation Plan for
                                                                                           tion plan may increase the p o p u l a t i o n of the Red-
the area in w h i c h our study site is located, however,
                                                                                           c o c k a d e d W o o d p e c k e r , but at the cost of possibly
c o n c e r n habitat i m p r o v e m e n t for the Red-cockaded
                                                                                           threatening Bachman's Sparrow locally. O v e r the long
                                                                                           term (50 years), the strategy outlined in the operation

  g                                                                                        Table 5. Variance analysis of the interaction effects among three
                                                                                           variables (P~ P~ and 133-5)* on population size (mean ±
                                                                                           standard error).
                                                                                           Source                         df       M e a n square                 F**
  e~                                                                                       PA                              1           13388.30                  590.28
                                                                                           Pj                              1           52876.20                 2331.29
                                                                                           ~3--5                           1            2973.21                  131.09
                                                                                           PA X    Pj                      1            2710.63                  119.51
                                                                                           PAX     ~3--5                   1             549.53                  24.22
                            0                           0.5                                Pj X    ~3--5                   1             680.46                   30.01
                   [~3(Female Offspring / Pair in 3-5 yr Old Clearcuts)                    PAX     Pj X ~3--5              1             106.99                    4.72
F i g u r e 7. I n f l u e n c e o f r e p r o d u c t i v e s u c c e s s i n o l d e r   * P^ is the adult survitcTmhip, Pj is thejuvenile survivorship, and ~3-~
                                                                                           is the reproductive success (female offsprlng per pair) in pine stands
c l e a r c u t s ( 3 - to 5 - y e a r s t a n d s ) o n a v e r a g e t o t a l p o p -   of 3-5 year o l d
ulation siza                                                                               ** p < 0.01 for all but PAx Pj X ~3-5 which is p < 0.05.

                                                                                                                                        Volume9, No. 1, February1995
74                l~pBdt Models and ForestManagement
            Spa~al~                                                                                                                              Llu et al.

            a)                                                                              b)
     1200"                                                                          200"
                                                                                                      =    .   .   .    .    =   .           •
     1000                                                          /
                                                                               =o    150"

             ""~                                        MatureStands
                                                                               .=   100"

     200"          ;."
       0                           ,                •              ,

       ,99o                   ~   ~o~o          ~o~o             ~o3o   ~o             0

                                       Time(year)                                                         Areaof Matureand YoungStands(ha)

Figure 8. A n n u a l change o f area o f y o u n g (1- to 5-year) a n d m a t u r e stands ( 8 0 + years) u n d e r clustered har-
vesting (a), a n d the linear relationship between the sparrow p o p u l a t i o n size a n d the total area o f m a t u r e
( 8 0 + yettrs) a n a y o u n 3 stands { l- to 5:years) f~).

plan should support sparrow populations that eventu-                                pine community. Our simulations suggest that expan-
ally meet management goals. Over the short term, how-                               sion of this program would be the most direct m e t h o d
ever, the management strategy may cause the sparrow                                 of mitigating the possible negative effects on the spar-
population to decrease substantially. Our simulation re-                            row caused by the reduction in clearcut habitat under
sults show that the short-term objectives (10 years in                              the operation plan.
the operation plan) can rarely be achieved. During the                                 By linking models such as ECOLECON with GIS land-
first several decades of the operation plan, our results                            scapes of a particular region, managers could test the
suggeslm that lhe sparrow population may remain v e r y                             etfect U/a particular managemem strategy on a variety
l o w or "pert)me eXunc~, esJ~e~m~y "fi)~-~e~t: s m ~ ' ~ -                         ~ s y ~ e s . ~ath~ b y m ~ e s - ~ y m ~ t ~ e x g ~ t)~ ~ne mt~heh
ship is low. If a stochastic event caused the sparrow                               would have to include a flexible life-history, subroutine
population to drop even further in the v a / / e y ~ t h e tran-                    that could be parameterized for the species in question
sitional p e r [ o d ~ t h e chances of local extinction would                      and a landscape scaled t o the species' movements and
be greatly increased because Mw population size in-                                 dispersal abilities. In addition, because animals in the
creases the chance of extinction (ShaPer & Samson                                   real world interact with one another, there is a need for
1985). The loss of mature pine stands due to insect                                 models that consider the interactions and population
cutbrea/~ tornado, oc hurricane ( D u n n i n g & Watts                             dyrmmics of multiple sgccies on a sgecilic landscape.
1991) is an example of such a stochastic event. Under                               Such a flexible combination of GIS analysis and popula-
some simulated scenarios, this potential threat is large:                           tion modeling could provide managers with the ability
gogulations ~r[ o v e c 8(?fg o6 simalacion re~lic~ce~ w e n t                      co e x s ~ i a c macragcment c/f~¢ccs o n a varicef o f sVecies
extinct, if ~ pol3ui~¢/ocl b ~ o m t s tx~i~x¢~ i~x ¢~¢ exposi-                     zsxgt ¢o ~ ¢ h     6oc wzys ~3~miCig~cAxg ¢~¢ ¢ff¢c~ o~x spe-
tional period, the mature pine stands available later are                           d e s of management c o n c e r n caused by implementation
meaningless for the sparrow unless some individuals                                 of plans designed to improve habitat for single species at
from other locations immigrate into the new habitats.                               the top of the priority list.
    Our simulations suggest that management of the mid-

tation conditions similar to those in mature forest is                              Acknowledgments

~gement g o a / f o r B a c ~ ' s Sparrows. The Forest Ser-                         W e thank .Joe Ferris, Liz Kramer< K u r t Snarl B r e n t
vice has initiated a pilot program of converting middle-                            Danielson, John Blake, Glen Gaines, Susan Loeb, and oth-
                                                                                    ers~ ~ ~ M e ~ _           rock ~h-e~te .'X¢em,e-cety.~ge~e-
Woodpeckers. This p r o g r a m includes mechanically                               ful for the criticisms and suggestions from two anony-
~te',rrtt Q~ cea~.io~n. t.t~ , .b ~,7~ eg .'dr~ .~cb'z~, .ar~,rl ,'-..e             wzfo-a~.ze~rewe~-ar~ ~z¢i~ ~ ~ I .         Tfz~ .pzojeet , w ~
burning of the understory and ground vegetation. Stands                             supported in part by the Savannah River Field Office

intended to produce areas of local range expansion for                              Department of Energy grant DE--FG09--89ER60881 to
the ~ o o d p e c g e r and ~he te-es~abii~tmaem o f ~h¢ longlg~f                   8. R. F~lli2wa, ~ d b y tht: <3ditto Resc',urth A ~ t ~ J.

Volume 9, No. 1, February1995
ldu et al.                                                                                     Spatially Egplicit Models and Forest Management      75

Liu. Support for field studies of Bachman's Sparrow was                       Gobris, N. 1992. Habitat occupancy during the breeding season by
received from National Science Foundation grant BSR-                              Bachman's Sparrow at Piedmont National Wildlife Refuge in cen-
                                                                                  tral Georgia. M.S. thesis. University of Georgia, Athens.
8817950 and U.S. Environmental Protection Agency                              Haggerty, T.M. 1986. Reproductive ecology of Bachman's Sparrow
grant C R 8 2 0 6 6 8 - 0 1 - 2 to H. R, Pulliam and J. B. Dun-                   (Aimophila aestivalts) in central Arkansas. Ph.D. dissertation. Uni-
ning, grants from the Savannah River Ecology Labora-                              versity of Arkansas, Fayetteville.
tory, and U.S. Forest Service Cooperative Research                            Haggerty, T. M. 1988. Aspects of the breeding biology and productiv-
Agreement grant 1 2 - 1 1 - 0 0 8 - 8 7 6 Supplements 163 and                     ity of Bachman's Sparrow in central Arkansas. Wilson Bulletin
192. We thank the Savannah River Forest Station of the                        Hunter, W.C. 1990. Handbook of nongame bird management and
U.S. Forest Service for a prepublication copy of the op-                          monitoring in the Southeast Region. U.S. Fish and Wildlife Service,
eration plan, compartment maps and GIS databases of                               Atlanta.
the Savannah River Site, and other materials, and for                         JCrgensen, S. E. 1986. Fundamentals of ecological modelling. Elsevier,
frequent consultations. Research and manuscript prep-                             Amsterdam.
                                                                              Ligon, J. D., P. B. Stacey, IL N. Conner, C. E. Bock, and C. S. Adldsson.
a r a t i o n w e r e s u p p o r t e d by C o n t r a c t D E - A C 0 9 -
                                                                                   1986. Report of the American Ornithologists' Union Committee
76SROO-819 between the U.S. Department of Energy                                  for the Conservation of the Red-cockaded Woodpecker. Auk
and the University of Georgia's Savannah River Ecology                            103:848-855.
Laboratory.                                                                   Liu, J. 1992. ECOLECON: A spatially-explicit model for ecological eco-
                                                                                  nomics of species conservation in complex forest landscapes.
                                                                                  Ph.D. dissertation. The University of Georgia, Athens, Georgia.
Literature Cited                                                              Liu, J. 1993aL Discounting initial population sizes for predicting ex-
                                                                                  tinction probabilities in spatially-explicit environments. Ecological
Borland International. 1991. Borland C + + 2.0. Borland International,            Modelling 70:51-61.
   Scotts Valley, California.                                                 Liu, J. 1993b. ECOLECON: An ECOLogical-ECONomic model for spe-
Clutter, J. L, J. C. Fortson, L V. Pienaar, G. H. Brister, and R. L Bailey.       cies conservation in complex forest landscapes. Ecological Mod-
   1983. Timber management: A quantitative approach. John Wiley &                 elling 70:63-87.
   Sons, New York.                                                            Liu, J., F. Cubbage, and R. Pulliam 1994. Ecological and economic
Conroy, M.J., Y. Cohen, F. C. James, Y. G. Matsinos, and B. A. Maurer.            effects of forest structure and rotation lengths: Simulation studies
   1995. Parameter estimation, reliability, and model improvement                 using ECOLECON. Ecological Economics 10:249.
   for spatially-explicit models of animal populations. Ecological Ap-        Murphy, D.D., and B. R. Noon. 1992. Integrating scientific methods
   plications, in press.                                                          with habitat conservation planning: Reserve design for Northern
Dukes, E.K. 1984. The Savannah River plant environment: DP-1642.                  Spotted Owls. Ecological Applications 2:3-17.
   E. I. du Pont de Nemours, Savannah River Plant, Aiken, South Caro-         Ott, L. 1988. An introduction to statistical methods and data analysis,
   lina-                                                                          3rd edition. PWS-Kent, Boston.
Dunning, J.B. 1993. Bachman's sparrow. Pages 1-16 in A. Poole, P.             Pulliam, H.R. 1988. Sources, sinks, and population regulation. Amer-
   Stettenheim, and F. Gill, editors. The birds of North America, no.             ican Naturalist 132:652--661.
   38. The Academy of Natural Sciences, Philadelphia, and The Amer-           Pulliam, H. IZ, J. B. Dunning, Jr., and J. Liu. 1992. Population dynamics
   ican Ornithologists' Union, Washington, D.C.                                   in complex landscapes: a case study. Ecological Applications
Dunning, J. B., and B. D. Watts. 1990. Regional differences in habitat            2:165-177.
   occupancy by Bachman's Sparrow. Auk 107:463-472.                           Savannah River Forest Station. 1992. Savannah River Site wildlife, fish-
Dunning, J. B., and B. D. Watts. 1991. Habitat occupancy by Bachrnan's            eries, and botany operation plan. Final draft. U.S. Forest Service,
   Sparrow in the Francis Marion National Forest before and after                 Savannah River Forest Station, New Ellenton, South Carolin~
   Hurricane Hugo. Auk 108:723-725.                                           Shaffer, M. L., and F. B. Samson. 1985. Population size and extinction:
Dunning, J. B., D.J. Stewart, B.J. Danielson, B. R. Noon, R. Root, R. H.          A note on determining critical population sizes. American Natu-
   Lamberson, and E.E. Stevens. 1995. Spatially-explicit population               ralist 125:144-152.
   models: Current forms and future uses. Ecological Applications, in         Sun Microsystems, Inc. 1991. Sun System user's guide (Desktop
   press.                                                                         Sparc). Sun Microsystems, Mountain View, California.
Dunning, J. B., R. Borgella, K. Clements, and G. K. Merle. Patch isola-       Thomas, J. W., E. D. Forsmann, J. B. Lint, E. C. Meslow, B. R. Noon, and
   tion, corridor effects, and colonization by a resident sparrow in a            J. Verner. 1990. A conservation strategy for the Northern Spotted
   managed pine woodland. Conservation Biology, in press.                         Owl. Interagency Scientific Committee to Address the Conserva-
Fahrig, L. 1988. A general model of populations in patchy habitats.               tion of the Northern Spotted Owl. 1990-791-171/20026. United
   Applied Mathematics and Computation 27:53--66.                                 States Government Printing Office, Washington, D.C.

                                                                                                                        Conservation Biology
                                                                                                                        Volume 9, No. 1, February 1995

To top