ArticlePDF Available

Facilitating Adaptive Forest Management under Climate Change: A Spatially Specific Synthesis of 125 Species for Habitat Changes and Assisted Migration over the Eastern United States

Authors:
  • USDA Forest Service

Abstract and Figures

We modeled and combined outputs for 125 tree species for the eastern United States, using habitat suitability and colonization potential models along with an evaluation of adaptation traits. These outputs allowed, for the first time, the compilation of tree species’ current and future potential for each unit of 55 national forests and grasslands and 469 1 × 1 degree grids across the eastern United States. A habitat suitability model, a migration simulation model, and an assessment based on biological and disturbance factors were used with United States Forest Service Forest Inventory and Analysis data to evaluate species potential to migrate or infill naturally into suitable habitats over the next 100 years. We describe a suite of variables, by species, for each unique geographic unit, packaged as summary tables describing current abundance, potential future change in suitable habitat, adaptability, and capability to cope with the changing climate, and colonization likelihood over 100 years. This resulting synthesis and summation effort, culminating over two decades of work, provides a detailed data set that incorporates habitat quality, land cover, and dispersal potential, spatially constrained, for nearly all the tree species of the eastern United States. These tables and maps provide an estimate of potential species trends out 100 years, intended to deliver managers and publics with practical tools to reduce the vast set of decisions before them as they proactively manage tree species in the face of climate change.
HQ and CL output table for the Allegheny National Forest, Pennsylvania, sorted in decreasing order of current species abundance (FIAsum). Range indicates whether species is wide or narrow (W/N) in distribution, dense or sparse (D/S) in frequency of presence within FIA plots, and high or low (H/L) in importance value when it is found. MR refers to model reliability (see [21] for explanation). %Cell refers to the percentage of DISTRIB-II cells with the species present. FIAiv is the average importance value for the species when present on FIA plots. ChngCl45 or 85 presents the change classes (increase, decrease, or no change) of habitat suitability by 2100, according to RCP 4.5 (low emissions) or 8.5 (high emissions). Adapt is a class of adaptability of the species according to the ModFacs. Abund is an abundance class based on FIAsum. Capabil45 or 85 is the capability of the species to cope with the climates of RCP 4.5 or 8.5 at 2100, based on abundance, change classes, and adaptability. SHIFT45 and 85 show two classes of Infill (+ or ++) to indicate these species are currently found rarely in the NF and likely to expand in the next 100 years, while Migrate (+ or ++) indicate that the species did not occur on FIA plots, but SHIFT (RCP 4.5 or 8.5) did indicate potential for colonization in the NF within 100 years. We also show two classes of Likely (+ or ++) to show, because of signals within SHIFT outputs, that the species is likely present in the unit even though it was not found on the FIA plots within the unit. SSO is species selection option to assist in decisions regarding promoting the species, where 1 indicates the species is currently present and has at least a fair capability to cope, 2 indicates the species is rare or close to the NF boundary and has a good chance of spreading into the NF, 3 indicates the species is not recorded in FIA plots but does have some chance of getting colonized within 100 years, and 0 indicates further evaluation may be required. Finally, the N column simply is a counter. The table is explained fully in Supplementary File 3-Explanation of species tables.
… 
Content may be subject to copyright.
Article
Facilitating Adaptive Forest Management under
Climate Change: A Spatially Specific Synthesis of
125 Species for Habitat Changes and Assisted
Migration over the Eastern United States
Louis R. Iverson 1, *, Anantha M. Prasad 1, Matthew P. Peters 1and Stephen N. Matthews 1,2
1USDA Forest Service, Northern Research Station, Northern Institute of Applied Climate Science,
359 Main Road, Delaware, OH 43015, USA; anantha.prasad@usda.gov (A.M.P.);
matthew.p.peters@usda.gov (M.P.P.); matthews.204@osu.edu (S.N.M.)
2
School of Environment and Natural Resources, Kottman Hall, 2021 Coey Road, The Ohio State University,
Columbus, OH 43210, USA
*Correspondence: louis.iverson@usda.gov; Tel.: +1-740-368-0097
Received: 6 September 2019; Accepted: 30 October 2019; Published: 6 November 2019


Abstract:
We modeled and combined outputs for 125 tree species for the eastern United States, using
habitat suitability and colonization potential models along with an evaluation of adaptation traits.
These outputs allowed, for the first time, the compilation of tree species’ current and future potential
for each unit of 55 national forests and grasslands and 469 1
×
1 degree grids across the eastern
United States. A habitat suitability model, a migration simulation model, and an assessment based
on biological and disturbance factors were used with United States Forest Service Forest Inventory
and Analysis data to evaluate species potential to migrate or infill naturally into suitable habitats
over the next 100 years. We describe a suite of variables, by species, for each unique geographic
unit, packaged as summary tables describing current abundance, potential future change in suitable
habitat, adaptability, and capability to cope with the changing climate, and colonization likelihood
over 100 years. This resulting synthesis and summation eort, culminating over two decades of work,
provides a detailed data set that incorporates habitat quality, land cover, and dispersal potential,
spatially constrained, for nearly all the tree species of the eastern United States. These tables and
maps provide an estimate of potential species trends out 100 years, intended to deliver managers and
publics with practical tools to reduce the vast set of decisions before them as they proactively manage
tree species in the face of climate change.
Keywords:
suitable habitat; migration; dispersal model; range shifts; decision-support tools; adaptive
management; DISTRIB-II; SHIFT
1. Introduction
Human-induced rapid climate change will have profound impacts [
1
,
2
], including large
impacts on the earth’s biodiversity [
3
]. A critical need towards better understanding potential
impacts and adaptation strategies for biodiversity conservation hinges on the development of good
predictive models, yet basic biological information, especially related to species interactions, dispersal,
demography, physiology, and evolution, is sorely lacking for most of the earth’s biodiversity [
4
].
Species distribution models, by themselves, often do not adhere to consistent standards for model
building, biological data incorporation, and model evaluation [
5
], and do not adequately account for
these biological attributes [
6
]. Even for well-studied species, critical information is often lacking for
the development of more realistic models, though such models are beginning to emerge for selected
Forests 2019,10, 989; doi:10.3390/f10110989 www.mdpi.com/journal/forests
Forests 2019,10, 989 2 of 25
species that provide more realism via hybrid models that combine aspects of species distribution
models and process models, and that can assist in forest management [
7
,
8
]. However, tools that help
manage the entire suite of species in the face of a changing climate are vital [9].
Our objective was to move beyond the individual evaluation of tree species or forest types by
synthesizing the disparate species-based results and tying them to unique geographic units, then
providing a comprehensive report of current and potential future configurations of forest communities
impacted by climate change. In doing so, we are likely to gain new understanding of both the
current configuration of forest communities but also be in a more robust position to evaluate how
climate change will continue to exert macroscale pressures across the continent. Our approach has
been to combine the results of a habitat suitability model (DISTRIB-II, for projecting potential future
suitable habitats), a probabilistic model (SHIFT, for estimating natural migration into projected suitable
habitat within 100 years), a literature-based set of modification factors (ModFacs, for evaluating
traits associated with adaptation), and a recent assessment of Forest Inventory and Analysis data
(FIA, www.fs.fed.us/fia) [10,11].
We use two spatial subdivisions of the eastern US in this approach to assess the potentials of
the tree species to resist, adapt, or migrate throughout this century—469 units comprising a grid of
1×1
degrees of latitude and longitude, and 55 units comprising the eastern US’ national forests and
grasslands. Within each unit, we evaluate many aspects of tree species’ potential behavior in light of
the changing climate and tabulate them in an information-packed table. Within each table, a user will
be able to assess the dynamics of the forest species in their unit, both currently as well as potentially
out to century’s end. We focus on evaluating species for their capacities to persist or migrate in the
changing climate. By persist, we mean a species’ capability to resist or adapt to the added stresses
associated with the changing climate, based on a species abundance and innate attributes (i.e., traits).
Because our models cannot detect which mechanism (resist or adapt) is in play to allow the species
to persist, we generalize to infer where conditions may allow the species to persist into the future.
With migration, we mean a species’ tendency to migrate (most often northward) to follow suitable
climate conditions into the future under the changed climate, and we do estimate migration with our
models. Though migration can also mean movement of species upslope, we cannot address that in this
study due to the coarse level of analyses.
With this combined species approach, our objective is to focus on providing tables of information
for any location in the eastern US, along with any eastern national forest or grassland. This combination
of the species distribution modeling of DISTRIB-II and the migration simulations of SHIFT allow,
for the first time, a credible picture of what tree species conditions are like now as well as potential
species changes into the future. The intention is to provide guidelines and tools enabling managers
and publics alike to engage proactively in adaptation eorts in the face of the changing climate. With
this information aggregated spatially, we are also able to map spatial trends of any of the tabulated
data from the 1
×
1 degree units or national forests and grasslands within the eastern US. This mapping
approach allows a comprehensive evaluation of a full suite of tree species and their spatial trends, both
currently and potentially into the future.
The intention is to create a set of outputs geared to assist in decision support for local and regional
forest management in the face of a changing climate. In keeping with our earlier products associated
with the Northern Institute of Applied Climate Science [
12
18
], this undertaking is an ambitious step
forward: to provide decision tools to managers by comprehensively synthesizing the habitat and
colonization potential of 125 tree species in the eastern United States.
2. Material and Methods
2.1. Study Area
This study encompasses the United States east of the 100th meridian. To summarize model
outputs, the study region was sliced into 469 1
×
1 degree grids (hereafter 1
×
1), which provides a
Forests 2019,10, 989 3 of 25
wall-to-wall coverage for any location in the region to assess general trends at a coarsely summarized
scale (Figure 1). In addition, we subset only the 55 national forests and grasslands (hereafter NF) to
exemplify the capability to assess any set of geographic entities within the region (Figure 1). These two
spatial units provide unique geographic footprints to (1) provide a continuous grid where transitions
and patterns in species can be quantified; and (2) the federal lands provide a spatially distributed set of
focal sites where dierences and similarities can be evaluated from a management perspective.
Forests 2019, 9, x FOR PEER REVIEW 3 of 25
the capability to assess any set of geographic entities within the region (Figure 1). These two spatial
units provide unique geographic footprints to (1) provide a continuous grid where transitions and
patterns in species can be quantified; and (2) the federal lands provide a spatially distributed set of focal
sites where differences and similarities can be evaluated from a management perspective.
Figure 1. The National Forests and Grasslands (NF), superimposed on the 1 × 1 degree grid of the
eastern US. Those NF locations occupying <8000 km2 were buffered to attain at least that size, in order
to access sufficient FIA plots for reasonable model outputs. The names of the forests or grasslands to
which the numbers correspond are presented in Table 3. The 469 1 × 1 degree units are also depicted
and provide a wall-to-wall coverage. The individual files are named by their southeast boundary,
such that, for example, a GPS coordinate of 42.44°N latitude, 82.55°W longitude will be named
S42_E82.
2.2. The Synthetic Approach
The overall synthesizing approach uses the results of the previously mentioned DISTRIB-II,
SHIFT, and ModFacs model components; details on these components are provided in an online
supplement as well as previous papers (Figure 2, Supplementary File 1Climate Change Atlas
description) [11,1921]. DISTRIB-II provides estimates of habitat suitability (or habitat quality, HQ),
by species, to century’s end according to various scenarios of climate change. A hybrid lattice of 20 ×
20 or 10 × 10 km cells was used for DISTRIB-II modeling with the cell size determined according to
the number of forested FIA plots in each cell. SHIFT estimates colonization likelihood, CL into new
habitats, for each 1 × 1 km cell over approximately 100 years, by running the algorithm 100 times,
with each time a cell gets colonized counting as a 1% probability of colonization. ModFacs uses a
literature approach to assess 12 disturbance and 9 biological attributes of species towards their
adaptability (Adap) to deal with additional stresses likely under climate change.
Figure 1.
The National Forests and Grasslands (NF), superimposed on the 1
×
1 degree grid of the
eastern US. Those NF locations occupying <8000 km
2
were buered to attain at least that size, in order
to access sucient FIA plots for reasonable model outputs. The names of the forests or grasslands to
which the numbers correspond are presented in Table 3. The 469 1
×
1 degree units are also depicted
and provide a wall-to-wall coverage. The individual files are named by their southeast boundary, such
that, for example, a GPS coordinate of 42.44
N latitude,
82.55
W longitude will be named S42_E82.
2.2. The Synthetic Approach
The overall synthesizing approach uses the results of the previously mentioned DISTRIB-II, SHIFT,
and ModFacs model components; details on these components are provided in an online supplement as
well as previous papers (Figure 2, Supplementary File 1—Climate Change Atlas description) [
11
,
19
21
].
DISTRIB-II provides estimates of habitat suitability (or habitat quality, HQ), by species, to century’s
end according to various scenarios of climate change. A hybrid lattice of 20
×
20 or 10
×
10 km cells
was used for DISTRIB-II modeling with the cell size determined according to the number of forested
FIA plots in each cell. SHIFT estimates colonization likelihood, CL into new habitats, for each
1×1 km
cell over approximately 100 years, by running the algorithm 100 times, with each time a cell gets
colonized counting as a 1% probability of colonization. ModFacs uses a literature approach to assess
12 disturbance and 9 biological attributes of species towards their adaptability (Adap) to deal with
additional stresses likely under climate change.
Forests 2019,10, 989 4 of 25
Forests 2019, 9, x FOR PEER REVIEW 4 of 25
Figure 2. Flowchart depicting workflow to produce the DISTRIB-II and SHIFT outputs. DISTRIB-II
predicts current and future habitat suitability/quality (HQ) and SHIFT calculates colonization
likelihoods (CL) and ModFacs estimates adaptability (Adap).
A focus of this paper is a tabular output for a specific 1 × 1 or NF unit that provides species-level
details on current status, potential future status, and features related to possible management such
as assisted migration; the outputs are derived from combinations of the three model components
(Figure 3). Tables, such as that produced in Microsoft Excel for the Allegheny NF (Figure 4,
Supplementary File 2Allegheny table) and identical to those produced for the 1 × 1 degree units,
use the headers described below (in bold print) but also include, in seven accompanying excel sheets,
a much more inclusive set of data along with suggestions on how to interpret the tables, variable
descriptions, and a set of questions that can be asked of the data. A full explanation of the table
components is also provided in Supplementary File 3Explanation of species tables. These tables
were generated for each of the 55 NF units and the 469 1 × 1 units. These tables are available for
download at https://doi.org/10.2737/Climate-Change-Atlas-Combined-v4.
Figure 3. Flowchart depicting workflow for outputs of DISTRIB-II (habitat quality, HQ) and SHIFT
(colonization likelihood, CL) to produce capability and candidate species for Infill (the number of
species that can be considered likely to expand under a RCP, and thus higher candidates for infill
planting), Likely (the species was not found in FIA plots, but there is a relatively high probability that
the species exists in the unit), or Migrate (the number of species with at least some chance of migrating
naturally into each unit, and thus higher candidates for artificial migration). HQCL is the combination
of habitat quality (HQ) and colonization likelihood (CL). %OccCol pertains to the percentage of the
unit that is either already occupied or with at least a 50% probability of becoming potentially occupied
(according to SHIFT) within 100 years. %50Col (or %2Col) pertains to the percentage of the unit not
already occupied that has at least a 50% (or 2%) probability of getting colonized within 100 years.
Figure 2.
Flowchart depicting workflow to produce the DISTRIB-II and SHIFT outputs. DISTRIB-II
predicts current and future habitat suitability/quality (HQ) and SHIFT calculates colonization likelihoods
(CL) and ModFacs estimates adaptability (Adap).
A focus of this paper is a tabular output for a specific 1
×
1 or NF unit that provides species-level
details on current status, potential future status, and features related to possible management such as
assisted migration; the outputs are derived from combinations of the three model components (Figure 3).
Tables, such as that produced in Microsoft Excel for the Allegheny NF (Figure 4, Supplementary
File 2—Allegheny table) and identical to those produced for the 1
×
1 degree units, use the headers
described below (in bold print) but also include, in seven accompanying excel sheets, a much more
inclusive set of data along with suggestions on how to interpret the tables, variable descriptions,
and a set of questions that can be asked of the data. A full explanation of the table components is
also provided in Supplementary File 3—Explanation of species tables. These tables were generated
for each of the 55 NF units and the 469 1
×
1 units. These tables are available for download at
https://doi.org/10.2737/Climate-Change-Atlas-Combined-v4.
Forests 2019, 9, x FOR PEER REVIEW 4 of 25
Figure 2. Flowchart depicting workflow to produce the DISTRIB-II and SHIFT outputs. DISTRIB-II
predicts current and future habitat suitability/quality (HQ) and SHIFT calculates colonization
likelihoods (CL) and ModFacs estimates adaptability (Adap).
A focus of this paper is a tabular output for a specific 1 × 1 or NF unit that provides species-level
details on current status, potential future status, and features related to possible management such
as assisted migration; the outputs are derived from combinations of the three model components
(Figure 3). Tables, such as that produced in Microsoft Excel for the Allegheny NF (Figure 4,
Supplementary File 2Allegheny table) and identical to those produced for the 1 × 1 degree units,
use the headers described below (in bold print) but also include, in seven accompanying excel sheets,
a much more inclusive set of data along with suggestions on how to interpret the tables, variable
descriptions, and a set of questions that can be asked of the data. A full explanation of the table
components is also provided in Supplementary File 3Explanation of species tables. These tables
were generated for each of the 55 NF units and the 469 1 × 1 units. These tables are available for
download at https://doi.org/10.2737/Climate-Change-Atlas-Combined-v4.
Figure 3. Flowchart depicting workflow for outputs of DISTRIB-II (habitat quality, HQ) and SHIFT
(colonization likelihood, CL) to produce capability and candidate species for Infill (the number of
species that can be considered likely to expand under a RCP, and thus higher candidates for infill
planting), Likely (the species was not found in FIA plots, but there is a relatively high probability that
the species exists in the unit), or Migrate (the number of species with at least some chance of migrating
naturally into each unit, and thus higher candidates for artificial migration). HQCL is the combination
of habitat quality (HQ) and colonization likelihood (CL). %OccCol pertains to the percentage of the
unit that is either already occupied or with at least a 50% probability of becoming potentially occupied
(according to SHIFT) within 100 years. %50Col (or %2Col) pertains to the percentage of the unit not
already occupied that has at least a 50% (or 2%) probability of getting colonized within 100 years.
Figure 3.
Flowchart depicting workflow for outputs of DISTRIB-II (habitat quality, HQ) and SHIFT
(colonization likelihood, CL) to produce capability and candidate species for Infill (the number of
species that can be considered likely to expand under a RCP, and thus higher candidates for infill
planting), Likely (the species was not found in FIA plots, but there is a relatively high probability that
the species exists in the unit), or Migrate (the number of species with at least some chance of migrating
naturally into each unit, and thus higher candidates for artificial migration). HQCL is the combination
of habitat quality (HQ) and colonization likelihood (CL). %OccCol pertains to the percentage of the
unit that is either already occupied or with at least a 50% probability of becoming potentially occupied
(according to SHIFT) within 100 years. %50Col (or %2Col) pertains to the percentage of the unit not
already occupied that has at least a 50% (or 2%) probability of getting colonized within 100 years.
Forests 2019,10, 989 5 of 25
Forests 2019, 9, x FOR PEER REVIEW 5 of 25
Figure 4. HQ and CL output table for the Allegheny National Forest, Pennsylvania, sorted in
decreasing order of current species abundance (FIAsum). Range indicates whether species is wide or
narrow (W/N) in distribution, dense or sparse (D/S) in frequency of presence within FIA plots, and
high or low (H/L) in importance value when it is found. MR refers to model reliability (see [21] for
explanation). %Cell refers to the percentage of DISTRIB-II cells with the species present. FIAiv is the
average importance value for the species when present on FIA plots. ChngCl45 or 85 presents the
change classes (increase, decrease, or no change) of habitat suitability by 2100, according to RCP 4.5
(low emissions) or 8.5 (high emissions). Adapt is a class of adaptability of the species according to the
ModFacs. Abund is an abundance class based on FIAsum. Capabil45 or 85 is the capability of the
species to cope with the climates of RCP 4.5 or 8.5 at 2100, based on abundance, change classes, and
adaptability. SHIFT45 and 85 show two classes of Infill (+ or ++) to indicate these species are currently
found rarely in the NF and likely to expand in the next 100 years, while Migrate (+ or ++) indicate that
the species did not occur on FIA plots, but SHIFT (RCP 4.5 or 8.5) did indicate potential for
colonization in the NF within 100 years. We also show two classes of Likely (+ or ++) to show, because
of signals within SHIFT outputs, that the species is likely present in the unit even though it was not
found on the FIA plots within the unit. SSO is species selection option to assist in decisions regarding
promoting the species, where 1 indicates the species is currently present and has at least a fair
capability to cope, 2 indicates the species is rare or close to the NF boundary and has a good chance
of spreading into the NF, 3 indicates the species is not recorded in FIA plots but does have some
chance of getting colonized within 100 years, and 0 indicates further evaluation may be required.
Finally, the N column simply is a counter. The table is explained fully in Supplementary File 3
Explanation of species tables.
Figure 4.
HQ and CL output table for the Allegheny National Forest, Pennsylvania, sorted in decreasing
order of current species abundance (FIAsum). Range indicates whether species is wide or narrow (W/N)
in distribution, dense or sparse (D/S) in frequency of presence within FIA plots, and high or low (H/L)
in importance value when it is found. MR refers to model reliability (see [
21
] for explanation). %Cell
refers to the percentage of DISTRIB-II cells with the species present. FIAiv is the average importance
value for the species when present on FIA plots. ChngCl45 or 85 presents the change classes (increase,
decrease, or no change) of habitat suitability by 2100, according to RCP 4.5 (low emissions) or 8.5 (high
emissions). Adapt is a class of adaptability of the species according to the ModFacs. Abund is an
abundance class based on FIAsum. Capabil45 or 85 is the capability of the species to cope with the
climates of RCP 4.5 or 8.5 at 2100, based on abundance, change classes, and adaptability. SHIFT45 and
85 show two classes of Infill (+or ++) to indicate these species are currently found rarely in the NF and
likely to expand in the next 100 years, while Migrate (+or ++) indicate that the species did not occur on
FIA plots, but SHIFT (RCP 4.5 or 8.5) did indicate potential for colonization in the NF within 100 years.
We also show two classes of Likely (+or ++) to show, because of signals within SHIFT outputs, that the
species is likely present in the unit even though it was not found on the FIA plots within the unit. SSO
is species selection option to assist in decisions regarding promoting the species, where 1 indicates the
species is currently present and has at least a fair capability to cope, 2 indicates the species is rare or
close to the NF boundary and has a good chance of spreading into the NF, 3 indicates the species is not
recorded in FIA plots but does have some chance of getting colonized within 100 years, and 0 indicates
further evaluation may be required. Finally, the N column simply is a counter. The table is explained
fully in Supplementary File 3—Explanation of species tables.
Forests 2019,10, 989 6 of 25
The FIA plot data were tabulated within each unit (either 1
×
1 or NF) to yield a ranked list of
tree species, by importance value (IV) quantified equally between total basal area and number of
stems. The actual IVs, under
FIAsum
(bolded elements are headers for the output tables for each
unit, example shown in Figure 4; see also Supplementary File 3—Explanation of species tables), were
summed for each 10
×
10 or 20
×
20 km cell within the unit and recalibrated to the standardized area of
10,000 km
2
(the approximate area of a 1
×
1 degree grid at 36
N latitude), so that values are normalized
and comparisons among units can be made on species importance; they then were assigned to one of
four abundance classes (noted as
Abund
) according to these breakpoints: abundant (FIAsum >75),
common (FIAsum 5–75), rare (FIAsum >0–5), and absent (FIAsum =0). The IVs were also averaged for
only those cells (noted as
%Cell
) that contained the species to yield the importance of the species only
where it is present, not summed over the entire unit (noted as
FIAiv
). An examination of the IV for each
species across its range allowed us to assign a
Range
classification. The Range field provides a quick
indication if the species of interest is narrow or widely dispersed across North America (
N vs. W
), with
Narrow meaning the species is found across <10% of DISTRIB-II grid cells in the eastern US and Wide
>10%. The species is also categorized as found commonly (Dense if
40% of FIA plots among grid
cells with IV >0) or rarely (Sparse if <40% of FIA plots have the species) within its overall distribution
among eastern US FIA records. Finally, the ecological importance or abundance for each species can be
indicated by the average importance value among the plots where the species is present, with High
if average IV
6.0 (the median of mean IV across all species) and Low for values <6.0 [
21
]. We also
assigned three levels of model reliability (
MR
) [
21
]. To do so, we evaluated and combined five model
performance variables into a single rating: (1) a pseudo-R
2
obtained from the RandomForest (RF)
model; (2) a Fuzzy Kappa comparing the imputed RF map to the FIA-derived map [
22
]; (3) a tree
skill statistic of the imputed RF, after removing records with very high coecient of variables (CV);
(4) the deviance of the CV among 30 regression trees via bagging [
23
,
24
]; and (5) the stability of the top
five variables from 30 regression trees [25].
By ratioing suitable habitat (IV) in future to suitable habitat at present, we can generate five classes
of change (noted as
ChngCl45
or
ChngCl85
, depending on representative concentration pathways
(RCP) 4.5 and RCP 8.5; Figure 3, [
21
]) according to the following break points in future:actual IV
ratio: large decrease (<0.5), small decrease (0.5–0.8), no change (0.8–1.2), small increase (1.2–2.0),
large increase (>2.0). For rare species, those occupying <10% of the region, the following break
points were used: large decrease (<0.2), small decrease (0.2–0.5), no change (0.5–4.0), small increase
(4.0–8.0), and large increase (>8.0). Ecologically, the assumption is that if the species is projected have
increased (or decreased) summed IVs in the future, conditions will be more (or less) favorable for the
species and is categorically represented by the change classes. The ModFacs assessment, based on
12 disturbance and 9 biological traits, provided a baseline rating as to the species’ adaptability (
Adap
)
to the changing climate (Figure 3, [
19
]). The disturbance factors address, based on literature surveys,
how well the species is expected to cope with additional stresses from drought, flood, invasives, wind,
fire, and the like into the future, while the biological factors address the capability of the species to
regenerate vegetatively or via seed, its shade tolerance, its edaphic and habitat specificity, and the
like. Adaptability scores were generated via weighted summations of climate-related biological and
disturbance scores and classified into low, medium, and high adaptability [
19
] (see also Supplementary
File 1—Climate Change Atlas). Notably, adaptability scores are based on the species’ attributes
across their entire range, but they may vary in their response to disturbances in certain locations;
managers may adjust individual species scores according to local knowledge. The SHIFT model, which
calculates colonization likelihood (CL) [
11
,
26
], was calibrated to approximate 50 km migration per
century, an optimistic assumption across species [
27
]. Though migration is dependent on the dispersal
characteristics of the species, and a single migration rate of 50 km/100 years is a broad assumption,
insucient information exists to defensively assign rates to individual species. We therefore use a
historically defensible rate based on Holocene estimates [27,28].
Forests 2019,10, 989 7 of 25
2.3. Rules for Capability to Cope with a Changing Climate
Besides habitat suitability (ChngCl45 or ChgnCL85), additional information is desired to assess
the species at a more specific spatial unit—the 1
×
1 or NF. To do so, we developed a rating scheme to
include suitability, adaptability, and abundance to derive a capability for each species to withstand the
challenges (and expand with opportunities) posed by the changing climate. We calculate a five-class
capability rating (
Capabil45
or
Capabil85
) based on Abund, Adap, and ChngCl45 or ChngCl85
(Table 1). Following the capability scoring presented in Table 1, if Abund was ‘Abundant’, we enhanced
the capability by one class (e.g., a rating of ‘Good’, became ‘Very Good’ in the final class). Similarly,
if Abund was ‘Rare’, we decreased the rating by one class, and if Abund was ‘Common’, no alteration
occurred. We somewhat arbitrarily promote or demote the capability one class based on abundance.
We assume that within the large area (roughly 8000–10,000 km
2
) of landscape within each area of
interest, there will be spatial variability of climate driven by topographic or geomorphic diversity [
29
]
which will increase the probability of persistence under a changed climate. If the species is abundant
now, we assume there is a greater potential for specific habitats to be suitable and/or genetic variants
to be more resistant. We realize we do not have direct data to support these assumptions although
landscape diversity has been linked to biodiversity and resilience [
30
]; we can only logically infer the
assumption will be true for at least a portion of the species we model.
Table 1.
Initial capability class ratings defined as colored classes, depending on the change class and
adaptability of the species, and other classes for selected species. For final capability rating (Very
Good to Very Poor): if abundance was ‘Abundant’, move up one class; if ‘Rare’ move down one class;
if ‘Common’ stay in class.
Adaptability
Change Class High Medium Low
Large increase Very Good Very Good Good
Small increase Very Good Good Fair
No change Good Fair Poor
Small decrease Fair Poor Poor
Large decrease Fair Poor Very Poor
Other capability classes:
Lost All suitable habitat lost
New Habitat New habitat appearing
FIA Only
Unacceptable model for future, only FIA reported
NNIS Non-native invasive species, only FIA reported
Unknown Modeled as present, unknown
2.4. Rules for Expand, Migrate, and Likely, and Species Selection Options
We used CL in conjunction with HQ to generate HQ-CL classes of species for consideration for
planting or other forms of promotion within the areal units of analyses (Figure 3). These are included in
the fields
SHIFT45
(RCP 4.5) or
SHIFT85
(RCP 8.5) in the table outputs. ‘Infill’ represents those species
that are currently rare in the unit but have high potential to expand within its boundaries. Two classes
(Infill+and Infill++) were generated based on the capability and the percent occupancy currently, plus
those areas with at least a 50% CL within 100 years (Table 2). The ‘Likely’ classes indicate that the
species was not found in FIA plots, but that there is a relatively high probability that the species exists
in the unit, because either the species has some area with at least 50% CL (Likely+) or that it has both
HQ-CL and at least 2% of the area has at least 50% CL (Likely++). The ‘Migrate’ classes indicate that
the species was not found by FIA, but that the DISTRIB-II model indicates new habitat appears in
the unit, with either low HQ and CL (HQ-CL =1, Migrate+) or higher levels of these two components
(HQ-CL >1, Migrate++). Admittedly, this is a very liberal assessment for ‘Migrate’, as the area qualifies if
only a fraction of the area of interest has only a 2% probability of colonization within 100 years.
Species selection options (
SSO
) provides options to assist in decisions regarding promoting the
species, where 1 indicates the species is currently present and has at least a fair capability to cope with
Forests 2019,10, 989 8 of 25
the changing climate—thus suited for planting; 2 indicates the species is rare or in close proximity
(SHIFT shows potential to reach the area within 100 years) to the 1
×
1 grid or NF boundary and has a
good chance of spreading into the area—thus also suited for planting but perhaps not as suited as #1;
3 indicates the species is not recorded in FIA plots but does have some chance of getting colonized
within 100 years—thus could be planted similar to ‘Migrate+or Migrate++’ mentioned above; and 0
indicates none of the above but further evaluation may be required before eliminating consideration of
the species for planting.
Table 2.
Criteria for classes of Infill, Likely, and Migrate, according to the HQ and CL outputs.
Capability and HQ-CL are determined separately for RCP 4.5 and RCP 8.5. %OccCol pertains to the
percentage of the unit that is either already occupied or with at least a 50% probability of becoming
potentially occupied (according to SHIFT) within 100 years. %2Col pertains to the percentage of the
unit not already occupied that has at least a 2% probability of getting colonized within 100 years.
Class Capability HQCL %OccCol %2Col
Infill+Fair, Poor 1>0–50 any
Infill++ Good, Very Good 1>10–50 any
Likely+New Habitat, Unknown any >0 any
Likely++ New Habitat, Unknown 1>2 any
Migrate+New Habitat =1 0 >0
Migrate++ New Habitat >1 0 >0
2.5. Mapping Aggregated 1 ×1 Outputs
The 1
×
1 degree units represent approximately 10,000 km
2
at 36
latitude. Because of the Earth’s
spherical shape, the east–west width (and area) is smaller north of that line and larger south of that line.
For smaller units of analysis (e.g., national forests or grasslands represented here), the unit was buered
so that a minimum of 8000 km
2
was represented for analysis (Figure 1). Thus, whenever we refer to a
NF, we refer to this buered (if necessary to reach 8000 km
2
minimum size) area. This minimum size
was required so that an adequate number of FIA plots could be represented for each forest within the
modeling grid [
21
]. The 1
×
1 degree units also represent ~100 10
×
10 km cells or
~25 20 ×20 km cells
,
or a mixture; the buered NFs represent at least 80 10
×
10 km cells or 20 20
×
20 km cells—these
numbers of cells represent sucient samples for meaningful summing or averaging across the unit,
and are the reason for choosing such a coarse analysis. As mentioned above, the FIAsum values were
all standardized to equate to the 10,000 km2to enable cross-table comparisons of species importance;
comparisons could therefore be made among the 469 1 ×1 units or the 55 NF units.
By combining information across the 469 1
×
1 tables, mapping of particular characteristics could
be accomplished. To do so, counts were made of the number of species matching the criterion being
queried for each 1
×
1 degree unit and aggregated across units to populate the maps. Though any of
the tabulated variables described above could be mapped, we present below: (1) the number of total
species quantified in each unit; (2) the number of oak (Quercus spp.) species, now and potentially in the
future; (3) the number of species with at least some chance of migrating naturally into each unit under
RCP 8.5, and thus higher candidates for artificial migration (Migrate+plus Migrate++); and (4) the
number of species that can be considered likely to expand under RCP 8.5, and thus higher candidates
for infill planting (Infill+plus Infill++).
3. Results
3.1. Mapping Habitat Quality and Colonization Likelihood Outputs
The combination of HQ and CL is exemplified for eastern hemlock (Tsuga canadensis) in Figure 5.
First, the current condition is established using FIA data within the modeling grid (Figure 5a). Then
potential suitable habitat (habitat quality, or HQ) by 2100 was mapped under RCP 8.5 (Figure 5b).
The colonization likelihood (CL) map was independently created (Figure 5c), then intersected with the
Forests 2019,10, 989 9 of 25
HQ map to yield a map depicting those areas that have at least some chance of colonization within
100 years (Figure 5d). Both the HQ and CL maps were reclassified into three possible classes (low,
medium, high), with IV breakpoints of HQ at 5 (break between low and medium) and 15 (between
medium and high) and CL breakpoints at 10 and 50 percent. Merging the two outputs can produce up
to nine combinations; for eastern hemlock, there was no high HQ so only six HQ-CL combinations
are shown (Figure 5d). Finally, the combined HQ-CL information was used to identify those areas
where the species could ‘infill’, or fill in cells that have no FIA evidence that it exists now, but is closely
surrounded by the species (Figure 5, Infill inset map). Or, if the species could have some potential
to move into new habitat not previously occupied (most often to the north of the current boundary),
those areas can be noted as ‘migrate’ areas (Figure 5, Migrate inset map). We also can identify those
species that are ‘likely’ present but missed by FIA (these locations would appear within the ‘infill’
zones on the map but cannot be precisely located). Thus, the combination of HQ and CL results not
only identifies potential changes in suitable habitat under various scenarios of climate change, but also
provides, for each species present currently or potentially in the future, estimates of CL from natural
migration within 100 years.
Forests 2019, 9, x FOR PEER REVIEW 9 of 25
100 years (Figure 5d). Both the HQ and CL maps were reclassified into three possible classes (low,
medium, high), with IV breakpoints of HQ at 5 (break between low and medium) and 15 (between
medium and high) and CL breakpoints at 10 and 50 percent. Merging the two outputs can produce
up to nine combinations; for eastern hemlock, there was no high HQ so only six HQ-CL combinations
are shown (Figure 5d). Finally, the combined HQ-CL information was used to identify those areas
where the species could ‘infill’, or fill in cells that have no FIA evidence that it exists now, but is
closely surrounded by the species (Figure 5, Infill inset map). Or, if the species could have some
potential to move into new habitat not previously occupied (most often to the north of the current
boundary), those areas can be noted as ‘migrate’ areas (Figure 5, Migrate inset map). We also can
identify those species that are ‘likely’ present but missed by FIA (these locations would appear within
the ‘infill’ zones on the map but cannot be precisely located). Thus, the combination of HQ and CL
results not only identifies potential changes in suitable habitat under various scenarios of climate
change, but also provides, for each species present currently or potentially in the future, estimates of
CL from natural migration within 100 years.
Figure 5. Process flow to intersect HQ and CL for eastern hemlock. (a) Current importance value of
hemlock as determined by FIA data; (b) habitat quality (HQ) at RCP 8.5, year 2100, reclassed into low,
medium, high HQ (there was no high HQ for hemlock), at a resolution of 10 × 10 or 20 × 20 km; (c)
colonization likelihood (CL) into RCP 8.5 habitat, and reclassed into low, medium, high CL, at a
resolution of 1 × 1 km; (d) combination of (ac), yielding locations with null to high CL on top of low
to medium HQ, as well as currently occupied cells; (infill inset) detail of locations where infilling is
primary; (migrate inset) detail of locations where migrating is primary.
3.2. Species Summaries by National Forest
Figure 5.
Process flow to intersect HQ and CL for eastern hemlock. (
a
) Current importance value of
hemlock as determined by FIA data; (
b
) habitat quality (HQ) at RCP 8.5, year 2100, reclassed into
low, medium, high HQ (there was no high HQ for hemlock), at a resolution of
10 ×10 or 20 ×20 km
;
(
c
) colonization likelihood (CL) into RCP 8.5 habitat, and reclassed into low, medium, high CL,
at a resolution of 1
×
1 km; (
d
) combination of (
a
c
), yielding locations with null to high CL on top of
low to medium HQ, as well as currently occupied cells; (infill inset) detail of locations where infilling is
primary; (migrate inset) detail of locations where migrating is primary.
3.2. Species Summaries by National Forest
We present our primary tabular output for one NF, the Allegheny NF (coded ‘1’ in Figure 1and
record 1 in Table 3), but similar tables have been prepared for each 1
×
1 degree area and are available
Forests 2019,10, 989 10 of 25
for download. Other units of analysis will eventually be available including national parks, watersheds,
ecoregions, states, and others. Our combination of HQ, CL, Adap, and current FIA estimates of
importance value allow a detailed presentation of (1) species importance currently; (2) the potential
changes in suitable habitat by 2100; (3) the adaptability of each species to the changing climate;
(4) the capability of each species to cope with the 2100 climate based on adaptability and abundance
currently within the NF; (5) the likelihood of each species to naturally migrate into the NF; and
(6) an assessment of the potential for the species to be used for planting or otherwise promoting within
the NF. These are all presented within an information-packed, but easily unpacked table (Figure 4).
Table 3.
Identity code (noted in Figure 1), coordinates, actual area, and buered area (km
2
) of each
national forest or grassland. National grasslands are noted as ‘NG’; all others are national forests. Also
shown are results for number of species recorded by FIA within the buer region, and the number of
species potentially infilling, likely, and migrating according to the RCP 8.5 scenario of climate change
(see text for explanation). For simplicity, only RCP 8.5 data are shown.
Identifier Name Longitude Latitude Area,
km2
Buer
Area, km2
Number
Species
Number
Infill
Number
Migrate
1 Allegheny 79 41.7 2996.5 8200 43 6 11
2 Angelina 94.4 31.3 1602.8 8800 61 22 1
3 Apalachicola 84.7 30.2 2555.9 8044 64 18 1
4 Bienville 89.5 32.3 1572 8900 67 16 2
5 Black_Kettle NG 99.5 35.7 982 8196 18 2 6
6 Caddo NG 96 33.6 277.3 9100 46 13 0
7 Chattahoochee 84.2 34.7 5950.2 9700 72 24 9
8 Chequamegon 90.8 46 4209.7 11,815 40 5 9
9 Cherokee 83.1 35.8 5950.2 9600 74 35 9
10 Chippewa 94.1 47.4 6465.5 8800 34 6 11
11 Conecuh 86.6 31.1 692.9 8500 60 18 4
12 Croatan 77.1 35 1243.7 8034 62 17 2
13 Daniel_Boone 83.9 37.3 8282.7 14,100 77 10 4
14 Davy_Crockett 95.1 31.3 1602.8 8400 55 15 1
15 Delta 90.8 32.8 490.7 8000 62 30 3
16 DeSoto 89 31.1 3252.4 8700 67 18 0
17 Finger_Lakes 76.8 42.6 58.4 8300 62 11 8
18 Francis_Marion 79.8 33.3 1696.1 8221 70 24 1
19 George_Washington 79.3 38.2 7270.9 14,300 73 16 10
20 Green_Mountain 73 43.4 2543 8800 42 6 14
21 Hiawatha 86 46.2 5197.3 9549 35 7 14
22 Holly_Springs 89.2 34.7 1906 8500 70 13 4
23 Homochitto 90.9 31.5 1540.6 8400 72 17 2
24 Hoosier 86.5 38.5 2612 8100 65 11 4
25 Huron 84 44.6 2787.3 8078 39 6 12
26 Jeerson 81.2 37.1 4926.7 15,100 76 11 8
27 Kisatchie 92.8 31.7 4195.7 9800 74 12 2
28 Lyndon_B_Johnson NG 97.6 33.4 465.5 9200 32 14 1
29 Manistee 85.9 43.9 5377.5 8566 47 7 6
30 Mark_Twain 91.7 37.3 12,237.4 23,900 66 15 5
31 Midewin NG 88.1 41.4 104.3 8350 31 19 12
32 Monongahela 79.9 38.6 7270.9 10,500 61 8 11
33 Nantahala 83.6 35.2 5950.2 9700 65 17 9
34 Nicolet 88.7 45.7 3890.7 8200 39 8 6
35 Ocala 81.8 29.2 1791.6 8177 42 4 5
36 Oconee 83.5 33.4 1049.8 8700 62 10 3
37 Osceola 82.5 30.4 932.4 8400 44 12 6
38 Ottawa 89.2 46.5 3890.7 9559 36 7 11
39 Ouachita 93.9 34.6 9682.3 15,000 62 16 11
40 Ozark 93.5 35.7 9682.3 11,800 70 13 11
41 Pisgah 82.4 35.8 5950.2 10,100 70 12 12
42 Sabine 93.9 31.5 1602.8 8500 63 22 3
43 Saint_Francis 90.7 34.7 120.5 8700 63 45 3
44 Sam_Houston 95.4 30.5 2005.3 8200 63 18 1
45 Shawnee 88.9 37.5 3501 8000 73 13 5
46 Sheyenne NG 97.2 46.4 551.6 8400 9 6 7
47 Sumter 82.1 34.3 5950.2 8200 72 24 1
48 Superior 91.6 47.8 13,204.5 16,415 32 11 14
49 Talladega 86.5 33.2 3042.6 9700 68 15 7
50 Tombigbee 89.2 33.7 713.1 8700 69 20 2
51 Tuskegee 85.6 32.5 63.2 8100 70 21 3
52 Uwharrie 79.9 35.4 889.2 8200 73 18 4
53 Wayne 82 39.3 3451.8 10,800 71 13 5
54 White_Mountain 71.4 44.2 3511.8 8800 37 9 14
55 WB_Bankhead 87.3 34.2 1409.3 8900 71 0 0
Forests 2019,10, 989 11 of 25
3.3. Trends in Area and Species Counts for National Forests and Grasslands
Besides identifying the locations (labels for Figure 1) for the 55 national forest and grasslands,
Table 3presents the geographic coordinates and size information of both the NF itself and the buer
area used in evaluating the species for this eort. Ranging from 104.3 km
2
for the Midewin NG
to 13,204.5 km
2
for the Superior NF, there is a 126-fold variation in area from smallest to largest.
To dampen this large variation and provide sucient area for multiple FIA plots, we added area to
accumulate a minimum of 8000 km
2
surrounding each NF, so the range in buered area was 8000 km
2
for the Delta NF to 23,900 km
2
for the Mark Twain NF (consisting of several units merged together
in the buering process). Though some species–area curve impacts could be present, a correlation
between the buered area and species count was only 0.14 (NS).
Tree species counts (according to FIA) for the various NFs ranged from 9 (for the Sheyenne NG
in North Dakota) to 77 (Daniel Boone NF in Kentucky), and were negatively correlated to latitude
(
r=0.57
,p<0.001); southern NFs generally had higher species richness (Table 3). Also recorded are
the number of ‘infill’, ‘likely’, and ‘migrate’ species (either +or ++) under RCP 8.5. The table indicates
that up to eight species were likely present in the NF but missed by FIA for the Green Mountain NF,
up to 45 species (St. Francis NF) have potential to infill, or expand importance within that forest,
and up to 14 species (Green Mountain NF) have potential to migrate into this buered NF from points
south (Table 3). The count of potential species to migrate was highly correlated with latitude (r=0.72,
p<0.001),
indicating that there is a greater potential for assisted migration of tree species into the more
northerly NFs, should that be desired.
3.4. Questions and Answers for Example NF: Allegheny NF
To simplify interpretation of the table outputs, we have developed a series of 17 questions that
can be asked of the data table. Users can peruse the questions and explanations to see what might
align with their interests. The methods to (a) extract the answers and (b) provide answers for one unit,
the Allegheny NF, are presented in Appendix A. From this example, many insights on the current and
potential state of tree species on the Allegheny NF can be gleaned. A subset of the questions are as
follows, which can be asked of any 1 ×1 degree or national forest area:
1.
How many species are known (by FIA) to be present currently, what are these species, and how
abundant are they?
2. By 2100, how might suitable habitat change for trees under low and high emission scenarios?
3. How adaptable are the tree species to conditions expected to change under climate change?
4.
How capable might the species be with regards to coping with the changing climate, and what
species may be most (or least) vulnerable?
5. Within the next 100 years, what species may be able to naturally migrate into the area?
6. What species, now rare, are suited to expand or infill in importance over the century?
7. What species are likely present in the area, though missed by FIA plots?
8.
What species might be useful for selecting as candidates for planting in the face of a
changing climate?
For the Allegheny example, we show that 43 species are currently present (6 abundant, 15 common,
and 22 rare species), though one additional species is likely there, and an additional 14 species have
new habitat appearing by end of century. Of these, 13 are likely to lose suitable habitat under RCP 8.5,
while 17 gain habitat. Still, 25 species appear to have good or very good capability to cope with the
changing climate, while the most vulnerable species appear to be red spruce, red pine, black ash, pin
cherry, quaking aspen, and white spruce. Nine species show good potential to migrate into the area
under the high emission scenario, while six additional species are marked as potentials to infill into
surrounding locations from their currently relatively uncommon status. Using the species selection
options, we found 30 species common and well suited for planting, four species uncommon but with
Forests 2019,10, 989 12 of 25
good adaptability, and 12 species potentially suited for planting, though with more risk of planting
failure. Further details are provided in Appendix Aand with the full table available as Supplementary
File 2—Allegheny table.
By providing answers to the 17 questions for any national forest or 1
×
1 degree grid, a user is
able to delve into many aspects of their forest attributes, both now and potentially into the future.
Managers and the casual user alike will have species lists showing which are abundant, common,
and rare within their area of interest. Managers will have data to support (or not) their decisions
related to adaptation to the changing climate. For example, selecting species that may be candidates
for planting may be better justified if the HQ-CL data suggest that the suitable habitats are ripe for the
species in their area of interest. This is especially true for selecting species not now identified from their
area of interest—picking species with ‘Migrate+’ or ‘Migrate++’ would seem more justified in that the
habitat will likely be present in coming decades and the species is close enough now to provide some
chance of natural migration into the area even without assistance.
3.5. Mapped Summaries for 1 ×1 Degree Grid
Tree species relative richness, as recorded on FIA plots and summed for each 1
×
1 degree
grid, shows a remarkable variation in both latitude and longitude (Figure 6A). These patterns show
maximum diversity in the southeastern portion of the U.S., with the maximum count of 82 species
found on the southern Mississippi–Alabama border. Lowest species richness was found, as expected,
in the prairie regions of North and South Dakota, where very few FIA plots were placed in forested
(mostly riparian) zones.
Forests 2019, 9, x FOR PEER REVIEW 13 of 25
Figure 6. (A) The number of species recorded by FIA by 1 × 1 degree cell across the eastern US. (B)
The number of oak (Quercus spp.) species recorded by FIA by 1 × 1 degree cell across the eastern US.
(C) The number of oak (Quercus spp.) species projected to have suitable habitat (HQ) and at least some
chance of colonization (CL) at 2100 according to RCP 4.5. (D) The number of oak (Quercus spp.) species
projected to have suitable habitat (HQ) and at least some chance of colonization (CL) at 2100 according
to RCP 8.5. Figures (C,D) use same figure legend as (B).
The number of species with potential to migrate, by 1 × 1 cell and for RCP 8.5, is depicted in
Figure 7A. There is a large north to south gradient in species numbers; the correlation with latitude
is 0.72 (p < 0.001). The Plains States and Corn Belt states also have fewer species available for
migration. The maximum number of potential species, according to this analysis, that may be
appropriate to plant in an assisted migration mode, is 19, from the Adirondack Region of New York,
with many other cells in the far north also with high numbers of species with potential to migrate in
(Figure 7A).
The number and distribution of species with potential to infill provides a major contrast to the
migration map, in that numbers can range up to 54 species, and with the highest numbers in the
highly agricultural zones (Figure 7B). The maximum 1 × 1 values appear in the southern Mississippi
floodplain, in highly agricultural zones (with few forested FIA plots so only a small total area was
sampled) near locations with relatively high species richness. The correlation between Infill85 and
Migrate85 was −0.34 (p < 0.01), further indicating the inverse relationship between the two, i.e., if the
unit is highly forested, there are less species ‘available’ for infilling as they have already been
captured by the FIA sampling.
Many other maps depicting trends among the 1 × 1 degree grids have been generated, and these
reveal patterns associated with the numbers of abundant or rare species, the numbers of species
within several common genera, the numbers of species most capable to deal with the changing
climate, the top three species within each cell, and information related to the range distribution,
among others. These maps and others will be made available for users to spatially compare among 1
× 1 degree grids relative to their areas of interest.
Figure 6.
(
A
) The number of species recorded by FIA by 1
×
1 degree cell across the eastern US.
(
B
) The number of oak (Quercus spp.) species recorded by FIA by 1
×
1 degree cell across the eastern
US. (
C
) The number of oak (Quercus spp.) species projected to have suitable habitat (HQ) and at least
some chance of colonization (CL) at 2100 according to RCP 4.5. (D) The number of oak (Quercus spp.)
species projected to have suitable habitat (HQ) and at least some chance of colonization (CL) at 2100
according to RCP 8.5. Figures (C,D) use same figure legend as (B).
Forests 2019,10, 989 13 of 25
Diving further into genus-level analyses, the oaks (Quercus spp.) show highest richness in the far
south central portion of the country, maxing out at 21 species; lowest oak richness was in both the
northwestern and the northeastern portions of the study area, as well as southern Florida (Figure 6B).
To demonstrate potential changes under climate change, we also present projected counts of the
number of oaks with suitable habitat and colonization potential under RCP 4.5 (Figure 6C) and RCP 8.5
(Figure 6D). In comparison with the current oak distribution (Figure 6B), the future (~2100) models
show potentials for subtle decreases in oak species counts in the south and slight increases towards the
north. This trend is more prominent with RCP 8.5 (Figure 6D).
The number of species with potential to migrate, by 1
×
1 cell and for RCP 8.5, is depicted in
Figure 7A. There is a large north to south gradient in species numbers; the correlation with latitude is
0.72 (p<0.001). The Plains States and Corn Belt states also have fewer species available for migration.
The maximum number of potential species, according to this analysis, that may be appropriate to plant
in an assisted migration mode, is 19, from the Adirondack Region of New York, with many other cells
in the far north also with high numbers of species with potential to migrate in (Figure 7A).
Forests 2019, 9, x FOR PEER REVIEW 14 of 25
Figure 7. (A) The number of species that may potentially migrate into each 1 × 1 degree cell (from
elsewhere), within 100 years across the eastern US (Migrate+ plus Migrate++), under RCP 8.5. (B) The
number of species with potential to infill (Infill+ plus Infill++) into each 1 × 1 degree cell, within 100
years, across the eastern US.
4. Discussion
This work brings together two lines of effort that our group has pursued for over two decades:
that of the projected changes in habitat suitability under the changing climate [19,20,23,25,3140] and
that of the potential colonization likelihood into suitable habitat over the course of a century
[11,26,4143]. Advances in each line were achieved, and often necessitated, by enhancements or
updates in data, climate projections, software, and computing capabilities. By combining these lines
of effort for most of the tree species in the eastern US, we provide a mechanism to evaluate, with
more reality imbedded, the potential outcomes at end of century. The intention of this work is to
assist managers, both at the scale of regions or the entire eastern US as well as individual national
forests or locations within 1 × 1 degree units, in adapting to a changing climate. We hope that the
information outlined here can be used to assist in the silvicultural treatments available to managers,
such as plantation, assisted migration, thinning, or natural regeneration, to help forests mitigate and
adapt to the impacts of climate change [44]. This is the focus of the Northern Institute of Applied
Climate Science, and the Climate Change Response Framework (www.forestadaptation.org) [45]. As
laid out by Millar and Stephenson [46], adaptation pathways lead to resistance, resilience, or response
(or facilitation, or transition). A national experiment is testing these approaches at various sites across
the continent [47], and the data provided here can help in ascertaining species appropriate for each
of these pathways, anywhere in the eastern United States. However, caution is advised when
applying these results because no model can capture the full range of complexity acting on species
through uncertain times, and although species suggested here as potentials for assisted migration
may be adapted climatically to future climate conditions, other factors such as competition or
herbivory may override other traits and restrain regeneration success [48] (see also Supplementary
File 1Climate Change Atlas). Additional caveats more specific to these results include the
averaging over large land areas (>8000 km2), the assumption of an average 100-year migration rate
for all species of 50 km in uninterrupted forest, and the use of projected future:actual IVs as a proxy
for future importance of the species. Thus the need exists for multiple, adaptive strategies, and
interpretation and species selections by local experts, as we move into this changing, but uncertain
future [49,50].
4.1. Applications for Summary Tables
This indices-based, tabular approach, while quite formidable upon first glance, is a
comprehensive way to combine the outputs from 125 species models plus 23 non-modeled species
(but with FIA information) into succinct, spatially explicit summaries for users of the data as a tabular
Figure 7.
(
A
) The number of species that may potentially migrate into each 1
×
1 degree cell (from
elsewhere), within 100 years across the eastern US (Migrate+plus Migrate++), under RCP 8.5.
(
B
) The number of species with potential to infill (Infill+plus Infill++) into each 1
×
1 degree cell,
within 100 years, across the eastern US.
The number and distribution of species with potential to infill provides a major contrast to the
migration map, in that numbers can range up to 54 species, and with the highest numbers in the
highly agricultural zones (Figure 7B). The maximum 1
×
1 values appear in the southern Mississippi
floodplain, in highly agricultural zones (with few forested FIA plots so only a small total area was
sampled) near locations with relatively high species richness. The correlation between Infill85 and
Migrate85 was
0.34 (p<0.01), further indicating the inverse relationship between the two, i.e., if the
unit is highly forested, there are less species ‘available’ for infilling as they have already been captured
by the FIA sampling.
Many other maps depicting trends among the 1
×
1 degree grids have been generated, and these
reveal patterns associated with the numbers of abundant or rare species, the numbers of species within
several common genera, the numbers of species most capable to deal with the changing climate, the top
three species within each cell, and information related to the range distribution, among others. These
maps and others will be made available for users to spatially compare among 1
×
1 degree grids
relative to their areas of interest.
Forests 2019,10, 989 14 of 25
4. Discussion
This work brings together two lines of eort that our group has pursued for over two decades: that
of the projected changes in habitat suitability under the changing climate [
19
,
20
,
23
,
25
,
31
40
] and that
of the potential colonization likelihood into suitable habitat over the course of a century [
11
,
26
,
41
43
].
Advances in each line were achieved, and often necessitated, by enhancements or updates in data,
climate projections, software, and computing capabilities. By combining these lines of eort for most
of the tree species in the eastern US, we provide a mechanism to evaluate, with more reality imbedded,
the potential outcomes at end of century. The intention of this work is to assist managers, both at
the scale of regions or the entire eastern US as well as individual national forests or locations within
1×1 degree
units, in adapting to a changing climate. We hope that the information outlined here can be
used to assist in the silvicultural treatments available to managers, such as plantation, assisted migration,
thinning, or natural regeneration, to help forests mitigate and adapt to the impacts of climate change [
44
].
This is the focus of the Northern Institute of Applied Climate Science, and the Climate Change Response
Framework (www.forestadaptation.org) [
45
]. As laid out by Millar and Stephenson [
46
], adaptation
pathways lead to resistance, resilience, or response (or facilitation, or transition). A national experiment
is testing these approaches at various sites across the continent [
47
], and the data provided here can help
in ascertaining species appropriate for each of these pathways, anywhere in the eastern United States.
However, caution is advised when applying these results because no model can capture the full range
of complexity acting on species through uncertain times, and although species suggested here as
potentials for assisted migration may be adapted climatically to future climate conditions, other factors
such as competition or herbivory may override other traits and restrain regeneration success [
48
]
(see also Supplementary File 1—Climate Change Atlas). Additional caveats more specific to these
results include the averaging over large land areas (>8000 km
2
), the assumption of an average 100-year
migration rate for all species of 50 km in uninterrupted forest, and the use of projected future:actual IVs
as a proxy for future importance of the species. Thus the need exists for multiple, adaptive strategies,
and interpretation and species selections by local experts, as we move into this changing, but uncertain
future [49,50].
4.1. Applications for Summary Tables
This indices-based, tabular approach, while quite formidable upon first glance, is a comprehensive
way to combine the outputs from 125 species models plus 23 non-modeled species (but with FIA
information) into succinct, spatially explicit summaries for users of the data as a tabular substitute for
the multiple individual maps. In this way, users can glean comprehensive information about the tree
species currently present, or potentially present in the future, in their area of interest. Though our group
and other groups have been modeling potential species changes of tree habitats for quite some time,
this approach gives a condensed summary of species’ current and potential future status for any area of
interest. Beyond estimates of habitat suitability changes or range boundary extensions, this work adds
reality in that natural species migrations will greatly lag the changing climatic conditions [
42
,
51
54
].
Estimates for end-of-century habitats and potential shifts provide information that managers can use
now as they ready their forests for end-of-century climates.
The 1
×
1 summary tables are available for download or from the authors for any geographic
location in the eastern US. The 469 1
×
1 degree files provide a wall-to-wall coverage and are named by
their southeast boundary, such that, for example, a GPS coordinate of 42.44
N latitude,
82.55
W
longitude will be named S42_E82. Refer also to Figure 1for assistance in locating the appropriate
table(s). Tables for the national forests and grasslands are also available. We intend to make tables
available as well for national parks, watersheds, ecoregions, and other locations; the purpose is to
provide a wealth of sortable information for managers, researchers, landowners, and interested publics.
The set of questions and answers for the Allegheny NF (see Appendix A) can be asked of any of
the tables and represents some of the ways in which users may apply the data. Users interested in the
current biodiversity status in their area of interest may glean the number, abundance, area occupied,
Forests 2019,10, 989 15 of 25
and identity of tree species that have been reported by FIA, or likely to be there even if not reported by
FIA. The Range field provides a quick indication if the species of interest is narrow or widely dispersed,
and how common or ecologically important it is within its range. Users interested in how species
may generally react to the changing climate in their area of interest would focus more on the habitat
suitability (ChgCl45 or ChgCl85) and modification factors (Adap).
The capability analysis, which incorporates habitat suitability, abundance, and adaptability,
provides an assessment of the species’ capability to withstand and cope under the expected conditions
modeled for RCP 4.5 or 8.5. Thus, users can have a better idea of how the species may fare to 2100 in
their particular area of interest.
Users interested in planting in the face of the changing climate also have resources available from
the tables, in fields SHIFT45, SHIFT85, and SSO. Those users wanting to “hedge their bets” on planting
species already reported by FIA in the area but quite uncommon (i.e., rare), can use the Infill markers
to identify candidate species. Because of the hit or miss nature of sampling trees in highly agricultural
regions like the Corn Belt or Mississippi floodplain, many of the Infill species listed for those areas will
be quite common in the fencerows, woodlots, or riparian regions of the area, just missed by the few
FIA plots landing on forests. Thus, an extra level of screening or targeted sampling may be required
for these areas to account for the fewer plots, keeping in mind also that the model reliability varies
among species represented on the tables. Of course, these highly agricultural zones will likely remain
highly agricultural and the potential to Infill naturally is highly improbable into those regions. In these
cases, managers could choose among the Infill species should they wish to aorest some of these
agricultural areas.
The Migrate classification reveals those species that FIA did not report for the area, but the habitat
will likely be suitable by 2100 and the species has some, often very limited, potential to migrate into
the area within 100 years. The assumption is if the species could have any chance of migrating into
the area naturally, it is likely a better candidate for assisted migration than if the species has potential
future suitable habitat but the sources for natural migration are long distances away. Thus, we would
encourage managers to look at these ‘Migrate’ species first if they wish to assist in transitioning
their area for adaptation to climate change. Forestry assisted migration has long been practiced by
silviculturalists [
55
], though it must be practiced with care [
50
]. The Infill and Migrate classes, along
with the SSO classes, provide some guidance for species selection, but again, we emphasize that these
lists are a starting place for decision making, in need of expert interrogation.
Further information regarding species selection options (SSO) is also provided via 0–3 scoring.
Some species will be common already in the area of interest, and have qualities—of fair, good, or very
good capability class—that provide the expectation that the species will also do fine even under the
changed climate of 2100 (SSO =1). An additional set of species may be present or near the area of
interest currently, are usually quite rare, yet are in a position to potentially expand over time (SSO =2).
Because these species are ‘rare’, it is a similar, but a more restrictive set of species as compared to ‘Infill’
mentioned above as they can be ‘common’ species. A third set of species (SSO =3) are those not present
currently according to FIA (‘new habitat’ capability class) but may actually be there (‘likely’ species)
or have a least a minimal potential to colonize over 100 years (‘migrate’ species). Users can tighten
the selection criteria by using another variable available in the long file, %2Col, which only counts
1×1 km cells
as colonized if the SHIFT model outputs at least 2% of the area of interest as colonized
over 100 years (i.e., the cell gets colonized at least 2 times over the course of 100 runs from SHIFT).
For example, one could consider only those species that have at least 5% of the area with at least a
2% probability of colonization (%2Col >5, see long definitions on spreadsheet, e.g., Supplementary
File 2—Allegheny table).
We emphasize that these analyses are only to be used as general guidelines for species selection.
The models are built from FIA data across the eastern US, and for certain species, local influences
(e.g., lake eects) will override the general tendencies across the entire eastern US. The models also
necessarily are built from coarse-level data and are unable to zero in on special or rare habitats, or may
Forests 2019,10, 989 16 of 25
have low model reliability (MR). Therefore, decision makers should use local knowledge to select
species, even if they are not coded 1, 2, or 3 on the Species Selection Option (SSO), as they may still
be suited for particular niches in a project area that helps meet overall objectives. Conversely, each
species coded 1, 2 or 3 for SSO should be evaluated in the local context as they may not be suitable for
particular sites.
Important also is consideration of the context of migration, and the role of migration, assisted or
not, in overall biodiversity management under climate change. Species will be moving but not all will
be perceived as valuable to the communities into which they establish. Though not so problematic
for the 125 tree species presented here, an evaluation of each species for persecuting, protecting,
or ignoring under the changing climate is warranted [56].
The spreadsheets available for the NFs and 1
×
1 degree grids have a total of eight worksheets
for each unit which enable the user to better interpret the dense amount of information contained
therein. These include the short-table and definitions (discussed above), questions to ask of the data,
interpretation help, and some references. Also included is a long table with definitions, which includes
all the short table fields but also a large number of fields that are available for users to delve into the
actual numbers used to derive many of the categorical fields in the short table (note: this information
is also presented in Supplementary File 3—Explanation of tables).
4.2. Mapped Summaries for 1 ×1 Degree Grids
The mapped summaries shown here, Figures 6and 7, provide examples of the kind of outputs
made possible by this wall-to-wall analysis. Any of the fields discussed above can be mapped to derive
spatial patterning across the eastern US. Some patterns are expected, such as the species richness
pattern (Figure 6A), which shows diminishing richness moving westward and northward. This pattern
has been explored previously (e.g., [
57
,
58
]), with richness in concert with energy and moisture patterns.
Concomitant with overall richness is the pattern of oak species richness, a very diverse genus of
up to 21 species per 1
×
1 degree grid in the south central states of Louisiana, Mississippi, and Alabama.
Oaks are a foundational species in much of the eastern forests [
59
], and an extremely important genus
for wildlife and botanical diversity [
60
,
61
] as well as a valuable timber resource [
62
]. For example,
Tallamy and Shropshire [
63
] reported that more than 500 Lepidopteran species in the mid-Atlantic
region use oaks. It is unfortunate that oaks, almost throughout their eastern North American range,
are experiencing a regeneration problem whereby they are being replaced by more mesophytic species
like maples (Acer spp.) that are more tolerant to shade but less tolerant to fire (that now rarely
occurs) [
64
66
]. Though silvicultural treatments of repeated fire and partial harvest are tools that have
been shown to promote oaks [
67
,
68
], these tools are dicult (i.e., expensive) to use across large swaths
of land, especially when the majority of eastern forestlands are in private ownership. In contrast,
models of climate change, such as those reported by our group since 1998 [
20
,
25
,
36
] and others [
69
71
]
have consistently reported that oaks as a group should do well in a warmer and more drought-prone
future. Nonetheless, it is imperative to conduct the silviculture to maintain a thriving oak component
now so that propagules are available into the future [
72
] and enable the perpetuation of the rich oak
diversity revealed in this 1 ×1 degree examination.
The number of species with potential to migrate reveals the dramatic but expected trend of more
potential species for migrations north vs. south (Figure 7A). Obviously, there are fewer species to draw
from locations further south in the southern latitudes; the Gulf of Mexico marks the end of terrestrial
habitat. In contrast, the northern locations are somewhat cumulative with regards to possible sources
of species to the south, in that the species-rich sections of the south also have less fragmented corridors
for northward movement [
73
] and some species have potentials for long-distance dispersal (up to
500 km within the SHIFT algorithm). The highly fragmented forests within the Plains and Corn Belt
of the Midwest also reduce the potential for species to migrate, i.e., the SHIFT model does not have
many propagule sources when most of the land is under cultivation. Also noticeable when viewing
the juxtaposition of the national forests (Figure 1) with the potential ‘migrate’ species of Figure 7A,
Forests 2019,10, 989 17 of 25
is the probable large influence of the NFs in providing possible sources of species for migration.
Higher number of species with a potential to migrate are found in proximity to NFs in especially the
southern Appalachians, the southern tier of NFs across Missouri, Indiana, Illinois, and Ohio, and the
Northwoods of Minnesota, Wisconsin, and Michigan.
Species may also expand, or ‘infill’ in suitable locations within their current distribution. With up
to 54 species depicted in 1
×
1 degree grids in some southern locations, and with very high numbers
also in the highly agricultural zones (Figure 7B), there is a major contrast with the migration map. These
values are high because forests are sparse and forested FIA plots are infrequent so that either (1) FIA
missed the disparate locations that have the species; or (2) land-use conversion and fragmentation
have eliminated the species from that unit. Managers can select from these lists for species potentially
capable of dealing with the changing climate and are already in the vicinity albeit in low numbers.
Of course, any species identified via these tables must be vetted by local experts as to their
suitability according to local ecological conditions. Because of the scale of analysis (8000 km
2
for
national forests or 10,000 km
2
for 1
×
1 degree grids), there will always be large variations in soils,
topography, land use, and hydrology within the unit that precludes or includes species as candidates.
The online Climate Change Atlas website (www.fs.fed.us/atlas), as well as many other documents and
web sites, also include ecological characteristics of species to help sort out the applicability of species
for particular planting sites.
5. Conclusions
This study is a synthetic eort of over two decades of research and focuses on combinations of
tree species occupying, currently or potentially in the future, each national forest and each
1×1 degree
grid throughout the eastern US. We have combined 125 modeled species, both with regard to
potential changes in suitable habitat and capability, but also with regard to the potential to migrate
or infill naturally (according to historic migration rates) over the next 100 years. This resulting
summarization eort provides an enormous data set, here described for 469 1
×
1 degree grids and
55 national forests and grasslands (and we are compiling them for hydrologic units, ecoregions, states,
and national parks as well). These data are available from authors and available for download at
https://doi.org/10.2737/Climate-Change-Atlas-Combined-v4. We emphasize that these tables and maps
are only the first-line estimate of potential species trends, intended to provide the managers with
some tools to reduce the vast set of decisions before them as they proactively manage in the face of
climate change.
Supplementary Materials:
The following are available online at http://www.mdpi.com/1999-4907/10/11/989/
s1, Supplementary File 1—Climate Change Atlas, Supplementary File 2—Allegheny table, Supplementary
File 3—Explanation of tables.
Author Contributions:
All four authors cooperated on conceptualization, methodology, validation, formal
analysis, investigation, data curation, visualization, review and editing, and preparing proposals for funding.
L.R.I. prepared the original draft, and supervised and administered the overall project beginning in 1995.
Funding:
This research was funded by the USDA Forest Service, Northern Research Station, the Northern Institute
of Applied Climate Science, and the USDA Climate Hubs as part of their appropriated funding.
Acknowledgments:
The authors are indebted to the hundreds of FIA staresponsible for acquiring and processing
data to make it available to researchers like ourselves. Funding for this project was provided by the USDA
Forest Service Northern Research Station; this research did not receive any specific gran from funding agencies
in the public, commercial, or not-for-profit sectors. Thanks to the reviewers of earlier drafts: Andrew Maday,
Bryce Adams, Patricia Leopold, and Chris Swanston. Special thanks are due the reviewers of the original submitted
manuscript; your comments were invaluable in strengthening the paper.
Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the
study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to
publish the results.
Forests 2019,10, 989 18 of 25
Appendix A
Appendix A.1. Questions and Answers for Example NF: Allegheny NF
An assessment of the output data for the Allegheny NF provides several observations both on
current and potential future species. We present them in the form of answers to questions that may be
asked of the data, then the way to extract the answer (a), and the answer for the Allegheny NF (b).
These questions are all presented within the excel tables (on a sheet labeled ‘Questions of tables’)
accessible with the online database. Though an extensive list of questions, it exemplifies the breadth of
information possible to be gleaned for any areal unit via these tables; we have grouped the questions
according to learning about the current status, the potential change in suitable habitat, and the potential
change in migration status over 100 years.
Appendix A.1.1. Current Status of Species
1. Of the species modeled, how many are known (by FIA) to be present in this NF?
a.
The column
N
counts the species in this table. Some are present and some are modeled
to have suitable habitat appear by 2100. Those with
%Cell
>0 are known by FIA to be
present. Some others may be present but were not found on FIA plots.
b. A total of 43 species identified by FIA for this NF.
2. What species are present in this area (according to FIA)?
a.
Look at
Common Name
or
Scientific Name
, those with
%Cell
>0 have FIA indication
of presence.
b. Red maple, black cherry, sugar maple, and American beech top the list of 43 species.
3.
How common is each species, when it is found, i.e., it may be common only in certain areas
within the unit, like bottomlands or plantations (according to FIA)?
a. FIAiv shows the average importance of the species, when it is found.
b.
Eastern hemlock ranks 5th in overall importance across the NF (FIAsum), but 4th in
importance when considering only cells where it is located—it has some particular habitats
that it prefers within the NF.
4. How abundant is each species, taken across the region of interest (according to FIA)?
a. Abund classifies FIAsum into: Abundant, Common, Rare, Absent
b. The Allegheny NF has 6 abundant species, 15 common species, and 22 rare species.
Appendix A.1.2. DISTRIB-II Model Outputs for Year 2100
5. How much confidence do we have in the models?
a. ModRel gives classes of model reliability: Low, Medium, High
b.
The Allegheny NF table shows the following species ratings: 18 high, 24 medium, and 15
low model reliability. Three others are labeled FIA, meaning that we only provide
information from FIA, and no models were developed because too few data were available
for acceptable models.
6.
What do the models suggest may happen to habitat suitability for the species by the year 2100,
according to a high (or low) emissions scenario?
Forests 2019,10, 989 19 of 25
a. ChngCl45
and
ChngCl85
provide, for RCP 4.5 (low emissions) and RCP 8.5 (high emissions)
respectively, the potential change in suitable habitat for the species by year 2100. It is
important to note that a change in suitable habitat does not mean the species importance
will actually change in that area by 2100, only that the habitat is expected to increase,
decrease, or remain unchanged in suitability for that species over time.
b.
The Allegheny NF table shows 15 species with large or very large decrease in suitable
habitat according to RCP 8.5, but also 4 species each for small decrease, small increase,
or no change, 13 species for large increase, and 14 species with new habitat appearing by
century’s end under high emissions.
7.
According to literature, how adaptable are the species to the direct and indirect impacts of a
changing climate?
a. Adapt
shows a color-coded rating (Low—pink, Medium—yellow, High—green) of a
compilation of modification factors that estimate the adaptability of the species across
their range.
b. For Allegheny NF, 15 species are classed as highly adaptable, 35 as medium, and 8 as low
in adaptability.
8.
How capable might the species be for coping with the changing climate, within the region of
interest, under a low (or high) emissions scenario?
a. Capabil45
(or
Capabil85
) gives an estimate of the species capability to cope, based on its
change in habitat suitability (
ChngCl45
or
ChngCl85
), its
Adapt
, and its
Abund
in the
region of interest. If the species is abundant locally, it is assumed to also be in refugia and
niches somewhat buered from the climate impacts. Ranks are coded Very Good,Good,Fair,
Poor,Very Poor,FIA only (no model, no Capability assigned), NNIS (no model, non-native
invasive species), Unknown (insucient data to model), and New Habitat (potential to
migrate into the region).
b.
Under a scenario of RCP 8.5, the Allegheny table shows 7 species with very good and 18
with good capability, 6 with fair capability, and 3 with poor and 6 with lost capability, along
with 14 with new habitat, 2 non-native invasive species, and 1 species with only FIA data.
9.
What species might be considered most (or least) vulnerable to the changing climate in this area?
a.
Those species listed under
Capabil45
(or
Capabil85
) as Poor,Very Poor, or Lost will be
vulnerable according to our assessments; those listed as Very Good,Good, or Fair are treated
as less vulnerable.
b.
Most vulnerable (lost) species for the Allegheny NF under RCP 8.5 are red spruce, red
pine, black ash, quaking aspen, pin cherry, and white spruce. Least vulnerable (very good)
are: chestnut oak, northern red oak, white oak, white ash, black oak, eastern white pine,
and American basswood. Of course, white ash is under severe threat of emerald ash borer,
which the models could not adequately consider.
Appendix A.1.3. SHIFT Model Outputs for Year 2100
10.
Natural migration of the species has been modeled with SHIFT—it predicts where the species
may migrate within the next 100 years that is currently unoccupied by the species. Assuming a
generous migration rate of ~50 km/century, where is the species likely to colonize?
a. SHIFT45
or
SHIFT85
highlights those species that have the best chance of migrating into
the region within 100 years. Those species labeled ‘Migrate+’ have both new suitable habitat
and some probability of colonization within 100 years, while those labeled ‘Migrate++
have a stronger potential for colonization according to the models.
Forests 2019,10, 989 20 of 25
b. Species with potential to migrate into the Allegheny NF within 100 years (under RCP 8.5)
include: post oak, shortleaf pine, sweetgum, sycamore, and eastern redcedar (all Migrate++),
flowering dogwood, sourwood, Virginia pine, and eastern redbud (all Migrate+). Under
RCP 4.5, only sycamore and sourwood appear on the list.
11.
What species are relatively rare in the region, but are poised to expand or infill within the region
under the changing climate?
a. SHIFT45
or
SHIFT85
labels those relatively rare species that could expand within the
region as ‘Infill+’ or ‘Infill++’, depending on the strength of the modeled indicators.
b.
The table for the Allegheny shows mockernut hickory, bitternut hickory, scarlet oak,
and shagbark hickory as the best candidates for infilling (Infill++), followed by green ash
and American elm (Infill+). Of course, these last two species would need to be resistant
varieties to withstand current plagues against them.
12.
What rare species are likely present in the region based on our model outputs, but the forest
inventories missed them?
a. SHIFT45
or
SHIFT85
labels those relatively rare species are likely present within the region
as ‘Likely+’ or ‘Likely++’, depending on the strength of the modeled indicators.
b.
The data predict that black walnut, not found in FIA plots, are now somewhere on the
Allegheny NF or its buer.
13.
What species might be useful for selecting as candidates for planting in the face of a
changing climate?
a. SSO
is the species selection option to assist in decisions regarding promoting the species,
where 1 indicates the species is currently present and has at least a fair capability to cope,
2 indicates the species is rare or close to the NF boundary and has a good chance of
spreading inside the NF, 3 indicates the species is not recorded in FIA plots but does have
some chance of getting colonized within 100 years into the NF, and 0 indicates further
evaluation may be required. Generally, managers might consider selecting species first
with SSO =1, then SSO =2, then SSO =3, then SSO =0.
b.
For the Allegheny NF, 30 species present now have at least a fair capability to cope with the
changing climate (SSO =1); the models suggest these species are most suited for planting
in the area. Three species are rare inside the NF (bitternut and mockernut hickory, slippery
elm) and have potential to expand and could be suited for planting (SSO =2) (green ash is
also listed but of course would not be recommended due to emerald ash borer). Twelve
additional species are marked as having some chance (>0% chance, SSO =3) of naturally
migrating there within 100 years and could be suited for planting, though with more risk
of failure as compared with SSO 1 or 2. The remaining 14 species have SSO =0, and are
generally not suited for planting, though they still can be considered as candidates for
planting under certain local conditions or manager preference.
14. What portion of the area has at least a 2% chance of getting colonized over 100 years? (note: these
variables are only on the long file available as Supplementary File 2—Allegheny table)
a. %2Col
estimates, for any species, the percentage of the region of interest with at least 2%
chance of colonization, and is not already occupied according to FIA. One cutowe use for
potential planting guidelines is that the species requires at least 5% of the region of interest
to have >2% chance of being colonized; often coded “Migrate+” or “Migrate++” under
SHIFT45/SHIFT85, and coded ‘3’ under SSO.
Forests 2019,10, 989 21 of 25
b.
The long form of the Allegheny table lists 49 species with at least 5% of the NF having
at least a 2% chance of getting colonized over 100 years. These include species that are
already present and potentially spreading to unoccupied locations within the NF.
15.
What portion of the area has greater than 0% chance of getting colonized over 100 years? (note:
these variables are only on the long file available as Supplementary File 2—Allegheny table)
a. %0Col
provides the percentage of the area, not already occupied according to FIA, that has
>0% chance of getting colonized.
b.
For the Allegheny NF, 54 species (out of 60) show some area with at least some chance of
getting colonized over 100 years. These also include species that are already present and
potentially spreading to unoccupied locations within the NF.
16.
What portion of the area has at least a 50% chance of getting colonized over 100 years? (note:
these variables are only on the long file available as Supplementary File 2—Allegheny table)
a. %50Col
provides the percentage of the area, not already occupied according to FIA, that
has >50% chance of getting colonized.
b.
Within the Allegheny, 41 species have at least some territory with >50% chance of
colonization. These also include species that are already present and potentially spreading
to unoccupied locations within the NF.
17.
What species may have the highest (or at least some) chance of migrating and finding suitable
habitat within 100 years (under low and high emissions)? (note: these variables are only on the
long file available as Supplementary File 2—Allegheny table)
a. HQCL45
(low emissions) and
HQCL85
(high emissions) are indices that use a weighted
average calculation (based on habitat quality and colonization likelihood), to assign a score
for the potential to migrate into suitable habitat—the higher the value, the greater the chance,
with values 1 or greater indicating the presence of both suitable and colonizable habitats.
b.
Within the Allegheny NF, of the 14 species with new habitat (question #8), 12 species have
HQCL85 greater than or equal to 1.0 (indicating the presence of both suitable habitat and
colonizable habitat), while for HQCL45, the number is only 3 species.
By providing answers to these 17 questions for any national forest or 1
×
1 degree grid, a user
is able to delve into many aspects of their forest attributes, both now and potentially into the future.
Managers and the casual user alike will have species lists showing which are abundant, common,
and rare within their area of interest. Managers will have data to support (or not) their decisions
related to adaptation to the changing climate. For example, selecting species that may be candidates
for planting may be better justified if the HQ-CL data suggest that the suitable habitats are ripe for
the species in their area of interest. This is especially true for selecting species not now identified
from their area of interest–picking species with ‘Migrate+’ or ‘Migrate++’ would seem more justified
in that the habitat will likely be present in coming decades and the species is close enough now to
provide some chance of natural migration into the area even without assistance. Of course, any species
identified via these tables must be vetted by local experts as to their suitability according to local
ecological conditions. Because of the scale of analysis (8000 km
2
for national forests or 10,000 km
2
for
1×1 degree
grids), there will always be large variations in soils, topography, land use, and hydrology
within the unit that precludes or includes species as candidates. The online Climate Change Atlas
website (www.fs.fed.us/atlas), as well as many other documents and web sites, also includes ecological
characteristics of species to help sort out the applicability of species for particular planting sites.
Forests 2019,10, 989 22 of 25
References
1.
IPCC Summary for Policymakers. In Global Warming of 1.5
C. An IPCC Special Report on the Impacts of Global
Warming of 1.5
C Above Pre-Industrial Levels and Related Global Greenhouse Gas Emission Pathways, in the
Context of Strengthening the Global Response to the Threat of Climate Change, Sustainable Development, and Eorts
to Eradicate Poverty; World Meteorological Organization: Geneva, Switzerland, 2018; p. 32.
2.
USGCRP. Impacts, Risks, and Adaptation in the United States: Fourth National Climate Assessment; Global Change
Research Program: Washington, DC, USA, 2018; Volume II: Report-In-Brief, p. 186.
3.
IPBES. Summary for Policymakers of the Global Assessment Report on Biodiversity and Ecosystem Services
of the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services; Intergovernmental
Science-Policy Platform on Biodiversity and Ecosystem Services: Bonn, Germany, 2019; p. 44.
4.
Urban, M.C.; Bocedi, G.; Hendry, A.P.; Mihoub, J.-B.; Pe’er, G.; Singer, A.; Bridle, J.R.; Crozier, L.G.; De
Meester, L.; Godsoe, W.; et al. Improving the forecast for biodiversity under climate change. Science
2016
,
353, aad8466. [CrossRef] [PubMed]
5.
Ara
ú
jo, M.B.; Anderson, R.P.; M
á
rcia Barbosa, A.; Beale, C.M.; Dormann, C.F.; Early, R.; Garcia, R.A.;
Guisan, A.; Maiorano, L.; Naimi, B.; et al. Standards for distribution models in biodiversity assessments.
Sci. Adv. 2019,5, eaat4858. [CrossRef] [PubMed]
6.
Higgins, S.I.; O’Hara, R.B.; Römermann, C. A niche for biology in species distribution models. J. Biogeogr.
2012,39, 2091–2095. [CrossRef]
7.
Dullinger, S.; Gattringer, A.; Thuiller, W.; Moser, D.; Zimmermann, N.E.; Guisan, A.; Willner, W.; Plutzar, C.;
Leitner, M.; Mang, T.; et al. Extinction debt of high-mountain plants under twenty-first-century climate
change. Nat. Clim. Chang. 2012,2, 619. [CrossRef]
8.
Zurell, D.; Thuiller, W.; Pagel, J.; Cabral, J.S.; Münkemüller, T.; Gravel, D.; Dullinger, S.; Normand, S.;
Schiers, K.H.; Moore, K.A.; et al. Benchmarking novel approaches for modelling species range dynamics.
Glob. Chang. Biol. 2016,22, 2651–2664. [CrossRef] [PubMed]
9.
Reynolds, K.M. Integrated decision support for sustainable forest management in the United States: Fact or
fiction? Comput. Electron. Agric. 2005,49, 6–23. [CrossRef]
10.
Iverson, L.; McKenzie, D. Tree-Species range shifts in a changing climate—Detecting, modeling, assisting.
Landsc. Ecol. 2013,28, 879–889. [CrossRef]
11.
Prasad, A.M.; Iverson, L.R.; Matthews, S.N.; Peters, M.P. A multistage decision support framework to
guide tree species management under climate change via habitat suitability and colonization models, and a
knowledge-based scoring system. Landsc. Ecol. 2016,31, 2187–2204. [CrossRef]
12.
Brandt, L.; He, H.; Iverson, L.; Thompson, F.; Butler, P.; Handler, S.; Janowiak, M.; Swanston, C.; Albrecht, M.;
Blume-Weaver, R.; et al. Central Hardwoods Ecosystem Vulnerability Assessment and Synthesis: A Report from the
Central Hardwoods Climate Change Response Framework Project; U.S. Department of Agriculture, Forest Service,
Northern Research Station: Newtown Square, PA, USA, 2014; p. 254.
13.
Handler, S.; Duveneck, M.; Iverson, L.; Peters, E.; Scheller, R.; Wythers, K.; Brandt, L.; Butler, P.; Janowiak, M.;
Swanston, C.; et al. Minnesota Forest Ecosystem Vulnerability Assessment and Synthesis: A Report from the
Northwoods Climate Change Response Framework; Gen. Tech. Rep. NRS-133; U.S. Department of Agriculture,
Forest Service, Northern Research Station: Newtown Square, PA, USA, 2014.
14.
Handler, S.; Duveneck, M.J.; Iverson, L.; Peters, E.; Scheller, R.; Wythers, K.; Brandt, L.; Butler, P.; Janowiak, M.;
Swanston, C.; et al. Michigan Forest Ecosystem Vulnerability Assessment and Synthesis: A Report from the
Northwoods Climate Change Response Framework; Gen. Tech. Rep. NRS-129; U.S. Department of Agriculture,
Forest Service, Northern Research Station: Newtown Square, PA, USA, 2014.
15.
Janowiak, M.K.; Iverson, L.R.; Mladeno, D.J.; Peters, E.; Wythers, K.R.; Xi, W.; Brandt, L.A.; Butler, P.R.;
Handler, S.D.; Shannon, P.D.; et al. Forest Ecosystem Vulnerability Assessment and Synthesis for Northern
Wisconsin and Western Upper Michigan: A Report from the Northwoods Climate Change Response Framework
Project; Gen. Tech. Rep. NRS-136; U.S. Department of Agriculture, Forest Service, Northern Research Station:
Newtown Square, PA, USA, 2014; p. 247.
16.
Brandt, L.A.; Derby Lewis, A.; Scott, L.; Darling, L.; Fahey, R.T.; Iverson, L.; Nowak, D.J.; Bodine, A.R.;
Bell, A.; Still, S.; et al. Chicago Wilderness Region Urban Forest Vulnerability Assessment and Synthesis: A Report
from the Urban Forestry Climate Change Response Framework Chicago Wilderness Pilot Project; U.S. Department of
Agriculture, Forest Service, Northern Research Station: Newtown Square, PA, USA, 2017; p. 142.
Forests 2019,10, 989 23 of 25
17.
Butler-Leopold, P.R.; Iverson, L.R.; Thompson, F.R., III; Brandt, L.A.; Handler, S.D.; Janowiak, M.K.;
Shannon, P.D.; Swanston, C.W.; Bearer, S.; Bryan, A.M.; et al. Mid-Atlantic Forest Ecosystem Vulnerability
Assessment and Synthesis: A Report from the Mid-Atlantic Climate Change Response Framework Project; Gen. Tech.
Rep. NRS-181; U.S. Department of Agriculture, Forest Service, Northern Research Station: Newtown Square,
PA, USA, 2018; p. 365.
18.
Janowiak, M.K.; D’Amato, A.W.; Swanston, C.W.; Iverson, L.; Thompson III, F.; Dijak, W.; Matthews, S.;
Peters, M.; Prasad, A.; Fraser, J.; et al. New England and Northern New York Forest Ecosystem Vulnerability
Assessment and Synthesis: A Report from the New England Climate Change Response Framework Project; Gen. Tech.
Rep. NRS-173; U.S. Department of Agriculture, Forest Service, Northern Research Station: Newtown Square,
PA, USA, 2018; p. 234.
19.
Matthews, S.N.; Iverson, L.R.; Prasad, A.M.; Peters, M.P.; Rodewald, P.G. Modifying climate change habitat
models using tree species-specific assessments of model uncertainty and life history factors. For. Ecol. Manag.
2011,262, 1460–1472. [CrossRef]
20.
Iverson, L.R.; Peters, M.P.; Prasad, A.M.; Matthews, S.N. Analysis of climate change impacts on tree species
of the eastern US: Results of DISTRIB-II modeling. Forests 2019,10, 302. [CrossRef]
21.
Peters, M.; Iverson, L.; Prasad, A.; Matthews, S. Utilizing the density of inventory samples to define a hybrid
lattice for a macro-level species distribution model. Ecol. Evol. 2019,9, 8876–8899. [CrossRef] [PubMed]
22.
Hagen-Zanker, A. Comparing Continuous Valued Raster Data: A Cross Disciplinary Literature Scan; Research
Institute for Knowledge Systems: Bilthoven, The Netherlands, 2006.
23.
Iverson, L.R.; Prasad, A.M.; Matthews, S.N. Modeling potential climate change impacts on the trees of the
northeastern United States. Mitig. Adapt. Strateg. Glob. Chang. 2008,13, 487–516. [CrossRef]
24.
Prasad, A.M.; Iverson, L.R.; Liaw, A. Newer classification and regression tree techniques: Bagging and
random forests for ecological prediction. Ecosystems 2006,9, 181–199. [CrossRef]
25.
Iverson, L.R.; Prasad, A.M.; Matthews, S.N.; Peters, M. Estimating potential habitat for 134 eastern US tree
species under six climate scenarios. For. Ecol. Manag. 2008,254, 390–406. [CrossRef]
26.
Prasad, A.M.; Gardiner, J.; Iverson, L.; Matthews, S.; Peters, M. Exploring tree species colonization potentials
using a spatially explicit simulation model: Implications for four oaks under climate change. Glob. Chang.
Biol. 2013,19, 2196–2208. [CrossRef]
27.
McLachlan, J.S.; Clark, J.S.; Manos, P.S. Molecular indicators of tree migration capacity under rapid climate
change. Ecology 2005,86, 2007–2017. [CrossRef]
28.
Davis, M.B.; Shaw, R.G. Range shifts and adaptive responses to quaternary climate change. Science
2001
,292,
673–679. [CrossRef]
29.
Morelli, T.L.; Daly, C.; Dobrowski, S.Z.; Dulen, D.M.; Ebersole, J.L.; Jackson, S.T.; Lundquist, J.D.; Millar, C.I.;
Maher, S.P.; Monahan, W.B.; et al. Managing Climate Change Refugia for Climate Adaptation. PLoS ONE
2016,11, e0159909. [CrossRef]
30.
Anderson, M.G.; Clark, M.; Sheldon, A.O. Estimating climate resilience for conservation across geophysical
settings. Conserv. Biol. 2014,28, 959–970. [CrossRef]
31.
Iverson, L.; Matthews, S.; Prasad, A.; Peters, M.; Yohe, G. Development of risk matrices for evaluating
climatic change responses of forested habitats. Clim. Chang. 2012,114, 231–243. [CrossRef]
32.
Iverson, L.; Prasad, A.M.; Matthews, S.; Peters, M. Lessons learned while integrating habitat, dispersal,
disturbance, and life-history traits into species habitat models under climate change. Ecosystems
2011
,14,
1005–1020. [CrossRef]
33.
Iverson, L.; Prasad, A.M.; Matthews, S.; Peters, M. Chapter 2. Climate as an agent of change in forest
landscapes. In Forest Landscapes and Global Change: Challenges for Research and Management; Azevedo, J.,
Perera, A., Pinto, M., Eds.; Springer: New York, NY, USA, 2014; pp. 29–50.
34.
Iverson, L.R.; Prasad, A.; Scott, C.T. Preparation of forest inventory and analysis (FIA) and state soil
geographic data base (STATSGO) data for global change research in the eastern United States. In Proceedings,
1995 Meeting of the Northern Global Change Program; General Technical Report NE-214; Hom, J., Birdsey, R.,
O’Brian, K., Eds.; U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station:
Radnor, PA, USA, 1996; pp. 209–214.
35.
Iverson, L.R.; Prasad, A.M. Potential changes in tree species richness and forest community types following
climate change. Ecosystems 2001,4, 186–199. [CrossRef]
Forests 2019,10, 989 24 of 25
36.
Iverson, L.R.; Prasad, A.M. Predicting abundance of 80 tree species following climate change in the eastern
United States. Ecol. Monogr. 1998,68, 465–485. [CrossRef]
37.
Iverson, L.R.; Prasad, A.M. Potential redistribution of tree species habitat under five climate change scenarios
in the eastern US. For. Ecol. Manag. 2002,155, 205–222. [CrossRef]
38.
Iverson, L.R.; Prasad, A.M.; Hale, B.J.; Sutherland, E.K. Atlas of Current and Potential Future Distributions of
Common Trees of the Eastern United States; General Technical Report NE-265; Northeastern Research Station,
USDA Forest Service: Radnor, PA, USA, 1999; p. 245.
39.
Iverson, L.R.; Prasad, A.M.; Schwartz, M.W. Predicting potential changes in suitable habitat and distribution
by 2100 for tree species of the eastern United States. J. Agric. Meteorol. 2005,61, 29–37. [CrossRef]
40.
Matthews, S.N.; Iverson, L.R.; Peters, M.P.; Prasad, A.M.; Subburayalu, S. Assessing and comparing risk
to climate changes among forested locations: Implications for ecosystem services. Landsc. Ecol.
2014
,29,
213–228. [CrossRef]
41.
Iverson, L.R.; Schwartz, M.W.; Prasad, A. How fast and far might tree species migrate under climate change
in the eastern United States? Glob. Ecol. Biogeogr. 2004,13, 209–219. [CrossRef]
42.
Iverson, L.R.; Schwartz, M.W.; Prasad, A.M. Potential colonization of new available tree species habitat
under climate change: An analysis for five eastern US species. Landsc. Ecol. 2004,19, 787–799. [CrossRef]
43.
Iverson, L.R.; Prasad, A.M.; Schwartz, M.W. Modeling potential future individual tree-species distributions
in the Eastern United States under a climate change scenario: A case study with Pinus virginiana.Ecol. Model.
1999,115, 77–93. [CrossRef]
44.
Bouchard, M.; Aquilu
é
, N.; P
é
ri
é
, C.; Lambert, M.-C. Tree species persistence under warming conditions:
A key driver of forest response to climate change. For. Ecol. Manag. 2019,442, 96–104. [CrossRef]
45.
Swanston, C.W.; Janowiak, M.; Brandt, L.; Butler, P.; Handler, S.D.; Shannon, P.D.; Lewis, A.D.; Hall, K.R.;
Fahey, R.T.; Scott, L.; et al. Forest Adaptation Resources: Climate Change Tools and Approaches for Land Managers,
2nd ed.; Gen. Tech. Rep. NRS-GTR-87-2; U.S. Department of Agriculture, Forest Service, Northern Research
Station: Newtown Square, PA, USA, 2016; p. 161.
46.
Millar, C.; Stephenson, N.L. Climate change and forests of the future: Managing in the face of uncertainty.
Ecol. Appl. 2007,17, 2145–2151. [CrossRef] [PubMed]
47.
Nagel, L.M.; Palik, B.J.; Battaglia, M.A.; D’Amato, A.W.; Guldin, J.M.; Swanston, C.W.; Janowiak, M.K.;
Powers, M.P.; Joyce, L.A.; Millar, C.I.; et al. Adaptive silviculture for climate change: A national experiment
in manager-scientist partnerships to apply an adaptation framework. J. For. 2017,115, 167–178. [CrossRef]
48.
Woodall, C.W.; Westfall, J.A.; D’Amato, A.W.; Foster, J.R.; Walters, B.F. Decadal changes in tree range stability
across forests of the eastern U.S. For. Ecol. Manag. 2018,429, 503–510. [CrossRef]
49.
Catanzaro, P.; D’Amato, A.W.; Hu, E.S. Increasing Forest Resiliency for an Uncertain Future; University of
Massachusetts, Cooperative Extension Landowner Outreach Pamphlet: Boston, MA, USA, 2016; p. 32.
50.
Guldin, J.M. Silvicultural options in forests of the southern United States under changing climatic conditions.
New For. 2019,50, 71–87. [CrossRef]
51. Loarie, S.R.; Duy, P.B.; Hamilton, H.; Asner, G.P.; Field, C.B.; Ackerly, D.D. The velocity of climate change.
Nature 2009,462, 1052–1055. [CrossRef] [PubMed]
52.
Serra-Diaz, J.M.; Franklin, J.; Ninyerola, M.; Davis, F.W.; Syphard, A.D.; Regan, H.M.; Ikegami, M. Bioclimatic
velocity: The pace of species exposure to climate change. Divers. Distrib. 2014,20, 169–180. [CrossRef]
53.
Carroll, C.; Lawler, J.J.; Roberts, D.R.; Hamann, A. Biotic and climatic velocity identify contrasting areas of
vulnerability to climate change. PLoS ONE 2015,10, e0140486. [CrossRef] [PubMed]
54.
Sittaro, F.; Paquette, A.; Messier, C.; Nock Charles, A. Tree range expansion in eastern North America fails to
keep pace with climate warming at northern range limits. Glob. Chang. Biol.
2017
,23, 3292–3301. [CrossRef]
55.
Pedlar, J.H.; McKenney, D.W.; Aubin, I.; Beardmore, T.; Beaulieru, J.; Iverson, L.R.; O’Neill, G.A.; Winder, R.S.;
Ste-Marie, C. Placing forestry in the assisted migration debate. BioScience 2012,62, 835–842. [CrossRef]
56.
Scheers, B.R.; Pecl, G. Persecuting, protecting or ignoring biodiversity under climate change. Nat. Clim.
Chang. 2019,9, 581–586. [CrossRef]
57.
Currie, D.J.; Paquin, V. Large-Scale biogeographical patterns of species richness of trees. Nature
1987
,329,
326–327. [CrossRef]
58.
Currie, D.J. Energy and large-scale patterns of animal-and plant-species richness. Am. Nat.
1991
,137, 27–49.
[CrossRef]
Forests 2019,10, 989 25 of 25
59.
Hanberry, B.B.; Nowacki, G.J. Oaks were the historical foundation genus of the east-central United States.
Quat. Sci. Rev. 2016,145, 94–103. [CrossRef]
60.
McShea, W.; Healy, W. Oak Forest Ecosystems: Ecology and Management for Wildlife; The Johns Hopkins
University Press: Baltimore, MD, USA, 2002.
61.
McShea, W.J.; Healy, W.M.; Devers, P.; Fearer, T.; Koch, F.H.; Stauer, D.; Waldon, J. Forestry matters: Decline
of oaks will impact wildlife in hardwood forests. J. Wildl. Manag. 2007,71, 1717–1728. [CrossRef]
62.
Johnson, P.; Shifley, S.; Rogers, R. The Ecology and Silviculture of Oaks; CABI Publishing: New York, NY,
USA, 2009.
63.
Tallamy, D.W.; Shropshire, K.J. Ranking lepidopteran use of native versus introduced plants. Conserv. Biol.
2009,23, 941–947. [CrossRef]
64.
Loftis, D.L. Upland Oak Regeneration and Management; U.S. Department of Agriculture, Forest Service, Southern
Research Station: Asheville, NC, USA, 2004; pp. 163–167.
65.
Nowacki, G.J.; Abrams, M.D. The demise of fire and ‘’mesophication” of forests in the eastern United States.
Bioscience 2008,58, 123–138. [CrossRef]
66.
Knott, J.A.; Desprez, J.M.; Oswalt, C.M.; Fei, S. Shifts in forest composition in the eastern United States.
For. Ecol. Manag. 2019,433, 176–183. [CrossRef]
67.
Hutchinson, T.F.; Yaussy, D.A.; Long, R.P.; Rebbeck, J.; Sutherland, E.K. Long-Term (13-year) eects of
repeated prescribed fires on stand structure and tree regeneration in mixed-oak forests. For. Ecol. Manag.
2012,286, 87–100. [CrossRef]
68.
Iverson, L.R.; Hutchinson, T.F.; Peters, M.P.; Yaussy, D.A. Long-Term response of oak-hickory regeneration to
partial harvest and repeated fires: Influence of light and moisture. Ecosphere 2017,8, e01642. [CrossRef]
69.
Clark, J.; Iverson, L.; Woodall, C.W.; Allen, C.D.; Bell, D.M.; Bragg, D.C.; D’Amato, A.W.; Davis, F.W.;
Hersh, M.H.; Ibanez, I.; et al. Impacts of increasing drought on forest dynamics, structure, diversity,
and management. In Eects of Drought on Forests and Rangelands in the United States: A Comprehensive Science
Synthesis; Gen. Tech. Rep. WO-93b; Vose, J., Clark, J., Luce, C., Patel-Weynand, T., Eds.; U.S. Department of
Agriculture, Forest Service, Washington Oce: Washington, DC, USA, 2016; pp. 59–96.
70.
Vose, J.; Elliott, K.J. Oak, fire, and global change in the Eastern USA: What might the future hold? Fire Ecol.
2016,12, 160–179. [CrossRef]
71.
Iverson, L.R.; Thompson, F.R.; Matthews, S.; Peters, M.; Prasad, A.; Dijak, W.D.; Fraser, J.; Wang, W.J.;
Hanberry, B.; He, H.; et al. Multi-model comparison on the eects of climate change on tree species in the
eastern U.S.: Results from an enhanced niche model and process-based ecosystem and landscape models.
Landsc. Ecol. 2017,32, 1327–1346. [CrossRef]
72.
Iverson, L.; Peters, M.; Matthews, S.; Prasad, A.; Hutchinson, T.; Bartig, J.; Rebbeck, J.; Yaussy, D.A.; Stout, S.;
Nowacki, G. Adapting oak management in an age of ongoing mesophication but warming climate. In Oak
Symposium: Sustaining Oak Forests in the 21st Century Through Science-Based Management; e-Gen. Tech. Rep.
SRS-237; Clark, S.L., Schweitzer, C.J., Eds.; U.S. Department of Agriculture Forest Service, Southern Research
Station: Asheville, NC, USA, 2019; pp. 35–45.
73.
Riitters, K.H.; Wickham, J.D.; O’Neill, R.V.; Jones, K.B.; Smith, E.R.; Coulston, J.W.; Wade, T.G.; Smith, J.H.
Fragmentation of continental United States forests. Ecosystems 2002,5, 815–822. [CrossRef]
©
2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
... For the Pest-Pathogen goal, family names are shown. Iverson et al. 2019. 1957, Ontario Ministry of Natural Resources 2004. ...
... Based on species present in our 141 Data (Table 1, and other size classes [C. R. Henry, unpublished data]), plus other species that may be reasonable candidates for climate change-inspired assisted migration efforts (Iverson et al. 2019), we identified groups of species for two full stocking goals, seven component stocking goals, and a group deemed unacceptable for stocking (Table 1). ...
... Exacerbating the problem of A. balsamea dominance, the species is considered highly sensitive to climate change in the region we assessed (fs.fed.un/nrs/atlas, Iverson et al. 2019). Silvicultural interventions to create more evenness in the Wildlife: Conifer group could include more intensive harvesting, which could favor less shade tolerant P. strobus, if locally present (Henry et al., unpublished data). ...
Article
Well-stocked natural tree regeneration is critical to sustainable management of northern hardwood forests (NHF) by selection silviculture, but explicit goals and stocking criteria for sapling recruit size classes are lacking. For stems 0–2 in. (0–5.1 cm), we define 171 and 691 stems ac−1 (422.5 and 1,707.5 ha−1) as lower and upper full stocking thresholds for basic NHF management goals (wood and biomass productivity) and overabundant/pest–pathogen-challenged species. We define component stocking targets as 15% of full stocking targets for six additional wildlife (e.g., nut producers) and resilience (e.g., pest–pathogen) goals and individual species representation. Applying our system to 141 managed stands in Michigan, USA, we found several areas of concern. At low threshold, 33% of stands were understocked for biomass and wood goals, whereas 67% were fully stocked with overabundant/pest–pathogen-challenged species. Among goals, component stocking (low threshold) was as low as 5% (nut producers), with more than half the stands component stocked for ≤ 2 goals and ≤ 2 individual species. Stocking varied geographically and decreased with increasing site quality, overstory basal area, and deer use. Study Implications The stocking goals and density criteria we developed can help managers objectively identify areas where current silvicultural practices are ineffective and need modification. Our system is based on an interactive platform that can be readily modified to accommodate new information and management goals. Several management suggestions emerged from applying our system to selection-managed NHF stands in Michigan. First, highest quality sites have the greatest stocking challenges. Stocking outcomes could likely be improved via more intense overstory harvests and mitigation of deer impacts. Second, most stands meeting stocking goals are vulnerable to future changes in climate and pest–pathogen outbreaks, given stocking is often dominated by a small subset of candidate species. Planting, including the incorporation of assisted migration species candidates could help ameliorate this problem, as species could be evaluated based on their contribution to multiple component stocking goals.
... In the USA there is a high level of concern among scientists, law academics, and practitioners over AM (see Hoegh-Guldberg et al. 2008;Ricciardi and Simberloff 2009;Richardson et al. 2009;Sandler 2009;Camacho 2010;Kabaz-Gomez 2012;Schwartz et al. 2012;Lopez 2015;Iverson et al. 2019). Climate change has harmed biodiversity from Alaska to Florida. ...
... Among the examples are the translocation of the conifer Torreya Taxifolia from Florida to the northern parts of the country (Torreya Guardians (nd)) and the relocation of bull trout to cooler waters (Joyce 2015). The AM of other species is also being focused on by policymakers and scientists (Iverson et al. 2019). However, the success of possible translocations depends not only on climatic stressors but also on the competition of the translocated organisms with other species as well as other factors (ibid). ...
Article
Full-text available
The negative impact of climate change on biodiversity will continue to escalate rapidly. While some species will naturally migrate to more suitable areas or adapt to the new climatic environmental conditions in different fashions, for others doing so may prove to be problematic or impossible. Against this backdrop, scientists and environmentalists have proposed implementing plans for Assisted Migration (AM)—meaning the translocation of plants and animals to areas outside their natural habitats to conserve their species under the new emerging climatic conditions. This article seeks to identify legal approaches towards AM considering not only possible benefits from using this tool but also a necessity to minimize related risks. With regard to its stated purpose, this article also compares legal and policy documents relevant to AM issues from the United States, Australia, and the European Union. In conclusion, we have found, and this article shows, that while existing legal and policy documents leave room for manoeuvreing in regard to climate-related translocations and even sometimes explicitly mention AM as a possible tool for conservation, there exists a need for the further development of concrete legal mechanisms and their balancing with the predominant ideas and goals brought about by the necessity to protect native biota.
... Although these changes will happen gradually, even under the most pessimistic climate scenarios, their impacts on biodiversity and ecosystems will, over the long term, be very great. Modelling studies designed to predict changes in habitat suitability and species distribution for future scenarios of climate change will be important for informing scientists, managers, and policy-makers about risks to vulnerable species and ecosystems, and support proactive efforts to minimize the most negative effects, while potentially taking advantage of some positive changes that may accrue from climate change (de Rigo et al., 2017;Garcia et al., 2013;Gilani et al., 2020a;Iverson et al., 2019;Peng et al., 2019). ...
... In southern Eurasia, significant habitat reductions are predicted for beeches including the common hornbeams, oaks, maples, alders, and other economically important species (Taleshi et al., 2019;Arslan et al., 2020;Çoban et al., 2020). In North America, Garza et al. (2020) found significant negative effects of climate change on the future distribution of the endangered shrub Manihot walkerae in Southern Texas and northern Mexico, and another study predicted habitat gains and losses for more than 100 tree species in the eastern United States, with gains exceeding losses (Iverson et al., 2019). In sub-Saharan Africa, several species of miombo trees (genus Brachystegia etc.) were found to be threatened in climate change scenarios (Jinga and Ashley, 2019). ...
Article
There is an urgent need to conduct scientific research in order to comprehend the impact of climate change on the geographical distribution of species that constitute an ecological niche. To meet this need, the current and future distribution areas of a species can be predicted by applying machine learning techniques. This study aims to model the current and future geographical distribution of common hornbeam under different climate scenarios. Furthermore, we focus on the differences between the predicted current and future potential distribution areas of the species in terms of area and location by means of change analysis. 15 bioclimatic variables obtained for presence data from sample points representing the natural distribution area of common hornbeam were reduced to six variables to prevent high correlation and multi-collinearity. Next, the CNRM-ESM-1 climate change model was used to determine how the distribution areas of the species will be affected by climate change and the potential distribution areas of the species under the SSP2 4.5 and SSP5 8.5 scenarios in the periods 2041–2060 and 2081–2100 were modelled by MaxEnt 3.4.4 software. The change analysis comparing the current and future potential distribution areas suggested that common hornbeam would be affected by climate change, expanding its range in the north. The findings can be used effectively to address biodiversity conversation planning and management as a whole and develop new strategies. In this way, future risks can be diminished.
... Indeed, while 24-26 species are forecasted to decline in future habitat compatibility (Peters et al., 2020), 11-19 species have already been identified for assisted migration and 1-12 are expected to infill naturally over time (Iverson et al., 2019), highlighting the potential importance of adaptation plantings. ...
... While modifications in seed zones will be required to maintain contemporary assemblages with climate-adapted genotypes, inter-species comparisons offer critical insights into historic and future climate-adapted species relationships and biophysical drivers affecting no-analogue assemblages. This is important for species-rich temperate forested regions like the northeastern US and elsewhere where more future climate-adapted species will require assisted migration (11-19 species) relative to those currently onsite that may infill naturally (1-12 species; Iverson et al., 2019). While it is possible alterations in provenance may improve the climate match for species tested in our investigation, warranting further study, the broader implications of inter-species comparisons better reflect potential changes for mixed-species systems and highlight challenges for assisted migration under current climate conditions. ...
Article
Full-text available
Species distribution models predict shifts in forest habitat in response to warming temperatures associated with climate change, yet tree migration rates lag climate change, leading to misalignment of current species assemblages with future climate conditions. Forest adaptation strategies have been proposed to deliberately adjust species composition by planting climate‐suitable species. Practical evaluations of adaptation plantings are limited, especially in the context of ecological memory or extreme climate events. In this study, we examined the three‐year survival and growth response of future‐climate adapted seedling transplants within operational‐scale silvicultural trials across temperate forests in the northeastern US. Nine species were selected for evaluation based on projected future importance under climate change and potential functional redundancy with species currently found in these ecosystems. We investigated how adaptation planting type (“population enrichment” vs. “assisted range expansion”) and local site conditions reinforce interference interactions with existing vegetation at filtering adaptation strategies focused on transitioning forest composition. Our results show the performance of seedling transplants is based on species (e.g., functional attributes and size), the strength of local competition (e.g., ecological memory), and adaptation planting type, a proxy for source distance. These findings were consistent across regional forests but modified by site‐specific conditions such as browse pressure and extreme climate events, namely drought and spring frost events. Synthesis and applications. Our results highlight that managing forests for shifts in future composition represents a promising adaptation strategy for incorporating new species and functional traits into contemporary forests. Yet, important barriers remain for the establishment of future climate‐adapted forests that will most likely require management intervention. Nonetheless, the broader applicability of our findings demonstrate the potential for adaptation plantings to serve as strategic source nodes for the establishment of future‐adapted species across functionally connected landscapes.
... Each method has strengths and weaknesses, and a frontier in vulnerability science is developing hybrid methods that integrate or combine features to improve both the efficiency and the accuracy of forecasts. For example, Iverson et al. (2019) combined climate envelope models, a correlational method, with a trait-based assessment of climate change adaptation potential to assess climate vulnerability of 125 eastern U.S. tree species, improving evaluations over each method used independently. ...
Chapter
Full-text available
Vulnerability assessments of National Park Service units and resources including infrastructure, natural resources, and cultural resources were evaluated to better understand the “state of the science” among these resource groups. While approaches are diverse, methods for evaluating infrastructure and natural resource vulnerability assessments were found to be more well established than what is known or available for design and development of cultural resources assessments. Consistent challenges were identified along with best practices and recommendations based on the literature reviews in each chapter and this synthesis.
... Assisted migration, or the introduction of a species outside its native range (McLachland et al. 2007), has been suggested for selected species of conservation concern (Barlow 2011) as a hedge against extinction resulting from climate change (Schwartz et al. 2012), but also to enhance survival of tree species not necessarily at risk of extinction (Pedlar et al. 2012;Iverson and McKenzie 2013;Williams and Dumroesse 2013;Handler et al. 2018). Experimental plantings under controlled conditions could be considered to test performance of tree species identified as candidates for migration (Table 5) to more northern locations (Iverson et al. 2019b). This effort is currently underway with multiple Adaptive Silviculture for Climate Change sites across the United States (Nagel et al. 2017;ASCC 2021) and other demonstrations assisted by the Northern Institute of Applied Climate Science (NIACS 2021). ...
Article
Full-text available
There have been dramatic changes to forest lands since the end of the last ice age, about 14,000 years before present, when boreal ecosystems were eventually replaced by deciduous forest and grassland. In Illinois at the time of Euro-American Settlement (circa 1820), forest lands, including fire-maintained woodlands and savannas, comprised about 42% of the land area. Habitat destruction, fire absence, livestock grazing, and infestations of non-native species have altered forests since the 1800s. Currently, forest land cover statewide is about 13.5%, mostly (83%) in private ownership and predomi-nately (68%) classified as oak-hickory cover type. Further modifications can be expected due to climate change, predicted for Illinois over the next 100 years to include warmer winter temperatures, warmer and wetter springs, and hotter, drier summers. Models predicting potential futures for 113 tree species as a response to climate change over the next 100 years were generated for ten primary Illinois ecoregions. Results indicate that there are likely to be increases in habitat suitability and capability for some species and decreased habitat suitability and capability for others with variability across ecoregions. Many species demonstrate differential responses to changing climate from north to south in the state. The dominant species in the oak-hickory cover type generally are projected to have fair to good capabilities, with some notable exceptions; however , Acer saccharum, a competitor in many oak-hickory stands, also is projected to have fair to good capability. Dominant species in mesic upland and bottomland forests include a rich variety of species about evenly split between those with fair-to-good capabilities and those expected to have poor capability. Potential 'New Habitat' and 'Migrate' species also are identified. New Habitat species are those that have potential habitat appearing in the state within 100 years; Migrate species have some potential for natural distribution to the state within 100 years and could be considered as candidates for assisted migration northward. Considerations for conservation and management of forest lands are discussed.
... These indices can broadly categorize or rank species or habitat vulnerability, but results can vary depending on which method is used (Lankford et al., 2014) and are rarely spatially explicit. Species habitat suitability and/or distribution models are spatially explicit but often fail to account for important behavioral and ecological species traits (Franklin, 2010), although recent methodological advances partially address these limitations (Iverson et al., 2019). Considering both habitat suitability projections and trait-based assessments will provide a more complete understanding of potential vulnerability than either approach alone. ...
Article
Full-text available
Abstract Climate change poses significant challenges to protected area management globally. Anticipatory climate adaptation planning relies on vulnerability assessments that identify parks and resources at risk from climate change and associated vulnerability drivers. However, there is currently little understanding of where and how protected area assessments have been conducted and what assessment approaches best inform park management. To address this knowledge gap, we systematically evaluated climate‐change vulnerability assessments of natural resources in U.S. National Parks. We categorized the spatial scale, resources, methods, and handling of uncertainty for each assessment and mapped which parks have assessments and for what resources. We found that a few broad‐scale assessments provide baseline information—primarily regarding physical climate change exposure—for all parks and can support regional to national decisions. However, finer‐scale assessments are required to inform decisions for individual or small groups of parks. Only 10% of parks had park‐specific assessments describing key climate impacts and identifying priority resource vulnerabilities, and 37% lacked any regional or park‐specific assessments. We identify assessment approaches that match the scale and objectives of different protected area management decisions and recommend a multi‐scaled approach to implementing assessments to meet the information needs of a large, protected area network like the National Park system.
... 536 ► Habitat suitability modeling projections: Modeled projections for native species were summarized from the Climate Change Atlas website under low and high emissions for the 1-degree latitude/longitude grid cell that covers Rhode Island (east of 71W and south of 41N). [537][538][539] ► Adaptability: Adaptability scores were generated for each species based on literature describing its tolerance to disturbances such as drought, flooding, pests, and disease, as well as its growth requirements such as shade tolerance, soil needs, and ease of nursery propagation. Scores were assigned to Rhode Island species using methods developed in an urban forest vulnerability assessment for Chicago. ...
... In some instances, species have moved to higher elevations or latitudes, largely in response to changing temperatures; in others, species have shifted longitudinally or to lower elevations, tracking shifts in precipitation patterns and moisture availability (e.g., Crimmins et al. 2011, Bell et al. 2014, Lenoir and Svenning 2015, Fei et al. 2017. Taking a variety of factors into consideration, model-based projections suggest that further range shifts among tree species are likely in forests across North America (e.g., Iverson et al. 2008Iverson et al. , 2019Woodall et al. 2009;Iverson and McKenzie 2013;Clark et al. 2014;Matthews et al. 2014;Prasad et al. 2020;Rehfeldt et al. 2020;Toot et al. 2020). ...
Technical Report
Full-text available
A rapidly changing climate, including rising temperatures, changing precipitation patterns, and more extreme storms, is having profound consequences for America’s national forests. Climate-related impacts on forest systems include larger and more severe disturbances (e.g., wildfires, drought, and insect outbreaks), shifts in tree species ranges and forest composition, and changes in forest dynamics and regeneration capacity. Many of our national forests have been significantly modified by past management and land use, and forest managers are contending with ongoing threats from invasive species, disease outbreaks, and other challenges. With the added impacts on forest systems from climate change, an enormous mismatch exists between the level of restoration work currently underway and the scale of the challenge. As a result, there is a need to substantially increase the pace, scale, and quality of restoration on our national forests, and to ensure that this restoration is carried out in an ecologically appropriate and climate-smart manner. Continuing and accelerating climatic changes, and their associated impacts, have significant implications for the effectiveness of traditional forest restoration efforts, including reliance on historical conditions as benchmarks for restoration outcomes. Drawing on a growing body of evidence, research, and experimentation, this science review and synthesis looks at how climate change is inspiring an important evolution in approaches for national forest restoration and management. Over the past decade, the U.S. Forest Service has made considerable progress in understanding the effects of a changing climate on forest ecosystems and working to incorporate climate considerations into its planning and management. Nonetheless, varying perspectives on what climate change means for ecological restoration in practice and how to navigate potential trade-offs continue to pose challenges to integrating climate adaptation and mitigation in national forest planning and management. Addressing this challenge would benefit from a shared understanding among agency staff and stakeholders of what constitutes a forward-looking and climate-smart approach to national forest restoration. To this end, this report reviews and summarizes recent advances and ongoing evolution in how the concepts and principles of climate adaptation and mitigation can help promote the development and application of climate-smart forest restoration.
... To aid assisted migration and support the tree species choice for forest management planning, species distribution models have been an important tool to find tree species better adapted to future climate conditions (Falk and Mellert, 2011;Falk and Hempelmann, 2013;Thurm et al., 2018) and numerous approaches have been consequently developed to improve model predictions (Booth, 2018;Gobeyn et al., 2019;Pecchi et al., 2019). Species distribution models (SDMs) have been criticized for their limitations, such as the inability to account for epigenetic effects and local adaptions (Booth, 2018;Iverson et al., 2019) that may play important roles discovering best adapted provenances. Nevertheless, SDMs can be a useful decision-support tool (Pecchi et al., 2019) and recent improvements in the use of species and environmental data will enhance their accuracy in the future, such as the inclusion of species traits as potential drivers of climatic adaptability (Garzón et al., 2019). ...
Article
Species distribution models (SDMs) are a standard tool for predicting species occurrence under climate change. Despite its limitations, SDMs have been widely used to assist forest management decisions in their choice of future tree species. The accuracy of SDMs is often affected by heterogeneous occurrence data caused by different forest inventory schemes, historical processes, climate and insect calamities and more. These processes bias the relationships modelled between species occurrence and climate. However, numerous studies have shown that explicit modelling of spatial effects may improve model accuracy. In this study, we applied the Integrated Nested Laplace Approximation (INLA) and Stochastic Partial Differential Equations (SPDE) algorithms as a means to accomplish Bayes inference for Generalized Additive Models (GAMs) with explicit modelling of spatial effects. Our results show that including spatial effects in GAM-based SDMs improved model performance and accuracy leading to more reliable predictions for the species Favorability. Conditional predictions that remove spatial effects from the models allow us to distinguish core and marginal ecological ranges with less bias through forest management, which may support the tree-species choice for climate-resilient forests.
Article
Full-text available
Abstract Species distribution models (SDMs) provide useful information about potential presence or absence, and environmental conditions suitable for a species; and high‐resolution models across large extents are desirable. A primary feature of SDMs is the underlying spatial resolution, which can be chosen for many reasons, though we propose that a hybrid lattice, in which grid cell sizes vary with the density of forest inventory plots, provides benefits over uniform grids. We examine how the spatial grain size affected overall model performance for the Random Forest‐based SDM, DISTRIB, which was updated with recent forest inventories, climate, and soil data, and used a hybrid lattice derived from inventory densities. Modeled habitat suitability was compared between a uniform grid of 10 × 10 and a hybrid lattice of 10 × 10 and 20 × 20 km grids to assess potential improvements. The resulting DISTRIB‐II models for 125 eastern U.S. tree species provide information on individual habitat suitability that can be mapped and statistically analyzed to understand current and potential changes. Model performance metrics were comparable among the hybrid lattice and 10‐km grids; however, the hybrid lattice models generally had higher overall model reliability scores and were likely more representative of the inventory data. Our efforts to update DISTRIB models with current information aims to produce a more representative depiction of recent conditions by accounting for the spatial density of forest inventory data and using the latest climate data. Additionally, we developed an approach that leverages a hybrid lattice to maximize the spatial information within the models and recommend that similar modeling efforts be used to evaluate the spatial density of response and predictor data and derive a modeling grid that best represents the environment.
Article
Full-text available
Forests across the globe are faced with a rapidly changing climate and an enhanced understanding of how these changing conditions may impact these vital resources is needed. Our approach is to use DISTRIB-II, an updated version of the Random Forest DISTRIB model, to model 125 tree species individually from the eastern United States to quantify potential current and future habitat responses under two Representative Concentration Pathways (RCP 8.5 -high emissions which is our current trajectory and RCP 4.5 -lower emissions by implementing energy conservation) and three climate models. Climate change could have large impacts on suitable habitat for tree species in the eastern United States, especially under a high emissions trajectory. On average, of the 125 species, approximately 88 species would gain and 26 species would lose at least 10% of their suitable habitat. The projected change in the center of gravity for each species distribution (i.e., mean center) between current and future habitat moves generally northeast, with 81 species habitat centers potentially moving over 100 km under RCP 8.5. Collectively, our results suggest that many species will experience less pressure in tracking their suitable habitats under a path of lower greenhouse gas emissions.
Article
Full-text available
Demand for models in biodiversity assessments is rising, but which models are adequate for the task? We propose a set of best-practice standards and detailed guidelines enabling scoring of studies based on species distribution models for use in biodiversity assessments. We reviewed and scored 400 modeling studies over the past 20 years using the proposed standards and guidelines. We detected low model adequacy overall, but with a marked tendency of improvement over time in model building and, to a lesser degree, in biological data and model evaluation. We argue that implementation of agreed-upon standards for models in biodiversity assessments would promote transparency and repeatability, eventually leading to higher quality of the models and the inferences used in assessments. We encourage broad community participation toward the expansion and ongoing development of the proposed standards and guidelines.
Article
Full-text available
Changing climatic conditions add a measure of uncertainty to sustainable forest management in forest ecosystems of the southern United States. Increasing temperatures and decreasing patterns of precipitation especially in the Mid-South suggest that water stress, drought, and changing patterns of natural disturbance events will challenge managers in the twenty-first century. Efforts to manage southern forest stands in the face of changing climatic conditions will require a diversity of approaches including tactics to promote genetic diversity in natural and planted stands, encouragement of species diversity as new stands develop, and considering ways to promote diverse stand structures that encourage recruitment of new age cohorts within stands on a regular basis. With predicted changes in climatic conditions, forest ecosystems across the South will respond in different ways, depending upon whether or not they are currently being managed. Unmanaged stands will change in unpredictable ways that reflect the absence of management. But in managed stands, silvicultural treatments are available for foresters to apply to respond and adapt to maintain productive forests adapted to those changing conditions. Finally, one approach often advocated to deal with this uncertainty is a strategy for assisted migration, in which species are established in locations beyond their current range, where predicted climatic conditions are likely to occur at some point in the future within which those species will survive. This is basically an exercise in artificial regeneration, but will likely be more complicated than simply planting a few exotic seedlings and hoping for the best. The technical and practical challenges of planting species at the margins or beyond their natural range include a lack of research support especially for species not commonly planted in the region. Moreover, planting is costly, and because of that, intensive practices are more likely on institutional and government lands rather than family forests. In the end, all of these concepts fall within the practice of silviculture, and are tactics with which the profession is familiar.
Article
Full-text available
Forest managers in the United States must respond to the need for climate-adaptive strategies in the face of observed and projected climatic changes. However, there is a lack of on-the-ground forest adaptation research to indicate what adaptation measures or tactics might be effective in preparing forest ecosystems to deal with climate change. Natural resource managers in many areas are also challenged by scant locally or regionally relevant information on climate projections and potential impacts. The Adaptive Silviculture for Climate Change (ASCC) project was designed to respond to these barriers to operationalizing climate adaptation strategies by providing a multiregion network of replicated operational-scale research sites testing ecosystem-specific climate change adaptation treatments across a gradient of adaptive approaches, and introducing conceptual tools and processes to integrate climate change considerations into management and silvicultural decisionmaking. Here we present the framework of the ASCC project, highlight the implementation process at two of the study sites, and discuss the contributions of this collaborative science-management partnership.
Article
This Perspective describes persecution, protection and ignorance archetypes for managing and monitoring species redistribution under climate change, and argues for global shared governance agreements to cope with species shifts into new geopolitical areas.
Article
Forest ecosystems in the United States (U.S.) are facing major challenges such as climate change, exotic species invasions, and landscape fragmentation. It is widely believed that forest composition in the eastern U.S. is transitioning from shade-intolerant, fire-tolerant species to shade-tolerant, fire-intolerant species, but most evidence is anecdotal or localized. No comprehensive studies exist to quantify the shifts in forest composition across multiple genera at a regional scale. Here, we examined the genus-level compositional changes in eastern U.S. forests to: (1) quantify the extent and magnitude of this transition, and (2) assess the influence of shade and fire tolerance traits on abundance change. Genus-level data were compiled from the Forest Inventory and Analysis (FIA) database across 37 states in the eastern U.S. for the last three decades. We analyzed shifts in forest composition with three metrics—stem density, basal area, and importance value—for 10 of the most abundant genera (Acer, Betula, Carya, Fraxinus, Nyssa, Pinus, Populus, Prunus, Quercus, and Ulmus). In addition, we estimated density-weighted fire and shade tolerances for each genus using species-level published data, assessed the shifts in spatial patterns of these traits, and analyzed the associations between these traits and county-level abundance changes. In general, Acer, Fraxinus, Pinus, and Prunus increased in abundance during the study period. Acer experienced the largest increase in abundance across the study area. In contrast, Carya, Nyssa, Quercus and Ulmus decreased in abundance in the majority of the study region, with Quercus having the largest and most extensive decline. Although density-weighted shade and fire tolerances were correlated at the genus level, shade tolerance was a better predictor of genus-level abundance change than fire tolerance. Traits of fire and shade tolerance are not always interchangeable when used to predict the dynamics of a genus, and management decision making based on traits should focus at the species level when possible. Our analyses provide evidence that forest composition has shifted in the last three decades in the eastern United States across multiple genera, and the shifts are more closely related to species’ shade tolerance than fire tolerance.
Article
The monitoring of tree range dynamics has emerged as an important component of adaptive responses of forest management to global change scenarios such as extreme precipitation events and/or invasive species. Comparisons between the locations of adults versus seedlings of individual tree species using contemporary forest inventories is one tool widely used to assess the status of tree ranges in light of these changing conditions. With the consistent remeasurement of standard forest inventory plots across the entire eastern US occurring since the 2000s, the opportunity exists to evaluate the stability of tree ranges of focal species across a decade. Using said inventory, the northern range margins of tree distributions were examined by comparing differences (Holm-Sidak adjusted p-value = 0.2) in the 95th percentile locations of seedlings to adults (i.e., trees) by 0.5 degree longitudinal bands over nearly 10 years and by categories of canopy disturbance (i.e., canopy gap formation) for 20 study species. Our results suggest that range margins are stable for 85% of study species at both time one and at remeasurement regardless of canopy disturbance. For the very few species that had a significant difference in seedlings and adults at their range margins, there was nearly a 0.4 degree difference in latitude with seedlings being farther south irrespective of disturbance. Our findings of tree range stability across forests of the eastern US indicate a general propensity towards range contraction, especially for study species forecasted to lose range and located on disturbed sites, which may present substantial hurdles for adaptive management strategies focused on maintaining and enhancing forest ecosystem resilience in the context of global change and associated rapid climate change.
Article
Forest ecosystems across the Northwoods will face direct and indirect impacts from a changing climate over the 21st century. This assessment evaluates the vulnerability of forest ecosystems in the Laurentian Mixed Forest Province of northern Wisconsin and western Upper Michigan under a range of future climates. Information on current forest conditions, observed climate trends, projected climate changes, and impacts to forest ecosystems was considered in order to assess vulnerability to climate change. Upland spruce-fir, lowland conifers, aspen-birch, lowland-riparian hardwoods, and red pine forests were determined to be the most vulnerable ecosystems. White pine and oak forests were perceived as less vulnerable to projected changes in climate. These projected changes in climate and the associated impacts and vulnerabilities will have important implications for economically valuable timber species, forest-dependent wildlife and plants, recreation, and long-term natural resource planning.