Landscape Ecology

, Volume 32, Issue 7, pp 1385–1397 | Cite as

Recovery dynamics and climate change effects to future New England forests

  • Matthew J. Duveneck
  • Jonathan R. Thompson
  • Eric J. Gustafson
  • Yu Liang
  • Arjan M. G. de Bruijn
Research Article



Forests throughout eastern North America continue to recover from broad-scale intensive land use that peaked in the nineteenth century. These forests provide essential goods and services at local to global scales. It is uncertain how recovery dynamics, the processes by which forests respond to past forest land use, will continue to influence future forest conditions. Climate change compounds this uncertainty.


We explored how continued forest recovery dynamics affect forest biomass and species composition and how climate change may alter this trajectory.


Using a spatially explicit landscape simulation model incorporating an ecophysiological model, we simulated forest processes in New England from 2010 to 2110. We compared forest biomass and composition from simulations that used a continuation of the current climate to those from four separate global circulation models forced by a high emission scenario (RCP 8.5).


Simulated forest change in New England was driven by continued recovery dynamics; without the influence of climate change forests accumulated 34 % more biomass and succeed to more shade tolerant species; Climate change resulted in 82 % more biomass but just nominal shifts in community composition. Most tree species increased AGB under climate change.


Continued recovery dynamics will have larger impacts than climate change on forest composition in New England. The large increases in biomass simulated under all climate scenarios suggest that climate regulation provided by the eastern forest carbon sink has potential to continue for at least a century.


New England Recovery dynamics Climate change LANDIS-II Forests 


Northeastern Unites States’ forests are part of one of the largest forested biomes in North America. These forests provide critical goods and services at local to global scales, including timber, tourism, water filtration, and carbon sequestration (Schwenk et al. 2012; Blumstein and Thompson 2015), and are representative of aggrading mesic temperate forests worldwide. Northeastern forests continue to recover from a long period of intensive agricultural clearing and lumbering that peaked in the late nineteenth century (Thompson et al. 2013). Despite the pervasive influence of many novel agents of global change, recovery dynamics, the processes by which forests respond to past disturbance including accrual of aboveground biomass (AGB) and turnover toward long-lived shade tolerant tree species, continue to be the dominant process driving forest change in the region (Thompson et al. 2013; Eisen and Plotkin 2015; Wang et al. in press). Whether future patterns of forest change will continue to be driven by the legacy of historical land-use under changing climate is not well understood.

Climate change has the potential to alter the current recovery trajectory. Global circulation models (GCMs) agree that climate of the Northeastern US will become increasingly warmer, but future patterns of precipitation are less clear (Ning et al. 2015). Possible direct effects of future climate change include increased growth due to longer growing seasons (IPCC 2013; Keenan et al. 2014), decreased growth in the warmest months due to greater respiration and evapotranspiration (Buermann et al. 2013), and shifts in tree species ranges as they respond to shifts in the isotherms to which they are adapted (Farnsworth and Ogurcak 2015). Indirect effects of climate change, such as changing disturbance regimes and CO2 fertilization (Ollinger et al. 2008) will also impact forest composition and function.

We also expect regional variation in the response of recovery dynamics to climate change, associated with forest composition, age, latitude and topographical position. Northeastern forests are characterized by large spatial heterogeneity in the occurrence of forest types, largely corresponding to climate and disturbance gradients (Cogbill et al. 2002; Foster et al. 2006). Individual tree species within forest types are expected to respond differently to forest recovery dynamics and climate change given their life history traits (e.g. shade and thermal tolerance). Uncertainty remains about the fate of regionally critical tree species, especially those predicted to be vulnerable to future conditions (e.g. sugar maple, Acer saccharum) (Gavin et al. 2008; Iverson et al. 2008).

Temperate forest trees are long lived and resilient and their response to climate change can be delayed by centuries (Davis and Botkin 1985). Forecasting changes in northeastern forests requires an approach that can incorporate multiple processes operating at different scales. Unlike niche models (Iverson et al. 2008) that do not incorporate physiological, succession and dispersal processes, physiologically-based, spatially-explicit mechanistic models are well designed to evaluate forest response interactions with climate (de Bruijn et al. 2014; Gustafson et al. 2014). By increasing the use of mechanistic links between tree physiology and climate drivers, such models have increased robustness for study of the response of forested landscapes to the novel climate conditions of the future (Gustafson 2013). To date, most physiologically based spatial models aggregate tree species into functional groups and may therefore miss important species-level responses to climate change, which are often important determinates of ecosystem function and services (Chandler et al. 2012).

In this study we simulated forest recovery dynamics as they are affected by projected climate change throughout New England (Fig. 1). There is concern that northeastern forests will be unable to sustain current ecosystem services under climate change (Dukes et al. 2009; Mohan et al. 2009; Tang and Beckage 2010). Our regional-scale, species-level simulations provide an understanding of future critical ecosystem services. This research will inform for example, the potential future ability of northeastern forests to sequester carbon. New England is an ideal model system to examine these interactions because the land-use history and initial patterns of recovery are well known (Thompson et al. 2013). We addressed the question: How will continued regional forest recovery dynamics affect forest biomass and species composition and how might climate change alter this trajectory?
Fig. 1

EPA level-IV (numbered) ecoregions and climate regions (colored) within the New England study area. Stars indicate net ecosystem exchange calibration sites with both the Hemlock and EMS sites represented by the Harvard Forest star. Inset indicates study area (blue) within North American forests (grey)


Study area

We simulated biomass and compositional change throughout the 13 million hectares of forest within the six New England states (Maine, New Hampshire, Vermont, Massachusetts, Rhode Island, and Connecticut; Fig. 1). Forests cover approximately 80 % of land area and span a gradient of boreal forests in northern Maine to oak-maple forests in southern New England (Duveneck et al. 2015). Mean annual temperatures range from 3 to 10 °C across a north to south gradient. Mean annual precipitation ranges from 79 to 255 cm with more precipitation at higher elevations (Daly and Gibson 2002). The region was completely forested until the seventeenth century, when two centuries of logging and agricultural clearing removed more than half of the forest cover and cut over most of the rest (Thompson et al. 2013). Forest cover reached its nadir in the mid-nineteenth century, after which widespread farm abandonment and a reduction in the intensity of harvesting initiated a century of natural reforestation and forest growth. Although New England is a well-documented example of a great forest transition, New England’s land-use history is paralleled in many regions of the world (Mather 1992).

Simulation model

We simulated forest change using the LANDIS-II v6.0 forest landscape model (Scheller et al. 2007) and a mechanistic growth and succession extension (PnET-succession) (de Bruijn et al. 2014) based on the PnET-II ecophysiology model (Aber et al. 1995). LANDIS-II is a spatially-interactive forest process model that has been widely used to study northern temperate-boreal forests. The PnET-Succession extension v1.0 (de Bruijn et al. 2014; Gustafson et al. 2014) simulates the competition of tree cohorts for light and water for photosynthesis using ecophysiology algorithms upscaled to a grid cell by integrating incoming radiation and water consumption within stacked layers of the canopy. Cohort biomass is used as a surrogate for tree height to simulate canopy layers, which are added when the variation in biomass among cohorts exceeds a user-defined amount. Competition for light is modeled by allocating incoming radiation though each layer with a standard Lambert–Beer formula. Varying climate scenarios are incorporated with monthly precipitation, light, maximum and minimum temperature at a monthly timestep. Growth is simulated as a competition for light and water for photosynthesis among all the cohorts on each grid cell by integrating incoming radiation and water within stacked layers of the canopy. Water balance is calculated by tracking soil water within cells based on precipitation, transpiration by cohorts, evaporation, runoff, and percolation out of the rooting zone. Within cells, cohorts compete for light and water where older and larger (more biomass) cohorts receive higher priority for resources. Growth increases with available light, foliar N, and atmospheric CO2 concentration; and decreases with age and departure from optimal temperature. Production from net photosynthesis is partitioned to biomass pools of wood, root, foliage, and reserves (nonstructural carbon). Soil water and subcanopy light also regulate monthly species establishment probabilities which stochastically determine where new cohorts will regenerate.

We calibrated the model using data from six sites by comparing simulated biomass dynamics on individual 1 ha cells to long-term observations of annual biomass from Howland Forest in Maine (Hollinger et al. 2004), Black Rock Forest in New York (Schuster et al. 2008), Hubbard Brook Forest in New Hampshire (Battles et al. 2014), and the Hemlock tower (Hadley and Schedlbauer 2002), EMS tower (Urbanski et al. 2007), and Lyford plots (Eisen and Plotkin 2015) at the Harvard Forest in Massachusetts. At sites with eddy covariance flux tower instruments (Howland, Hemlock, and EMS), we also calibrated simulated monthly net ecosystem exchange (NEE) to flux measurements of NEE. NEE is a measurement of total carbon flux from the atmosphere to the forest, but our simulations of NEE do not include soil respiration. To account for this discrepancy, we added seasonally observed soil respiration as a proportion of ecosystem respiration to simulated NEE based on observations at the Harvard (Giasson et al. 2013) and Howland Forest (Davidson et al. 2006) flux towers. For our simulations of AGB at the landscape scale, we compared observed AGB from US Forest Service Forest Inventory Analysis (FIA) plots (Bechtold and Patterson 2005) to simulated AGB of pixels representing FIA plots. Duveneck et al. (2015) described an extensive validation of species composition used in initial conditions across New England. As an addition, we simulated 4118 1 ha cells where each cell’s initial forest composition was a representation of a single FIA plot using the same methods to develop landscape initial communities (Appendix I in supplementary material). We then compared observed AGB to simulated AGB at model year zero (spin-up values) using Pearson’s correlation coefficient (r) and root mean squared error (RMSE). Finally, we compared published estimates of total forest AGB within each New England state (Bechtold and Patterson 2005) to modeled AGB, aggregated for each New England state at year zero.

Experimental design

We simulated 100 years of forest dynamics (2010–2110) and evaluated modeled live AGB and compositional changes at 5 year time steps. We used a 250 m cell resolution resulting in >2 million active forested cells. Because our goal was to better understand the relative effect of continued recovery dynamics and anticipated climate change on future forests, we simulated natural gap-scale disturbances, but excluded all other biological and anthropogenic disturbance regimes (see Appendix I in supplementary material).

We compared simulations that assume a continuation of the current climate to simulations that depict future climate change. Simulations that continued the current climate used monthly 4 km gridded temperature and precipitation data provided by the parameter-elevation relationships on independent slopes model (PRISM; Daly and Gibson 2002) based on the period spanning 1981–2013. PRISM uses a large network of climate observations using additional spatial variation such as elevation to calculate gridded climate variables. Simulations of future climate change used downscaled climate grids developed based on a high-emission representative concentration pathway (RCP 8.5) provided by the International Panel on Climate Change (IPCC). There is considerable variation in GCM projections of the RCP 8.5. We compared four GCMs shown to most accurately simulate historical climate in the region (Sillmann et al. 2013): community climate system model v4.0 (CCSM4), community earth system model v1—community atmosphere model v5 (CESM1), Hadley global environment model v2—earth system (HADGE), and the Max Planck Institute earth system model—low resolution (MPIMLR). For each climate scenario, we used 12 km monthly downscaled projections of maximum temperature, minimum temperature, and precipitation obtained from the USGS Geo Data Portal (Stoner et al. 2013) (see Appendix I in supplementary material) as LANDIS-II inputs. For the current climate scenario, we kept CO2 constant at 390 ppm. Under the climate change scenarios, we used the RCP 8.5 projections of increasing CO2. We replicated each of the five scenarios (current climate + 4 GCM projections) five times; however because simulated disturbance were rare (see Appendix I in supplementary material), model outputs were virtually deterministic, so variability among replicates was extremely low, only one replicate was used for the analysis (Thompson et al. 2011; Duveneck et al. 2014).

To evaluate the effect of recovery dynamics and climate change on biomass, we compared average percent change in biomass of the current climate and climate change scenarios from 2010 to 2110 within Environmental Protection Agency (EPA) level-IV ecoregions (Environmental Protection Agency 2012) (Fig. 1). We compared changes in average AGB for each of the 33 tree species and six forest types throughout the 100 year time horizon. We also mapped and compared changes in the proportion of aggregated spruce-fir, northern hardwood, and pine species to assess the spatial distribution and persistence of forest types. We plotted these forest types as red, green, blue (RGB) composite maps at year 2010 and 2110 with the current and as an example, the HADGE climate change scenario. For all analyses we used the R statistical software (R Core Team 2013) and the Raster package for R (Hijmans 2014).



Calibration simulations showed high agreement in the magnitude and timing of observed AGB and monthly NEE at multiple scales (Fig. 2). At the Harvard Forest hardwood flux tower, simulated NEE largely reproduced growing season dynamics, but underestimated respiration during winter months (positive NEE) (RMSE = 71.07 g m−2, r = 0.66). At the Harvard Forest hemlock flux tower, simulated NEE largely represented observed monthly NEE (RMSE = 54.92, r = 0.63), however growing season growth (negative NEE) was overestimated after 2010 when photosynthetic activity began to decline due to an insect infestation that is not incorporated in our simulations. At the Howland Forest spruce flux tower, simulated NEE largely captured monthly observed NEE, but overestimated growing season NEE during a number of years (RMSR = 40.88 g m−2, r = 0.57). Spin-up AGB had moderate agreement with observed AGB across 4118 FIA plots (RMSE = 6.71 kg m−2, r = 0.29). No validation bias was observed across forest types. For individual New England states, total simulated AGB was largely consistent with published estimates of state total forest AGB (±30 %), but overestimated biomass in Maine where greater forested area was present.
Fig. 2

Simulated monthly net ecosystem exchange (NEE) in three New England sites with eddy-covariance observations. Negative NEE values indicate when more carbon uptake is sequestered into the forest than is released to the atmosphere

Aboveground biomass response

Simulated forest growth under all climate scenarios resulted in increases in AGB (Fig. 3). Under current climate the average simulated AGB across New England increased from 11.5 to 15.4 kg m−2 (34 % increase). Under the high emission GCMs, the average simulated AGB increased to 21.0 kg m−2 (82 % increase). Overall, biomass gains were larger in the eastern portion of the study region, especially gains attributable to climate change. At the ecoregion scale, the greatest increases in biomass from 2010 to 2110 were found in forests with the lowest initial biomass (Fig. 3), which were also typically the youngest (Appendix II in supplementary material).
Fig. 3

Average simulated aboveground biomass (AGB) kg m−2 at year 2010 (upper left panel) and percent change in (AGB) from year 2010 to 2110 within ecoregions under different climate scenarios

Simulated species response to recovery dynamics included increases in many moderately to highly shade-tolerant species such as American beech (Fagus grandifolia), red spruce (Picea rubens), and Eastern hemlock (Tsuga canadensis) (Fig. 4). Under climate change, enhanced growth of Eastern white pine (Pinus strobus), sugar maple, and red maple resulted in additional increases in simulated AGB. Spruce-fir species [e.g., balsam fir (Abies balsamea), and red spruce] with lower optimal temperatures for growth (Appendix I in supplementary material), also resulted in greater AGB under climate change scenarios. Responses of other species are found in Appendix III in supplementary material. Within forest types, climate change resulted in a greater proportion of pine abundance and less proportion of spruce-fir (Fig. 5a). However, both pine and spruce-fir forest types resulted in increases in AGB under climate change (Fig. 5b). In addition, large increases in northern hardwood AGB were simulated under climate change compared to current climate (Fig. 5b), although the proportion of northern hardwood AGB under climate change was very similar to current climate (Fig. 5a). Across scenarios, less dominant forest types experienced less biomass change under climate change than more dominant species.
Fig. 4

Mean aboveground biomass of dominant species simulated through time under five climate scenarios. Other species are found in Appendix II in supplementary material

Fig. 5

Proportional (a) and average (b) changes in aboveground biomass by forest type. NH Northern hardwood, SF spruce-fir, PE pine, AB aspen-birch, OK oak, TC tamarack-cedar-hemlock

Spatially, northern hardwood, pine, and spruce-fir species changed similarly in proportional abundance compared to current abundance (Fig. 6a) under both current (Fig. 6b) and a future climate scenario (Fig. 6c). Comparing current climate to climate change at 2110 showed few differences in the spatial dominance of forest types. For example, in both current and climate change, spruce-fire persisted in areas such as the high elevation White Mountains (near the center of the New England landscape). Likewise, species composition in areas originally dominated by northern hardwood forest types in southern New England stayed relatively constant across a climate change scenario (Fig. 6).
Fig. 6

Spatial distribution of northern hardwoods (NH), spruce-fir (SF), and pine (PE) forest types at year 2010 (a) and at year 2110 under current climate (b) and the HADGE climate change scenario (c). Red, green, bluecolor combinations represent proportional distribution of each forest type within each pixel. Numbered insets map locations are represented in corresponding black boxes


New England’s forests are accruing AGB at a rapid rate and the simulations suggest that the regional ecosystem will continue to serve as an important sink for at least another century. Our results corroborate other studies that suggest that these forests are still recovering from colonial land-use and remain well below their maximum biomass capacity (Hurtt et al. 2002; Thompson et al. 2011; Eisen and Plotkin 2015; Wear and Coulston 2015; Wang et al. in press). This has implications for critical local and global ecosystem services including carbon sequestration, tourism, habitat, biodiversity and the regional forest-based industries. The largest potential increases in AGB may be in eastern New England where land use history, climate, and soil properties have otherwise limited growth. Under climate change, additional increases in biomass accrual are expected in addition to those increases expected from recovery dynamics.

Climate change effects on New England forests

Our simulations suggest that 100 years of climate change will have modest effects on forest composition. Trees are both long-lived and slow to migrate (Davis and Botkin 1985; Zhu et al. 2012); and many of the individual trees currently on the landscape will persist for the next 100 years. In addition, the dominant tree species in New England are well north of their southern range boundary and thus are resistant to local extirpations due to shifting climatic conditions. For example, recent evidence suggests that spruce-fir species are not retreating in the New England mountains, as once thought, but instead are expanding downslope as they recover from intensive harvesting at the beginning of the twentieth century (Foster and D’Amato 2015). The stability of tree species distributions in New England demonstrated through mechanistic modeling is a stark contrast from the large shifts in tree species habitat that are derived from climate niche models. This discrepancy has potentially significant consequences when niche models are driving management decisions about a future that cannot be realized given the physiological and ecological processes involved. Indeed, mechanistic simulations done by Wang et al. (in press) were not able to reproduce the habitats portrayed in a niche model even after simulating tree migration to year 2300, a full two centuries longer than GCMs produce climate forecasts.

Our mechanistic approach to modeling climate change effects on species do suggest some species will benefit more than others. For example, throughout much of New England we projected large increases in white pine, sugar maple, and spruce-fir species under climate change. These increases are again in contrast to climate niche models, which project decreased future suitable habitat for these species (Iverson et al. 2008). Sugar maple is an important tree species with strong economic and cultural ecosystem service value to the region. Because sugar maple is a shade tolerant species with a native range extending far to the south of New England (Little 1971), the simulation inputs of warmer temperatures, minor disturbance, and little change in precipitation resulted in large increases in biomass. Although spruce-fir species have lower optimal temperatures than other species simulated (Appendix I in supplementary material), longer growing seasons, increased CO2, and low disturbance frequency resulted in greater AGB under climate change. For these species, we expect that rising temperatures will increase growth rates, but only to a threshold, beyond which growth rates may decline (Sendall et al. 2015; Gustafson et al. in prep.). In addition, species are sensitive to stressors that we did not simulate (e.g., timber harvest, acid rain, insect defoliation and ice storms), which could cause decline (Gavin et al. 2008; Bishop et al. 2015). Contrarily, we did not simulate a temperature mediated response to foliar nitrogen, which may increase growth under warmer conditions (Fisichelli et al. 2015). Nevertheless, we believe that our simulations depict the major physiological and dispersal processes that will determine tree species response to climate change and therefore are more reliable than approaches based on statistical distributions of climatic conditions within species’ ranges. Nonetheless, we acknowledge that climate projections under the RCP 8.5 scenario may produce unforeseen effects stemming from processes not included in this model (e.g., soil microbial activity, invasive insects or extreme weather).

Modeling strengths and limitations

Although our results do not represent the full suite of current and future processes affecting forests, we are confident in our findings, not as predictions, but as scenarios of plausible futures. By incorporating long-term measurements of NEE and a large sample of inventory plots, our calibration procedure provided confidence that successional trajectories of common New England forest types were moderately well represented. Compared to other landscape simulation modeling in the region (Scheller et al. 2011; Thompson et al. 2011), our calibration was more intensive (more data points) and our results showed similar agreement with empirical NEE and AGB. However, our calibration did not include every ecophysiological process. For example, the overestimation of summer month growth (negative NEE) for the hemlock eddy-flux tower at Harvard Forest was likely caused by damage from the hemlock woolly adelgid (Adelges tsugae) (Albani et al. 2010) that became established at the eddy-flux site. As PnET-Succession simulates monthly climate data, we did not include individual extreme weather events expected to be exacerbated by climate change (e.g., individual storms, or freeze events). These events have had, and will continue to influence New England forests. Nevertheless, the anomalies that we found can generally be explained, and we have demonstrated that PnET-Succession was successful at simulating New England forests under current climate at multiple scales. PnET-Succession is particularly strong because it does include climate extremes better than most models, including other variants of LANDIS that rely on average growth and establishment rates at an annual or decadal time step. The climate extremes that have the most effect on structuring forests (e.g., droughts, heat waves, cold winters) are robustly included at a monthly time-step.

Because our focus was on recovery dynamics and the role of climate change, we did not explicitly simulate forest harvesting or land use change, which would almost certainly impact recovery dynamics (Ordonez et al. 2014; Wear and Coulston 2015). The recovery dynamics in our simulations continue patterns of recovery from agricultural abandonment and exploitive harvesting in the nineteenth century as well as the recent land-use regime. Our initial conditions represent modern forests as measured by inventory plots. These forests are a product of the contemporary land-use regime. One consequence of “turning off” the modern disturbance regime may be an overestimate of the magnitude of recovery effects and an underestimate of the climate effects. However, our results are consistent throughout the region, even in areas that are subject to little or no timber harvest or other major disturbances such as federally-owned forests in Vermont and New Hampshire and the suburban forest of southern New England. We acknowledge that land-use interactions with climate change will play an important role in structuring future New England forests and this is an important research topic. Indeed, we are exploring additional interacting drivers such as forest conversion (Ordonez et al. 2014), timber harvesting (Nunery and Keeton 2010) and insects (Albani et al. 2010) that will likely have larger impacts to New England forests than climate change alone, and may interact with climate change in unexpected ways.

We conclude that over the next century, recovery dynamics will supersede climate change effects on New England forests in terms of changes in tree species composition. Most of the dominant tree species and forest types are likely to persist, and some will thrive over the next century. However, our results do not speak to even longer-term effects when extant trees senesce and regeneration becomes more of a driver under an altered climate.



This research was supported in part by the National Science Foundation Harvard Forest Long Term Ecological Research Program (Grant No. NSF-DEB 12-37491) and the Scenarios Society and Solutions Research Coordination Network (Grant No. NSF-DEB-13-38809). Additional funding was provided by an Agriculture and Food Research Initiative Competitive Grant No. 105321 from the USDA National Institute of Food and Agriculture to Purdue University. We thank David Foster and two anonymous reviewers that helped improve the manuscript.

Supplementary material

10980_2016_415_MOESM1_ESM.docx (589 kb)
Supplementary material 1 (DOCX 588 kb)


  1. Aber JD, Ollinger SV, Federer CA, Reich PB, Goulden ML, Kicklighter DW, Melillo JM, Lathrop RG (1995) Predicting the effects of climate change on water yield and forest production in the Northeastern United States. Clim Res 5:207–222CrossRefGoogle Scholar
  2. Albani M, Moorcroft PR, Ellison AM, Orwig DA, Foster DR (2010) Predicting the impact of hemlock woolly adelgid on carbon dynamics of eastern United States forests. Can J For Res 40:119–133CrossRefGoogle Scholar
  3. Battles JJ, Fahey TJ, Driscoll CT, Blum JD, Johnson CE (2014) Restoring soil calcium reverses forest decline. Environ Sci Technol Lett 1:15–19CrossRefGoogle Scholar
  4. Bechtold WA, Patterson PL (2005) The enhanced forest inventory and analysis program—national sampling design and estimation procedures. General technical report. SRS-80. Department of Agriculture, Forest Service, Southern Research Station, AshvilleGoogle Scholar
  5. Bishop DA, Beier CM, Pederson N, Lawrence GB, Stella JC, Sullivan TJ (2015) Regional growth decline of sugar maple (Acer saccharum) and its potential causes. Ecosphere 6:1–14CrossRefGoogle Scholar
  6. Blumstein M, Thompson JR (2015) Land-use impacts on the quantity and configuration of ecosystem service provisioning in Massachusetts, USA. J Appl Ecol 52:1009–1019CrossRefGoogle Scholar
  7. Buermann W, Bikash PR, Jung M, Burn DH, Reichstein M (2013) Earlier springs decrease peak summer productivity in North American boreal forests. Environ Res Lett 8:024027CrossRefGoogle Scholar
  8. Chandler CC, King DI, Chandler RB (2012) Do mature forest birds prefer early-successional habitat during the post-fledging period? For Ecol Manag 264:1–9CrossRefGoogle Scholar
  9. Cogbill CV, Burk J, Motzkin G (2002) The forests of presettlement New England, USA: spatial and compositional patterns based on town proprietor surveys. J Biogeogr 29:1279–1304CrossRefGoogle Scholar
  10. Daly C, Gibson W (2002) 103-year high-resolution temperature climate data set for the conterminous United States. The PRISM Climate Group, Oregon State University, CorvallisGoogle Scholar
  11. Davidson EA, Richardson AD, Savage KE, Hollinger DY (2006) A distinct seasonal pattern of the ratio of soil respiration to total ecosystem respiration in a spruce-dominated forest. Glob Change Biol 12:230–239CrossRefGoogle Scholar
  12. Davis MB, Botkin DB (1985) Sensitivity of cool-temperate forests and their fossil pollen record to rapid temperature change. Quat Res 23:327–340CrossRefGoogle Scholar
  13. de Bruijn A, Gustafson EJ, Sturtevant BR, Foster JR, Miranda BR, Lichti NI, Jacobs DF (2014) Toward more robust projections of forest landscape dynamics under novel environmental conditions: embedding PnET within LANDIS-II. Ecol Model 287:44–57CrossRefGoogle Scholar
  14. Dukes JS, Pontius J, Orwig D, Garnas JR, Rodgers VL, Brazee N, Cooke B, Theoharides KA, Stange EE, Harrington R, Ehrenfeld J, Gurevitch J, Lerdau M, Stinson K, Wick R, Ayres M (2009) Responses of insect pests, pathogens, and invasive plant species to climate change in the forests of Northeastern North America: what can we predict? This article is one of a selection of papers from NE Forests 2100: a synthesis of climate change impacts on forests of the Northeastern US and Eastern Canada. Can J For Res 39:231–248CrossRefGoogle Scholar
  15. Duveneck MJ, Scheller RM, White MA, Handler SD, Ravenscroft C (2014) Climate change effects on northern Great Lake (USA) forests: a case for preserving diversity. Ecosphere 5:1–26CrossRefGoogle Scholar
  16. Duveneck MJ, Thompson JR, Wilson BT (2015) An imputed forest composition map for New England screened by species range boundaries. For Ecol Manag 347:107–115CrossRefGoogle Scholar
  17. Eisen K, Plotkin AB (2015) Forty years of forest measurements support steadily increasing aboveground biomass in a maturing, Quercus-dominant Northeastern forest. J Torrey Bot Soc 142:97–112CrossRefGoogle Scholar
  18. Environmental Protection Agency (2012) Level IV ecoregions of EPA region 1. US EPA Office of Research and Development (ORD)—National Health and Environmental Effects Research Laboratory (NHEERL), CorvallisGoogle Scholar
  19. Farnsworth EJ, Ogurcak DE (2015) Biogeography and decline of rare plants in New England: historical evidence and contemporary monitoring. Ecol Appl 16:1327–1337CrossRefGoogle Scholar
  20. Fisichelli NA, Stefanski A, Frelich LE, Reich PB (2015) Temperature and leaf nitrogen affect performance of plant species at range overlap. Ecosphere 6:1–8CrossRefGoogle Scholar
  21. Foster DR, Oswald WW, Faison EK, Doughty ED, Hansen BCS (2006) A climatic driver for abrupt mid-Holocene vegetation dynamics and the hemlock decline in New England. Ecology 87:2959–2966CrossRefPubMedGoogle Scholar
  22. Foster JR, D’Amato AW (2015) Montane forest ecotones moved downslope in Northeastern US in spite of warming between 1984 and 2011. Glob Change Biol 21:4497–4507CrossRefGoogle Scholar
  23. Gavin DG, Beckage B, Osborne B (2008) Forest dynamics and the growth decline of red spruce and sugar maple on Bolton Mountain, Vermont: a comparison of modeling methods. Can J For Res 38:2635–2649CrossRefGoogle Scholar
  24. Giasson M-A, Ellison A, Bowden R, Crill P, Davidson E, Drake J, Frey S, Hadley J, Lavine M, Melillo J, Munger J, Nadelhoffer K, Nicoll L, Ollinger S, Savage K, Steudler P, Tang J, Varner R, Wofsy S, Foster D, Finzi A (2013) Soil respiration in a Northeastern US temperate forest: a 22-year synthesis. Ecosphere 4:1–28CrossRefGoogle Scholar
  25. Gustafson EJ (2013) When relationships estimated in the past cannot be used to predict the future: using mechanistic models to predict landscape ecological dynamics in a changing world. Landscape Ecol 28:1429–1437CrossRefGoogle Scholar
  26. Gustafson EJ, De Bruijn AMG, Miranda BR, Sturtevant BR (2016) Using first principles to increase the robustness of forest landscape models for projecting climate change impacts. TBDGoogle Scholar
  27. Gustafson EJ, De Bruijn AMG, Pangle RE, Limousin J-M, McDowell NG, Pockman WT, Sturtevant BR, Muss JD, Kubiske ME (2014) Integrating ecophysiology and forest landscape models to improve projections of drought effects under climate change. Glob Change Biol 21:1–14Google Scholar
  28. Hadley JL, Schedlbauer JL (2002) Carbon exchange of an old-growth eastern hemlock (Tsuga canadensis) forest in central New England. Tree Physiol 22:1079–1092CrossRefPubMedGoogle Scholar
  29. Hijmans RJ (2014) Raster: geographic data analysis and modeling. R package version 2.2-12Google Scholar
  30. Hollinger DY, Aber J, Dail B, Davidson EA, Goltz SM, Hughes H, Leclerc MY, Lee JT, Richardson AD, Rodrigues C, Scott NA, Achuatavarier D, Walsh J (2004) Spatial and temporal variability in forest-atmosphere CO2 exchange. Glob Change Biol 10:1689–1706CrossRefGoogle Scholar
  31. Hurtt GC, Pacala SW, Moorcroft PR, Caspersen J, Shevliakova E, Houghton RA, Moore B (2002) Projecting the future of the US carbon sink. Proc Natl Acad Sci USA 99:1389–1394CrossRefPubMedPubMedCentralGoogle Scholar
  32. IPCC (2013) Climate change 2013: the physical science basis, working group I contribution to the fifth assessment report of the Intergovernmental Panel on Climate Change, summary for policymakersGoogle Scholar
  33. Iverson LR, Prasad AM, Matthews SN, Peters M (2008) Estimating potential habitat for 134 eastern US tree species under six climate scenarios. For Ecol Manag 254:390–406CrossRefGoogle Scholar
  34. Keenan T, Gray J, Friedl M (2014) Net carbon uptake has increased through warming-induced changes in temperate forest phenology. Nat Clim Change 4:598–604CrossRefGoogle Scholar
  35. Little EL (1971) Atlas of United States trees: conifers and important hardwoods, vol 1. US Department of Agriculture Miscellaneous Publication 1146Google Scholar
  36. Mather AS (1992) The forest transition. Area 24:367–379Google Scholar
  37. Mohan JE, Cox RM, Iverson LR (2009) Composition and carbon dynamics of forests in northeastern North America in a future, warmer world. This article is one of a selection of papers from NE Forests 2100: a synthesis of climate change impacts on forests of the Northeastern US and Eastern Canada. Can J For Res 39:213–230CrossRefGoogle Scholar
  38. Ning L, Riddle EE, Bradley RS (2015) Projected changes in climate extremes over the Northeastern United States*. J Clim 28:3289–3310CrossRefGoogle Scholar
  39. Nunery JS, Keeton WS (2010) Forest carbon storage in the Northeastern United States: net effects of harvesting frequency, post-harvest retention, and wood products. For Ecol Manag 259:1363–1375CrossRefGoogle Scholar
  40. Ollinger SV, Goodale CL, Hayhoe K, Jenkins JP, Goodale SVOCL, Jenkins KHJP (2008) Potential effects of climate change and rising CO2 on ecosystem processes in Northeastern US forests. Mitig Adapt Strat Glob Change 13:467–485CrossRefGoogle Scholar
  41. Ordonez A, Martinuzzi S, Radeloff VC, Williams JW, Radelo VC (2014) Combined speeds of climate and land-use change of the conterminous US until 2050. Nat Clim Change 4:1–6CrossRefGoogle Scholar
  42. R Core Team (2013) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna. ISBN 3-900051-07-0.
  43. Scheller RM, Domingo JB, Sturtevant BR, Williams JS, Rudy A, Gustafson EJ, Mladenoff DJ (2007) Design, development, and application of LANDIS-II, a spatial landscape simulation model with flexible temporal and spatial resolution. Ecol Model 201:409–419CrossRefGoogle Scholar
  44. Scheller RM, Van Tuyl S, Clark K, Hom J, La Puma I (2011) Carbon sequestration in the New Jersey pine barrens under different scenarios of fire management. Ecosystems 14:987–1004CrossRefGoogle Scholar
  45. Schuster WSF, Griffin KL, Roth H, Turnbull MH, Whitehead D, Tissue DT (2008) Changes in composition, structure and aboveground biomass over seventy-six years (1930–2006) in the Black Rock Forest, Hudson Highlands, southeastern New York state. Tree Physiol 28:537–549CrossRefPubMedGoogle Scholar
  46. Schwenk WS, Donovan TM, Keeton WS, Nunery JS (2012) Carbon storage, timber production, and biodiversity: comparing ecosystem services with multi-criteria decision analysis. Ecol Appl 22:1612–1627CrossRefPubMedGoogle Scholar
  47. Sendall KM, Reich PB, Zhao C, Jihua H, Wei X, Stefanski A, Rice K, Rich RL, Montgomery RA (2015) Acclimation of photosynthetic temperature optima of temperate and boreal tree species in response to experimental forest warming. Glob Change Biol 21:1342–1357CrossRefGoogle Scholar
  48. Sillmann J, Kharin VV, Zhang X, Zwiers FW, Bronaugh D (2013) Climate extremes indices in the CMIP5 multimodel ensemble: part 1. Model evaluation in the present climate. J Geophys Res Atmos 118:1716–1733CrossRefGoogle Scholar
  49. Stoner AMK, Hayhoe K, Yang X, Wuebbles DJ (2013) An asynchronous regional regression model for statistical downscaling of daily climate variables. Int J Climatol 33:2473–2494CrossRefGoogle Scholar
  50. Tang G, Beckage B (2010) Projecting the distribution of forests in New England in response to climate change. Divers Distrib 16:144–158CrossRefGoogle Scholar
  51. Thompson JR, Carpenter DN, Cogbill CV, Foster DR (2013) Four centuries of change in Northeastern United States forests. PLoS One 8:e72540CrossRefPubMedPubMedCentralGoogle Scholar
  52. Thompson JR, Foster DR, Scheller R, Kittredge D (2011) The influence of land use and climate change on forest biomass and composition in Massachusetts, USA. Ecol Appl 21:2425–2444CrossRefPubMedGoogle Scholar
  53. Urbanski S, Barford C, Wofsy S, Kucharik C, Pyle E, Budney J, McKain K, Fitzjarrald D, Czikowsky M, Munger JW (2007) Factors controlling CO2 exchange on timescales from hourly to decadal at Harvard Forest. J Geophys Res 112:G02020CrossRefGoogle Scholar
  54. Wang WJ, He HS, Thompson III FR, Fraser JS, Dijak WD (in press) Changes in forest biomass and tree species distribution under climate change in the Northeastern US. Landscape EcolGoogle Scholar
  55. Wear DN, Coulston JW (2015) From sink to source: regional variation in US forest carbon futures. Sci Rep 5:16518CrossRefPubMedPubMedCentralGoogle Scholar
  56. Zhu K, Woodall CW, Clark JS (2012) Failure to migrate: lack of tree range expansion in response to climate change. Glob Change Biol 18:1042–1052CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media Dordrecht 2016

Authors and Affiliations

  • Matthew J. Duveneck
    • 1
  • Jonathan R. Thompson
    • 1
  • Eric J. Gustafson
    • 3
  • Yu Liang
    • 1
    • 4
  • Arjan M. G. de Bruijn
    • 2
    • 3
  1. 1.Harvard ForestHarvard UniversityPetershamUSA
  2. 2.Department of Forestry and Natural ResourcesPurdue UniversityWest LafayetteUSA
  3. 3.Institute for Applied Ecosystem Studies, Northern Research StationUSDA Forest ServiceRhinelanderUSA
  4. 4.Institute of Applied EcologyThe Chinese Academy of SciencesShenyangPeople’s Republic of China

Personalised recommendations