Warnings, Cautions and Questions

How to cite the tree atlas

Prasad, AM; Iverson, LR; Peters, MP; Matthews, SN. 2014. Climate change tree atlas. Northern Research Station, U.S. Forest Service, Delaware, OH. http://www.nrs.fs.fed.us/atlas.

How to cite the bird atlas

Matthews, SN; Iverson, LR; Prasad, AM; Peters, MP. 2014. Climate change bird atlas. Northern Research Station, U.S. Forest Service, Delaware, OH. https://www.nrs.fs.fed.us/atlas/bird/

We recommend the following publications be cited along with the atlas citation, depending on what you used:

For habitat suitability models on trees:

Iverson, L. R., A. M. Prasad, S. N. Matthews, and M. Peters. 2008. Estimating potential habitat for 134 eastern US tree species under six climate scenarios. Forest Ecology and Management 254:390-406. http://www.treesearch.fs.fed.us/pubs/13412

For Modification Factors/Adaptability of tree species:

Matthews, S. N., L. R. Iverson, A. M. Prasad, M. P. Peters, and P. G. Rodewald. 2011. Modifying climate change habitat models using tree species-specific assessments of model uncertainty and life history factors. Forest Ecology and Management 262:1460-1472. http://treesearch.fs.fed.us/pubs/38643

For habitat suitability models on birds:

Matthews, S. N., L. R. Iverson, A. M. Prasad, and M. P. Peters. 2011. Potential habitat changes of 147 North American bird species to redistribution of vegetation and climate following predicted climate change. Ecography 260:1460-1472. http://treesearch.fs.fed.us/pubs/39841

Read the following before interpreting the maps and tables

While we believe that our tree atlas is an example of predictive vegetation mapping and contributes to the understanding of tree distribution under changed climate, we want to emphasize what it is not. In order to avoid the mis-interpretation of our atlas, we want everyone to read the following section before making sense of the maps.

First of all, the results of our modelling effort give potential habitat distributions for future General Circulation Model (GCM)scenarios (2100) for 134 tree species. By potential distribution we mean that the habitat becomes suitable for a species to colonize, provided that the GCM predicted climate of the future is accurate and our model captures all relevant attributes pertaining to the current distribution of the species.

Please note here that we merely use the results of the GCM-climate scenarios - we have no control on its outputs - also, like with all models, ours has several assumptions and limitations.

GCM climate scenarios: We used the data for two emission scenarios: the A1fi (high emissions - which assume that the current emission trends continue into the future without modification) and the B1 (significant conservation and reduction of CO2 emissions). These two emissions scenarios bracket most of the future emissions as outlined by the Intergovernmental Panel on Climate Change’s evaluation of emission scenarios, and end the century at roughly double (550 ppm-B1) and triple (970 ppm-A1fi) the pre-industrial levels for CO2. We also averaged the three models for each emission scenario to yield an average high and average low emission set of climate predictors.

So, if species x has increased its range in our maps on one of the GCM scenarios, it would be accurate to assume that: if climate were to change as defined by that GCM model, then the suitable habitat for colonization for x could expand according to our model. We would like to stress here that our model is not predicting migration of species x - but rather the movement of suitable habitat for that species. In order to predict migration, our model has to take into account several additional factors like fragmentation of landscapes, competition with other species, and other possible inhibiting and accelerating factors, which are beyond the scope of our model.

In addition to the geographic range-shifts, our model predicts the potential change in abundance (importance value) of species. Some species could potentially increase in abundance in some areas and decrease in other areas under future climates. The same caveats outlined above for geographic distributions apply to interpreting changes in abundance.

Frequent Asked Questions

Getting started; where to begin?

If this is your first time viewing the atlas or you would like more information about the contents offered, we suggest that you spend some time watching the tutorial videos provided on the Climate Change Atlas main page, http://www.nrs.fs.fed.us/atlas/. The tutorials are located on the right side and are each 4 to 6 minutes in length.

I don’t see the species I’m interested in.

Currently, we have modeled potential suitable habitat at a resolution of 20x20 km for 134 tree species east of the 100th meridian, using forest inventory data and 38 environmental variables. While we know other species are present within the eastern U.S., there is insufficient data to accurately model these species. As more data become available and our models are updated with finer-scaled data, we may increase the number of species.

What is DISTRIB?

DISTRIB is an empirical statistical modeling framework that utilizes machine learning algorithms (such as random forest and bagging trees) to predict values based on training data. We use DISTRIB to model the potential distribution of suitable habitat for 134 tree species. More details about the modeling methods are provided at: treeatlas_intro.html as well as multiple publications listing in Citations.html.

What is SHIFT?

SHIFT is a spatially explicit cell-based model, which considers landscape fragmentation, abundance of the species, and historical migration rates to determine the probability of colonization for newly suitable habitat modeled by DISTRIB. This allows us to examine how much of new suitable habitat beyond its current range boundary could be colonized by a species.

What are ModFacs?

Output from DISTRIB does not consider many biological or disturbance factors which favor or limit tree establishment, growth, or mortality. We have scored each species by 12 disturbance factors and none biological factors, based on information obtained from the literature, to help interpret DISTRIB models. For example, the life history characteristics of red maple allow it to thrive under most conditions, and likely will do better under climate change than the outputs of the DISTRIB models suggest. The Modification Factors, or ModFacs, are just one way to locally integrate our models into management decisions.

What is meant by current?

The term current is relative and can differ among variables. For example, the current Land Cover Data used in our models was derived from 1992 Landsat imagery. With the climate data, we use current to refer to the 30 year period of climate (1961-1990) used as a baseline. For more information about the variables used in the models, see predictors_help.html.

What is meant by future?

The term future here refers to a future time into which our models attempt to forecast potential suitable habitat for trees, based on estimated climates for 2010-2040, 2040-2070 or 2071-2100. In these models, the non-climate environmental variables are assumed to remain constant into the future.

What is the difference between weather and climate?

The major difference is the measure of time. Weather refers to events that occur over a short period of time, while climate refers to trends over a longer period of time. Our models use average climate values for a 30 year period.

What is a General Circulation Model?

A General Circulation Model (GCM) is a complex mathematical model parameterized with information about the atmosphere, Earth, and oceans to simulate hourly or daily climatic conditions. GCMs are run for a length of time prior to the period of future simulations in order to assess the accuracy of known events. The Intergovernmental Panel on Climate Change (IPCC) has included results from many GCMs into their assessments. These GCMs differ in their sensitivity of climate to respond to various levels of CO2 in the atmosphere. We present the results for three GCMs here.

What is the difference between a General Circulation Model (GCM) and an emission scenario?

GCMs simulate climatic conditions based on information related to the atmosphere, Earth, and ocean, whereas emission scenarios are part of the information included in the GCM to provide future data on greenhouse gas emissions. Each emission scenario estimates the amount of atmospheric greenhouse gases emitted to the atmosphere, based on various conditions related to population growth, technological advances, conservation efforts, and other factors. These ‘storylines’ of emission were developed by the Intergovernmental Panel on Climate Change. The combination of GCM and emission scenario dictates the climate conditions predicted into the future.

Why were the B1 and A1fi emission scenarios chosen?

Of the various emission scenarios included in the Intergovernmental Panel on Climate Change (IPCC) Assessment Report 4 (2001), scenario B1 is the most conservative where, by 2050, emission levels are substantially reduced due to technological advances and conservation efforts. Scenario A1fi assumes that the current emission trends continue into the future, resulting in a tripling of CO2 by the end of the century. Thus, these two scenarios represent the low and high extremes of greenhouse gas emissions into the future and thus are intended to capture a wide range of possibilities.

Why were the PCM, GFDL, and HadleyCM3 models chosen?

These three GCMs represent a wide range in sensitivity to CO2 in the atmosphere. Each has been run with the A1fi and B1 emission scenarios. Thus, with the combination of GCMs and emission scenarios selected, we intend to provide a realistic range of possibilities for changes in suitable habitat. In particular, the HadleyCM3 and A1fi combination is the most severe, while the PCM and B1 combination is the least severe with respect to predicted future climate.  

What assumption do the GCMs use?

Each GCM makes various assumptions about the complex processes which drive climatic events. We do not attempt to model future climate, but rather accept the GCM data as is. For specific details related to the three GCMs used to model suitable habitat, we refer you to the papers referenced on the GCM Model page, treeatlas_gcms.html.

Why don’t some GCMs appear on certain pages?

You may notice that for some analyses, the PCM low and Hadley high models were used, while other analyses use the GCM3 average low and high models. Depending on the analysis, we have shown the potential extremes with the PCM low and Hadley high models or the average of the three GCM models, under high or low emissions, to show the results with relatively less variability in the climate values.

How current is the FIA data and how is it used to calculate importance values?

Forest Inventory and Analysis records from >100,000 plots surveyed during 1980-1993 were used to calculate Importance Values (IVs) for 134 tree species. Importance Values(X) = (50 * basal area(X) / basal area(all species)) + (50 * number of stems(X) / number of stems(all species)), where X is a single species. Thus within a 20x20 km cell, the IV for a particular tree species represents the relative abundance for its potential suitable habitat.

From what time period is the NLCD (National Land Characterization Database), used to classify the landscape?

Variables derived from the NLCD (Fragmentation Index, percent cropland, forest, non-forest, and water) were prepared by Riitters et al. (2002) from ~1992 Landsat TM data.

Which species were used to classify forest types?

Forest types are combinations of species that occur in various abundances within the landscape. We define 10 forest types based on the USDA Forest Service classification, where the dominant species are used in combination to determine the type. More information is provided within the context sensitive help on the forest type pages, curr_fortypes_help.html.

What is Little’s range?

Elbert L. Little developed range boundaries for many tree species across North America and published these ranges in a series of atlases during the 1970’s. The ranges defined by Little used data from field surveys, herbarium records, and expert knowledge to delineate boundaries to encompass the distribution of a tree species. We use “Little’s Boundaries” in many of the maps represented here as one estimate of the range boundary for the species. Even though these boundary estimates are now quite old, these data still remain the primary estimate of range boundaries for the trees of North America.

How realistic are the DISTRIB models?

DISTRIB uses the statistical processes of Random Forest, Bagging Trees, and Regression Tree Analysis to correlate 38 environmental predictors across the eastern U.S. with tree species Important Values derived from forest inventory data. Because we use many predictor variables other than the seven climate variables, DISTRIB is not a ‘climate envelope model’, but could be referred to as a ‘niche model’ as a subset of ‘species distribution models (SDMs). All statistical models lack the capability to include various disturbance and biological features of the species being modeled, and assume that the species is in equilibrium with its environment and has integrated those other features over a long time (e.g., capacity to withstand drought, ice, competition) to allow the species to survive in particular places. As such, the need to uncover every physiological or ecological relationship, as is needed in mechanistic models, is thwarted. For more information about our modeling refer to the publications on the citations page, treeatlas_citations.html.

How reliable is a model with low model reliability?

Models with low model reliability are designated as such because poor performance has been observed among the statistical processes used to predict the current importance values of suitable habitat. Often poor models are the result of low sampling abundance within the forest inventory data; the species is relatively rare or occupies a small geographic area. For these species, it is difficult for DISTRIB to properly detect variations among predictor variables and accurately predict suitable habitat. However, even models with low reliability are useful when examining the potential effects of climate change.

Are the GIS data used to create the maps in the atlas available for my own use?

For each tree species, we provide a link on the species main page under external links to view each model with Google Earth. GIS data are provided upon request to the authors, which allow us to keep records of who is using the data, and possibly offer suggestions on how to better analyze the data.

Where are the models for the western United States?

Our modeling framework, DISTRIB, has proven to perform well in regions with minimal elevational differences. Though we have not attempted to model western species, the predictor variables related to elevation would not be expected to perform in the same way. Basically, we are from the East and know the Eastern tree species and Forest Types pretty well. We cannot say the same for the West. Thus we only offer information for the eastern U.S. (east of the 100th meridian). For the West, we refer you to another group’s web site (http://forest.moscowfsl.wsu.edu/climate/). They provide modeled outputs, using a climate envelope approach, for many western species. Because the methods, input parameters, and resolutions differ, interpretations also differ between systems.

Why do the models stop at the U.S./Canadian border?

Ecosystems do not follow political boundaries, however, datasets collected and managed by various agencies and governments often do not use the same methods for sampling or definitions for reporting. For reasons related to data incompatibility between the USDA Forest Service Forest Inventory and Analysis data and the Canadian equivalent, our models do not yet extend north of the U.S./Canadian border. We hope to extend models across the border one day.

Is information offered at a smaller extent, say State level?

All data and results offered within this atlas have been derived from various aggregates of 20km grids. Regional, State-level, Federal land, and ecoregion analyses that are presented were performed using the minimum number of grid cells for interpretation. It is inappropriate to focus on changes for a single grid cell, or only a few cells, because of the scale and uncertainty of the climate data. Therefore, the modeled suitable habitat within this atlas is better suited for broader landscape-scale assessments. We are working on methods, like Modification Factors (Modfacs), to help local managers better interpret our models based on local site conditions.

 What is the smallest area that should be used for analysis or interpretation?

We recommend that at least 10 20x20 grid cells be considered when interpreting the model outputs. This number of cells, when averaged, will allow reasonable confidence in the outputs, that spurious model outliers are not driving the results. Though we do here provide results for sites down to two grid cells (Hot Springs National Park and Tuskegee National Forest), extra caution is require for such sites. Among all Federal lands that we provide results for, the number of grid cells range from 2 – 68 with a mean of 19 cells.

Is there an easier way to navigate between species?

It really depends on what you are interested in. For most browsing among specific species pages you can easily change the species code within the URL web address in the location bar. Species codes use FIA codes to represent a species, so if you are at www.nrs.fs.fed.us/atlas/tree/ RFtreemod_531.html (American beech) and you want to switch to flowering dogwood (491) you need only replace 531 with 491 in the address. This can be performed with any address that contains a species code. Should you want to move from one species page to another you can also type the FIA code, common name or scientific name into the search box on the right side of the page. This search box uses suggestive typing, so as you type, all entries containing your entered text will appear.