What is SeedSmart?

SeedSmart was developed as a pilot restoration tool, and is designed for use in the Ridge and Valley of the Appalachians, from central Pennsylvania west to eastern West Virginia, and south through Maryland, Virginia, Tennessee, Georgia, and Alabama. The tool is designed to easily expand to other regions of the United States and Canada.

As a pilot, this tool has not yet been tested. We are seeking feedback from users of this tool in the Ridge and Valley.


Setting a goal for restoration is one of the most important, and also one of the most challenging, decisions practitioners face. Our guide recommends the use of ecosystems to inform that decision. NatureServe developed a standardized classification of ecosystems in North America (Comer 2003) that include species composition, range, environmental setting, and other information. This classification was mapped nationwide in partnership with LandFire, GAP, and The Nature Conservancy.

Two versions of the map are updated with new information and maintained by NatureServe (NatureServe 2009). The Existing Vegetation map illustrates natural vegetation (ecosystems), as well as disturbed vegetation and other land uses as it occurs now. The other map, Potential Ecosystem, illustrates natural vegetation presumed to occur in the absence of disturbance. If a site is in poor condition but retains some of its natural vegetation such that the ecosystem can still be identified, the Existing Vegetation map, in combination with a key, can aid the user in identifying the ecosystem(s) at the restoration site. If a site has been completely altered such that no original vegetation remains, the Potential Ecosystem map can be used as a guide.

It is recommended that the user make use of the key to ecosystems at the restoration site. We do this because 1) the map was derived from remote sensing with ground data, but has had limited field verification, and 2) the resolution of the maps (90-m pixel size for Potential Ecosystem, and 30-m pixel size for Existing Vegetation) shows vegetation types as squares, which is, of course, not how vegetation occurs. This is particularly true for ecosystems that occur naturally at smaller scales, or for ecosystems that form a complex of types that are not easily mapped individually.


Sourcing seeds and other plant materials

Understanding one’s goal is an important first step in restoration, and using plants characteristic of the ecosystem that naturally occurs at the site will likely increase the chances of successful establishment. The ecosystem description provides information on the dominant vegetation, including geographic variations. Other variations are typical within an ecosystem, such as ephemeral seepage areas within a forest, or openings within the forest where soils are more shallow. The description also provides information on other expected ecosystems in the area. Natural disturbance regimes that maintain ecosystems are also described.

SeedSmart provides guidance in locating areas supporting the same vegetation type(s) as those found at the restoration site. An ecosystem type is characterized by a particular environmental setting that differentiates it from others. Many plant species occupy a number of different ecosystem types, exhibiting tolerance to a range of settings. However, plants become adapted to local conditions (Doherty et al. 2017; Beaulieu et al. 2004; Leites et al. 2012), and those plants collected from a different example of the same ecosystem type are potentially better adapted to those conditions than those that are collected in a different ecosystem. The Existing Vegetation map can be used to locate other examples of the ecosystem types found at the restoration site if local materials are not available.

SeedSmart also analyzes the recent climate conditions of the restoration site (1984 to 2014) to determine whether climate is already changing at the site. If the climate has changed over the past 30 years, the tool then locates other examples of the ecosystem type where the climate of the previous 30 years is most similar to that of the restoration site. Plants in these examples have already survived climate conditions similar to the changing climate at the restoration site, and collection of those materials as opposed to those of more local examples may confer a greater potential for successful establishment and survival in the near term.  

Important information

“Potential ecosystem” is not a prediction of how the ecosystem is predicted to change with changing climate. This map uses the environmental characteristics of the site to interpret what ecological system would be present there NOW if there had not been human disturbance. We are not currently attempting to predict how ecosystems will transform in mid-century. E.g. “Northeastern Interior Dry-Mesic Forest may transition to Central Appalachian Dry Oak-Pine Forest”.  Our approach is to increase the likelihood of successful restoration over the next decade or so by using plants can adapt to the restoration site by growing in the same environmental setting (same ecosystem elsewhere) as the target, and by also finding areas where the past climate is similar to that of the most recently changing climate at the restoration site.

SeedSmart is one approach to restoration. There is a growing body of knowledge and tools, and our recommendation is to make use of complementary tools and data. Two additional tools available in this region are in West Virginia and North Carolina. Both use environmental information at your restoration site to identify previously collected vegetation plot data to inform planting lists. These are particularly useful in areas where an abundance of plot data exists.



Q: When I click on the Potential Ecosystem button, then on the Existing Vegetation button, sometimes it shows two different natural types at the same location. Why is this? Does it mean that the existing natural type is expected to become the potential type in the future?

A: NO. The two different results are due to map resolution only. The Potential Ecosystem was mapped at a coarser (90-m) resolution than the Existing Vegetation map is (30-m). Smaller 30 square meter patches modeled on the existing vegetation map are not represented on the Potential Ecosystems map, which analyzed the same data for a 90 square meter patch.


Q: Why do we need the key?

A: Maps are models, and are not always perfect reflections of what’s on the ground. It is best to confirm what is mapped is what is actually there by using the descriptors you see in the field.


Q: What if my site doesn’t have any natural vegetation?

A: You can refer to the Potential Ecosystem map to identify your target. You can even use the key on surrounding areas that do have some remnant vegetation to make sure that the Potential Ecosystem makes sense for your location.


Q: some of the ecosystems have very sharp boundaries coinciding with geographical boundaries. Why is this?

A: In some cases, the transition from one type to another is difficult to map. In these cases, we imposed an arbitrary (geographic) boundary. We suggest confirming the type on the ground by comparing the two descriptions, or verifying with the key. Both maps are actively updated as new information becomes available.


Desired functions not yet included

  1. Predicted climate: the tool informs decisions based on most recent climate, and does not provide information about predicted climate change.
  2. Transparent climate results: The climate models are ecosystem-specific, and are based on 14 bioclimatic variables. For each ecosystem, the 6 least correlated variables were used to generate the models. The climate models as currently displayed show only where climate is similar, but does not show what the characteristics of the changing climate are. This information is valuable for planning beyond the next few decades, when climate climate is predicted to change considerably. Understanding the most important variables impacting individual ecosystems will allow for planting species that respond more favorably, or are at least more tolerant of, the changing climate characteristics. Successional pathways: mature forests do not arise as such. Early stages are usually dominated by rapidly-growing light-requiring species that eventually recede as longer-lived trees take hold, which in turn shade the understory and provide for more shade-tolerant herbs.


Beaulieu, J., M. Perron, and J. Bousquet. 2004. Multivariate patterns of adaptive genetic variation and seed source transfer in Picea mariana. Canadian Journal of Forest Research 34: 531-545.

Comer, P., D. Faber-Langendoen, R. Evans, S. Gawler, C. Josse, G. Kittel, S. Menard, M. Pyne, M. Reid, K. Schulz, K., Snow, and J. Teague. 2003. Ecological Systems of the United States: A Working Classification of U.S. Terrestrial Systems. NatureServe, Arlington, Virginia.

Doherty, K.D., J. Butterfield, and T.E. Wood. 2017. Matching seed to site by climate similarity: techniques to prioritize plant materials development and use in restoration. Ecological Applications 27: 1010-1023.

Ferree, C. and M.G. Anderson.  2013.  A map of terrestrial habitats of the Northeastern United States: Methods and approach.  The Nature Conservancy, Eastern Conservation Science, Eastern Regional Office.  Boston, MA.

Leites, L.P., A.P. Robinson, G.E. Rehfeldt, J.D. Marshall, and N.L. Crookston. 2012. Height-growth response to climatic changes differs among populations of Douglas-fir: a novel analysis of historic data. Ecological Applications 22: 154-165.

NatureServe. 2009. Terrestrial Ecological Systems and Land Cover of the Conterminous United States. NatureServe, Arlington. VA.

Rollins, M.G. 2009. LANDFIRE: a nationally consistent vegetation, wildland fire, and fuel assessment. International Journal of Wildland Fire 18:235-249.

Yang, J.,  J.H. Pedlar, D.W. McKenney, and A. Weersink. 2015. The development of universal response functions to facilitate climate-smart regeneration of black spruce and white pine in Ontario, Canada. Forest Ecology and Management 339: 34-43.