The CNPS Approach to Classification

A Focus on Floristics, Rarity, and Representativeness

The floristic components of a classification are those that refer to the plant taxa making up the vegetation of a given area. We believe that the most important units of conservation in a vegetation hierarchy are the floristically based lower units of the alliance and the association. Broader physiognomic units are largely synthetic, including units lacking common species. For example, botanists classify open stands of Pinus monophylla and Quercus engelmannii as woodlands, yet these woodlands share few if any plant taxa. Recognizing both stands as woodland does not express the level of biological diversity that interests conservationists and land managers.

Although we recognize the value of higher levels of vegetation in the national classification, we do not emphasize them in this edition. Appendix 3 displays the hierarchy, in seven of the eight levels (to alliance level), that is being adopted by the National Vegetation Classification (FGDC 2008). This new classification system arises from the earlier National Vegetation Classification System codified first by the Federal Geographic Data Committee (FGDC 1997).

The new national classification addresses many of the concerns brought about by the first version. It places more floristic emphasis on the levels of the hierarchy immediately above the alliance (e.g., macrogroup or group; Table 1). The new classification relies less on rigid structural rules and more on regional floristic and abiotic factors than the original scheme did. This adjustment in the new classification stems from the need to use and apply functional units at various scales more easily for mapping and regional conservation. Although a given alliance may be rare and localized, we are now able to use the new hierarchy to interpret floristic and geographical relationships of that alliance to other similar alliances that may be more common.

Example of a Macrogroup

In the warm temperate zone, where winter temperatures are not as cold as in the higher mountains, forests and woodlands exist in California that are largely dominated by species endemic to the California Floristic Province (Hickman 1993). These include species of oaks (Quercus), pine (Pinus), or cypress (Hesperocyparis). This large group of alliances is the California forest and woodland macrogroup. This macrogroup can be divided further by ecological and floristic criteria into three smaller, but distinct groups of alliances. One includes all of the alliances characterized by broadleaf species and is called the California broadleaf forest and woodland (Quercus agrifolia-Q. douglasii-Q. wislizeni) group. The parenthetical species list the main diagnostic tree species, but the list of alliances and special stands includes:

TABLE 1. Summary of National and International Vegetation Classification Levels and Criteria

Upper Levels: Physiognomy Plays a Predominant Role
Class Broad combinations of general dominant growth forms that are adapted to basic temperature (energy budget), moisture, and substrate conditions. Examples: Shrubland & grassland and Polar & high montane vegetation classes.
Subclass Combinations of general dominant and diagnostic growth forms that reflect global macroclimatic factors driven primarily by latitude and continental position, or that reflect the aquatic environment. Examples: Mediterranean scrub & grassland and Temperate & boreal alpine vegetation subclasses.
Formation Combinations of dominant and diagnostic growth forms that reflect global macroclimatic conditions as modified by altitude, seasonality of precipitation, substrates, and hydrologic conditions. Examples: Mediterranean scrub and Alpine scrub, forb meadow & grassland formations
Middle Levels: Floristics and Physiognomy Play Predominant Roles
Division Combinations of dominant and diagnostic growth forms and a broad set of diagnostic plant species that reflect biogeographic differences in composition and continental differences in mesoclimate, geology, substrates, hydrology, and disturbance regimes. Examples: Mediterranean California scrub and Western North American alpine scrub forb meadow & grassland divisions.
Macrogroup Combinations of moderate sets of diagnostic plant species and diagnostic growth forms that reflect biogeographic differences in composition and subcontinental to regional differences in mesoclimate, geology, substrates, hydrology, and disturbance regimes. Examples: California chaparral (Adenostoma fasciculatum, Heteromeles arbutifolia macrogroup) and Rocky Mountain alpine scrub meadow & grassland macrogroups.
Group Combinations of relatively narrow sets of diagnostic plant species (including dominants and codominants), broadly similar composition, and diagnostic growth forms that reflect regional mesoclimate, geology, substrates, hydrology, and disturbance regimes. Examples: California xeric chaparral (with Adenostoma fasciculatum and Ceanothus cuneatus), Rocky Mountain and North Pacific alpine turf & fell-field (Carex breweri, Carex spectabilis and Festuca brachyphylla) groups.
Lower Levels: Floristics Plays a Predominant Role
Alliance Diagnostic species, including some from the primary layer, which has moderately similar composition that reflects regional to subregional climate, substrates, hydrology, moisture/nutrient factors, and disturbance regimes. Examples: Arctostaphylos viscida alliance, Calamagrostis muiriana alliance.
Association Diagnostic species, usually from multiple growth forms or layers, which have similar composition that reflects topo-edaphic climate, substrates, hydrology, and disturbance regimes. Examples: Arctostaphylos viscida / Salvia sonomensis shrubland association, Calamagrostis muiriana-Vaccinium cespitosum herbaceous association.

Source: Faber-Langendoen et al. 2007.

These alliances share similarities in their ecology by requiring moderate winter rainfall with little snow and frost, being well adapted to summer-drought conditions, typically sprouting after fire, and generally preferring mesic to dry site conditions. These alliances range from relatively mesic, such as the Quercus lobata, Quercus kelloggii, and Quercus wislizeni alliances, to the Quercus douglasii and Aesculus californica alliances, which have special adaptations of summer deciduousness to deal with pronounced summer drought.

The following group contrasts with the preceding one by exhibiting needle-leaves and cones in the Californian evergreen coniferous forest and woodland (Pinus attenuata-P. muricata-P. sabiniana-Hesperocyparis sargentii-H. pigmaea group). This group includes the following alliances or special stands:

This group characterizes many fire-adapted (often closed-cone, or serotinous) conifer species that range from mesic to rather xeric conditions, and they are mostly found below the zone of heavy winter snow. These examples are indicative of the focus of the new middle level of the classification hierarchy. These woodlands may vary from open to dense overstory cover and may be relatively tall to relatively short in stature. Physical structure is de-emphasized compared to ecological similarity. What is more important is that they associate based upon biological, environmental, and geographical similarities.

Rarity in California Vegetation

As with species, we can think of vegetation types in terms of floristic units that range from extremely common to extremely rare. Common and rare vegetation types have similarities to common and rare species. In fact, many rare vegetation types defined by overstory dominance of a rare species under locally favorable conditions form stands of vegetation (e.g., Arctostaphylos myrtifolia, Pinus torreyana, and various cypress stands). As with the huge range of rare species in the state, there is a concomitant range of rare vegetation. Some examples of rare California alliances are the following:

Some rare alliances in California are extensive elsewhere in North America, whereas others are endemic to the state. Some once extensive alliances are reduced now to a small part of their original range; and others were never extensive. Some are diverse and include several associations, while others have a single association. The process of sampling and classifying rare vegetation is often long and detailed. For example, work by the CNDDB on the Platanus racemosa alliance required two seasons of field sampling and several months of data analysis. Only then were ecologists able to define and map three associations, two of which are very rare (Keeler-Wolf et al. 1997). We hope that placing focus upon describing the rarest vegetation types will aid in their protection.

Broad-Scale Vegetation Assessments

The early focus on rarity does not imply that we ignore vegetation types that are common and extensive. The main way that we can understand the rare types is in the context of all of the state’s vegetation. In an introduction to the first edition of this book, we asked the question, “How do we build a quantitative, data-driven classification without collecting and analyzing data from throughout the state?” Our initial answer was that we draw upon existing regional classifications of the state’s vegetation, especially those done by the Forest Service’s ecological type classification work (Allen 1987, Gordon and White 1994, Jimerson 1994, Smith 1994, and others). In addition, we chose to define some well-known alliances from largely anecdotal information because little quantitative data existed. We now have data to support and modify some of our earlier speculations. These data come from many recent synoptic treatments of vegetation types within small to very large assessment areas. Many areas, including the national parks, many large state parks and wildlife areas, and heavily affected zones important for their biodiversity such as the South Coast, Sacramento-San Joaquin River Delta, and northern Sierra foothills, have been mapped.

The use and widespread replication of either the intensive rarity focus seen in the sycamore woodland study or the extensive regional approach of an area such as western Riverside County (Evens and Klein 2005) will increase our understanding of rare vegetation types. In some cases, both intensive and extensive assessments have been useful in defining the alliance and in identifying the conservation value of each of these patterns.

For example, we know that the Quercus lobata alliance is widespread (Allen et al. 1991). Although clearing of this oak woodland for agriculture and urban development has reduced acreage over the past 200 years, considerable area still exists where Q. lobata dominates. However, acreage of closed-canopy valley oak forest in the riparian zone is very small today. Agricultural clearing, gravel mining, and other streambed alterations have substantially reduced its extent. This subtype of the alliance is visually distinct, and Holland (1986) called it Great Valley valley oak riparian forest. By targeting stands dominated by Q. lobata and others not dominated by the oak in the riparian zone for sampling and analysis, Hickson and Keeler-Wolf (2007) verified that the valley oak riparian forest is a unique association within the Quercus lobata alliance. In the recent study of the northern Sierra Nevada foothills, Klein et al. (2007) identified several other riparian or semi-riparian associations of the Quercus lobata alliance including the Quercus lobata / herbaceous semi-riparian association, the Quercus lobata / Rhus trilobata association, the Quercus lobata / Rubus armeniacus association, and the Quercus lobata-Alnus rhombifolia association.

Each of these associations tells a slightly different tale of the alliance. Alnus rhombifolia with Q. lobata suggests narrow valley streams with a permanent summer flow. The Rubus armeniacus (the non-native Himalaya berry) association suggests unnatural or repeated understory disturbance within or immediately adjacent to the stands. Information compiled from study to study shows the range of conditions under which these plants tend to associate. This approach enables us to mark vegetation types just as a geneticist marks the genes of the organisms. By tracking the array of well-defined markers, a vegetation ecologist can identify and map the extent of these biologically defined components of the state’s natural landscape. Through mapping and remapping, we can track the contractions, expansions, and change of these basic components of all ecosystems, or vegetation types. Rapidly developing methodologies are enabling us to map and identify changes in film or satellite imagery very precisely. Some land management agencies are funding regular updates at 5- or 10-year intervals.

Dominance, Existing Vegetation, and Refinements

We typically define alliances based on the dominant plants in the uppermost layer as determined by their relative cover; however, in some cases, we use indicator species that ecologists describe as “representative,” “diagnostic,” or “characteristic” (Mueller-Dombois and Ellenberg 1974, Kent and Coker 1992, FGDC 1997, Grossman et al. 1998, Jennings et al. 2004). Their presence alone sometimes suffices to define an alliance even though the plants may be low in cover (e.g., Yucca brevifolia alliance). In most cases, however, the dominant plants are also the ones that are characteristic of the set of environmental conditions associated with the alliance.

Our classification is for vegetation that exists today. We make no assumptions that stands of a particular alliance will naturally always stay in that alliance or that they will not be replaced. Ecologists know the dynamics (seral states) of a number of forest types in California, but our principal focus in this manual is to identify any vegetation type that currently exists as a visually distinct and repeated entity. At the same time, we recognize that vegetation change and development are real, and we place value upon understanding the different states of a particular area. Understanding allows for wise resource management, including that of endangered species requiring special habitat conditions.

In comparison, earlier classifications developed for California by the Forest Service (e.g., Fites 1993, Jimerson 1994) were based on the concept of a “potential natural community,” not on one of the existing vegetation types. A potential natural community is a type that would establish if all successional sequences of its ecosystem were completed without human-caused disturbance, under present environment conditions (Forest Service Manual). We cannot provide a uniform view of such a potential for the state’s alliance, but we can incorporate results of these studies in the manual.