USGS Home
Contact USGS
Search USGS

Swink and Wilhelm (1979, 1994) developed and later refined a system for assessing the quality of native plant communities in the Chicago, Illinois region of the United States. Their system, floristic quality assessment, was based on the concept that plant species display varying degrees of tolerance to disturbance, as well as varying degrees of fidelity to specific habitat integrity. They termed this tolerance and fidelity “species conservatism” and assigned each native plant species in the Chicago region a coefficient of 0 to 10 based on its conservatism relative to other native species in the region. A very conservative species (i.e., one with very low tolerances to disturbance and high fidelity to habitat integrity) was assigned a coefficient of 10, while a species that tolerates almost any disturbance and can be found in almost any habitat type was assigned a coefficient of 0. Species with conservatisms falling between the two extremes were assigned appropriate coefficients ranging between 0 and 10 based on the professional judgment of botanists familiar with the region's flora. Subsequently, coefficients of conservatism were assigned to plant species of Michigan, Missouri, the remainder of Illinois, northern Ohio, and most recently North Dakota, South Dakota, and their adjacent grasslands (Ladd 1993, Andreas and Lichvar 1995, Herman et al. 1997, Taft et al. 1997, Northern Great Plains Floristic Quality Assessment Panel 2001).

Given the assumption that the floristic quality of an area is directly related to its richness in conservative species (Wilhelm and Ladd 1988), the assigned coefficient of conservatism (C) values, together with a list of native plant species present in an area, can be used to evaluate the floristic quality of that area (Swink and Wilhelm 1994). A mean C () can be calculated by summing the C values for each native species present in the survey and dividing the summation by the total number of native species present (N). Thus, represents the average conservatism of the native plant community. If habitat quality of an area has degraded, the first plants lost from the plant community will be conservative species (i.e., those with the highest C values). decreases as conservative species are replaced by less conservative species (i.e., those with lower C values), non-native weeds, or no plants at all. A floristic quality index (FQI) can also be calculated. FQI is a weighted species richness estimate that uses a square root transformation of N to incorporate species richness into the index (Swink and Wilhelm 1994, Taft et al. 1997). It is possible for two plant communities, one with 2 species and one with 20 species, to have the same , but the FQI for the community with 20 species would be much higher. Introduced species are not used in the calculations of either or FQI, but their impact is measured indirectly to the extent that their occurrence is related to a diminishment in the number of conservative, native species present (Swink and Wilhelm 1994).

A common criticism of floristic quality assessment is that the C values are assigned subjectively by a relatively small group of individuals. This concern persists even though the subjectivity comes from a group of experts intimately familiar with the region's flora that assigns C values before floristic quality is evaluated. In our study, we used the C values assigned by the Northern Great Plains Floristic Quality Assessment Panel (2001) to assess the quality of plant communities of one natural and three restored wetland complexes in North Dakota. C values assigned by the panel reflect the panel members' opinions of the conservatism of each species throughout North and South Dakota, excluding the Black Hills. In addition, we performed a second independent assessment using a more objective approach in which we derived C values from data collected from 204 wetlands randomly distributed throughout the prairie pothole region of the United States (Gleason and Euliss, unpublished data; see http://www.npwrc.usgs.gov/wetland). We then compared the results obtained using both the “subjectively” assigned and “objectively” estimated sets of C values.