STAND MANAGEMENT COOPERATIVE

 

 

Back to Newsletter

Articles of Interest

Biodiversity Indices: Roadmaps to Future Forests? JoEllen Kassebaum, Graduate Student, Evergreen College, Former SMC Field Crew Member

Managed forests have come under scrutiny with regards to decreased biodiversity (Pitkänen, 1998). Yet, there does not appear to be an established quantitative measurement system(s) to address this concern (Mazzotti & Morgenstern, 1997; Kangas & Pukkala, 1996; Halpern & Spies, 1995; Roberts & Gilliam, 1995; Underwood, 1995; Fairweather, 1993; Ehrlich & Ehrlich, 1992; Murphy & Noon, 1992; Hansen et al, 1991; Thomas & Salwasser, 1989).

Recently, Stand Management Cooperative understory vegetative data collected from tree-spacing research plots were utilized to test existing biodiversity measures using a standard spreadsheet program. These data were collected from several sites in British Columbia, Washington, and Oregon using two different systems for species identification (life form characterization or botanical genus and species) and percent cover to estimate species abundance. Species identification (termed species richness) and abundance are integral parts of biodiversity calculations (Magurran, 1988; Pielou, 1975; Poole, 1974; Whittaker, 1972).

Using the SMC-collected understory vegetation data from research sites with known sampling methods, eight alpha diversity measures (simple species richness and abundance counts, Margalef’s and Menhinick’s species richness indices, Shannon-Wiener index, Simpson index, Berger-Parker index, Q Statistics, and Hill’s order of indices system) were evaluated for 1) appropriateness to a conventional understory vegetation sampling method, 2) ease of calculation, 3) applicability to typical computer spreadsheet use, and 4) information produced.

Comparing the Stand Management Cooperative sampling methods against established methods (see Gauch, 1989; Causton, 1988; Ludwig & Reynolds, 1988; Southwood, 1978; Mueller-Dombois & Ellenberg, 1974) indicates both the initial competition sampling grid and permanent survey plots used an establishment method similar to stratified random sampling. The environment, size and shape of each sample quadrant, and plant community structure under consideration are consistent. The field data collection methods are standardized and protocols are printed in the field manual available to the field crew each season. To this point, the sampling methods seem consistent with most ecological research procedures.

The sampling methods need improvement in three areas:

  1. Each season new vegetation surveyors are hired for the data collection period. Survey crew familiarity with vegetation identification and code names, estimation of percent cover, and thoroughness of quadrant examination are variables. These inconsistencies will skew the biodiversity data. What is called a generic “forb” one season may be called by its genus name the following survey season. This directly affects the species richness numbers in biodiversity calculations. As an example, the vegetation from the LaVerne Park installation was surveyed in 1989 in which 80 entries labeled “forb” and 51 entries labeled “grass” were recorded. The next survey in 1995 showed only four unknowns labeled, x-1, x-2, etc. Since it was not noted whether “forb” represents one species seen 80 times or 80 different species, the diversity calculations will have to reflect that this was one species. The same can be said of other recordings. The LaVerne Park data may reflect an increase in biodiversity even if no new species established between survey years. Clearly, comparing life form designations with genera and species names is not a legitimate comparison. Analysis using this data set in its current state could not derive logical inference or conclusions.

  2. The thoroughness in investigating vegetation quadrats has the same shortcoming. These are samples of a larger population. If species are missed in the quadrant survey, it affects both species richness and abundance figures as well as not being truly representative of the population. Percent cover estimates are rarely precise but adequate if the estimations are consistent among the data. Calibrating human surveyors is not easily accomplished. Variation in accuracy of cover estimation will affect the abundance figures in diversity calculations. Employing, at least, one surveyor consistently to train each year’s survey crew can reduce the abundance and identification errors to some degree.

  3. Using percent cover is considered appropriate given the life form parameters of plants (Magurran, 1988; Goldsmith et al, 1986; Pielou, 1975; Mueller-Dombois & Ellenberg, 1974). However, percent cover can sum to more than 100% for any given sample area. Vegetation foliage overlaps, consequently more than one species can project a shadow upon the same portion of soil area. While this is reality, biodiversity quantification is based upon abundance proportions that sum to 1.00 or 100%. Deciding upon a conversion method to adjust the overlapping percent cover figures to sum to 1.00 is a necessity.

Biodiversity calculations would be performed more easily within a statistical computer program. Most indices are applicable to typical spreadsheet programs, but the calculations require repetitions of multiplication and/or division. Tedious, but achievable. The same is true of squaring procedures within calculations. With larger biological data sets, the calculations require a computer with high memory capacity and a processor with some speed (i.e., greater than a 486). The entire Stand Management Cooperative database with 70,000 records does not translate to a normal-sized spreadsheet. Larger data sets are even more of a challenge. Additionally, the charting tool in the spreadsheet program (used to display some calculations) was limiting in the scale that could be chosen, the differentiation of data display bars, and the placement of titles. There may be more flexibility with a statistical program that is specifically written with statistical graphic displays in mind. Although it is possible to accomplish the diversity calculations on a typical spreadsheet program, a statistical software program or biodiversity program would reduce the time involved and the possibility of errors considerably. Prince (1986, p. 370) offers a table comparing the various software packages useful for ecological assessments. Biodiversity calculation packages are also available on the Internet.

These diversity calculations produce unitless numbers that usually increase with increasing diversity. (Technically, they produce nats or nat bels if using a natural log in the calculation. However, use of any unit was not evident in the literature.) As such, there is no universal standard to compare against; a biodiversity index result needs to be compared to other results of the same index, either over time or space. Graphed over time they can indicate canopy closure, changes in habitat, management regime, etc. by the increases or decreases in their calculation number. Their use as a monitoring tool by comparing one year to the next might allow funds and personnel to be utilized more efficiently. A drastic decrease/increase in the index number for a site over a period could signal the need to investigate management practices, habitat changes, etc. Slight index variations might be interpreted as continued health and indicate that no immediate investigation is needed.

Since the Stand Management Cooperative study is long-term, comparison of the research plots over time would produce characteristic data set patterns that could be correlated to specific stand health, which might allow development of a rating system similar to the Index of Biological Integrity used for streams (See Karr et al., 1986).

For some resource managers, the possibility of assessing the fluctuation of plant communities with animal communities in the same area is of interest. Index evaluation might help establish correlation between management practices and animal populations and be useful with regard to Habitat Conservation Plans. Plant and animal communities could not be directly compared to one another, but if one shows an increase in diversity, it might be of interest to see if the other is also showing an increase. Using the same accepted diversity measurement system (e.g., Shannon-Wiener, Simpson, or species richness) for both populations would seem to reduce inconsistencies that might arise by using two different quantification systems.

Another monitoring application might be tracking the fluctuation of the native understory community verses the invasive weed and/or noxious weed communities. High populations of one or more of these communities might be indicative of stand health, success of management regimes, or change in environmental conditions.

Biodiversity measures are useful tools for resource management. They are not a definitive answer to diversity management but more one method to describe, monitor, or compare specific sites. Choice of diversity measure will be determined by the project goals or sampling methods. The Shannon-Wiener index is used in a variety of disciplines (communications, psychology, animal behavior study, engineering) but has some flaws when applied to ecological data (Magurran, 1988; Ludwig & Reynolds, 1988; Pielou, 1974; Poole, 1974). Specifically for the SMC data set, the limiting factor is the sampling methods under current practice. A well-trained survey crew with extensive botanical knowledge and a consistent sampling schedule would provide results that are more accurate. Quantifying biological functions or systems has the inherent problem of reducing complex relationships, interactions, or life processes into a more abstract concept (in this case, a number). Information is lost. Underlying mechanisms, health, or associations are not reflected. Although it seems more unbiased to assess sites quantitatively, resource management decisions involving biological entities should not be made solely on quantitative results. These indices are tools to describe, monitor, or compare a part of a whole system. As such, they can be indicators or guides to help discern areas or management regimes that need more thorough evaluation, suggest overall health or plant-animal relationships, reflect a general trend, or signal responses from silviculture and/or horticulture practices. They should not be interpreted as a definitive answer or proof. Although the Shannon-Wiener index appeared to be the most reflective of this data set and useful under a variety of other circumstances, there is no one perfect measure. Different diversity measures may be required at different times or for different purposes.

As a potential monitoring device in resource management, a suggested general protocol might be:

  1. Decide on a resource management goal, e.g., monitoring, comparison of management practices, etc.
  2. Implement and consistently apply established ecological sampling methods which include appropriate plant identification, standardized cover estimates, data verification, and accurate database records.
  3. Construct species-abundance graphs to verify the data set conforms to characteristic biological patterns (see Poole, 1974 or Magurran, 1988 for examples).
  4. Calculate the Shannon-Wiener and Simpson indices (see Magurran, 1988, Worked examples section). Examine the data set using Hill’s order of indices system.
  5. Graph index results at specific sites over time to develop a site evaluation system (See Karr et al., 1986).
  6. Evaluate index results in relation to species lists, management regimes, canopy closure, animal populations, etc.
  7. Investigate any indicated abnormalities by appropriate qualitative evaluation. These indices are tools not definitive answers.
References

Causton, D. R. 1988. An Introduction to Vegetation Analysis. London: Allen & Unwin, Inc.

 
Ehrlich, P. R. and A. H. Ehrlich. 1992. The Value of Biodiversity. Ambio, 21(3):210-225.

Fairweather, P. G. 1993. Links between ecology and ecophilosophy, ethics and the requirements of environmental management. Australian Journal of Ecology, 18:3-19. Gauch, Jr., H. G. 1989. Multivariate Analysis in Community Ecology. Cambridge: Cambridge University Press. Goldsmith, F. B., C. M. Harrison, and A. J. Morton. 1986. Description and analysis of vegetation. Methods in Plant Ecology. 2nd ed. Moore, P. D. and S. B. Chapman, eds. Oxford: Blackwell Scientific Publications, 437-524. Halpern, C. H. and T. A. Spies. 1995. Plant Species Diversity in Natural and Managed Forests of the Pacific Northwest. Ecological Applications, 5(4):913-934. Hansen, A. J., T. A. Spies, F. J. Swanson, and J. L. Ohmann. 1991. Conserving Biodiversity in Managed Forests. BioScience, 41(6):382-392.

 

Back to SMC Newsletter