Can biodiversity be measured using composite indices?

Abstract

The search for globally accepted measures for biodiversity evaluation has received increasing attention (for a review Heywood 1995). Many advocates of biodiversity conservation assume that we have - or will have - a precise scientific measure of biodiversity and that measure, once articulated, will tell us which elements of nature to 'target' in conservation effort. A rather tempting approach is the construction of composite indices using subjective weightings, summing or multiplication. Such indices have been widely used in sciences such as economics -e.g. gross national product and consumer price indices (Eichhorn 1978)-, psychology -e.g. IQ and other test scores (Goldstein and Hersen, 1984)-, and in environmental management -e.g. air and water quality indices (Ott 1978)-. The construction of composite indices in all these fields required the development of a strong theoretical basis. No such theoretical basis has yet been established for conservation evaluation (Smith & Theberge, 1987). An alternative approach is to derive composite indices from empirical ecological measurement (e.g., Goldsmith 1975, Adamus & Clough 1980, Williams 1980, and more recently Onofre & Borralho 1993, Araújo 1994, Gomes et al. 1995). Here criteria and indicators perceived as being important are pooled together into an algorithm in order to produce a singular output (usually a number). The significance of these outputs in decision making have been questioned (for a review e.g., Adams & Rose 1978; Given & Norton 1993).
In this study a single data set on distribution and abundance of birds is used to derive biodiversity scores for three criteria inc. ten indicators. Biodiversity scores were averaged by land-use categories and compared. The potential use of these criteria and indicators in linear ranking procedures (e.g., composite indices) is analyzed and discussed.