Ecological footprint (EF), a measure of sustainability, represents the areas of land and water required to produce the goods and services for satisfying the consumption of a defined human population or economy and to provide environmental assimilation capacity for keeping its environmental quality at a given level. Production land footprint, a main component of the EF, represents the lands used by production sectors for producing the goods and services for the final consumption of a given human population. There have been, at least, two alternatives in estimating the production land footprint. One is based on the estimates of 'apparent consumption' of each type of resource, along the line of the methodology proposed by Wackernagel and Pees (Wackernagel, M., Pees, W.E., 1996) and its later refinement. The other, suggested by Bicknell et al. (1998) is based on the monetary values of products that are delivered to domestic final demand, the information recorded in an input-output table. The former faces the challenge of distinguishing the commodities used as intermediate input from those as final consumption. Although its recent development has tried to calculate the raw materials used directly in traded manufactured products (Wackernagel, M., Lewan, L., Hansson, C.B., 1999a) it is not an easy job because the indirect input requirements are also involved. Using input-output analysis, which has been widely used to consider the direct and indirect input requirements of supporting final demand, may alleviate this difficulty. Bicknell et al. (1998) adopted input-output analysis to estimate the footprint of New Zealand. In their calculation, they used land multipliers to obtain the production land areas required to produce the outputs for domestic final consumption directly and indirectly and presented the estimates under three land categories, agricultural land, forest land, and degraded land. However, this presentation causes errors in the concept and calculation. This is because that the area requirement of sector j's production land is obtained by multiplying sector j's land multiplier with its output for domestic final consumption. The result would contain not only sector j's own production land, but also those of other sectors whose outputs sector j uses as inputs. Thus the ensuing estimates on sectoral production lands cannot be presented by land category. Using the hypothetical data in the article of Bicknell et al. (1998), this paper explains the reasons for using the composition of land multipliers instead of land multipliers in estimating the production land footprint so that the calculated areas can be expressed by land category. The revised calculation is then applied to Taiwan for estimating her two footprint components, production land and energy land in 1991. (C) 2001 Elsevier Science B.V. All rights reserved.
