Sampling of land types by protected areas: three measures of effectivenessapplied to western New South Wales

At the end of 1997, the Western Division of New South Wales had 22 reserves with a total extent of 9458 km(2) or 2.9% of the region. We used five meas ures to follow the effectiveness of the reserve system as it developed betw een 1960 and 1997. Two of the measures - number and total extent of reserve s - are basic statistics in any review of protected areas. The other three measures concern how well the reserve system sampled the region's land type s (e.g. ecosystems, vegetation types), defined here as land systems mapped at 1:250,000. The first of these measures was representativeness - the numb er of land systems sampled to a threshold level. The second was efficiency - the proportion of the reserve system contributing to, but not in excess o f, conservation targets set for each land system. The third measure of samp ling effectiveness was vulnerability bias - the extent to which reserves ha ve been dedicated in parts of the region with most risk of vegetation loss. The representativeness of the reserve system at the end of 1997 was very l ow. Results for efficiency showed that a substantial part of the reserve sy stem was not contributing to conservation targets. This partly reflected ex tensions of reserves to improve their design, highlighting the trade-off be tween design and efficiency. Values for vulnerability bias were close to th ose expected if reservation had been indifferent to risk of vegetation loss from clearing or cropping. Higher values would be expected if reservation had been intended to secure good examples of the more vulnerable land syste ms before clearing or cropping compromised conservation targets. Fluctuatio ns in efficiency and vulnerability bias since 1960 can be related to the es tablishment and extension of individual reserves. We finish the paper by pl acing our measures of effectiveness in the context of a more comprehensive list needed to deal with issues such as environmental gradients and species ' requirements for long-term persistence. (C) 2001 Elsevier Science Ltd. Al l rights reserved.