Identifying recorder-induced geographic bias in an Iberian butterfly database

A database with comprehensive butterfly faunistic information from the Iberian Peninsula and the Balearic Islands was used to estimate inventory completeness as well as the environmental, spatial, and land-use effects on sampling intensities, on a 50x50 km UTM grid. The degree of sampling effort was assessed by means of accumulation curves based on the Clench function. Using the General Linear Model regression procedure, the effects of 22 variables on the estimated sampling efforts were assessed. This combination of methods is proposed as a preliminary step in biodiversity studies, in order to evaluate not only the degree of geographic coverage of existing faunistic data, but also the amount and nature of the bias on the faunistic work done throughout the last two centuries. The degree of spatial effects on the data was greater than the effects of environmental or land-use variables, although the latter two proved to be locally relevant. The results confirm previous findings that collecting is often skewed by relatively simple factors that affect collector activity, such as accessibility and attractiveness of sampling sites. With regard to Iberian and Balearic butterflies, adequate inventories on the scale investigated may probably suffice for further studies of the diversity of this insect group. Additionally, the results enabled us to develop general guide lines for the design of further faunistic work in the area.