Abstract
Practical, scientific criteria are needed for selecting reserve areas. Therefore, we evaluated conservation priority to determine the most important habitats of grassland birds. We constructed habitat models to predict the distributions of 29 grassland birds in Chiba Prefecture using bird survey records and habitat data (areas of seven grassland habitat types and neighborhood effects) at 57 sites. We used patches as a unit of analysis to consider connectivity. Logistic regression models were constructed based on species presence/absence at every site as the response variable and the combination of habitat types at seven connection distances as explanatory variables. Then, we selected the best models using the Akaike Information Criterion (AIC). The observed and predicted showed good agreement (≥70% accuracy for 22 species). Moreover, previous knowledge indicated that this method extracted the essential species habitat. We proposed the Conservation Priority Index (CPI) as a criterion for evaluating habitat importance. Then, we plotted the CPI as a map. The high CPI values of paddy fields indicated their importance for grassland birds. As a result, we propose to conserve the paddy fields in Katori and Tako, with the highest CPI. This study proposed an improved species distribution model and the CPI, a new criterion for selecting reserve areas.
