SEARCH

SEARCH BY CITATION

In the paper by C.S. Mantyka-Pringle et al., 2012, doi: 10.1111/j.1365-2486.2011.02593.x, published in Global Change Biology, 18, 1239–1252, Figs 3–5 were incorrectly published due to a data processing error that was identified by C.S. Mantyka-Pringle in December of 2012. The authors apologize for any inconvenience that has been caused and the corrected figures are presented below. Consequently, some of the coefficient signs and effect sizes have changed. The authors have therefore rewritten sections of the results and present them below, but the overall conclusions of the paper are unchanged. Readers may refer to the online version of this article to view the article with the corrected figures and results.

Results

  1. Top of page
  2. Results

The influence of climatic factors on habitat loss and fragmentation effects

Paragraph 5 has changed to:

The model-averaged coefficients revealed that maximum temperature was positively related to habitat loss/fragmentation impacts and had the strongest effect compared to the other climatic factors (Fig. 3). The positive relationship indicates that, as maximum temperature increases, the probability of negative habitat loss/fragmentation impacts also increases. On the other hand, the effect size for precipitation change was negative. A negative relationship indicates that the probability of a negative habitat loss/fragmentation effect is lowest where rainfall has increased most over time. Coefficients for minimum precipitation, temperature change and habitat amount were small in magnitude.

image

Figure 3. Coefficient averages from the logistic regression models in Table 2 explaining the variation in habitat loss and fragmentation effects on biological populations as influenced by current climate, climatic change and amount of habitat. Max temperature represents maximum temperature of warmest month; Min precipitation represents precipitation of driest month; Mean precipitation change represents annual average precipitation difference; Mean temperature change represents the annual average temperature difference; Habitat amount represents proportion of the area covered by suitable habitat.

Download figure to PowerPoint

Taxonomic group and habitat type variation

Paragraphs 6 & 7 have changed to:

There were very few differences in the habitat loss/fragmentation effect sizes among taxa (Fig. 4). Apart from arthropods, habitat loss/fragmentation effects on species were large and positive for maximum temperature (Fig. 4a). This indicates that, in general, as temperature increases, the chance of a species being negatively affected by habitat loss/fragmentation also increases, especially for reptiles, which had the largest coefficient of 1.5. For arthropods, the effect size was still relatively large but, as temperature increases, the chance of a species being negatively affected by habitat loss/fragmentation declines. On average, all taxonomic groups (excluding reptiles) had relatively large negative coefficients for precipitation change; in general, habitat loss/fragmentation effects are lowest in areas where rainfall has increased most over time (Fig. 4c). Reptiles, on the other hand, displayed no response to precipitation change. The coefficients for all taxa in relation to minimum precipitation and temperature change were small (Fig. 4b and d).

image

Figure 4. Logistic regression coefficients for each taxonomic group averaged across all models and combined with the fixed-effect model-averaged coefficients. Positive associations exist between habitat loss/fragmentation effects and (a) maximum temperature of warmest month, (b) minimum precipitation of driest month, (c) mean annual precipitation difference, or (d) mean annual temperature difference for taxonomic groups with coefficients greater than zero. Negative associations exist for those taxonomic groups with coefficients less than zero. Coefficients for habitat amount were too small to present.

Download figure to PowerPoint

Effect sizes for different habitat types showed several distinct differences (Fig. 5). First, the coefficients were more variable than for taxonomic groups, indicating that habitat type probably drives most of the variation in the dataset. Coefficients for forest, savanna/grassland, rainforest and wetland habitats were large and positive for maximum temperature, indicating that, in areas with high current temperature, the chance of a species being negatively affected by habitat loss/fragmentation in these habitats is also high (Fig. 5a). Coefficients for woodland and shrubland/heathland were negative, but smaller in magnitude, suggesting that the effects of current temperature on habitat loss/fragmentation effects in these habitats were relatively minor. Other habitats displayed no response to maximum temperature. For precipitation change, the majority of coefficients (excluding wetlands and other habitats) were negative; habitat loss/fragmentation effects in these habitats were lowest in areas where rainfall has increased most over time (Fig. 5c). The wetland and other habitats coefficients were positive, indicating that habitat loss/fragmentation effects in these habitats were highest in areas where rainfall has increased most over time. The coefficients for minimum precipitation were all small (Fig. 5b). Effect sizes varied considerably for temperature change (Fig. 5d). Woodland, shrubland/heathland and rainforest displayed negative coefficients; habitat loss/fragmentation effects were lowest in areas where temperature has increased most over time. In contrast, habitat loss/fragmentation effects on species in savanna/grassland, wetlands and other habitats were highest in areas where temperature has increased most. Forest coefficients were small.

image

Figure 5. Logistic regression coefficients for each habitat type averaged across all models and combined with the fixed-effect model-averaged coefficients. Positive associations exist between habitat loss/fragmentation effects and (a) maximum temperature of warmest month, (b) minimum precipitation of driest month, (c) mean annual precipitation difference, or (d) mean annual temperature difference for habitats with coefficients greater than zero. Negative associations exist for those habitats with coefficients less than zero. Coefficients for habitat amount were too small to present.

Download figure to PowerPoint