Survey data were collected from 36 NFI sites and 107 monitoring locations across 8 regions. Surveys were conducted between 2022-07-11 and 2022-09-13. Total survey effort comprised 7129 hours over 56 nights (944 ‘survey nights’ in total). Survey locations and timelines for all sites are shown in appendix A1.
We compiled the tables below using the results of our automated classification algorithms. Our algorithm allows us to automatically identify bat species from the recordings collected based on their sound emissions. We have set cautious thresholds for accepting a classification result as true, however no algorithm can be 100% accurate, so even with cautious thresholds a small proportion of classifications may be incorrect. Our approach to automatically and manually checking classification results is being continually revised to improve accuracy. Note, due to difficulties in automating Myotis species classifications, we report classifications to genus level for these species.
Table 1. Species detected by survey region (ticks indicate presence, crosses indicate absence).| Species | East Midlands | Eastern England | North-East England | North-West England | South-East England & London | South-West England | West Midlands | Yorkshire & Humber |
|---|---|---|---|---|---|---|---|---|
| Barbastelle | \(\checkmark\) | \(\checkmark\) | \(\times\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\times\) | \(\checkmark\) |
| Serotine | \(\checkmark\) | \(\checkmark\) | \(\times\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) |
| Myotis | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) |
| Noctule | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) |
| Common pipistrelle | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) |
| Soprano pipistrelle | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) |
| Brown long-eared bat | \(\times\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\checkmark\) | \(\times\) |
| Greater horseshoe bat | \(\times\) | \(\checkmark\) | \(\times\) | \(\times\) | \(\checkmark\) | \(\checkmark\) | \(\times\) | \(\times\) |
| Lesser horseshoe bat | \(\times\) | \(\checkmark\) | \(\times\) | \(\checkmark\) | \(\times\) | \(\checkmark\) | \(\checkmark\) | \(\times\) |
| Species | Broadleaved | Coniferous |
|---|---|---|
| Barbastelle | \(\checkmark\) | \(\checkmark\) |
| Serotine | \(\checkmark\) | \(\checkmark\) |
| Myotis | \(\checkmark\) | \(\checkmark\) |
| Noctule | \(\checkmark\) | \(\checkmark\) |
| Common pipistrelle | \(\checkmark\) | \(\checkmark\) |
| Soprano pipistrelle | \(\checkmark\) | \(\checkmark\) |
| Brown long-eared bat | \(\checkmark\) | \(\checkmark\) |
| Greater horseshoe bat | \(\checkmark\) | \(\checkmark\) |
| Lesser horseshoe bat | \(\checkmark\) | \(\checkmark\) |
We calculated species richness, evenness, and diversity per NFI plot.
Species richness describes the number of species detected. Figure 1 shows the distribution of species richness values across all regions, sites, plots, and dominant woodland types. Figure 2 shows the distribution of species richness values by region. Figure 3 shows the distribution of species richness values by dominant woodland type.
Figure 1. Distribution of species richness values across all NFI 2022 survey plots. Black points and bars represent mean values for all NFI plots and 95% confidence intervals.
Figure 2. Distribution of species richness values across all NFI 2022 survey plots by region. Black points and bars represent mean values for all NFI plots and 95% confidence intervals.
Figure 3. Distribution of species richness values across all NFI 2022 survey plots by dominant woodland type. Black points and bars represent mean values for all NFI plots and 95% confidence intervals.
We found a statistically significant main effect of region on species richness values (Likelihood ratio test (LRT); χ2 = 92.287, degrees of freedom (df) = 7, p = 0). Post-hoc Tukey’s HSD (honestly significant difference) tests revealed numerous statistically significant differences between regions (Figure 4).
Figure 4. Species richness values across all NFI 2022 survey plots by region. The model’s fitted values are represented by coloured points, and their respective 95% confidence intervals are depicted as coloured error bars. Annotations (brackets and p values) indicate statistically significant pairwise differences between regions.
Species evenness describes the balance between the abundance of each detected species. Figure 5 shows the distribution of species evenness values across all regions, sites, plots, and dominant woodland types. Figure 6 shows the distribution of species evenness values by region. Figure 7 shows the distribution of species evenness values by dominant woodland type.
Figure 5. Distribution of species evenness values across all NFI 2022 survey plots. Black points and bars represent mean values for all NFI plots and 95% confidence intervals.
Figure 6. Distribution of species evenness values across all NFI 2022 survey plots by region. Black points and bars represent mean values for all NFI plots and 95% confidence intervals.
Figure 7. Distribution of species evenness values across all NFI 2022 survey plots by dominant woodland type. Black points and bars represent mean values for all NFI plots and 95% confidence intervals.
We also found a statistically significant main effect of region on species evenness values (LRT; χ2 = 18.576, df = 7, p = 0.01). Post-hoc Tukey’s HSD tests revealed a statistically significant difference in species evenness between North East England and South East England and London (Figure 8).
Figure 8. Species evenness values across all NFI 2022 survey plots by region. The model’s fitted values are represented by coloured points, and their respective 95% confidence intervals are depicted as coloured error bars. Annotations (brackets and p values) indicate statistically significant pairwise differences between regions.
Species diversity index (Shannon’s) describes the number of species present, taking into account the abundance of each species; higher values indicate greater species diversity. Figure 9 shows the distribution of species diversity values across all regions, sites, plots, and dominant woodland types. Figure 10 shows the distribution of species diversity values by region. Figure 11 shows the distribution of species diversity values by dominant woodland type.
Figure 9. Distribution of species diversity values across all NFI 2022 survey plots. Black points and bars represent mean values for all NFI plots and 95% confidence intervals.
Figure 10. Distribution of species diversity values across all NFI 2022 survey plots by region. Black points and bars represent mean values for all NFI plots and 95% confidence intervals.
Figure 11. Distribution of species diversity values across all NFI 2022 survey plots by dominant woodland type. Black points and bars represent mean values for all NFI plots and 95% confidence intervals.
We found no statistically significant effects of region or dominant woodland type on species diversity values (Figure 12).
Figure 12. Species diversity values across all NFI 2022 survey plots by region. The model’s fitted values are represented by coloured points, and their respective 95% confidence intervals are depicted as coloured error bars.
Bat activity can be measured in many different ways.
Here we show relative bat activity as the number of minutes per hour that bats were detected as present in the audio recordings across all NFI survey sites. After controlling for variations in survey effort between regions, sites, plots, and dominant woodland type, we found a statistically significant main effect of survey hour on bat activity levels (LRT; χ2 = 240.303, df = 9, p = 0) (Figure 13).
Figure 13. Bat activity levels (number of minutes per hour that bats were detected) across all NFI 2022 survey plots by hour. The model’s fitted values are represented by coloured points, and their respective 95% confidence intervals are depicted as black error bars. Grey dots represent raw data points, shaded according to frequency (dark indicates many observations, light indicates few).
Here we show relative bat activity as the number of calls per minute per hour that were detected in the audio recordings across all NFI survey sites. After controlling for variations in survey effort between regions, sites, plots, and dominant woodland type, we also found a statistically significant main effect of survey hour on bat calls per minute (LRT; χ2 = 1877.733, df = 9, p = 0) (Figure 14).
Figure 14. Bat activity levels (number of calls per minute per hour) across all NFI 2022 survey plots by hour. The model’s fitted values are represented by coloured points, and their respective 95% confidence intervals are depicted as black error bars. Grey dots represent raw data points, shaded according to frequency (dark indicates many observations, light indicates few).
Figure A1.1. NFI 2022 survey locations.
Figure A1.2. NFI 2022 survey timelines by region and plot.