C6. Insects of the wider countryside (butterflies)

Notes:

1. The line graph shows the unsmoothed trend (dashed line) and the smoothed trend (solid line) together with its 95% confidence interval (shaded).
2. The figure in brackets shows the number of species included in the index.
3. The bar chart shows the percentage of species within the indicator that have shown a statistically significant increase, a statistically significant decrease or no statistically significant change (little change).
4. Since 2017, an improved analysis method has been used to derive the species indices and in 2020, improvements were made to better account for the colonisation of sites (see ‘Background’ section for further information). Therefore, the charts presented here are not directly comparable to those presented in releases prior to 2020.

Source: Butterfly Conservation; UK Centre for Ecology & Hydrology; Defra; British Trust for Ornithology; Joint Nature Conservation Committee.

Habitat specialist species, which are vulnerable to semi-natural habitat loss and fragmentation, have not recovered from declines experienced in the late 1970s. These declines were mainly attributed to the knock-on effects of the drought conditions experienced in 1976. Overall, the unsmoothed habitat specialist index declined by 61% between 1976 and 2020 (Figure C6ia). Analysis of the smoothed trend shows a statistically significant reduction in relative abundance over the period 1976 to 1979. The index showed an increase over the period 2015 to 2020 from 32% to 39% of the 1976 level, however, analysis of the smoothed trend showed no statistically significant short-term change.

Individual butterfly species fare differently within the overall trend. Habitat specialists showing the greatest long-term decline since 1976 include heath fritillary; wood white; Lulworth skipper; grayling; and small pearl-bordered fritillary. No species show a significant short-term decline since 2015. Silver-spotted skipper; large heath; black hairstreak; silver-washed fritillary; and dark green fritillary show the largest significant increases over the long term, whilst silver-studded blue is the only species of habitat specialist butterfly to show a statistically significant increase since 2015.

Notes:

1. The line graph shows the unsmoothed trend (dashed line) and the smoothed trend (solid line) together with its 95% confidence interval (shaded).
2. The figure in brackets shows the number of species trends included in the index.
3. This indicator includes individual measures for 25 species of butterflies; the wider countryside index, however, only includes 24 trends. This is because an aggregate trend is used for small skipper (Thymelicus lineola) and Essex skipper (Thymelicus sylvestris) and has been retained for indicator consistency; these two species were combined due to historical difficulties with distinguishing between them in the field.
4. The bar chart shows the percentage of species trends within the indicator that have shown a statistically significant increase, a statistically significant decrease or no statistically significant change (little change).
5. Since 2017, an improved analysis method has been used to derive the species indices and in 2020, improvements were made to better account for the colonisation of sites (see ‘Background’ section for further information). Therefore, the charts presented here are not directly comparable to those presented in releases prior to 2020.

Source: Butterfly Conservation; UK Centre for Ecology & Hydrology; British Trust for Ornithology; Defra; Joint Nature Conservation Committee.

The unsmoothed species of the wider countryside index showed a decrease of 22% between 1976 and 2020 (Figure C6ib); however, analysis of the smoothed trend shows the change was not statistically significant. The index showed an increase over the period 2015 to 2020, from 69% to 78% of the 1976 level, however, analysis of the smoothed trend also showed this short-term change is not statistically significant.

Individual butterfly species again fare differently within the overall trend. Species of the wider countryside showing the greatest declines since 1976 include wall; small tortoiseshell; white-letter hairstreak; small heath; and gatekeeper. No wider countryside butterfly species show a significant short-term decline since 2015. Ringlet; comma; speckled wood; and marbled white show the largest significant increases over the long term. Small Heath shows a significant short-term increase since 2015.

The year-to-year fluctuations in butterfly numbers are often linked to natural environmental variation, especially weather conditions. Therefore, in order to identify underlying patterns in population trends, the assessment of change is based on smoothed indices. The smoothed trend in the multi-species indicator is assessed by structural time-series analysis. A statistical test is performed using the software ‘TrendSpotter’ to compare the difference in the smoothed index in the latest year versus other years in the series. Within the measures, each individual species trend is given equal weight, and the annual figure is the geometric mean of the component species indices for that year.

Populations of individual species within each measure may be increasing or decreasing, irrespective of the overall trends. The bar chart provided alongside each trend graph above shows the percentage of species within that indicator that have shown a statistically significant increase, a statistically significant decrease or no statistically significant change (little change). A list of species included within each index, together with a summary of the estimated long-term and short-term changes for each species and an assessment of the individual species trends can be found in the accompanying dataset.

In 2017, the method for compiling species annual indices was improved. Indices are now calculated for all species using the Generalised Abundance Index (GAI) method presented in Dennis et al. (2016) with an additional modification that the data from each site in each year is weighted in the final stage relative to the proportion of the species flight period surveyed that year for that site. This weighting is necessary as the GAI extrapolates from observed data to estimate the total count across the season, accounting for gaps in the recording, and ensures that the observed data have a stronger effect upon the final indices than the extrapolated data.

The improved method uses butterfly count data collected at both UKBMS butterfly transect and reduced effort sites (3,768 in total) and randomly selected 1 km squares of the Wider Countryside Butterfly Survey (2,149 in total) – see Figure C6ii for the locations of these sites. All butterfly counts in a season are used to estimate the seasonal pattern of butterfly counts for that year, using a concentrated likelihood method (see Dennis et al. (2016)); the resulting indices and species trends are similar to those generated through previous analysis methods.

In 2020, further methodological improvements were made to better model trends for species that have expanded in range and colonised new sites (UKBMS transects and WCBS squares) by adding pre-colonisation zero abundance counts for species at sites they have colonised, where the site was being monitored prior to colonisation. At the time, the effect of these changes was most notable for species such as Essex skipper and purple hairstreak, where there was a slight reduction in their population indices for earlier years relative to latter years.

As there are delays in data submission, data for previous years are also updated retrospectively. This means that the species indices for individual years may vary from previous publications.

Further details of the methods used can be found on the UKBMS website and in the technical background document for this indicator.

Since 2019, two additional multi-species indices have been compiled for butterflies in the UK: the first for butterflies of the wider countryside on farmland, the second for butterflies of the wider countryside in woodland. Although these habitat-specific measures are not included within the formal assessment for ‘Butterflies in the United Kingdom: habitat specialists and species of the wider countryside’, they are presented here in order to give a more complete picture of the trends for UK butterfly populations.

These habitat indices include 23 species of wider countryside butterflies surveyed on farmland and 25 species surveyed in woodland; the farmland and woodland indices, however, only include 22 and 24 trends respectively. This is because an aggregate trend is used for small skipper (Thymelicus lineola) and Essex skipper (Thymelicus sylvestris); these two species have been combined due to historical difficulties with distinguishing between them in the field. The indices use data collected at 4,143 locations with farmland habitat and 2,871 locations with woodland habitat (these farmland and woodland locations are subsets of the 5,917 locations shown in Figure C6ii). A list of species included within each index, together with a summary of the estimated long-term and short-term changes for each species and an assessment of the individual species trends can be found in the accompanying dataset.

The assessment of change for these habitat measures is made on an analysis of the smoothed trends, however, the base year for the long-term trends is 1990 rather than 1976. This is because prior to this date, there are insufficient data for a number of species included within the two indices. Since 1990, UK populations of wider countryside butterflies on farmland show no significant long-term change, whilst in woodland, butterflies have declined significantly. Since 2015, trends for both farmland and woodland butterflies show no statistically significant short-term change.

Butterflies of the wider countryside on farmland

The unsmoothed butterflies of the wider countryside on farmland index for the UK increased by 2% between 1990 and 2020. The index also increased in the five years to 2020, from 4% below its 1990 level in 2015 to 2% above this level in 2020 (Figure C6iii). Analysis of the smoothed trends shows that neither increase is statistically significant.

Notes:

1. The line graph shows the unsmoothed trend (dashed line) and the smoothed trend (solid line) together with its 95% confidence interval (shaded).
2. The figure in brackets shows the number of species trends included in the index.
3. This indicator includes individual measures for 23 species of butterflies; the UK farmland index, however, only includes 22 trends. This is because an aggregate trend is used for small skipper (Thymelicus lineola) and Essex skipper (Thymelicus sylvestris) and has been retained for indicator consistency; these two species were combined due to historical difficulties with distinguishing between them in the field.
4. The bar chart shows the percentage of species trends within the indicator that have shown a statistically significant increase, a statistically significant decrease or no statistically significant change (little change).
5. In 2020, improvements were made to the analytical techniques to better account for the colonisation of sites (see ‘Background’ section for further information). Therefore, the charts presented here are not directly comparable to those presented in releases prior to 2020.

Source: Butterfly Conservation; UK Centre for Ecology & Hydrology; British Trust for Ornithology; Defra; Joint Nature Conservation Committee.

Individual butterfly species fare differently within the overall stable trend. Species in significant long-term decline on farmland include small tortoiseshell; wall; Scotch Argus; gatekeeper; and Essex/small skipper. No species decreased on farmland over the short term (since 2015). Three species, the ringlet; speckled wood; and brimstone, increased significantly on farmland over the long term, whilst small heath and small copper have increased significantly over the short term.

Butterflies of the wider countryside in woodland

The unsmoothed butterflies of the wider countryside in woodland index for the UK decreased by 42% between 1990 and 2020 but increased in the five years to 2020, from 51% of its 1990 level in 2015 to 58% of its 1990 level in 2020 (Figure C6iv). Analysis of the smoothed trends shows a statistically significant decline since 1990 but no significant change in the short term.

Notes:

1. The line graph shows the unsmoothed trend (dashed line) and the smoothed trend (solid line) together with its 95% confidence interval (shaded).
2. The figure in brackets shows the number of species trends included in the index.
3. This indicator includes individual measures for 25 species of butterflies; the UK woodland index, however, only includes 24 trends. This is because an aggregate trend is used for small skipper (Thymelicus lineola) and Essex skipper (Thymelicus sylvestris); these two species have been combined due to historical difficulties with distinguishing between them in the field.
4. The bar chart shows the percentage of species trends within the indicator that have shown a statistically significant increase, a statistically significant decrease or no statistically significant change (little change).
5. In 2020, improvements were made to the analytical techniques to better account for the colonisation of sites (see ‘Background’ section for further information). Therefore, the charts presented here are not directly comparable to those presented in releases prior to 2020.

Source: Butterfly Conservation; UK Centre for Ecology & Hydrology; British Trust for Ornithology; Defra; Joint Nature Conservation Committee.

The long-term decline of woodland butterflies is thought to be due to a lack of woodland management and loss of open spaces in woods. Species showing the greatest long-term declines in woodland include wall; small tortoiseshell; small copper; Essex/small skipper; and gatekeeper. No species show a significant decline in the short term. Ringlet displayed the same trends in woodland as on farmland, increasing significantly over the long term, but showing no significant change since 2015. Small heath and white-letter hairstreak show a significant short-term increase.