Distribution statistics

Our analytical methodology explained

This major analysis of distribution statistics was written in 2016. The analysis consists of a page of information per species, each with a distribution map, together with two tables, looking respectively at long term and short term statistics.

For distribution information on each butterfly species see our Butterfly Atlases.

Dorset butterfly surveys 1970-2014

The table below summarises the various survey periods and publication of the data (if any).

SURVEY YEARS KM squares with 1 or more records % of 2,915 KM squares covered Number of species per square Map of the recording effort Hard copy publication sources ?
1970-84
[15 years]
2,624 90% na   Butterflies of Dorset 1970-84 
1980-94
[15 years]
2,857 98% na   New Atlas of Dorset Butterflies 1980-94
1990-2004
[15 years]
2,639 91% 12.7   Not published
2000-14
[15 years]
2,671 92% 14.9   Not published
1995-99
[5 years]
2,465 85% 10.5   Dorset's Millenium Butterfly Population 1995-99
2000-04
[5 years]
1,826 63% 10.5   Dorset's Butterfly Population 2000-04
2005-09
[5 years]
1,964 67% 11.1   Not published
2010-14
[5 years]
2,365 81% 11.7 Map Dorset Butterfly Atlas 2010-14 on Dorset web

The information in the table above shows that 2010-2014 is the most recent of many surveys dating back to 1970. To find out what changes have taken place in the distribution of Dorset butterflies we need to find a way of combining these statistics.

Since the end of the twentieth century we have nationally recorded butterflies in five-year periods. The first was 1995-1999, published in the national ‘Millennium Atlas of Butterflies in Britain & Ireland’ (Jim Asher, Martin Warren et al. Oxford University Press, 2001. 456 pages.)

The Dorset statistics for our latest five-year survey, 2010-14 compare very favourably with those of 1995-99, as each covered over 2,000 kilometre squares (kms), representing more than 80% of the total number of kilometer squares (kms) in Dorset (2,915kms). The two surveys of 2000-04 & 2005-09 recorded a lower number and percentage of Dorset kms and are so best left out of the comparison.

However, before these surveys, Dorset had already completed two overlapping 15-year map surveys of its own under the leadership of Dr Jeremy Thomas and published in two books (details are shown on the Butterfly Atlases index page. Stretched over much longer periods of time, both the 1970-84 and 1980-94 projects attained over 90% coverage.

Comparing 1970-1994 statistics with later statistics

Comparing 1970-1994 statistics with later statistics to show what changes have taken place in the distribution of butterfly species in Dorset.

Unfortunately technical difficulties make it impossible to capture the data from the older 1970-1994 atlases for entry into this atlas. However the format of overlapping 15 years is a sound method of measuring long term changes in butterfly distribution, as it smooths over any short-term changes brought about by unusual weather or poor recording and gives more time for the region to be properly covered.

It has been possible to reconstruct the later data into the same overlapping 15 year intervals, thus producing the additional results for 1990-2004 and 2000-2014 shown in the table above.

It would make sense to select the years 1970-84 for comparison with 2000-14 but only if the recording effort is broadly similar. An increase in recording effort might create an illusion of rapid expansion, while a drop in recording might suggest a decline in the number of squares occupied by a species. Fortunately the percentage of km squares covered by the surveys is broadly similar at 90% compared to 92%.

It must be remembered that just one visit recording one butterfly would mean that a square is ‘filled’. The figure for the number of species per square is therefore a useful statistic, to show the real density of recording, but unfortunately we have no data for the number of species per square in 1970-84. The figure of 14.9 species per square for 2000-14 is the highest recorded in the table. It is therefore just possible that the 1970-84 recording effort may not have been comparable with 2000-14. The statistics showing growth in numbers of occupied squares could simply be due to more recording effort rather than any real expansion of species, and so need to be looked at with caution.

Because long term comparisons tend to conceal more recent but possibly important short term trends it has been decided to also choose a set of five- year comparisons. Contrasting 1995-99 with 2010-14 would appear to work well with a close comparison of 85% squares covered against 81%.The periods 2000-04 and 2005-09 are much less suitable with only 63% and 67% of squares covered. However the increase of species per square from 10.5 in 1995-99 to 11.7 in 2010-14 might suggest that again the recording effort in the later period has exaggerated the expansion of some species.

The tables for individual species which follow must therefore be treated with caution. However any species which add or lose large percentages of squares on both long and short term comparisons might be reliably accepted as showing real signs of improvement or decline.

The individual species reports which follow display the long term 15-year and the short term 5-year comparisons and use colour coding to highlight clear changes in distribution.

Distribution vs. abundance

We are aware that distribution - in how many places a species is found - as covered on these pages, is only one way of assessing how various species are faring. The other measure is abundance, i.e. how many are found. We hold data on this, but are unable to undertake an analysis at this time.