Introduction: Pollinator Decline in the UK

Issue Addressed

  • Decline of pollinator populations across the UK
  • Bees, hoverflies, butterflies, and other insects are essential for:
    • Ecosystem stability
    • Pollination of food crops
    • Biodiversity functioning
  • Evidence shows declines linked to:
    • Habitat loss
    • Climate change
    • Pesticide use

Context and Motivation

  • Understanding how pollinator abundance and diversity change over time and space is crucial for:
    • Informing conservation policy
    • Protecting ecosystem services
  • This study uses real ecological monitoring data from the UK Pollinator Monitoring Scheme (UK PoMS)
  • We analyse:
    • Temporal trends (2017–2021)
    • Spatial patterns and hotspots
    • Differences across UK countries
    • Comparison of two monitoring methods (FIT Count and pan-traps)

Methodology Overview

Data Source & Preparation

  • Dataset: UK Pollinator Monitoring Scheme (UK PoMS)
    • 1km FIT Count data (2017–2021)
    • Countries: England, Scotland, Wales, Northern Ireland
  • Data cleaning included:
    • Handling missing values
    • Addressing Northern Ireland’s different coordinate system
  • Initial exploratory analysis performed using:
    • Summary statistics (mean, median, standard error)
    • Distribution checks and missingness assessment

Analytical Approach

  • Temporal trends analysed using line plots
    • Mean abundance per year
    • Confidence ribbons (± SE)
  • Hotspot detection:
    • Grouping by 1km grid squares
    • Calculating mean insect counts
  • Cross-country comparison:
    • Conducted using ANOVA
  • Method comparison (FIT Count vs Pan-trap):
    • t-tests and boxplots to assess differences
    • Evaluation of consistency across years and countries
Sample Counts per UK Country
country n
England 880
Northern Ireland 6
Scotland 605
Wales 444
Formatted Summary of Dataset Structure
Variable Type Missing Values Example Value
sample_id numeric 0 17277106
country character 0 Northern Ireland
location_code character 0 NI09
location_name character 551 near West Itchenor
x1km_square character 0 J3565
sample_projection character 0 OSIE
land_cover character 0 semi-natural
date POSIXct 0 2021-09-11
year numeric 0 2021
digitised_by numeric 0 207506
recorder_type character 30 I am confident in identifying the commonly-occurring pollinating insects to species level
habitat character 0 Garden
habitat_other_detail character 1719 rough grassland with thistle patches
habitat_type character 0 garden
target_flower character 0 Other - please describe below
target_flower_corrected character 0 Other - please describe below
target_other_name character 761 rose
target_other_name_corrected character 758 rose
target_flower_family character 6 Rosaceae
flower_structure character 2 open
flower_cover character 0 Target flowers cover less than half of patch
floral_unit_count numeric 5 6
floral_unit character 5 individual flower
flower_context character 0 Growing in a larger patch of the same flower
count_start_time POSIXct 0 1899-12-31 10:28:00
cloud_cover character 0 All or mostly cloud
sunshine character 0 Entirely shaded
wind_speed character 0 Leaves still/moving occasionally
bumblebees numeric 0 0
honeybees numeric 0 0
solitary_bees numeric 0 0
wasps numeric 0 0
hoverflies numeric 0 2
other_flies numeric 0 12
butterflies_moths numeric 0 0
beetles numeric 0 0
insects_small numeric 0 0
insects_other numeric 0 0
all_insects_total numeric 0 14 

## [1] 6
## [1] 6
## [1] 3
Projection Systems Used
OSGB OSIE
England 880 0
Northern Ireland 0 6
Scotland 605 0
Wales 444 0 
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   0.000   3.000   9.500   7.667  11.500  14.000
Key Considerations When Working with Northern Ireland FIT Count Data
Issue Details
Different coordinate system Northern Ireland uses OSIE projection, while other UK countries use OSGB. Projection(s): OSIE
Missing location names 6 out of 6 NI records have missing or ‘NA’ location names.
Small sample size Only 6 samples compared with much larger sample sizes in England, Scotland and Wales.
Potential data inconsistencies Some NI fields contain text ‘NA’ instead of true missing values, requiring cleaning.

Dataset Overview (Question 1)

Distribution of Pollinator Counts by Country

Sample Counts per UK Country

  • England: 880 samples
  • Scotland: 605 samples
  • Wales: 444 samples
  • Northern Ireland: 6 samples

Key Considerations for Northern Ireland

  • Uses OSIE / Irish Grid coordinate system
  • Missing or incomplete location names
  • Very small sample size
  • Requires extra cleaning for mapping and UK-wide comparisons
Annual Pollinator Count Statistics for England
year n_samples mean_insects median_insects sd_insects se_insects q1_insects q3_insects
2017 169 20.31 9 32.50 2.50 5 20
2018 175 12.46 10 12.21 0.92 4 17
2019 182 18.24 12 30.12 2.23 5 21
2020 105 14.31 11 15.61 1.52 6 18
2021 249 12.02 8 19.43 1.23 4 14

Trends in Pollinator Counts (England, 2017–2021)

Pollinator Hotspots in England (Question 2.2)

(Question 2.2)

Top 5 pollinator hotspots

Top 20 Hotspots

Grid Square Mean n
ST9351 38.9 18
TR3149 33.8 6
SU3986 32.0 23
TR0534 26.0 21
SU1466 25.6 18
TG4403 24.2 12
SE1845 21.6 36
SO4718 21.4 26
SO6087 20.5 31
SK9826 19.3 24
SO8911 17.7 28
SX7048 16.8 13
SP4951 16.5 33
ST4543 15.4 29
TQ1827 15.4 19
SX8144 15.1 16
SK4706 14.9 10
SP1430 14.9 19
SU7800 14.4 30
SE1582 14.2 22

Key Insights:

  • Hotspots are 1 km squares with consistently high pollinator counts
  • Likely contain good floral resources and diverse habitats (hedgerows, meadows, flower-rich farmland)
  • Priority areas for conservation efforts and habitat management research
  • Comparison with low-count areas helps identify key drivers of pollinator abundance

Dominant pollinator group in the top 20 English hotspots (1km squares)
x1km_square mean_insects n_samples group mean_count
ST9351 38.94 18 Other_flies 5.11
TR3149 33.83 6 Other_flies 2.00
SU3986 31.96 23 Other_flies 3.61
TR0534 25.95 21 Other_flies 7.95
SU1466 25.56 18 Other_flies 6.50
TG4403 24.25 12 Other_flies 12.17
SE1845 21.56 36 Other_flies 9.61
SO4718 21.38 26 Hoverflies 2.35
SO6087 20.55 31 Hoverflies 6.42
SK9826 19.29 24 Other_flies 3.96
SO8911 17.71 28 Other_flies 6.11
SX7048 16.85 13 Other_flies 3.46
SP4951 16.45 33 Other_flies 4.76
ST4543 15.45 29 Other_flies 5.72
TQ1827 15.42 19 Hoverflies 3.74
SX8144 15.12 16 Other_flies 8.12
SK4706 14.90 10 Hoverflies 4.20
SP1430 14.89 19 Other_flies 4.21
SU7800 14.37 30 Hoverflies 2.77
SE1582 14.23 22 Other_flies 4.00

Composition of Insect Groups (Question 2.3)

Total Pollinator Group Counts

Pollinator Group Total
Other flies 3,414
Hoverflies 2,255
Bumblebees 1,007
Beetles 642
Honeybees 611
Solitary bees 467
Wasps 368
Butterflies/moths 307

Dominant Groups by Square

Group N %
Other flies 24 64.9
Hoverflies 12 32.4
Bumblebees 1 2.7

Dominant group distribution across 1km squares (2017–2021)

Cross-Country Comparison of Pollinator Abundance

(Question 3.1)

Mean pollinator abundance by country

Summary Statistics

Country n Mean SE
England 880 15.3 0.80
Wales 444 9.2 0.46
Northern Ireland 6 7.7 2.38
Scotland 605 7.3 0.39

ANOVA Results:

  • F-statistic: 27.637 | p-value: < 1e-04
  • Conclusion: Significant differences between countries (p < 0.001)
  • England shows highest abundance (15.3 insects/sample), ~2× higher than other countries

Temporal Trends in Pollinator Counts Across UK Countries

Comparing Sampling Methods: FIT Count vs Pan-trap (Question 4)

Summary Statistics by Method

Method N Mean Median SE
FIT Count 1935 11.4 7 0.4
Pan-trap 4186 80.4 45 2.6

T-test Results

Statistic Value
t-statistic -26.594
p-value 0.000
Mean Difference 69.000
95% CI Lower -74.090
95% CI Upper -63.920

Interpretation: Pan-traps capture significantly more insects than FIT Counts (p < 0.001), with a mean difference of ~19 insects per sample.

Key Findings: Pan-traps show higher median counts and greater variability. Both methods capture similar range of low counts, but pan-traps detect more high-abundance sites.

4.4 + part I r## Country Comparison: FIT Count vs Pan-trap

(Question 4.4)

Mean counts by country and method

Summary Statistics

Country Method n Mean SE
England FIT Count 880 15.3 0.80
England Pan-trap 2163 92.4 3.27
Northern Ireland FIT Count 6 7.7 2.38
Northern Ireland Pan-trap 30 98.9 9.56
Scotland FIT Count 605 7.3 0.39
Scotland Pan-trap 1061 48.8 1.90
Wales FIT Count 444 9.2 0.46
Wales Pan-trap 932 87.8 8.28

Key Insights:

  • England shows highest counts for both methods, particularly with FIT Count
  • Method consistency varies by country—some show similar results, others differ substantially
  • Pan-trap method generally produces lower but more consistent counts across countries
  • Sample sizes vary considerably between countries and methods

Insect Group Composition: Method Comparison (Part I)

(Question 4.5 - Total Counts)

Total counts by insect group and method

Composition Table

Method Group Count %
FIT Count Other Flies 6,285 40.2
FIT Count Hoverflies 4,346 27.8
FIT Count Bees 3,041 19.5
FIT Count Beetles 956 6.1
FIT Count Wasps 529 3.4
FIT Count Butterflies & Moths 461 3.0
Pan-trap Beetles 30,129 53.3
Pan-trap Wasps 9,145 16.2
Pan-trap Hoverflies 8,837 15.6
Pan-trap Bees 7,330 13.0
Pan-trap Butterflies & Moths 1,048 1.9
Pan-trap Other Flies 0 0.0

Key Insights:

  • Other Flies dominate FIT Count but absent from Pan-trap classification
  • Bees and Hoverflies captured in substantial numbers by both methods
  • FIT Count generally records higher absolute counts across most groups
  • Different taxonomic resolution between methods affects group comparisons

Method Comparison: Composition & Temporal Trends

Conclusions

Key Findings

  1. Temporal Trends (England): Pollinator counts show volatility with a notable spike in 2019 followed by decline, suggesting declining/stabilizing populations.

  2. Hotspots: Top pollinator hotspots identified in England, likely containing rich floral resources and diverse habitats—priority areas for conservation.

  3. Composition: “Other flies” and Hoverflies dominate total counts in England. Other flies and Hoverflies are the dominant groups across 1km squares (65.4% and 32.4% respectively).

  4. Cross-Country Patterns: England shows significantly higher pollinator abundance (~15.3 insects/sample) compared to other UK countries (~7-9 insects/sample) (p < 0.001).

  5. Method Comparison: Pan-traps capture significantly more insects than FIT Counts (mean difference ~19 insects, p < 0.001), with similar temporal patterns but different taxonomic biases.

Implications for Conservation

  • Spatial targeting: Focus conservation efforts on identified hotspots
  • Monitoring strategy: Both methods valuable—FIT Counts for trend monitoring, Pan-traps for comprehensive surveys
  • Regional differences: England’s higher counts may reflect habitat quality, sampling effort, or landscape differences
  • Declining trends: Urgent need for habitat protection and restoration across UK