Woodland expansion and deer management shape tick abundance and Lyme disease hazard
Data files
Nov 25, 2024 version files 301.09 KB
-
Dataset_archive.xlsx
289.62 KB
-
README.md
11.47 KB
Abstract
The largest land use change in Europe is woodland expansion, through planting and natural regeneration. Unforeseen consequences of this could include changes in environmental hazards, such as exposure to parasites and pathogens.
Tick-borne Lyme disease is the most prevalent tick-borne disease in the northern hemisphere and is often associated with woodlands. Therefore, to inform the planning and management of expanding woodlands, we test how land covers that reflect different types and stages of the woodland expansion process, along with their deer management, impact tick densities and Lyme disease hazard (density of infected nymphs). We also test whether differences in rodent abundance play a mechanistic role in explaining differences in Lyme disease hazard.
In Northwest Scotland, a touristic area undergoing woodland expansion, we recorded deer management, rodent densities, Ixodes ricinus nymph densities, pathogen prevalence and Lyme disease hazard between: open moorland, young pine and mature pine, birch and spruce. These represent pre-, early and late stage woodland establishment, and the three woodland types in the region.
Rodents, ticks, pathogen prevalence and Lyme disease hazard were generally lowest in moorland and young pine and highest in mature woodland, especially birch, although variability was high. Deer management reduced tick densities and, marginally, Lyme disease hazard. There was insufficient evidence for rodents increasing Lyme disease hazard, but rodents augmented tick densities and the most abundant Lyme disease pathogen was that transmitted by rodents.
Practical implication:
Woodland expansion could, once mature, eventually lead to higher tick densities, pathogen prevalence and Lyme disease hazard. Importantly, an environmental solution could be to control deer populations.
https://doi.org/10.5061/dryad.jh9w0vtn0
Description of the data and file structure
You can open the .xls file in microsoft Excel. The first tab includes a key for all the data
so you can see what each column is. Tab number 2 is the dataset used for the tick density and nymphal infection prevalence models and the third
tab is the dataset used the density of infected nymphs, adding a value of one in a pine habitat so the models could run (see methods for details).
If you need the R script, please let me know.
Sara
Files and variables
Cells with “NA” in the dataset are cells for which no data were collected/available. For instance, NAs in the column “rodentsper100trapnights19” indicates that rodent trapping was not done for this specific site/year. NAs in the prevalence column indicates that no ticks were tested as fewer than 50 nymphs were collected.
File: Dataset_archive.xlsx
Tick density_prevalence tab | |
---|---|
Column name | what is it? |
year | Year |
session | Session (spring, mid-summer, summer etc.) |
plot | Site identification - BM corresponds to mature birch, H to heather, PM to mature pine, PY to young pine and SM to mature spruce. The numbers corresponds to the site number within this landcover type. |
dragnr | Transect number |
landcover | Landcover for the site |
DON_drag | Number of Ixodes ricinus nymphs counted (transect level) |
totalnymphswithextrasmin50_plot | Total number of nymphs tested for B. burgdorferi s.l. at the plot/year level |
totalBinfectednymphswithextrasmin50_plot | Total number of nymphs infected with B. burgdorferi s.l. at the plot/year level |
totalBuninfectednymphswithextrasmin50_plot | Total number of nymphs uninfected with B. burgdorferi s.l. at the plot/year level |
B.afzelii_plot | Number of B. afzelii positive nymphs |
B.garinii_plot | Number of B. garinii positive nymphs |
B.sensustricto_plot | Number of B. burgdorferi s.s. positive nymphs |
B.valaisiana_plot | Number of B. valaisiana positive nymphs |
prevalenceB_plot | Prevalence of B. burgdorferi s.l. (plot/year level) |
DINBper1000plot | Density of infected nymphs per 1000 m² |
DINBper1000intplot | Density of infected nymphs per 1000 m² rounded to nearest integer |
DINBArea1000plot | Area surveyed for offset |
deermgmt | Deer management strategy - none: no deer management at the site level, managed: deer management at the site level (fencing, culling, hunting) |
vegdens | average vegetation density at the transect level |
vegdens_plot | average vegetation density at the plot/year/visit level |
trapnights19 | Number of trap nights for rodents in 2019 |
newrodents19 | Number of individual rodents trapped in 2019 |
rodentsper100trapnights19 | rodents per 100 trap nights in 2019 |
DIN tab | |||
---|---|---|---|
Column name | what is it? | ||
year | Year | ||
session | Session (spring, mid-summer, summer etc.) | ||
plot | Site identification - BM corresponds to mature birch, H to heather, PM to mature pine, PY to young pine and SM to mature spruce. The numbers corresponds to the site number within this landcover type. | ||
landcover | Landcover for the site | ||
DINB1000m2_session | Density of infected nymphs per 1000 m² at the plot/session/year level | ||
DIN1000int | Density of infected nymphs per 1000 m² at the plot/session/year level rounded to nearest integer | ||
DINArea1000 | Offset for the density of infected nymphs per 1000 m² at the plot/session/year level | ||
DINBper1000plot | Density of infected nymphs per 1000 m² at the plot/year level | ||
DINBper1000intplot | Density of infected nymphs per 1000 m² at the plot/year level rounded to nearest integer | ||
DINBArea1000plot | Offset for the density of infected nymphs per 1000 m² at the plot/year level | ||
vegdens_plot | Vegetation density at the plot/session/year level | ||
rodentsper100trapnights19 | Individual rodents per 100 trap nights in 2019 | ||
deer_mgmt | Deer management (none, managed) |
Code/software
Microsoft Excel
We gathered data on deer management, rodent abundance, questing I. ricinus nymph density, B. burgdorferi s.l prevalence and Lyme disease hazard in Wester Ross, northwest Scotland 2018 - 2020 in plots over five land cover types: (1) heather/grass moorland (dominated by Erica and Calluna species); (2) young Scots pine aged 13 to 30 years old; and (3) mature (at least 60 years old) Scots pine woodland; (4) deciduous birch woodland (Betula spp.) and (5) commercial Sitka spruce (Picea sitchensis) plantations.
Information on deer management (culling, exclusion fencing, or both) for all 40 plots (8 plots from each land cover type) was gained from the landowner or manager.
We collected data on Ixodes ricinus nymph density at all plots in 2018 by the standard blanket drag transect method: a 1m x 1m square of blanket was dragged along the ground for 10m and all ticks counted, collected and stored in 1.5ml Eppendorf tubes containing 70% ethanol to await pathogen identification. Each 10 m2 transect was separated by at least 20 m and we adopted a stratified random sampling strategy to ensure representation of different parts of each woodland. We conducted 20 transects per plot per visit, with three visits per plot, in May, late June/early July and August. We conducted further tick collections in 2020 for pathogen prevalence estimation so we could link rodents, which were captured in 2019, to the pathogen prevalence in nymphs the following year.
Along the same 10 m x 1 m tick survey blanket drag transects, we measured the height-density index of the ground vegetation using a sward stick with coloured bands every 5 cm. We counted the number of bands hidden by the vegetation at the start, middle and end of each transect.
Ticks were analysed for infection with Borrelia burgdorferi sensu lato as follows: The DNA was extracted from each individual nymph (n=6463 nymphs) using ammonium hydroxide (Wielinga et al., 2006). Each DNA extract was tested for the presence of B. burgdorferi s.l. using qPCR (Heylen et al., 2013). To identify which genospecies of B. burgdorferi s.l. was present, positive samples were subject to conventional PCR of the intergenic spacer region and then Sanger sequenced (Coipan et al., 2013). We used positive controls and negative water controls on every plate.
Rodent density index (number of new captures per trap-night) was estimated at 20 of the 40 plots (four replicates of each land cover type) in August 2019 using 50 Longworth live-traps per plot over two nights. Traps had holes that allow shrews, but not voles or mice, to escape. Traps were deployed in a 40 m x 40 m grid, with two traps at each 10 m point in the grid.
Statistical analysis:
We performed all statistical analyses in R (R Core Team, 2020) using the glmmTMB, , emmeans packages to run Poisson and negative binomial GLMMs and Kruskall-Wallace tests. We used Akaike Information Criterium (AIC) comparisons (smaller is better fit) for model selection using the dredge function in the MuMIn package, and Tukey post-hoc tests to understand the nature of multiple comparisons.