Skip to main content

Habitat availability alters the relative risk of a bovine tuberculosis breakdown in the aftermath of a commercial forest clearfell disturbance

Cite this dataset

Murphy, Kilian J et al. (2022). Habitat availability alters the relative risk of a bovine tuberculosis breakdown in the aftermath of a commercial forest clearfell disturbance [Dataset]. Dryad.


1. Human modification of landscapes and associated disturbances may facilitate the emergence and spread of zoonotic diseases. Policy-makers need better understanding of the link between anthropogenic disturbances and wildlife disease hosts at the interface of human society and the natural environment, e.g. agriculture, forestry and aquaculture. Empirical research is strongly needed for the control of novel zoonoses which might emerge, as well as the management of existing zoonoses with significant economic repercussions such as bovine tuberculosis (bTB).

2. We aimed to examine the link between ecological disturbance and relative bTB risk using Ireland as a case study. We analysed clearfell forestry operations and assessed bTB breakdowns within cattle farms across different spatio-temporal scales over multiple years, examining how ecological conditions may modulate this relationship using conditional logistic regression models.

3. We found a significant effect of the interaction between the extent of clearfell forestry removed and the extent of natural grassland and mixed forestry present on relative bTB risk. This interaction was dynamic, leading to an increase or decrease of the relative bTB risk depending on where (between 2 and 6 km from the farm) and when (between 0 and 36 months prior to the bTB outbreak) the clearfell operations occurred.

4. Our study provides empirical evidence of the link between mechanised forestry operations and fluctuating relative bTB risk in cattle farms, although the mechanism behind it is yet to be elucidated. Given our data, we hypothesise that wildlife hosts may abandon the area subjected to clearfell when disturbance is highest (during active operations and shortly afterward) but are subsequently attracted back to the site as they regenerate, potentially affecting the contact rates with livestock and thus, relative bTB risk.

5. Our analysis demonstrates that landscape modification is correlated with a change in relative bTB risk that is dynamic in time and space, allowing managers to understand the risk in landscape modification and inform policy accordingly. Landscape-level studies are necessary to unveil subtle ecological processes, shifting research and management efforts away from cattle herd-centric and toward macroecological surveillance of wildlife hosts and longitudinal assessment of bTB risk.


The relevant bTB testing and breakdown history data for all farms in Ireland was collected from the Department of Agriculture, Food and the Marine (DAFM) in collaboration with UCD Centre for Veterinary Epidemiological Risk Analysis (CVERA). Every cattle herd in Ireland has at least one herd test per year. All cattle present on the farm on the day of the test are tested, with the exception of calves aged under 6 weeks which were born on the farm. The test used is the single intradermal comparative tuberculin test (SICTT). The SICTT test is the main diagnostic test used in Ireland, though it is imperfect. The  sensitivity and specificity of the test is known to vary with estimates of 52.9-60.8% and 99.2-99.8% respectively from field trials in Ireland (Clegg et al, 2011).  The animals showing a positive reaction to the test are known as reactors - due to the imperfections of the SICCT test it is possible for false positives to be detected in herds showing only one reactor (see below for more details on how we took this bias into account during modelling).

Spatial and administrative data for these farms were also provided by DAFM-CVERA via the Land Parcel Information System (LPIS). All relevant spatial and administrative forestry data including private forestry, public forestry, clearfell sites and forest boundaries were collected from Coillte, Ireland's semi-state forest custodian. Coillte manage 57% (440,000 hectares) of Ireland's forestry, which is predominantly conifer high forest in clearfell rotations. We collected data on the date of each forest clearfell event, the administrative boundary of the forest and the ID of each forest subplot felled within the property.

We collected habitat data relevant to known bTB hosts in Ireland from the Corine Land Cover (2018) dataset (© European Union, Copernicus Land Monitoring Service , European Environment Agency). We removed all habitat classifications which did not meet our a priori expectation of influencing resource selection by bTB disease hosts in Ireland. We used mixed forestry, which is defined as coniferous tree species mixed with broad-leaved deciduous tree species with height > 5-7 m and canopy closure around 30 % (always > 15 %). We used natural grassland, which is defined as areas with herbaceous vegetation (maximum height is 150 cm and gramineous species are prevailing) covering at least 50 % of the surface. The term ”natural” indicates that vegetation is developed under a minimum human interference i.e not mowed, drained, irrigated, sown, fertilised or stimulated by chemicals, which might influence production of biomass. We also used broad-leaved forest, defined as vegetation formation composed principally of trees, including shrub and bush understorey, where broad-leaved species predominate. The predominant classifying parameter for this class is a crown cover density of > 30 % or a minimum 500 subjects/ha density, with broad-leaved trees representing > 75 % of the formation. The minimum tree height is 5 m.

These data then underwent geospatial analysis to create the databases needed for modelling. We computed a spatial buffer for each scale (ranging from 1km - 6km) and extracted the summary statistics which we would use as covariates in the models, such as the number of land parcels on the farm, the elevation and the area of habitats available in the buffer. See Supplementary Material 1 for further information on the geospatial data analysis and full list of data extracted from the buffers.

Usage notes

All necessary data to re-run the analysis included here.