Monitoring of breeding birds provides insight into age-specific bird-tick interactions

Description

This is the data and scripts associated with the paper: Monitoring of breeding birds provides insight into age-specific bird-tick interactions which is currently in review at Ticks and Tick-borne Diseases. Abstract: Ixodes ricinus (Family Ixodidae), a widespread generalist tick in Europe, feeds on a large number of host species and acts as a major vector for various tick-borne pathogens. Immature stages of I. ricinus are often found feeding on various bird species, but questions remain open regarding ecological interactions between these organisms. Utilizing ongoing monitoring schemes during the breeding season in northern Germany, we studied the occurrence of ticks on birds caught in two independent forest areas during three subsequent seasons (2022-2024) to study relationships between tick infestation and bird-specific traits (i.e., age, body condition). Tick infestation differed significantly between bird species and we observed significant interactions with bird age and body condition. Adult birds showed lower tick burdens in comparison to juvenile conspecifics and tick infestation showed complex interactions with body condition; highlighting multiple mechanisms are most likely influencing our results. Through morphological identification of attached ticks, we could show that I. ricinus larva and nymphs differ in their interactions with bird traits, with only the number of larvae differing with bird age and neither life-stage alone correlating with bird body condition. Taken together, we supported life-stage specific infestation as a potential driver of higher tick burdens in juvenile birds and show how standardized monitoring can provide insight into these interactions.

Steps to reproduce

Data analysis All analysis was performed in R v4.5.1 (R Core Team, 2024). To assess individual bird body condition, we calculated the scaled mass index (SMI) as described by Peig and Green, (2009). To be able to calculate body condition for all years we used wing length as our standard body measurement. The scaled mass index was calculated individually per bird species. We restricted the analysis to the 11 most common bird species (accounting for >85% of the full dataset. To analyze the impacts of various factors on tick infestation we fitted generalized linear mixed effects models (GLMM) using the function glmer from the R-package lme4 (Bates et al., 2015). We ran two models, (i) a binomial model to determine effects of a bird to be infested (1) or non-infested (0) (i.e., infestation probability); and (ii) a negative binomial model on tick burden of all birds to assess the effects on the absolute number of ticks (i.e., tick burden). Further, two negative binomial models were run to predict the number of I. ricinus larva and nymphs, respectively, on the infested birds. For all models, the fixed and random effect structure was identical including both a linear and quadratic SMI term, bird age (EURING codes 3, 5, 6), sampling location, sampling year, and day of capture (Julian day) as fixed effects. The life-stage model additionally included tick-burden as a fixed effect. All continuous variables were mean-centered using the scale function prior to running the models. Bird species and ring number (i.e., individual) were included as random effects in both models. Due to the interaction with bird age categories in the overall model, we further aimed to test if this could be due to among- or within-individual changes as described (van de Pol and Verhulst, 2006). For this, age categories were translated into years for all re-caught individuals where possible. In this case, birds caught as juveniles (EURING Code 3) were coded as 0 years old in that year of capture and adult birds aged to the previous year (EURING Code 5) were coded as 1 year old in that year of capture. All following observations were determined based on the number of years passed since the last known age. For age, we calculated a mean age (average across all observations) per individual and also a delta age (deviation from mean age) for a specific observation. The model assumed a negative binomial error distribution and included mean age, delta age, and year as fixed effects. Ring number and species were included as random effects. All models were checked for overdispersion using a parametric Pearson residuals test and inclusion of terms was checked through reduction in AIC and BIC values, with terms decreasing AIC/BIC being retained. Additionally, mean estimates and their 95% credible intervals (CIs) for all models were estimated based on 5000 simulations using the function sim from the package arm (Gelman and Su, 2024).

Categories

Funders

• Deutsche Forschungsgemeinschaft

  North Rhine-Westphalia, Bonn

  Grant ID: SFB 1372 - Nav05 (Project #: 395940726)

Licence