Climate and tree seed production predict the abundance of the European Lyme disease vector over a 15-year period

Abstract

Background: To predict the risk of tick-borne disease, it is critical to understand the ecological factors that determine the abundance of ticks. In Europe, the sheep tick (Ixodes ricinus) transmits a number of important diseases including Lyme borreliosis. The aim of this long-term study was to determine the abiotic and biotic factors driving the annual abundance of I. ricinus at a location in Switzerland where Lyme borreliosis is endemic.

Methods: Over a 15-year period (2004 to 2018), we monitored the abundance of I. ricinus ticks on a monthly basis at three different elevations on Chaumont Mountain in Neuchâtel, Switzerland. We collected climate variables in the field and from nearby weather stations. We obtained data on beech tree seed production from the literature, as the abundance of Ixodes nymphs can increase dramatically two years after a masting event. We used AIC-based model selection to determine which ecological variables drive annual variation in tick density.

Results: We found that elevation site, year, seed production by beech trees two years prior, and mean annual relative humidity together explained 73.2% of the variation in our annual estimates of nymph density. According to the parameter estimates of our models, (i) the annual density of nymphs almost doubled over the 15-year study period, (ii) changing the beech tree seed production index from very poor mast (1) to full mast (5) increased the abundance of nymphs by 86.2% two years later, and (iii) increasing the field-collected mean annual relative humidity from 50.0 to 75.0% decreased the abundance of nymphs by 46.4% in the same year. Climate variables collected in the field were better predictors of tick abundance than those from nearby weather stations indicating the importance of the microhabitat.

Conclusions: From a public health perspective, the increase in nymph abundance is likely to have increased the risk of tick-borne disease in this region of Switzerland. Public health officials in Europe should be aware that seed production by deciduous trees is a critical driver of the abundance of I. ricinus, and hence the risk of tick-borne disease.

Keywords: Beech tree; Climate change; Fagus sylvatica; Ixodes ricinus; Lyme disease; Mast years; Tick population ecology; Tick-borne disease.

Fig. 1

Annual variation in log₁₀-transformed cumulative nymphal density (CND) and beech tree mast score 2 years prior. The log₁₀-transformed CND (dots and solid lines) and the beech tree mast score 2 years prior (grey bars) are shown over time for each of the three elevation sites (low, medium, high) on Chaumont Mountain. The log₁₀-transformed CND increased significantly over the 15-year study period (2004–2018). Years of high seed production by beech trees (beech masting index) are strongly positively correlated with high log₁₀-transformed CND two years later. The CND is an estimate of the annual abundance of I. ricinus nymphs per 100 m² and is calculated by integrating the area under the curve of the 12 monthly estimates of the number of questing nymphs collected by dragging an area of 100 m². Beech tree mast scores range from 1 to 5 (1, very poor mast; 2, poor mast; 3, moderate mast; 4, good mast; and 5, full mast)

Fig. 2

Effect of elevation on the log₁₀-transformed cumulative nymphal density (CND). According to the model parameter estimates, the CND (on the original scale) at the low elevation was 11.4% higher than the medium elevation and 43.1% higher than the high elevation (partial r² = 26.6%). The parameter estimates used to calculate the effect sizes were taken from the best model in Table 1, which had 76.0% of the support and explained 73.2% of the inter-annual variation in the log₁₀-transformed CND

Fig. 3

Effect of year on the log₁₀-transformed cumulative nymphal density (CND). According to the model parameter estimates, the CND (on the original scale) increased by 88.4% over the 15-year study period (partial r² = 14.8%). The parameter estimates used to calculate the effect sizes were taken from the best model in Table 1, which had 76.0% of the support and explained 73.2% of the inter-annual variation in the log₁₀-transformed CND

Fig. 4

Effect of beech mast score 2 years prior on the log₁₀-transformed cumulative nymphal density (CND). According to the model parameter estimates, increasing the beech mast score from 1 (poor mast) to 5 (full mast) increased the CND (on the original scale) by 86.2% (partial r² = 26.9%). The parameter estimates used to calculate the effect sizes were taken from the best model in Table 1, which had 76.0% of the support and explained 73.2% of the inter-annual variation in the log₁₀-transformed CND

Fig. 5

Effect of the field-collected mean annual relative humidity on the log₁₀-transformed cumulative nymphal density (CND). According to the model parameter estimates, increasing the field-collected mean annual relative humidity from 50.0% to 75.0% decreased the CND from the same year (on the original scale) by 46.4% (partial r² = 7.6%). The mean annual relative humidity was calculated for the same year as the CND (i.e. no time lag). The parameter estimates used to calculate the effect sizes were taken from the best model in Table 1, which had 76.0% of the support and explained 73.2% of the inter-annual variation in the log₁₀-transformed CND