Climate and environmental change drives Ixodes ricinus geographical expansion at the northern range margin

Abstract

Background: Global environmental change is causing spatial and temporal shifts in the distribution of species and the associated diseases of humans, domesticated animals and wildlife. In the on-going debate on the influence of climate change on vectors and vector-borne diseases, there is a lack of a comprehensive interdisciplinary multi-factorial approach utilizing high quality spatial and temporal data.

Methods: We explored biotic and abiotic factors associated with the latitudinal and altitudinal shifts in the distribution of Ixodes ricinus observed during the last three decades in Norway using antibodies against Anaplasma phagocytophilum in sheep as indicators for tick presence. Samples obtained from 2963 sheep from 90 farms in 3 ecologically different districts during 1978 - 2008 were analysed. We modelled the presence of antibodies against A. phagocytophilum to climatic-, environmental and demographic variables, and abundance of wild cervids and domestic animals, using mixed effect logistic regressions.

Results: Significant predictors were large diurnal fluctuations in ground surface temperature, spring precipitation, duration of snow cover, abundance of red deer and farm animals and bush encroachment/ecotones. The length of the growth season, mean temperature and the abundance of roe deer were not significant in the model.

Conclusions: Our results highlight the need to consider climatic variables year-round to disentangle important seasonal variation, climatic threshold changes, climate variability and to consider the broader environmental change, including abiotic and biotic factors. The results offer novel insight in how tick and tick-borne disease distribution might be modified by future climate and environmental change.

Figure 1

Geographic distribution of farms which were positive and negative for antibodies against A. phagocytophilum during timespan 1, timespan 2 and timespan 3 in the three regions (INLAND, COAST and FJORD). A positive farm is defined as a farm with one or more positive samples. A negative farm has no positive samples.

Figure 2

Changes over time in the specific climate variables, which was significantly associated with the outcome in the multivariable model. (A) INLAND, (B) COAST and (C) FJORD. Farm denotes the variables significant at farm level, whilst pasture denotes variables significant at rough grazing level. Changes are shown for timespan 2 (1990-1999) and timespan 3 (2000-2008) relative (in %) to timespan 1 (1980-1989). RRSum_May : Precipitation in May; SnoStartDays: Number of days from 1 September - 31 August to snow depth ≥2 cm, TDecr_÷5<Days_Jan-Dec :  Number of days per month where temperature decrease in Ground Surface Temperature (GST) from a day to the next day are >5 °C; RRSum_Mar : Precipitation in March; and BlackFrdays: Number of days from 1 September – 31 August with black frost, daily GST < 0°C and ground bare of snow or now depth < 2 cm.

Figure 3

Prediction of the presence of A. phagocytophilum at farm level using the final model. The colour black denotes timespan 1, red timespan 2 and green timespan 3. The black line is the regression line of observed versus predicted presence with an intercept of −0.013 and a slope of 1.023 which give an adjusted R2 of 0.80.