Modeling the vector-borne disease transmission potential in northern Europe with a special emphasis on microclimatic temperature

Abstract

Denmark experienced bluetongue outbreaks in 2007 and 2008 even though the Danish climate was previously considered to be too cold for the transmission of a number of vector-borne diseases (VBD), including the disease caused by bluetongue virus (BTV). BTV causes weight loss, abortion, reduced milk yield and ultimately death in ruminants (cattle and sheep), and is responsible for restricted trade of animals and animal products internationally. Schmallenberg is an emerging Culicoides-borne viral disease that can affect cattle, sheep and goats, and causes abortion and stillbirth with congenital malformation. Schmallenberg virus was detected in Denmark in 2012. Epidemiological modeling plays an important role in policy making and strategies for the prevention and control of VBD. Deciding what approach to take requires models able to generate accurate predictions, which in turn relies on an accurate knowledge of the natural lifecycle of vectors, including the daily survival probability, biting rate, extrinsic incubation period (EIP) of the virus, and vector competence. All these parameters are dependent on the temperature to which the insect vectors are exposed. Mathematical models used to quantify VBD transmission parameters are mostly driven by temperatures recorded by meteorological institutes. Insects are poikilothermic and they experience temperatures at a very local level, down to a scale of meters or even centimeters. Models that use meteorological temperatures do not take into account the temperature to which the insects are exposed in the microhabitat. Weather stations may be located as far as 50–100 km from the insect resting sites, and the lack of data from microclimatic environments hinders the use of microclimatic temperatures in disease modeling. The aim of this thesis was to improve our understanding of the transmission potential of VBD in northern Europe. We conducted four studies and systematically studied the microclimatic temperature of potential insect habitats at different heights for an entire warm season (May to October). We also systematically studied the role of micro climatic temperature on VBD transmission potential, as well as the variation and uncertainty associated with bluetongue virus transmission in Denmark. Finally, we studied the potential role of temperature on the occurrence of Setaria tundra outbreaks (a mosquito-borne filarioid disease) and on the intensity of worm transmission to reindeer in general. In the first step, we recorded the hourly temperature in six microclimatic habitats (dry meadow, wet meadow, hedges, trees, and cattle and horse fields) in Strødam and Faxe in Denmark from May to October, 2015. We did this by setting temperature data loggers in triplicate and at 2 or 3 different heights in each habitat. We also collected information on the hourly meteorological temperature, solar radiation, wind speed, humidity, and precipitation from the Danish Meteorological Institute (DMI)’s weather stations in the same areas. We performed multiple linear regressions to predict the microclimatic temperature of different habitats based on the parameters collected from meteorological weather stations. Finally, we estimated the development time of six vector-borne pathogens (bluetongue, Schmallenberg virus, West Nile virus, Dirofilaria, malaria and dengue) that were either reported in Denmark or likely to be introduced to the country, by developing ratesummation models for each pathogen. In the second step, we quantified the different land cover classes within a 500 m radius of all cattle farms in Denmark (N=22,004) using CORINE land cover, and regrouped them into four major land cover types: dry meadow, wet meadow, hedges, and forest. We then obtained the meteorological temperatures and other parameters (solar radiation, wind speed, humidity) near the farm between 2000 and 2016 from DMI. We calculated the hourly microclimatic temperatures of each farm based on their surrounding habitat types and meteorological parameters using our microclimatic temperature prediction models for those four major land cover types. We then modeled the daily EIP of Schmallenberg virus for each farm for the period 2000-2016 using both hourly DMI and hourly microclimatic temperatures, and calculated the mean EIP over the 17 years for each farm. In the third step, we used the microclimatic temperature data for all Danish cattle farms in a mechanistic transmission model (biological process model) to estimate the daily number of infectious bites (IB) with BTV using Culicoides midge abundance data based on 1,453 light trap collections from 2007-2016. We used published equations for the EIP, daily vector survival rate, daily vector biting rate, and host-to-vector transmission rate, and quantified the variation and sensitivity of each parameter associated with daily IB estimates.