A spatially explicit degree-day model of Rift Valley fever transmission risk in the continental United States

Abstract

A spatially explicit degree-day model was used to evaluate the risk of Rift Valley fever virus (RVFV) transmission by mosquitoes to humans and livestock within five target states in the continental United States: California, Minnesota, Nebraska, New York, and Texas. A geographic information system was used to model potential virus transmission based on a 12-day moving window assessment of the extrinsic incubation period theorized for RVFV in the United States. Risk of potential virus transmission in each state was spatially evaluated on a 10-km grid using average historical daily temperature data from 1994 to 2003. The highest levels of transmission risk occur in California and Texas, with parts of these states at risk of RVFV transmission for up to 8 months per year. Northern Minnesota, central New York, and most of coastal and high-elevation California are at low to null risk. Risk of impact to the livestock industry is greatest in California, Texas, and Nebraska. A standard global climate model was used to evaluate future risk in the year 2030 in Nebraska, and showed an increase of transmission risk days from approximately 3 to 4 months per year.

Cautioning the use of degree‐day models for climate change projections in the presence of parametric uncertainty

“…Yang et al (2006, 2010), Zhou et al (2008), Craig et al (1999), Konrad et al (2011)). Our purpose is to investigate the influence of parametric uncertainty on degree-day model output using techniques commonly applied, and thus we did not further validate model output.…”

“…However, even when variance estimates are available from previous work, parameter point estimates are often borrowed directly from the literature for use in degree-day model applications, with the influence of parametric uncertainty remaining unexplored (e.g. Yang et al (2006, 2010), Zou et al (2007), Konrad et al (2011)). …”

“…These confidence intervals are often sizable (Table 2), and such variance may have a substantial impact on predictions of organism emergence times or in determining the suitability of specific regions for organism establishment or persistence. However, even when variance estimates are available from previous work, point estimates are often used, with the influence of parametric uncertainty remaining unexplored (e.g., Yang et al 2006, 2010, Zou et al 2007, Konrad et al 2011.…”

Application of a degree-day model of West Nile virus transmission risk to the East Coast of the United States of America

“…Previously, we developed a geographical information system (GIS)-based model, which summed degreeday (DD) temperatures over the duration of the vector feeding period to spatially and temporally evaluate the risk of mosquitoes transmitting arboviruses (Zou et al, 2007;Konrad et al, 2009Konrad et al, , 2011. The model was initially created to work with WNV transmitted by Culex tarsalis mosquitoes, but can be adapted for other arboviruses and mosquito species by changing a few model parameters.…”

A geographical information system-based web model of arbovirus transmission risk in the continental United States of America

“…Rift Valley fever virus (RVFV) as show by Konrad et al (2011) and Konrad and Miller (2012a). Since RVFV has not appeared in USA, field calibration with North American mosquitoes is not currently possible.…”