Weather Regulates Location, Timing, and Intensity of Dengue Virus Transmission between Humans and Mosquitoes

Abstract

BackgroundDengue is one of the most aggressively expanding mosquito-transmitted viruses. The human burden approaches 400 million infections annually. Complex transmission dynamics pose challenges for predicting location, timing, and magnitude of risk; thus, models are needed to guide prevention strategies and policy development locally and globally. Weather regulates transmission-potential via its effects on vector dynamics. An important gap in understanding risk and roadblock in model development is an empirical perspective clarifying how weather impacts transmission in diverse ecological settings. We sought to determine if location, timing, and potential-intensity of transmission are systematically defined by weather.Methodology/Principal FindingsWe developed a high-resolution empirical profile of the local weather-disease connection across Peru, a country with considerable ecological diversity. Applying 2-dimensional weather-space that pairs temperature versus humidity, we mapped local transmission-potential in weather-space by week during 1994-2012. A binary classification-tree was developed to test whether weather data could classify 1828 Peruvian districts as positive/negative for transmission and into ranks of transmission-potential with respect to observed disease. We show that transmission-potential is regulated by temperature-humidity coupling, enabling epidemics in a limited area of weather-space. Duration within a specific temperature range defines transmission-potential that is amplified exponentially in higher humidity. Dengue-positive districts were identified by mean temperature >22°C for 7+ weeks and minimum temperature >14°C for 33+ weeks annually with 95% sensitivity and specificity. In elevated-risk locations, seasonal peak-incidence occurred when mean temperature was 26-29°C, coincident with humidity at its local maximum; highest incidence when humidity >80%. We profile transmission-potential in weather-space for temperature-humidity ranging 0-38°C and 5-100% at 1°C x 2% resolution.Conclusions/SignificanceLocal duration in limited areas of temperature-humidity weather-space identifies potential locations, timing, and magnitude of transmission. The weather-space profile of transmission-potential provides needed data that define a systematic and highly-sensitive weather-disease connection, demonstrating separate but coupled roles of temperature and humidity. New insights regarding natural regulation of human-mosquito transmission across diverse ecological settings advance our understanding of risk locally and globally for dengue and other mosquito-borne diseases and support advances in public health policy/operations, providing an evidence-base for modeling, predicting risk, and surveillance-prevention planning.

Associations between meteorological variables, vector indices and dengue hospitalizations in Can Tho, Vietnam: a field survey

“…This stood in contrast to Hanoi (annual average temperature of 23.6°C), where the opposite was observed. This relationship between dengue transmission, temperature and humidity was modeled in a high-resolution profile of these factors across space in Peru [17]. Our findings are consistent with this study, which demonstrated that dengue transmission potential was dependent on the duration within an optimal temperature range and was amplified exponentially by high humidity [17].…”

Research on Climate and Dengue in Malaysia: A Systematic Review

“…The average temperature is around 27 °C, and mean cumulative rainfall is about 2500 mm a year. The warm and wet weather coupled with high humidity is conducive to the development of Aedes mosquitoes, viral replication, and transmission of dengue year-round [3•, 18••, 19]. …”

Evaluating the ALERT algorithm for local outbreak onset detection in seasonal infectious disease surveillance data

“…Seasonal infectious disease surveillance data often shows regular patterns of onset, peak, and nadir . The goal of ALERT is to assign a static value to the incidence threshold level used to define epidemic onset.…”

“…Seasonal infectious disease surveillance data often shows regular patterns of onset, peak, and nadir. 24 The goal of ALERT is to assign a static value to the incidence threshold level used to define epidemic onset. Using historical information from a local surveillance system (eg, a hospital or city), ALERT assists in the choosing of an appropriate time to begin a particular intervention that would cover the period of highest seasonal respiratory virus activity.…”