Submitted by Dr. Rachel Lowe, Royal Society Dorothy Hodgkin Fellow, Associate Professor
21 Jun 2021 , last updated 21 Jun 2021

Linking early warnings to early action: compound hydrometeorological hazards and dengue outbreaks

Dengue fever is caused by a virus carried by mosquitoes and is considered one of the top ten threats to global health. Brazil has the greatest number of dengue cases in the world, reporting more than 2 million cases of dengue in 2019 alone. Aedes mosquitos, which are responsible for the transmission of dengue, thrive in warm and humid conditions, with rainfall increasing the number of outdoor breeding sites. However, drought conditions can also promote breeding, due to an increase in water storage containers in and around the home.

How do hydrometeorological hazards impact outbreaks of dengue?

Extreme hydrometeorological events, such as tropical storms, floods, and droughts, can result in outbreaks of water- and mosquito-borne diseases, such as dengue and leptospirosis. However, the timeframes in which hydrometeorological hazards might alter the timing and intensity of climate-sensitive diseases is poorly understood.

Increasing levels of severe droughts and flooding episodes due to climate change has led to interruptions in water supply networks in Brazil. The improvised water storage containers used to combat this become breeding grounds for mosquitoes (Fig 1).

Several drought severity indices have been developed by the meteorological community, assimilating data on rainfall and other water supply indicators. However, their potential use in dengue control and preparedness plans has been little explored.

In a recent study, we modelled dengue case data in 558 microregions of Brazil between January 2001 and 2019 and used information on droughts and wet conditions and assessed the dengue risk differences in urban and rural areas.

Our study showed that:

  • Both extremely wet conditions and extreme drought can increase dengue relative risk with different delays.
  • The risk of dengue was high in urban areas three to five months after extreme drought, whereas extremely wet conditions increased dengue risk in the same month and up to three months later.
  • In rural areas, dengue risk was more readily associated with very wet conditions (Fig 2).
This graphic shows associations between hydrometeorogogical extremes and dengue risk in urban and rural settings.
Fig 2: Associations between hydrometeorological extremes and dengue risk in urban and rural settings. The risk of dengue is high 3-5 months after drought conditions and 0-2 months after extremely wet conditions in urban areas. In rural areas the high risk of dengue is more immediate during extremely wet events.

This study confirms that initial findings from Barbados, which quantified the delayed effects of drought on dengue risk, are robust across a large gradient of climate and urbanisation conditions across Brazil.

Densely populated urban areas that suffer interrupted water supply during periods of water scarcity are more vulnerable to dengue outbreaks following drought conditions. The impact of extremely wet conditions on dengue risk is more immediate and exacerbated in rural areas, predominantly in the Amazon region.


Linking hydrometeorological information to dengue control and prevention

In Brazil, large outbreaks are typically observed after wet and warm periods and most interventions are targeted at these times. However, our findings raise awareness of the importance of targeting mosquito control activities according to geographical setting, with a focus in urban areas with intermittent water supply, not only during the warm, rainy season but also during drought periods (Fig 3). In an operational prediction framework, observed drought severity and temperature indicators could be combined with seasonal forecasts drought severity indicators to provide a probabilistic prediction 2-3 months ahead of time.

In the short term, a community effort is required to ensure improvised water storage containers do not serve as additional larval habitats in drought periods. During wet periods, outdoor water recipients, including discarded waste should be kept to a minimum.

In the longer term, governments must invest in local infrastructure to ensure permanent water supply and promote environmental hygiene in areas prone to dengue and other mosquito-borne disease epidemics.

This graphic shows a combination of observed and forecat temperature and drought severity indicators
Fig 3: Setting specific forecast scenarios for urban and rural settings. A combination of observed (3-5 month lag) and forecast (0-2 month lag with a 1-2 month lead time) temperature and drought severity indicators could be used to predict the risk of dengue several months in advance (e.g. issue a dengue forecast in Dec-Jan, valid for Mar). More weight could be given to a long lag drought observation in urban settings and a short lag extremely wet forecast in rural settings.

This blog post was written by Dr. Rachel Lowe, Royal Society Dorothy Hodgkin Fellow, Associate Professor, Faculty of Epidemiology and Population Health, Department of Infectious Disease Epidemiology, London School of Hygiene & Tropical Medicine. She can be contacted here.

Practitioners’ experiences anticipating Dengue outbreaks in Bangladesh and the Philippines

How are you anticipating Dengue outbreaks?

In Bangladesh, START Network FOREWARN Bangladesh conducted a dengue risk study in early 2020 that investigated several climate variables such as rainfall, temperature, humidity and historical case analysis from 2000 to 2019. This allowed the FOREWARN Health Expert Group to determine the peak time period of dengue outbreak and develop a surveillance model that allowed the team to discover dengue hotspots. The early actions implemented included a mosquito extermination program in dengue-hotspots, dengue diagnosis kits in local hospitals, training of medical and non-medical staffs on clinical management and a specialized dengue remedial ward for vulnerable communities  - read more about the learnings here.

In the Philippines, START Network in partnership with World Vision’s Dengue Project, developed the Dengue Risk Analysis Tool to monitor the rise of dengue cases in urban cities such as Manila, Malabon and Quezon City. The Dengue Risk Analysis Tool, can forecast the number of dengue cases in an area on a temporal basis. Users need to regularly log their data to forecast the possible rise of dengue cases in the identified or selected areas. The Department of Health (DOH) expressed their interest to use the Dengue Risk Analysis Tool and offered to share their national data on dengue to help improve the tool.

How do the study's key findings connect with your work? 

In Bangladesh, the risk analysis identified that urban areas are more at risk of Dengue in contrast to rural areas and that population density, open water sources and mass development projects undergoing in Dhaka city contributed to the outbreak occurring in the city. Similarly, to the above study, it found that increasing dengue risk is discovered after 0-3 months of wet season. When rainfall is detected in February in Dhaka city, it suggests a greater degree of Dengue risk. It has been reported that the peak of dengue cases occurs in July-August, thus early action measures should begin in the first week of March.

Findings from the development of the Dengue Risk Analysis Tool in the Philippines identified that there was a spike in Dengue cases, three months after specific environmental conditions, including high mean temperature, and increased precipitation and humidity, in the three cities in Metro Manila where data was gathered. It also found that the effect of weather factors is significant 1-2 months before outbreak initiation, and 3-5 months before outbreak peak.


Contribution by Marwa Tasnim, Anticipation Partnership Officer - FOREWARN Bangladesh, Ana Marie Dizon, CARE Philippines, Start Network - FOREWARN Coordinator and Dr. Jomar Rabajante Institute of Mathematical Sciences and Physics University of the Philippines Los Baños.