SIAM News Blog

Changing Temperatures, Changing Risks: Modeling Dengue Transmission in the Context of Climate Change

By Julie Allison Spencer and Kelsey Skogstad

Imagine what life with dengue fever would be like in the U.S. during the year 2050. You live with your family in Houston, Texas, and enjoy frequent fishing excursions at the nearby nature reserve. On average, days are about 5° Fahrenheit warmer than in the previous decade, and a summer dengue outbreak is underway. The local public health agency recommends using insect repellent, wearing long sleeves and pants, covering strollers with mosquito netting, and installing screens on windows and doors. It also suggests eliminating mosquito habitats near homes by emptying water-holding containers like planters, toys, tires, and trash receptacles.

You drive to the nature reserve and park your car at the lake. While settling your one-year-old daughter into her stroller, you remember that you left the mosquito netting at home. Just then, you notice a mosquito on your daughter’s arm, already swelling with blood. You immediately pack up to return home, but the bite is irreversible. A few days later, your daughter develops a fever and a rash. When she starts vomiting, you take her to urgent care.

You spend a sleepless night holding your young daughter in a hospital room as she is hydrated with IV fluids. In the morning, you are incredibly relieved when the doctor says that she will likely not develop severe dengue. As you drive home, you remember when mosquito bites were not as serious and fishing trips were easier. Times have changed.

The Current Threat of Dengue

Figure 1. Compartmental diagram for dengue virus transmission and subsequent disease progression across humans and mosquitoes. A susceptible human \((S_h)\) becomes exposed to dengue \((E_h)\) from the bite of an infectious mosquito \((I_v)\). Exposed individuals become infectious \((I_h)\) after an incubation period, then are removed \((R_h)\) once they clear the infection. Susceptible mosquitoes \((S_v)\) enter the system at the rate \(h_v(N_v)\), and either die or become exposed \((E_v)\) when they bite an infected person. Exposed mosquitoes become infectious \((I_v)\) after the extrinsic incubation period and leave the system though natural death. Figure courtesy of [2].
Almost half of the world’s population is currently at risk of contracting dengue virus, and around 25 percent of people who are infected with dengue become sick. About five percent of patients with symptoms develop severe dengue, which can become life-threatening within mere hours; this risk is higher for infants and pregnant women [1]. While dengue in the contiguous U.S. presently occurs in mostly isolated cases in travelers who are returning from the tropics, the Aedes mosquitoes that carry the virus have recently expanded their range from tropical to temperate regions. No dengue vaccine is approved for use in the U.S., and no medication that is specific to dengue is available at this time.

A New Dengue Model

A team of scientists at Los Alamos National Laboratory (LANL) recently investigated the effect of changing temperatures on dengue transmission within a climate change scenario [2]. The group extended previous models to develop a susceptible-exposed-infected-recovered (SEIR)-type mechanistic model that allows both mosquito lifespan \(({\mu_v}^{-1})\) and extrinsic incubation period \(({\nu_v}^{-1})\) to vary with temperature (see Figure 1). This new SEIR model captures the nonlinear effects of changing temperatures on mosquito dynamics and dengue transmission. Such an approach can explore realistic simulations of potential summer dengue outbreaks in southern U.S. cities where the primary dengue vector, Aedes aegypti, has already been observed.

The LANL researchers found that under model assumptions and with an average temperature increase of 3° Celsius (about 5° Fahrenheit), dengue risk would double in Los Angeles, Calif., and Houston but decrease in Phoenix, Ariz.; Brownsville, Texas; and Miami, Fla., due to the extreme heat (see Figure 2). Although the virus travels through a mosquito’s system to the salivary glands more quickly as temperatures rise, there exists a thermal maximum beyond which Aedes aegypti cannot survive. Sensitivity analysis indicates that as temperatures increase, the determinants of risk begin to shift from mosquito biting rate and carrying capacity to the duration of the human infectious period. This finding suggests that public health recommendations that adjust human behavior may become more important than those that control mosquitoes as temperatures get warmer.

Figure 2. The possible epidemic trajectory of dengue in five U.S. cities—Los Angeles, Calif.; Houston, Texas; Miami, Fla.; Brownsville, Texas; and Phoenix, Ariz.—at the current mean temperature from June to October for each city (top) and with a 3°C temperature increase from the current mean temperature (bottom). At current mean temperatures, the fastest disease progression and greatest number of infected humans are in Houston, Miami, and Brownsville. With a 3°C mean temperature increase, disease spread decreases in Miami and Brownsville and is eliminated in Phoenix. In all five of the cities that were modeled, an outbreak occurs at the current mean temperature. Figure courtesy of [2].

Improvements in the accuracy of scientists’ efforts to model and project vector-borne disease will be essential when informing communities of risk factors as temperatures rise. Future research could enable decision-makers to prepare for outbreaks and deploy resources where they are most needed.

Julie Allison Spencer delivered a minisymposium presentation on this research at the 2022 SIAM Conference on Mathematics of Planet Earth, which took place concurrently with the 2022 SIAM Annual Meeting in Pittsburgh, Pa., last year.

[1] Centers for Disease Control and Prevention. (2023). Dengue. Retrieved from
[2] Trejo, I., Barnard, M., Spencer, J.A., Keithley, J., Martinez, K.M., Crooker, I., … Manore, C. (2023). Changing temperature profiles and the risk of dengue outbreaks. PLOS Clim., 2(2), e0000115.

Julie Allison Spencer is a postdoctoral researcher in the Information Systems and Modeling Group at Los Alamos National Laboratory, where she collaborates on a broad range of projects at the intersection of data science and mathematical epidemiology. She received her Ph.D. in biology from the University of New Mexico in 2020. 
Kelsey Skogstad was a post-masters research associate in the Information Systems and Modeling Group at Los Alamos National Laboratory until February 2023. She is currently on hiatus from research to focus on family. She received her master's degree in public health from George Washington University in 2021. 
blog comments powered by Disqus