As hurricanes move slowly along their path, they can dump more than three feet of rain. When they finally dissipate, the resulting flooding and standing water wreak havoc in multiple ways—threatening drinking water, damaging critical infrastructure and washing out roadways. But they also result in another danger few people consider: a spike in mosquito populations. Receding floodwaters are an ideal habitat for mosquito breeding and, consequently, a rise in mosquito-borne diseases.

Mosquitoes cause millions of deaths every year by spreading diseases such as malaria, chikungunya, West Nile virus, Zika virus and others. Knowing where an increased population will arise could help communities prevent the spread of mosquito-borne diseases by providing them, as well as public health departments, with advance warning to respond to and prepare for potential outbreaks.

At Los Alamos National Laboratory, we’re studying the dynamics of mosquito populations to understand how they grow, how they change with the seasons and, in particular, how they spread infectious diseases to humans and to other animals. The goal is to create a computer-based model that will accurately simulate these populations based on data about precipitation, temperature, water levels and other environmental factors in a given area, so people will know ahead of time about an increased risk of disease transmission.

For this project, we’re specifically looking at West Nile virus, which birds transmit to humans via mosquitoes. We analyzed 15 years of data from several different locations in the United States and Canada, making it one of the largest modeling studies of mosquito populations over time ever conducted. Previous studies have looked at temperature and precipitation, but this is the first to use stream-gauge and water-level data.

Because standing water levels directly influence mosquito populations, these data turned out to be a crucial missing link in predictive models. Using all of this information, we trained the models on six years of existing data and then had the models predict the next several years of historical data as if they did not yet exist.

To train a model, we had to make some assumptions about mosquito lifecycles and how they respond to environmental variables such as temperature, precipitation and water levels. For example, juvenile mosquito stage development rates were assumed to depend on temperature. Available standing water levels were assumed to affect habitat for eggs and larvae, with reduced standing water resulting in fewer successfully hatching eggs and more competition among larvae. We also assumed that temperature affects adult mosquito lifespan. From those assumptions, we then looked at what the models predicted and compared them to the actual data.

The results were surprising: the predictive models very closely resembled the actual patterns. This is promising news. If the model can accurately predict populations before they surge, public health officials could, for the first time, be warned early that their communities might be at greater risk for the spread of mosquito-borne diseases. This can give them time to put in place preventive measures, such as spraying where mosquitoes are laying eggs and where they are seeking hosts, distributing personal protective measures including repellent or nets, and draining areas of standing water. Larvicide can be put into water bodies with mosquito larvae. The models provide a kind of early warning system that enables action to significantly reduce disease transmission.

This is particularly important given rising water levels, warming temperatures and more extreme precipitation events around the world—not just in the aftermath of hurricanes—which could mean even more widespread mosquito-borne diseases.

The potential to reduce the spread of these diseases is about more than keeping individual communities healthy; it also has far-reaching national and global security implications. Regions hit hard by illness are more vulnerable to economic woes, stressed health care systems and other destabilizing conditions that can weaken the social fabric of a community and foster political unrest.

While this might be truer in developing countries with less reliable social supports and infrastructure, developed countries like the U.S. also face the very real threat of rapid, unpredictable disease outbreaks. Missed work hours and medical costs all tax the economy, with lasting ripple effects. Predicting mosquito populations is one way to help mitigate those impacts.

Our modeling research also deepens our understanding of how diseases spread—which is critical to protecting U.S. citizens at home and abroad, including warfighters deployed overseas. Introduction of a new disease to a country or community, or the rapid spread of an existing disease, all pose risks. Knowing these patterns ahead of time will help officials develop the proper response and keep people safer.