A mathematical model developed by Texas A&M researchers can predict temperatures within mosquito breeding grounds, which can be used to estimate populations and track vector-borne diseases. 

With an impressive capability of drinking up to three times their body weight in a single blood meal, mosquitoes are formidable parasites. But to reach adulthood, mosquitoes need to be raised in environments where the temperatures are conducive to their breeding, growth, and development.

In a new study in the journal Scientific Reports, Texas A&M University researchers have developed a mathematical model based on machine learning to precisely predict the local or microclimatic temperature within the breeding grounds of the Aedes albopictus mosquitoes, carriers of the chikungunya and dengue viruses. Their algorithm also reveals that even in winter, the temperature may be warm enough in certain breeding grounds to allow mosquitoes to grow and thrive.

“Our goal is to develop accurate and automated mathematical models for estimating microclimatic temperature, which can greatly facilitate a quick assessment of mosquito populations and consequently, vector-borne disease transmission,” said Madhav Erraguntla, associate professor of practice in the Wm Michael Barnes ’64 Department of Industrial and Systems Engineering.

Responsible for around a million deaths globally, mosquitoes continue to wreak havoc to public health in many parts of the world. In addition to  water, temperature plays a critical role at different stages in mosquitoes’ life cycle. Furthermore, The mosquitoes’ development, reproduction and survival can be mathematically modeled on the basis of temperature.

Past studies have largely relied on ambient temperature, or general air temperature, to make predictions about mosquito populations. However, these calculations have not been precise since ambient temperatures can deviate from those within mosquito breeding grounds. Recognizing this shortcoming, scientists rely on sensors, called data loggers, to continually keep track of the temperature, light intensity and humidity within breeding grounds. Despite their advantages, these sensors are inconvenient due to their cost and long-term use.

“People have realized that the microclimatic conditions are important, but right now data loggers are the only way to keep track of temperature,” Erraguntla said. “We wanted to address this gap by automating the process of estimating microclimatic temperatures so that we can model the life cycle of mosquitoes accurately.”

For their experiments, the researchers placed sensors in common mosquito breeding grounds around Houston, including storm drains, shaded areas and inside water meters. In addition, they obtained information on ambient temperatures from the National Oceanic and Atmospheric Administration repository. With this data as training input to a machine learning algorithm, the computer model could predict the microclimatic temperatures for a variety of ambient temperatures and breeding grounds within 1.5 degrees centigrade. Further, the model now could even forecast microclimatic temperatures for any ambient temperature, precluding the need for sensors.

Next, they fed the values of the microclimatic temperatures to another mathematical model, called the population dynamic model, that tracks the life cycle of the mosquitoes. Based on the microclimatic temperature and other parameters, the population dynamic model could estimate the populations at different stages in the lifecycle including eggs, larvae, pupae, and adult Aedes albopictus mosquitoes.

The model also revealed that the insulated conditions of the storm drains could result in the survival of 84% of juveniles and eggs and 96% of adults during the winter months, a time of the year when mosquitoes are assumed to be dormant.

Although their climatic temperature prediction model has a high degree of accuracy, the researchers noted that additional research is needed to affirm if their model is applicable to places outside of Texas.

“Our work automates the prediction of microclimatic conditions, bypassing an otherwise expensive and time-consuming process of placing the sensors in different breeding spots, collecting the sensor data and analyzing it,” Erraguntla said. “From a public health context, this work will help epidemiologists better track mosquito-borne disease transmission and surges in mosquito abundances.”

WildfireDB contains over 17 million data points that capture how fires have spread in the contiguous United States over the last decade 

A team at UC Riverside led by computer science assistant professor Ahmed Eldawy is collaborating with researchers at Stanford University and Vanderbilt University to develop a dataset that uses data science to study the spread of wildfires. The dataset can be used to simulate the spread of wildfires to help firefighters plan emergency responses and conduct evacuation. It can also help simulate how fires might spread soon under the effects of deforestation and climate change, and aid risk assessment and planning of new infrastructure development.

The open-source dataset, named WildfireDB, contains over 17 million data points that capture how fires have spread in the contiguous United States over the last decade. The dataset can be used to train machine learning models to predict the spread of wildfires.

“One of the biggest challenges is to have a detailed and curated dataset that can be used by machine learning algorithms,” said Eldawy. “WildfireDB is the first comprehensive and open-source dataset that relates historical fire data with relevant covariates such as weather, vegetation, and topography.”

First responders depend on understanding and predicting how a wildfire spreads to save lives and property and to stop the fire from spreading. They need to figure out the best way to allocate limited resources across large areas. Traditionally, fire spread is modeled by tools that use physics-based modeling. This method could be improved with the addition of more variables, but until now, there was no comprehensive, open-source data source that combines fire occurrences with geospatial features such as mountains, rivers, towns, fuel levels, vegetation, and weather.

Eldawy, along with UCR doctoral student Samriddhi Singla and undergraduate researcher Vinayak Gajjewar, utilized a novel system called Raptor, which was developed at UCR to process high-resolution satellite data such as vegetation and weather. Using Raptor, they combined historical wildfires with other geospatial features, such as weather, topography, and vegetation, to build a dataset at a scale that included most of the United States.

WildfireDB has mapped historical fire data in the contiguous United States between 2012 to 2017 with spatial and temporal resolutions that allow researchers to home in on the daily behavior of fire in regions as small as 375-meter square polygons. Each fire occurrence includes the type of vegetation, fuel type, and topography. The dataset does not include Alaska or Hawaii.

To use the dataset, researchers or firefighters can select information relevant to their situation from WildfireDB and train machine learning models that can model the spread of wildfires. These trained models can then be used by firefighters or researchers to predict the spread of wildfires in real-time. 

“Predicting the spread of wildfire in real-time will allow firefighters to allocate resources accordingly and minimize loss of life and property,” said Singla, the paper’s first author. 

The paper, “WildfireDB: an open-source dataset connecting wildfire spread with relevant determinants,” will be presented at the 35th Conference on Neural Information Processing Systems (NeurIPS 2021) Track on Datasets and Benchmarks and is available here. A visualization of the dataset is available here. Eldawy, Singla, and Gajjewar were joined in the research by Ayan Mukhopadhyay, Michael Wilbur, and Abhishek Dubey at Vanderbilt University; and Tina Diao, Mykel Kochenderfer, and Ross Shachter at Stanford University.