With an estimated $300 million in annual recreational value and as a main source of drinking water for Metro Atlanta, Lake Lanier is an important natural resource to the state of Georgia. Dr. Nandita Gaur, Assistant Professor in Soil Physics, Crop and Soil Sciences at the University of Georgia, is examining exactly how nutrients from surrounding septic systems make their way to the lake—so that its waters can be kept clean for generations to come. 

“We’re trying to develop a model that will tell us how nutrients that come from septic systems, such as phosphorus and nitrogen, actually get channeled to the lake,” Gaur explained. 

In the past, models have based their information from isolated point measurements at shallow depths and assumed that the composition of underground layers of soil and weathered rock were the same throughout the modeling domain; they base their predictions on horizons, or the layers of rock and soil expected to lie beneath the surface. However, this approach is not always accurate. 

“In reality, the deep subsurface is heterogenous. Especially in Georgia, where we have a lot of saprolite, which is really weathered rock that can have high porosity or ability to store water and also transport water when saturated,” Gaur said. 

This project, which is funded by the Gwinnett Department of Water Resources and is in collaboration with Georgia Tech, uses a technique called tomography to create a more accurate assessment of how quickly the nutrients from septic systems get flushed into lakes. Our team collects the geophysics datasets and we are working with Dr. David Radcliffe- one of the PIs of the project and Georgia’s expert on modeling the transport of nutrients from septic systems to water bodies to develop the hydrological models. 

“What we do is inject current through the ground, and it gives you an image. It’s like x-raying the subsurface- only you are not using x-rays but electric current to image. From that, you can decipher the lithology, or the characteristics of the rock, including how the land is stratified. And what we found is that it’s very different from what we had previously assumed. There are regions that have weathered more than the surrounding areas, which potentially can allow nutrients and water to pass more readily but we are still investigating this,” Gaur said.  

Our research combines simple machine learning based post processing techniques to analyze sub-surface resistivity images that can improve estimation and prediction of hydrologic flux and nutrient transport by adding greater detail to the modeling domain. This can especially be advantageous when you have preferential flow paths in the sub-surface which are hidden and won’t be accounted for in a regular hydrologic model set-up. While older models might have assumed that it would take decades for nutrients from septic systems to reach Lake Lanier, we are investigating if the reality of nutrient reaching the Lake might be different than previously thought.

“If there is a preferential flow path, or porous area of rock that allows water to travel quickly, it could act like a pipe in transporting contaminants down to the lake,” Gaur said. 

Factors such as how quickly nutrients travel are important for city planners to take into account. 

“When you’re designing the system, if you think that you have 100 years before you have to be concerned about nutrient levels in the lake, you’ll have a big problem when nutrients travel more quickly. With a big rain storm, you’ll get a spike in nutrients and an algal bloom, and you may have no idea why it happened. Especially now that we’ve been experiencing extreme weather events—rainfall followed by periods of drought and then rainfall again. That can change how nutrients travel through the subsurface. It’s why it’s so important to determine the timing of these nutrients reaching the lake,” Gaur explained. 

Algal blooms are a result of eutrophication, or excessive richness of nutrients in a lake. When the algae die, they typically sink to the bottom and decompose.  This can create areas of low oxygen harmful to aquatic animals. Specific kinds of algal blooms, such as bluegreen algae, or cyanobacteria, are especially dangerous. 

“Cyanobacteria can be toxic. Lately, there have been reports of pets dying after drinking water from lakes during algal blooms across the country. People also use Lake Lanier for recreation, and it’s a drinking water source—its water quality has a direct impact on our lives. Better models would allow us to predict when these dangerous blooms may happen in the future, so that we can better protect people and their pets,” Gaur said. 

This technique for creating more accurate models could also be adapted to places outside of Lake Lanier. 

“The idea is that if we scale this model up, we can predict how water quality will respond when we have extreme weather events in the future. It will help us manage our water resources better and take steps to prevent these issues,” Gaur said.