Stink Wars

American farms have been growing moderately since the 1980s and, to keep up with demand, the number of animals per farm has also increased. More animals produce more manure, and not everyone is pleased with the outcome: smellier farms.

Most states have right-to-farm laws that can prevent the owners from being sued over their farms’ unpleasant smells; however those laws don’t always stop people from taking action.

Fangle Chang, a doctoral candidate in agricultural and biological engineering, explained, “Neighbors of farms can find smells more than just annoying. Research shows that odors emitted from agricultural production can make people anxious and uncomfortable. There have even been lawsuits and it’s hard to say who’s right and who’s wrong, especially if the farmer was there first.”

Chang believes there may be a benefit to classifying the odors emitted from dairy farms. Specifically, she is conducting research that combines biological and mechanical olfaction systems in an effort to bridge the gap between the two and create a model that will allow technology to predict how humans will respond to smells.  

To carry out her research, Chang collects solid cow manure samples and feed samples from two privately owned dairy farms in Pennsylvania and three Penn State dairy farms. She brings the samples back to her lab, where they are “smelled” by humans and technology.

Using a glass syringe, Chang collects a trace amount of a scent from each jar and presents it to the humans’ noses. The participants have been trained on a general hedonic (pleasantness) scale (ranging from -11 to 11), so they give consistent responses after smelling each sample.

Chang takes the average of the human responses for each sample to come up with a “target” level.

Next, she uses two instruments — an eNose and a zNose™ — to collect data.

The eNose measures change of resistance. Chang explains, “It’s kind of like a balloon. I insert a scent sample and the eNose will ‘swell’ more if the odor components are stronger.”

The zNose™ measures change of frequency in a similar fashion. Chang injects a scent sample into one end of the device and it measures the frequency in ‘peaks.’ “Sometimes we see bigger jumps in the peaks. That means a bigger frequency change and more components in the scent sample.”

Both the eNose and the zNose™ have software that records the respective data.

Chang said once all the data are collected it is important to connect the results from the instruments to the results from the humans’ responses. “We know what the resistance and frequency changes mean by the numbers but what does it reflect in terms of the odors? How can we use the aggregate data to predict how humans will respond to an odor?”

She is using her data to build a computer prediction model, utilizing artificial neural networks, that she hopes will accurately predict human responses to odors emitted from dairy operations.

Potentially, Chang’s research could lead to a consistent scent scale that the government might use to ensure all dairy farms are ranked equally. “If the smell is below a certain level, then the farmer won’t be at fault, but if the smell is over that level, then the farmer will have to find a way to reduce the odors,” she said.

Paul Heinemann, head of the Department of Agricultural and Biological Engineering at Penn State, is Chang’s adviser. He said, “Fangle came to Penn State with a computer science and engineering background. We wanted to be able to tap into her software and instrumentation skills, but we needed a project that was very applicable to agricultural and biological engineering.”

“Although Fangle didn’t initially know much about dairy operations when she arrived at Penn State, the integration of her background knowledge with an application area that was totally new to her has led to unique discoveries and a great broadening of her educational experience,” added Heinemann.

Chang has presented her work at two national American Society of Agricultural and Biological Engineers conferences.

Last Updated August 25, 2016