When farmers, ranchers and foresters manage the land for a harvest of life-essential products, from food to clothing to components of medicine, they are actually doing far more. Well-managed lands, and the wild lands private landowners conserve, are the key to clean water, healthy soils, climate resilience, carbon sequestration, and biodiverse wildlife habitats—the ecosystem services that society needs but doesn’t often value. At least not at the bank.

Most people want to be able to enjoy natural lands, to escape the city noise and traffic. But the reality is, natural lands are undervalued based on their natural assets. In Florida, one of fastest-growing states in the U.S., the value of land has skyrocketed for its development potential. While development is not bad, it’s difficult for a family to justify keeping agricultural land in production on thin margins when selling that land for development could yield great, albeit one-time, profits.

Landowners need to be compensated for being good stewards of the ecosystem—for managing the land in ways that conserves and enhances its ability to provide us all with a clean, productive environment.  

“But the key is you need to know what the value of those ecosystem services are,” says Joel Harley, Ph.D., associate professor of electrical and computer engineering in the University of Florida’s SmartDATA Laboratory. “That’s not easy.”

Harley and his team are part of the larger, AI-HARVEST team at the University of Florida led by Alina Zare, Ph.D., professor of electrical and computer engineering and associate dean for research and facilities in the Herbert Wertheim College of Engineering, and fellow collaborators Lincoln Zotarelli, Ph.D., professor of horticultural sciences, and Jose Dubeux, Ph.D., professor and associate chair of forages.

Through the AI-HARVEST project, researchers have begun developing the tools needed to quantify—and, hopefully, eventually put a dollar amount to—ecosystem services using artificial intelligence (AI). The one-year project focused on quantifying plant biodiversity, how much and how many different kinds of plants cover Florida lands. More biodiverse lands feature more plants and animals, each of which contributes a different benefit to the overall ecosystem.

To build an AI algorithm that can quantify something as complex as the ecosystem, the researchers have collected and integrated close to a million pieces of data on plant species and locations throughout Florida. They also have more than 50 million square miles of satellite imagery.

The researchers started their data collection by going straight to the source: landowners and land managers who are part of or friends with the Florida Climate Smart Agriculture Work Group, led by Jim Strickland, rancher, and Lynetta Usher Griner, co-owner of a cattle and timber operation. Students walked fields, pastures, and woodlands, noting forages, forbs and other plant species.

Satellite images give researchers a Florida-wide view of changes in the landscape each month. The researchers also look from the sky using drone images and tap into public and crowdsourced data. They use the popular iNaturalist app, which allows people to take pictures of and find identifications for plants they see in nature, as well as more comprehensive, research-centric databases from federal and state conservation land managers.

“It’s a massive undertaking,” Harley says. “But there’s a lot of potential, and a lot of potential value to a lot of people, especially in the state of Florida, where our natural lands are one of our biggest attractors.”

As of August, the AI-HARVEST team has developed its preliminary algorithm for quantifying ecosystem services, starting with biodiversity. Through interviews with agricultural producers, the researchers found one of their next key ecosystem services to quantify: water quality.

 “Water quality is a huge concern in Florida right now,” says Harley. “So, we’ve been looking into how we address that next.”

The researchers also want to eventually expand to quantify water quality, water quantity, water storage, and soil nutrients.

Long-term, this research, and the resulting AI tools, could be used by agricultural producers to inform decision-making, such as what crops to grow on what pieces of land.

It could also help:

• Build policy-based incentives for landowners and land managers to protect and enhance the health of agricultural and conservation lands.
• Create markets for ecosystem services beyond carbon.
• Give land buyers and sellers a better idea of the land’s value, beyond its obvious development potential.

For more information about AI-HARVEST, contact Joel Harley at joel.harley@ufl.edu.