A Cornell-led team of ecologists and computer scientists in collaboration with the Wildlife Conservation Society created an Artificial Intelligence program that tackles ecological impacts of hydropower dams in the Amazon basin — a sensitive region consisting of more than a third of South America.
Spearheaded by Prof. Carla Gomes, Director of the Institute of Computational Sustainability, Prof. Alex Flecker, evolutionary biology and ecology, and postdoctoral research fellow Rafael Almeida, the program could improve the way future hydropower dams are built by determining favorable locations for dam installation and ensuring maximum energy use and low carbon emissions.
Hydropower dams generate hydro-electricity by controlling the flow of water. Water flowing through the dam spins a turbine, which turns a shaft in an electric generator which produces electricity. If you’re familiar with Beebe Lake, then you have most likely seen the Triphammer Dam in action generating electricity for Cornell.
However, the construction of hydropower dams in the Amazon are often unsustainably built in poor locations, potentially undermining the ecological goals that hydropower aims to address. These poorly designed dams can emit up to ten times more greenhouse gases than thermoelectric plants.
Almeida explained that non-optimal locations of dams can emit dangerous levels of greenhouse gases. The team found that while 92% of planned highland hydroelectric dams in the Amazon would be sustainable over the long term, only 36% of the lowland dams would yield less emissions.
“If you don’t pick the right locations for dams, you can have a combination of dams that are dirtier than coal and natural gas [emissions],” Gomes said. According to Gomes, these greenhouse gas emissions are attributed to the decaying vegetation within reservoirs created by such poorly placed hydropower dams.
Flecker, specializing in Latin American ecology, described the focus of the team’s research: biodiverse places containing major tropical rivers that are located in countries with growing economies. In Brazil, there were many proposed hydropower dam locations in the Amazon basin yet a lack of history with constructing hydropower dams.
In addressing the ecological effects of hydropower, the team looked at different combinations of already existing dams, and developed a computational model that uses A.I. to find the most promising configurations of dam sites.
With this new program, researchers are able to determine where low-carbon hydropower dams would best be built.
Looking at every possible location would be “naive,” Gomes said. “The A.I. based approach can look at this problem in a very smart way and decompose it into smaller problems that eliminate … and rule out dams that are provably bad.”
Almeida further explained how, in the pursuit of accomplishing environmental and energy objectives, this model shows that strategic planning can minimize the environmental impacts of each desired energy generation target.
“It’s not just enough to look at the energy they produce,” Gomes explained. The effects of hydropower dams are more substantial than the service they provide. It’s important to “look at other aspects like how it’s going to impact fisheries, [and] how it’s going to impact populations,” Gomes said.
Overall, the team hopes that decision makers will plan carefully and look at the ecological effects of hydropower dams and energy. Gomes explained that often the “thinking behind hydropower energy means ‘Oh, this is clean energy’ but really the big message is ‘Well, not so fast.’”
In the pursuit of sustainable energy, the effects of the infrastructure are underlooked, as such it’s “not a guarantee that by just using hydropower you are using clean energy, you might end up with solutions that are worse than coal,” Gomes said.
The model shows the potential for A.I. in addressing sustainability efforts.
“A.I. can transform the world,” Gomes said — and this tech should be ethically used “to address key challenges and the protection of the environment so that we can really be on the path [to] a sustainable future.”