How AI Can Save a Guatamalan Lake From Destructive Algae | Fast Company

– My name is Africa Flores. I’m a Research Scientist at the University of
Alabama in Huntsville. I’m originally from Guatemala, but I’m also a NatGeo Explorer. With NatGeo and Microsoft,
we are working on a project in Lake Atitlan. Lake Atitlan is a hot
spot of cultural heritage and biodiversity for Guatemala. It’s the second most
visited site in the country and recently it has been
suffering from algae blooms. The largest one covers
about 40% of the lake and affected tourism greatly. There are a number of
towns around this lake. About 40% of these towns
are using the water directly from the lake without treatment. Some of the bacteria have the
potential to produce toxins so if the water is being
used for human consumption that definitely has an
impact on human health. The algae blooms mean
you have massive presence of one species, one or two species and this is covering
the surface of the lake. So, it’s affecting the whole eco-system because it’s preventing
sunlight to come in to the water and that affects the
rest of the life there. It’s a beautiful lake. It’s very well known because of the transparency of the water. In the ’70’s the transparency of the water used to be like 20 meters. Imagine that is like 10
floors of transparency. It’s a very deep lake. In the deepest point it’s 300 meters. It’s an enclosed system. Everything that happens in
that basin and so in the lake. And so every time that it rains, a lot of soil gets transported to the lake and this carries nutrients. At the end of the rainy season, we have less cloud cover so
we have more solar radiation and then the sani-bacteria
is spread already throughout the lake and
it has enough nutrients. And then we see an algae bloom. My background is in remote sensing and geo-special technologies so we are using all these
datas, satellite data, modular data, weather data,
to identify when algae blooms are going to happen in this lake. We are using the artificial
intelligence component, specifically machine learning to analyze these large
amounts of information to eventually train an algorithm that is going to be able to identify and then forecast algae blooms. In the past, it took me
hours, days, and months to process a years worth of data. Not to mention, 10, 20
years of information and now we can process hundreds of data, thousands of data for 30 years historical and have results for
that in a matter of hours or minutes, seconds. Our goal is to replicate
this in other lakes. Harmful algae blooms is not
an issue unique to Guatemala or to Lake Atitlan. This is happening worldwide and I think that through
this prior lake in Guatemala we are going to learn a
lot and expect to expand it to the rest of Central
America, Latin America and hopefully be a global application. (slow music)

Add a Comment

Your email address will not be published. Required fields are marked *