Home Artificial Intelligence AI used to fight hurricane-powered exotic invasions in the Everglades

AI used to fight hurricane-powered exotic invasions in the Everglades

play

Note to readers: The News-Press and Naples Daily News are providing special coverage this week of all things weather in Southwest Florida as a new hurricane season having started on June 1.

It’s not exactly robots fanning out across the Everglades to destroy alien invaders … except it kind of is.

Scientists are using Artificial Intelligence to combat habitat-swallowing invasive plants like Brazilian peppers.

And though it’s not metal claw-to-branch combat, it is a tool in a potentially life-and-death battle to keep Florida ecosystems from being overrun.

As a changing climate makes storms stronger, more frequent and less predictable, tracking the spread of exotics has become more challenging, since they’re spreading in ways they haven’t before.

Finding them is the first step. Artificial Intelligence can help with that by spotting them and predicting their spread, which can help land managers get rid of them before things get out of hand.

Researchers at the University of Florida are trying to understand the way hurricanes distribute non-native plants, something that may become even more urgent as the National Hurricane Center considers changing the start of the Atlantic hurricane to May 15 instead of June 1, two weeks earlier has traditionally begun. That makes sense, though, points out UF spokesman Brad Buck in “UF Researchers to Use AI to Predict how Hurricanes Spread Invasive Plants,” since Tropical Storm Arthur formed May 16 last year. Since 2003, 15 tropical cyclones have formed before June 1, including two that developed in April, USA Today Network Florida reports.

Stronger hurricanes might turbo-charge invasions

A UF research team is using Artificial Intelligence to gather data for a study that combines ground-level research with a technique called hyperspectral sensing. The goal: Understand how hurricanes impact the spread of two of the peninsula’s worst invaders – Brazilian pepper and Old World climbing fern.

Both appear on the “Dirty Dozen” list of Everglades invaders – species wreaking havoc on the Glades’ fragile natural system.

According to the National Park Service, “In terms of sheer magnitude, Brazilian pepper is the most widespread invasive plant … particularly abundant along the fringes of the mangroves. In some instances, individual stands of Brazilian pepper cover 4,000 to 6,000 acres.”

As its name implies, the fern scrambles up over native habitats, forming green shrouds on trees and can completely engulf the Everglades tree islands that are key habitat for native species and traditional homes for Seminole and Miccosukee people. Its long skirts can become ladders for wildfire, allowing blazes to climb to the top of tree canopies and destroy hugs stands of old-growth forest. It now infests more than 200,000 south Florida acres, according to the Florida Fish and Wildlife Conservation Commission.

Devastating as both plants already have been to the Big Cypress National Preserve, Everglades National Park and other wild lands throughout south Florida, stronger hurricanes may turbo-charge their damage in the future.

Led by Fulbright Scholar Luke Flory, a UF Institute of Food and Agricultural Sciences associate professor of ecology, the team also includes University of Iowa’s Susan Meerdink, an assistant professor of geographical and sustainability sciences, and Florida professors Alina Zare and Paul Gader as well as doctoral student Dylan Stewart, all with UF’s Herbert Wertheim College of Engineering.

The team intends to examine post-hurricane conditions with remotely sensed data that can be used to map the distribution of invasive plants, Buck says.

The idea is to get ahead of the plants by predicting future patterns, then planning a counterattack. If resources can be allocated before or right after a storm, invasions can be managed more effectively and efficiently, Flory says.

This is where Artificial Intelligence comes in. It’s quite a name, but a method called multitarget multiple-instance spectral match filter looks promising. Abbreviated to MTMI-SMF, it’s a machine-learning approach that “allows flexibility in training data (and) can detect highly problematic invaders using multispectral imagery,” said Meerdink in the abstract to the paper, “Dealing with imperfect data for invasive species detection using multispectral imagery,” published in the journal Ecological Informatics.

Also promising: “This pipeline could be readily adapted to other invaders and ecosystems, as well as different remote sensing image sources, and has low computational requirements. Our study demonstrates that remote sensing technologies and multiple-instance learning algorithms can provide managers with critical tools.”

How hurricanes help exotic plants spread

Hurricanes can shape plant invasions two ways:

  • Storms can make the landscape more receptive to invasives. If saltwater kills native plant communities or big trees fall and allow sunlight to penetrate where it hadn’t before, exotics can capitalize on that and move into new places and
  • Storms can spread exoric seeds and spores to new places.

Hurricanes can, in essence, create a perfect storm for invasions, Flory says.

Beating back exotic invaders: Audubon Corkscrew Swamp Sanctuary passes hard-fought habitat milestone

So, how to fight back?

The News-Press asked Flory about the team’s work and its potential applications. Answers have been edited for length.

What are the mechanics of the project?

FLORY: Very generally speaking, the goal was to connect data from satellites with data collected on the ground so that, ultimately, satellite-based data (i.e., remotely sensed data) could be used over the very broad scale of the Everglades to better understand the spread of invasive plants (two species in particular) and how their spread was impacted by hurricanes. The approach changed over time due to logistics, from using handheld and near-surface devices to using photos taken by a person (lab manager and PhD student Drew Hiatt) from a helicopter.

What, exactly, is remote sensing and how does it work?

FLORY: Remote sensing is simply the collection of data from afar – which could be from a drone, helicopter, or most often satellite.

What is hyperspectral sensing?

FLORY: Essentially hyperspectral sensors collect information as a set of ‘images’, each representing a narrow wavelength range of the electromagnetic spectrum, also known as a spectral band. We had planned to eventually (once data was available) use hyperspectral data to better detect invasive plants because at least some species have a unique detectable spectral signature compared to native species. We’ve initially used ‘multi-spectral’ data, which is similar to hyperspectral except that involves many fewer bands of data. Importantly, it still worked.

What are the handheld devices like? When you do the flyovers, is that on a drone?

FLORY: No, drones are not allowed in national parks, unfortunately. We ended up mostly using data from a handheld digital camera from the helicopter plus data collected using satellites.

Where are you doing the sensing? How big an area are you working in?

FLORY: In Everglades National Park – they provided funding and a lot of collaborative assistance, including the helicopter flights.

 

Reference

Denial of responsibility! TechCodex is an automatic aggregator of Global media. In each content, the hyperlink to the primary source is specified. All trademarks belong to their rightful owners, and all materials to their authors. For any complaint, please reach us at – [email protected]. We will take necessary action within 24 hours.
DMCA compliant image

Leave a Comment