Satellite images and computational models combat invasive species
Invasive species are the second greatest threat to biodiversity after the destruction of natural habitats. The proliferation of satellites and the data they produce helps to prevent the spread of invasive species and restore ecosystems that have already suffered damage. LUT University is contributing its expertise in applied mathematics and data analysis to this valuable work.
The raccoon dog in the forests of Finland and the water hyacinth in Lake Tana in Ethiopia – these two vastly different species have one thing in common: they have been classified as invasive species that negatively impact biodiversity beyond their native ecosystems.
Invasive species spread independently or are introduced by humans to new areas, harming the local ecosystem and threatening biodiversity.
In Finland, the raccoon dog jeopardises bird populations, as it eats their eggs and carries dangerous diseases and parasites. The water hyacinth, in turn, is a free-floating perennial aquatic plant that spreads exceptionally quickly. Many freshwater systems in Africa are suffering from its rapid distribution.
The water hyacinth, native to the South American tropical and subtropical zones, has been classified as one of the world's hundred most harmful invasive species (Global Invasive Species Database). Experts from a number of fields are collaborating with authorities to undo its damage to areas it has claimed.
"The water hyacinth blocks river deltas and other waterways. It hinders fishing and is not even fit for feeding animals: it gives cows liver damage, for instance," explains Associate Professor Lassi Roininen.
Roininen has last witnessed the devastation by the water hyacinth in Ethiopia in March 2020 right before global tourism took a nosedive due to the coronavirus. How is the work of Roininen, an associate professor of applied mathematics, connected to preventing the spread of invasive species in Africa?
Roininen says his work largely deals with data produced by satellites and upper atmosphere radar systems. Together with his researcher colleagues, Roininen develops computational models to improve data quality by eliminating inaccuracies in observations and measurement results.
"In this specific study, we are creating computational models that will improve the reliability of satellite images of Lake Tana. This research provides specific data to help prevent the spread of the water hyacinth. The actual prevention may take place manually or mechanically."
The research takes advantage of the freely available imagery of the American Landsat 8 satellite, and Roininen is also working on a scientific paper on the research with two of his international colleagues.
Computational models detect what the eye can't see
In assessing the spread of the water hyacinth based on satellite images, inaccuracies usually result from gaps between individual blossoms; compact clusters, however, can usually be identified reliably even with the naked eye.
"We're aiming to create a reliable computational model that can analyse each pixel and determine whether it represents a plant."
One of Roininen's key colleagues is Heikki Haario, LUT's long-time professor of applied mathematics. Through Haario, Roininen became involved in a study of great national importance that improves satellite image data with computational models.
"I have also engaged in increasing the accuracy of satellite data on the algal situation in the Baltic Sea. Clouds usually create blind spots over sea areas. Statistical models can create a computational model for each individual pixel to determine whether it contains algae. In other words, computational models see what the bare eye cannot," Roininen relates.
Roininen keeps referring to his fellow researchers and collaborating universities around the world and satellites and their global application areas. The highly networked researcher has many ongoing projects and ideas that soar sky-high.
"Satellite measurements and the utilisation of related data have multiplied. Especially the EU is openly campaigning for open-access satellite data and wants different sectors of society to be able to profit from it. Space-based data will provide researchers with work for years to come," Roininen concludes.