A predictive method can better identify which species are vulnerable

Destructive factors such as wildfires, flooding, pollution and overfishing can alter the balance of ecosystems and sometimes jeopardize the future of
entire species.
But assessing these ecosystems to determine which species are most at risk, so that conservation actions and policies can be focused where they are most needed is a challenging task
.

Most such studies assume that ecosystems are essentially in equilibrium, with external disturbances leading to temporary changes
until they eventually return to equilibrium.
But this assumption does not explain the fact that ecosystems are often in flux, and the relative abundance of their different components varies
with their own timelines.
Now, a team of researchers at MIT and elsewhere has come up with a better, predictable way to assess these systems in order to rank the relative vulnerability of different species and detect threatened but potentially overlooked species
.

They found that, contrary to current traditional methods of conducting such rankings, the species with the lowest populations or the fastest decline in numbers — the criteria commonly used today — were sometimes not the most at risk
.

The findings, published today in the journal Ecology Letters, were published
by Serguei Saavedra, associate professor of civil and environmental engineering at MIT, recent doctoral student Lucas Medeiros, and three other PhDs.

Saavedra said the new work, similar to Edward Lorenz's analysis of weather patterns decades ago, revolutionized the field
.
Lorentz's research shows that small disturbances can eventually lead to very large results – the famous formulation is that butterflies flap their wings in one place and eventually cause hurricanes
in another.
"Even very close initial conditions change a lot over a given period of time and therefore become unpredictable
," he said.
With that in mind, "We said, what would happen if we used the same lens to try to figure out which are the most sensitive species?"

In some cases, such as in weather forecasting, scientists understand the basic physics of phenomena and can generate equations that describe their dynamics to some extent
.
Complex ecosystems are not like that, he says, and we don't even have the basic equations for the dynamics of individual species, let alone entire systems
.
But over the last decade or so, he says, the team has developed mathematical techniques so that "we can describe dynamics without knowing the fundamental equations," as long as there is enough time series data to process
.

The team developed two different methods called expected sensitivity ranking and eigenvector ranking
.
Both methods performed well in tests using large amounts of simulated data, producing rankings that were very close
to those expected by the underlying assumptions of the simulation model.

Traditional attempts to rank species vulnerability tend to focus on metrics such as body size (larger species tend to be more vulnerable) and population size, both of which are useful indicators
in many cases.
But, as Saavedra points out, "these species are embedded in communities that have nonlinear suddenness, where a small change in one place completely changes other aspects of
the system in another way.
" ”

Species in an ecosystem can rise and fall in abundance, sometimes periodically, sometimes randomly, or determined by external forces, meaning that the exact timing of a given perturbation can make a big difference – something
that equilibrium models cannot explain.
Medeiros said: "Methods based on equilibrium dynamics have a static view
of the effects of species interactions.
" "Under non-equilibrium abundance fluctuations, the effects of these interactions change over time, affecting any given species' sensitivity
to perturbations.
"

For example, a species that is highly active in summer but dormant in winter may be strongly affected by summer wildfires or heat waves, but not at
all if the destruction occurs in winter.
Or, if the interaction between predator and prey changes over the course of the year, the timing of destruction may be more destructive
in some seasons than others.

Saavedra said the new analysis method is broadly applicable to any type of ecosystem, whether marine or terrestrial, tropical or Arctic
.
In fact, when applied to systems with many interactions and constant flow, these formulas are so common that some researchers have also successfully applied them to predict the dynamics
of financial markets.

"These techniques are suitable for any dynamical
system that is not linear dynamic or generally out of balance," Saavedra said.
One student in the group who had been working on the techniques ended up working for a hedge fund, he said, and another took a long vacation to work at
a foreign bank.
"He was basically able to apply these techniques, and they worked
.
"

But the main goal of this work remains to assess the vulnerability of species, and the findings are already being applied
.
For example, Medeiros, the paper's first author, is working at the University of California, Santa Cruz and the National Oceanic and Atmospheric Administration to apply these techniques to fisheries management
.
"Fisheries in particular, you have a lot of data series to see these populations rise and fall
over time," Saavedra said.
Using the data, he said, it is now possible to "accurately predict the species most sensitive to climate change or the
maximum catch quota ratio.
" ”

The research team also includes Stefano Allesina, who currently works at the University of Chicago and Northwestern University; Vasilis Dakos, currently working at the University of Montpellier, France; and George Sugihara
, now at the University of California, San Diego.
This work was supported
by the MIT Environmental Solutions Program, the Martin Family Association of Sustainability Fellows, the U.
S.
Department of Defense Environmental Research and Development Program, the National Science Foundation, the Department of the Interior, and the MIT Ocean Grant Program.