Modelling Ecosystem Services under Uncertainty with ARIES: Gary Johnson at TEDxUVM

Gary Johnson models the spatial flow of services from ecosystems to people and changes to these services by different management scenarios. His research interests include ecosystem service modelling and simulation, network flow optimization and substitutability, hierarchical modelling, stochastic processes and uncertainty propagation, decision and game theory, functional programming, and machine learning.

We all know that we receive
benefits from nature,

but have you ever tried
to list them out?

To identify them,
assign values to them

or actually trace them back, to particular
landscapes that give rise to them.

Most of us probably don't go
through this exercise

on a regular basis if ever.

But the answer to these questions
is fundamental to our ability

to manage our landscapes,
for both sustainability

and for improved quality of life.

But to answer these questions,
I need to know,

what values you assign

to wetlands,

streams, forests, fields?

And this question isn't particularly
easy to answer,

if you think about it.

We all have familiarity with assigning
a value to a pint of maple syrup

or a glass of water.

But, what's the value of the maple trees
that produced that syrup,

or the forest
where maple trees grow?

Is the value of the forest equal
to the value of the maple syrup?

Probably not.

Forests produce
a lot of other services,

and we could sit and think
about them for a minute.

We can think whether it is –
board field lumber it produces,

or they generate other food,
fuel-fiber type resources, firewood –

These all have market values,

so, again, it's relatively easy
to look up at the values

or think about them,
or think about trading them.

But what about the elements,

the services that we get
from this ecosystems

that aren't necessarily material,

that aren't part or the structure,

but rather functions

of the greater structural complexity
of these systems?

That is – what is the value
of a forest as a forest

as opposed to the value
as a piece of lumber?

So, that's an important question
to think about.

So if we can think about things like,

forests absorb carbon dioxide
from the air,

thereby medicating greenhouse
gas emissions and climate change –

they produce oxygen
that we can breathe,

they retain nutrients,
like phosphorus and nitrogen,

as well as sediment,
keeping them out from water ways

to keep them clear.

They provide habitat for biodiversity

and they provide endless recreation
opportunities for us.

We can think of all
these kinds of things.

So, maybe we can get
at a lower bound for value

for a given forest,
if we try to add up

the individual contributions
of each of these different elements

to our well being.

So we can try and do that.

But now we are still faced
with a more fundamental problem,

which is that – we're talking
about questions of value,

and we're talking about value,

we're talking about
people's perceptions of worth,

which, been held subjectively,
vary widely across populations,

culture, generation, ethnicity,
any number of things,

we can think about these axes.

So, that means extremely,
extremely difficult to assign

blanket values, generalized values
to a given landscape,

because the services
that they generate are valued

by different people in different places
at different times.

So, that's the problem space
that we want to play with here.

So, if we're thinking
about this localization of the problem,

maybe a more important question,
or a different way to phrase this,

is not to ask ourselves,
or is not to try and say,

"The value of a service
from forest is x,"

but rather to say,

"The value of this service from this forest
is x to these people,"

and to get specific.

So, in that spirit,
for the last few decades,

our researchers
in ecosystems services area

have been traveling
around the world

and surveying people
about the values

they assign to the services of nature.

But, obviously these things
are time consuming

and they are expensive.

So it's extremely difficult
to get very much data here.

There are databases built up
on these things,

specially in the last decade –

we've started to see some
databases emerging that you can query

and try to get an idea
of what the literature says

about some of these
different kinds of values –

these, again, socio-economic values
that we're playing with.

But, we don't think even
remotely close to global coverage,

nowhere near.
And at the same time,

especially in the last five to six years,
we've seen a major upswell

in institutions of both
the public and the private sector,

begging for global coverage
of ecosystems service information

that they can use
for their land management decisions,

and run scenarios against.

So, as we've already seen, we do have
a lot of [due] a spatial data now.

That's kind of a new big fun thing
in ecosystems services world –

we're not just limited to doing
these one off surveys,

because we can actually do
these secondary

meta-level evaluations of the data.

And what we get to do with this –
we get all this geospatial data,

we put it together, and now
what we can do in filling these gaps,

is we can actually try
and create functions

that go in and study the structure
in the data

of the landscapes and the people,
the cities, the community centers,

the roads,
all these kinds of structures,

and try to pull out
with these signature functions,

where services are likely
to be produced,

and where there's probably
demand for them.

But, once you've applied
these kinds of functions,

you still don't necessarily know –

if you know where
the supply might be,

and you know where
the demand might be

in any given landscape,
once you've run these functions –

you still don't know
if any service is being delivered.

So, what we have to do there –

is we take the landscapes,
we project this information

about likelihood of supply and demand
up on to a network,

and then we start flowing around,
we simulate in our computers

across all this geospatial data.

We simulate the flow of,
what we call "service carriers",

so things like bees
for pollination services,

or carbon dioxide moving around,
water moving for flooding

and wild-fire, water supply,
water quality, any number of things.

You move it across the landscape
and you try actually see –

given any particular
topographic variables,

what is the service flow topology,
any given area,

and thereby, you can finally answer
the question:

Who receives services from where
in any given landscape?

And that's extremely powerful,
if you have that kind of information.

So, the kinds of things
you can answer with that, now –

– that's not too bad –

So, for example, you can finally
show maps like this,

where the green areas here are like

[repeated sound of] on the top,
and over here you have a mountaineer,

and we're looking at scenic views –

So the impact of scenic views
on different properties –

for the red, on the top there,
is the city of Kent.

And so you can try and see:

Who receives services from where,
and to what degree?

And the yellow stuff is visual blight –

You can actually look at the degree
to which individual properties

are being impacted in terms
of their service,

because of the way
landscapes configure.

And you can run scenarios against this

to try and actually, really answer
questions about

who wins and who looses
on different management scenarios?

So you say,
on a development scenario one,

this group of people gained something,

a different group of people gains
a little bit more,

and this third group of people
gets hurt.

Whereas under development scenario two,
if I develop in this area,

well, it turns out that everybody
gets hurt a little bit,

but if I develop in the third area,
everyone benefits.

This is really becoming
very interesting

in the US, in particular,
the EPA has this entire research divisions

entirely turned around
to ecosystem service research these days;

the US GS has a very
big program in that;

the US GA has an office
of ecosystem services and market –

just started off a few years ago
under this administration,

and so on and so forth.

So, our government's all into it.

We're seeing a lot of
ecosystem service work,

starting to find its way
into public policy in the EU.

And I'm participating in some projects
in Africa as well,

for the Gund Institute right now,

where this stuff is also
coming into play.

So, we're hoping that taking this kind
of technology

to finally connect people
to the landscapes,

the actual landscapes,
which generate their services,

will really help us to better inform,

better land management
in the future for all of us.

Thank you.

(Applause)