Landsat in Action - Forestry Research with Warren Cohen

Video Transcript
Download Video
Right-click and save to download

Detailed Description

Warren Cohen with the USDA Forest Service talks about the value of Landsat's long history of observations in monitoring and assessing forests.


Date Taken:

Length: 00:03:20

Location Taken: Sioux Falls, SD, US

Video Credits

Producer, Steve Young


I'm Warren Cohen, I'm with
the USDA Forest Service.

Forest health issues are
boiling up. Forests have

always sort of seen insects
and diseases sporadically

and forests sort of succumb
in some cases but they

always grow back. But there's
been a lot more physiological

stress on the forest due to
climate, and by looking at

time series, and the denser
the better, we can begin to see

those changes, because they're
multi-year. They take time.

When you start to look beyond
the easy stuff, there's a lot of

information there, but it can
be mixed with noise. And

trying to figure out time series
noise, it's interesting to work

there and try to separate out
the signal from the noise.

It's not easy. Especially with
automated algorithms.

But it is possible, and we're
starting to get there I think.

Landsat was launched and
there were observations

beginning at that time.
Between '72 and '84 there

was a huge spike in population
growth. The beginning of the

population explosion. All of
that is kind of in the Landsat

record back then, if you want
to try to get it out, and what I

want, is what most people I
think want to use Landsat data

they want to be able to use TM,
ETM+, OLI, Sentinel, without

worrying about the data from
there and time series, they're

not worrying about what sensor
it came from. Fortunately there's

overlap, coincident images
acquired on Landsat 5 in

particular, so you got these
two datasets that are very

different from each other,
acquired at exactly the same time

that you can begin to work out
statistical relationships for

how to relate those to each
other. And so that's what

we've been doing. There's a
number of issues to get

even to that point that have
to do with geometry and

things like that, but the real
spectral harmonization is the

part that I really wanted to get
to. And we've developed a 

system to do that and I think
it works fairly well.

The big problem you have with
taking a single image of Landsat

and try to do anything with
forest biomass for example

is it saturates very early at
canopy closure. And you might

have 70, 80, 100 or more years
after that where everything is

slowly changing. Spectrally
maybe not changing very much,

but the forest structure and the
biomass increasing and changing

very dramatically. If you know
the history, right, if something

was disturbed 20 years ago
by a large amount, you then

say, well that's got to have
less biomass than a stand that

wasn't disturbed. And we can
go back to 1972 to look at

that stability. Those things can
really help you to predict

forest structure with
Landsat data.