>> From the Library of Congress in Washington, D.C.

>> Good morning.
I'm Michelle Cadoree Bradley, a reference specialist in the Science,
Technology and Business Division of the Library of Congress.
Welcome to today's program on mapping water use from space
in which we'll learn about the measurement
of water use via applications of space-based technologies.
Today's program is the fourth in a series in 2012 jointly sponsored
by the Science, Technology and Business Division
and the NASA Goddard Space Flight Center.
Access to fresh water is a daily issue of life or death in many parts
of world, and it means healthy crops, wherever they are grown.
Water is central to the economies of U.S. western states
where legal battles over fresh water have become common.
Now that the means to objectively measure and map water use
with observations from space-based satellites has been discovered,
and as it moves from innovation to everyday operations, it can be used
for practical water monitoring.
Research scientists have learned to measure the exchange of water vapor
between the land and the atmosphere,
which is called evapotranspiration, ET.
ET is made up of all the water that's evaporated from surfaces
and all water used by plants in the process of photosynthesis.
ET monitoring let's us know how much water is left behind to do things
like grow crops, provide drinking water
or feed the regional streams and river systems.
The process of ET can now be monitored
from space making it an invaluable tool
where Earth-based monitoring is either lacking or unavailable.
It can help with monitoring drought, managing water,
planning for irrigation, and predicting crop yield,
all key areas for farming and natural resource management.
Our speaker today is Dr. Martha C. Anderson.
Dr. Anderson is a research physical scientist in the U.S. Department
of Agriculture's Agricultural Research Service in the Hydrology
and Remote Sensing Laboratory in Beltsville, Maryland.
Her research interests focus on mapping water, energy,
and carbon land surface fluxes at field to continental scales,
using thermal remote sensing with applications in drought monitoring
and soil moisture estimation.
Dr. Anderson received a B.A. degree in Physics from Carleton College,
Northfield, Minnesota, and a Ph.D. in Astrophysics from the University
of Minnesota, Minneapolis.
Her interests in environmental issues have steered her towards
research in Earth observation using satellite-imaging systems.
This has led to international collaborations
in monitoring water resources around the world.
She served as a member
of the Landsat science team from 2006 to 2011.
NASA has graciously provided literature and handouts
on this research, which are in the lobby.
There are also some related books from the Library
of Congress collections that are on display,
which we encourage you to look at.   If you need more information,

NASA's Jeannie Allen is also here to provide help.
Today, please join me in welcoming Dr. Martha C. Anderson.
[ Applause ]
>> Well, thank you, thank you for coming today.
So, yes, I'm Martha Anderson, and I work for the USDA Research Service,
and I'm stationed in Beltsville, Maryland.
I work in the Hydrology and Remote Sensing Lab, not too far from here.
And today I am going to be talking about monitoring
and mapping water use and water availability at scales
from individual farm fields, all the way up to the nation,
continents, and up to the globe.
We do this by looking at the Earth's land surface temperature.
Very much like with the human body, temperature is a good diagnostic
of how healthy the planet is and how much water it's cycling.
And surface temperature is something we can measure fairly accurately
from space.
But to begin with, before I get into that,
just a little background on me, why I'm here.
I did my Ph.D., actually, in astrophysics,
completely unrelated to what I do today.
I studied acceleration of cosmic ray particles
in galactic supernova remnants, University of Minnesota,
and I used a very large array of radio telescopes,
which is down in New Mexico.
I really like this, I like making maps like this.
This is the Cas A supernova remnant.
But somewhere in the course of my graduate studies,
it became pretty clear to me that I was much more interested in things
that are kind of going on here on Earth than way up there.
So through a series of events, I found my way into the field
of Earth-remote sensing, which is the study of using satellites to map
out properties of the land surface and the ocean surface,
and it's turned out to be the perfect solution to my problems.
It allowed me to look at environmental issues,
but still use some
of the observational skills I had acquired in graduate school.
You know, in a lot of ways, these two fields are kind of similar.
You're just kind of observing things looking down instead of looking up.
In both cases, it is just amazing how much we can learn about objects
that are very far away from our sensing systems purely based
on their radiative properties.
And that's what I hope to talk to you about today.
So in my role at the USDA, I studied the distribution and use
of fresh water around the globe.
And we know that water use, water security issues are going
to become more and more important given the current trends
in population growth and climate change.
Of all the water that there is on the Earth,
only .7 percent is available for human use.
Fresh water is not tied up in the icecaps.
And of that little sliver of water right here, about 70 percent
of that available fresh water is used to support -- whoops --
agriculture, and that's why you have somebody from the USDA here talking
to you about water use today.
We know that these fresh water resources are not uniformly
distributed over the globe,
so this is a graphic showing the per capita availability of fresh water
in the year 2000, and you can see there are some discrepancies
in where the water is.
The countries with the most per capita fresh water happen
to be Surinam and Iceland, and those with the least,
Egypt and the United Arab Emirates.
Now, one kind of concerning aspect of this graphic is that some
of these regions with low per capita fresh water availability also are
areas that are undergoing some ongoing political unrest.
So there have been international risk security assessments conducted
with fresh water availability being one of the key factors.
This particular study made a ranking of countries in order
of extreme water security risk.   And as the report notes,

some of these countries are already experiencing internal
and cross border tensions due to limited water resources,
countries like Egypt and Sudan and Pakistan.
We expect that these tensions are only going become more acute
as the climate continues to change,
and as some of these global water resources are starting
to be unsustainably drained.
Now, it is therefore, I believe, in our kind of national interest,
both from a humanitarian standpoint, national security standpoint,
to develop means for monitoring fresh water availability
across the globe.
We can't necessarily count on individual nations
to give us a good accounting of their water use.
Sometimes it's not in their best interests
to give a good accounting of their water use.
But more often than not, these data just are not available
in any accurate form.
And that is why our group in Beltsville,
and other groups internationally, have been working on techniques
for monitoring water use, or evapotranspiration,
as we call it in the field, using Earth-observing satellites,
because these satellites give us kind of an objective view
on our global resources
that sometimes we can't obtain through any other means.
So briefly, what I hope to cover in this talk is first of all,
what is this water use process we call evapotranspiration,
or ET for short.
How do we monitor ET, or water use, using satellites.
I'll give you some examples of some real life applications
for satellite-based ET, and then we'll just wrap
up with what are the future prospects
for water resource monitoring using satellite data.
First of all, evapotranspiration, what is this?
This is a hydrology term.
If you talk to a water resource manager,
they may use a similar term, consumptive water use.
And this is a little movie that NASA put together
to demonstrate the concept of evapotranspiration.
I'll see if I can fire it up here.
Okay, there should be a rain cloud appearing, there it is.
It's coming over the land surface.   It's raining on the land surface.

Some of this moisture infiltrates into the soil,
and some of that is taken up by the root systems of these plants,
and that makes them happy.   They were a little stressed.

Some of the water is evaporating directly from the soil surface.
Much of that water that's taken up by the plants escapes
through tiny pores in the leaf's surface
in a metabolic process called transpiration.
That's the crop water use component.
So you get evaporation from the soil,
transpiration from the plants, evapotranspiration.
So it's the net exchange of water
between the land surface and the atmosphere.
We know that ET is a major component of the global water budget.
About 60 percent of the water that's applied to the land surface
in the form of precipitation is believed to be returned
to the atmosphere through ET.
So if we want to understand the global water budget,
how it's changing in time, we have to find ways to monitor ET
at large global spatial scales.   But this ET data is also useful

at much smaller spatial scales, as well.
If we wanted to know what the water use patterns were
over the central valley of California, it might look a lot
like this, if you can see that.
This is a map made with the Landsat satellite.
So what do we have here?
These green patches are irrigated agricultural fields,
and they're using a lot of water, typically quantified in terms
of millimeters of water use per unit time,
in this case a whole growing season.
You can also see the background moisture availability or water use
that would exist in this area if humans weren't in there irrigating.
This kind of spatial information is very valuable to kind
of a broad range of stakeholders.
The growers clearly want to know how much water their crops are using.
They want to know when they need to reapply water
in the form of irrigation.
Water resource managers need to know this information
so they can plan the future of water distribution infrastructure
for the state of California.
And I'll touch on some of these applications a little later
in the talk, but first of all, how do we make maps like this?
How do we map water use using satellite data?
And as I mentioned before, the key is for many,
for most satellite-based ET retrieval algorithms,
the key is the land surface temperature.
I'm going to use the human analogy over and over again today.
As with the human body, a cooler body, or cooler land surface,
tends to indicate a healthier, better hydrated land surface.
So essentially, we're making these maps using different size
thermometers flying around on satellite platforms.

You may have some familiarity with thermal imaging
if you've ever had a home energy audit.
Has anybody had a home energy audit before?
We have. I know we have.
They may give you a picture like this
of your house showing you all the leaky windows that you need
to replace, because they're too hot, and this guy has got some kind
of issue with the insulation in his attic.
He's losing a lot of heat through his attic.
Using thermal sensors that operate in these thermal wave bands,
which are much longer than our eyes can detect,
we can remotely determine the physical temperature
of an object purely based on its thermal radiation properties.
So if we use those same kind of thermal imaging sensors to look
at a vegetative land surface,
you may see some pretty strong temperature contrasts.
Healthy, green, vegetation tends to be much cooler
than the soil background, as you can see here.
Now, why is this?   It's because of evaporative cooling.

This looks like it was a pretty dry field at this time.
It's very sunny.   The soil surface has dried out.

It's not evaporating a lot of water right now,
and it's achieving some pretty high temperatures,
these yellow tones you can see here.
That's like 50 degrees Celsius, 120 degrees Fahrenheit.
In contrast, the corn plants have these rooting systems
that they can use to mine water from deeper in the soil profile,
so they can continue to transpire water long
after that surface layer has stopped evaporating.
As with sweat, the transpiration helps to keep the plants cool.
So in some sense, the temperature of the soil
and the vegetation components of this scene is giving us a good clue
as to how much water is being used by each component.
And we can get a quantitative estimate of that water use
or that water loss if we run these thermal images
through a modeling system that computes the surface energy balance.
And this is the one piece of science I'm going to show here.
It won't be very long.
If we know how much energy is delivered to the land surface
from the sun and from the atmosphere minus the back radiation off the
land surface, we can use surface energy balance
to compute how much transpiration and evaporation must be going
on to keep the soil and the canopy, the vegetation, at the temperature
that we observe from the thermal imaging satellite.
Surface energy balance conservation of energy,
they're fairly simple models.
Now there are a number of satellites currently in orbit
that are giving us this thermal imagery that we need
to run these surface energy balance algorithms.
And I'm listing here a couple of these instruments
that we tend to use in our work.
And I've ordered them in order of the spatial detail
that these sensors provide, from very coarse scale sensors like GOES,
which can really only see, or distinguish features
on the landscape that are very big, like 5, 10 kilometers large,
several miles in dimension.
Down to much finer scale sensors like the Landsat satellite
that can see features as small as like a football field.
Now it so happens, unfortunately, due to instrumental constraints,
data collection rates and so forth that as spatial resolution
of these sensors increases, the temporal resolution tends
to decrease the rate at which it can acquire imagery.
And that's too bad for water resources.
We really like to have a sensor that saw things
at 100 meters spatial resolution and gave us a map every 15 minutes,
but such a sensor doesn't exist right now.
So we have to make creative use of all the different types
of sensors we have available to us.
At the coarse end of the spectrum here,
we have these geostationary satellites, and these are run
in the United States by NOAA,
and they're used mostly for weather forecasting.
The geostationary satellites orbit the Earth at the same speed
that the Earth rotates,
so it's always getting the same hemispherical view
of the land surface, and that's really very useful,
because we can acquire images really rapidly
in that kind of configuration.
We can kind of make a movie of how ET is changing
over the continent during the course of the day.
But because the field of view of the sensor is so wide,
it's not giving us a lot of spatial detail on the ground.
At any given time, there are typically two GOES satellites
operating over the U.S. We've got GOES West giving us a view
of the conditions over the Pacific Ocean, and GOES East that's looking
at the land surface and the Atlantic Ocean.
We typically stitch together ET maps developed from each of these sensors
to give us fairly decent coverage over both the North
and South American continents.
At the other end of this spatial resolution scale,
we have the polar orbiting satellites like MODIS and Landsat.
And these satellites give us a lot finer spatial detail,
but the temporal frequency isn't so good.
They give us a snapshot of conditions maybe once per day
or once per every a couple of weeks.
And this is a little movie showing the image acquisition strategy
of the Landsat satellite.
It's in a sun synchronous polar orbit, so it's passing
over the same patch of land at about the same time of day every time
that orbit repeats over that patch,
and that's very useful for modeling purposes.
It's got a really narrow swath.
It's going to take this satellite a long time to image the whole planet.
But you can see the amount of spatial detail it's collecting
within that narrow swath.   So here we go.

It's showing us how much coverage a single Landsat acquires during the
course of a single day.
It's going to take 16 days for this satellite to cover the whole Earth.
It's building up successive orbital swaths.
If we have two Landsats in orbit at the same time in staggered orbits,
we can get an image every eight days, and that's quite valuable.
And we had that for a while when Landsats 5


Now the MODIS satellite, very similar.
It's in a polar orbit, sun synchronous,
but its swath is much wider than Landsat.
So it can actually cover the whole Earth's surface about once per day,
but the spatial detail is not as fine as Landsat.
We can make these maps of evapotranspiration using any
of these thermal imaging sensors that I've discussed.
And if we kind of stack them up in order of scale, one might envision
in the future a kind of a Google Earth for water resource monitoring
where you can zoom in and look in great detail at the distribution
of moisture availability in regions of specific interest.
Here at the continental scale,
this is a map stitched together from GOES East and West.
And you're seeing the strong east/west gradient
across the United States in precipitation
and vegetation cover amount.
We have much more rainfall, much denser vegetation
in the eastern U.S. leading to these higher ET rates, or water use rates,
indicated by the greener tones.
And these areas happen also to be cooler than the West.
If we zoom in a little closer using this MODIS polar orbiting satellite,
start to see some hydrologic features emerging in the landscape.
You can see, if you look closely,
there's some enhanced ET along these major river ways here.
Zooming in even further, looking at Landsat,
Landsat has the unique ability to differentiate water use
on a field-to-field basis.
That is incredibly useful for agricultural applications,
and that kind of detail is lost with the course of resolution sensors.
We can also make maps of how water use is distributed
within a given field if we have very high resolution thermal imaging,
maybe collected by an aircraft.
And that was the case here, where we're looking
at the corn field growing next to a soybean field.
The corn is using a lot more water at this point in the growing season.
You can also see there's kind of a strange --
let me get my mouse back here --
this linear feature that's running through this area.
This was an old railroad bed that used to run through Ames, Iowa,
and it was removed a long time ago,
but the soils are still compacted locally around that railroad bed,
and the crops don't grow very well there.
So you've got local reduction in T, the transpiration component of ET,
and we can see that very clearly in this thermal imagery
and this these retrieved ET maps.
So you get a lot of spatial detail if you go fine enough.
There is use for ET information at all these different spatial scales;
all these thermal images are very useful to us.
At the continental scale, moisture patterns, exchange of water vapor
between the land and the atmosphere is really important
in determining the regional weather patterns.
So NOAA is interested in ET information
at this kind of spatial scale.
At the state, or basin scale, water resource managers
like this kind of information.
They need it to run their hydrologic models
and do their regional water use assessments.
At the Landsat scale, there is just a host
of the applications monitoring water use at the scale
that it's actually being used.
This is a zoom in on a part of Orlando.
It's around the Disney World region, actually.
You might recognize this.
This is the Seven Seas Lagoon where all the resorts are,
and Disney World is where -- Magic Kingdom is right over here.
And this is the big parking and transportation center.
You leave your car there, you get on the monorail.
We can see it's very hot, very dry.
It's a paved surface, not a lot of ET there.
There's another interesting set
of landscape features over here to the west.
These happen to be rapid infiltration basins that are used
to dispose of the treated wastewater that's coming off the Magic Kingdom.
Now, they don't -- they're used periodically
when there's an overflow.
They don't appear to be in use at this late time in the tour season,
because they are also very hot and dry.
This is the level of detailed water information you can get
at the Landsat scale.
Landsat has a long history of collecting imagery around the world.
The first Landsat was launched back in 1972, Landsat 1.
And I remember reading about this in my --
my parents had National Geographic.
I remember seeing this article, like how cool are these images
that they're making of the Earth these days.
The first thermal imager was added to Landsat 4 back in the early 80s,
so right now we have 30 years worth
of near global water use information unmatched
by any other water use data archive available.
With this Landsat thermal image archive,
we can study how water use patterns have changed around the globe
over the past three decades as these urban areas have expanded,
and as forests have been converted into agricultural land use
and sometimes into subdivisions.
Even though Landsat gives us a snapshot only once every 16 days,
an instantaneous snapshot every 16 days, or maybe longer depending
on cloud cover conditions, scientists have figured out ways
to interpolate water use between those overpasses,
so we can get time continuous water use information.
For example, in our lab, we've been working on a technique
for fusing information from multiple different thermal imaging systems,
and this is work that's been done by several people in this room,
and I thank you for coming.
We use the GOES data, the geostationary data,
to derive the daily curve in the evaporative flux
to get the daytime total flux.
MODIS gives us a little bit higher spatial resolution snapshot once per
day, but it's the Landsat satellite that really brings us
down to water use features that are recognizable on the land surface.
What this algorithm does is it pairs MODIS Landsat image pairs on days
when they're available, and then it uses that spatial information
to disaggregate the MODIS images on all the intervening days.
So we end up with a time continuous mapping of water use


So what can we do with all this data?
We can collect all this water use data.
What are the applications?
I'll give you some examples in the areas of water management
in the western U.S., in monitoring drought and crop conditions,
and then also some more international applications in areas
of food and water security.
But first, to the western U.S. There are several states
in the western U.S. that are actively using Landsat ET
to inform better water management decision making right now.
Many of these states are depleting their water resources
at unsustainable rates, primarily to support irrigated agriculture.
They need to have the information on their water use.
And some of the examples I'll show in this section were generated
by Rick Allen, who's the researcher at the University of Idaho.
He's done this in direct support of operational needs
by the Idaho Department of Water Resources.
As my friend, Tony Morris, who used to work with the IDWR likes to say,
we can't manage what we can't measure,
and Landsat gives us the opportunity to measure water use
at the scale that it's being used.   Here's an example of this.

This is a map over the Snake, part of Snake River plane
in central Idaho, and each of these black boxes here, these polygons,
are associated with a water right, the maximum amount of water
that that right holder is allowed to extract
from the community groundwater pool.
And then all these little blue dots are groundwater wells
that are extracting the water to use
for irrigation, or maybe other purposes.
The IDWR's job is to figure
out whether these rights holders are complying with their rights.
Clearly, they can't use more water than they are allowed to, but also,
according to western water law,
it's kind of a use it or lose it situation.
So if the water is not being put to beneficial use,
that right can become under jeopardy.
So this is a clear example of the need for both historical
and current water use information at the scale of individual fields.
And this is a job that is uniquely well suited
for the Landsat thermal imager guide.
In the past, how IDWR used to do this is one technique was to look
at electric power meter records.
Essentially, they'd associate the electricity used by that pump,
to individual pumps, to an estimate of the total volume of water
that was pumped over some period of time.
Oh, great, I'm going -- thank you.   Security. Cancel, cancel, cancel.

All right.
So, you know, this was a really costly way of doing this.
It involved a lot of on-the-ground work,
often going around visiting different sites,
reading their power meters.
And it was also getting some fairly low accuracy estimates.
You can get some crazy seasonal water use estimates
through this technique, kind of well outside the range
of what one might expect from some normal cropping systems.
But in contrast, they found
that Landsat ET could be processed very cheaply, relatively cheaply.
The USGS is distributing the Landsat imagery free of charge,
so it's mostly just labor costs in interpreting the imagery.
They also found that it provided much higher accuracy.
The seasonal water use estimates were pretty well constrained
within the limits of what one might expect for these different crops,
and then you can just go and target some of these outliers
who seem to be using too much.
Idaho, other western states, are also using Landsat ET data
for other applications for resolving water rights disputes,
for deciding what to do in the case of groundwater call.
They're using it to negotiate and monitor interstate water compacts,
a lot of applications that Landsat is currently being used for.
Why do we need Landsat for this?
Because this is not going to hold up in a court of law.
There is no way to associate water use
with an individual water use user at this kind of spatial scale,
but we can do that at the Landsat scale.
The scale is very critical for these applications,
and Landsat is currently the only satellite providing thermal imagery


So Landsat ET, not just useful for water administrators,
also can be very useful for the growers, as well.
They can use these maps to kind of track
out how much water their different fields
and their landholdings are used.   Maybe use it to schedule irrigation,

when do different fields need to be irrigated.
Particularly critical in crop growing regions in semiarid parts
of United States, and this is an example.
This is an ARS facility in the Texas Panhandle outside
of Amarillo, Texas.
The Ogallala Aquifer, which feeds the groundwater for this region,
is declining very rapidly due to over extraction
for irrigated agriculture.
They have to use their water very sparingly and very wisely.
They need information on how to do this.
I made a little map here, or a movie actually,
of daily ET over this agricultural area that I'll show you.
You will see a couple of fields.
This is an unirrigated cotton field in the foreground
of this picture growing side by side with an irrigated cotton field
for experimental purposes.
And you'll see these showing up in the movie.
So I'll try to launch this.   Here we go.

To begin with, you're going to see some strong pulsing
between dry and wet and dry and wet.
And this is mostly due to differences in cloud cover,
clear days with a lot of ET, cloudy days with lesser ET.
But you're also going to see some shifts
in the management going on, as well.
This southern part of this pivot just turned on.
You're going to see this, this southern part turn on, too.
This is a big concentrated animal feeding operation,
and in the center is a huge manure lagoon that's evaporating a lot
of water.
This is a residential area of Bushland, Texas,
and the residents are likely irrigating, or watering their lawns.
This is this irrigated cotton field next
to the unirrigated cotton field, and you can see the great difference
in the cumulative water consumption
by these two different management techniques.
But I'll tell you, this unirrigated field did not do well
in this particular season.
The plants were stunted; it was very patchy.
Irrigation is critical to getting a good crop in these climates,
but the water has to be used as wisely as possible.
Now, our group in Beltsville has recently started up a collaboration
with some research scientists at the Ernest
and Julio Gallo Winery in California.
And these guys are also very interested in Landsat ET
for managing their irrigation scheduling.
But in this case, the issues of subtle differences in water use
from field to field and in crops for us is much more important
that it might be for a corn grower in Iowa or Nebraska.
We know that the best wine grapes are grown when the plants are
under some degree of water stress, but not extreme water stress,
so they're very, very carefully applying deficit irrigation
techniques in this region.
Very difficult to manage over the 200,000 acres
that these guys have to manage.
And that is why they are so interested
in these satellite ET retrieval techniques,
especially at the Landsat scale.
They can see the differential water use from field to field,
which may be due to differences in the soil type, maybe it's due
to differences in the kind of grapes that they're growing.
Even row direction can influence the water use.
Hill slope, microclimate, these are the types of features
that we can capture through these satellite-derived ET maps.
The final example in the western U.S., we have to be not only worried
or concerned about water use by agriculture, but also water use
by natural species, as well.
In this case, it's an invasive species.
And this is a map of water use
over the Rio Grande River Basin south of Albuquerque.
And a lot of this is irrigated agriculture out here,
but right along the river bed, that is consumptive use
by riparian vegetation, mostly cotton, woods, and willows,
and more recently, the salt cedar, or the tamarisk.
Now the tamarisk, salt cedar was introduced into North America
as a hardy ornamental, but it didn't have any natural predators.
It started to spread, and really, estimates of water use
by salt cedar were kind of alarming.
They appeared to be just giant water hogs, and people got concerned
that they're going to suck all the water out of the river systems.
So there was this campaign of massive salt cedar eradication
that was mounted in some regions, including the release of a beetle
that is known to defoliate salt cedar.
So this program has been going on for several years.
There has not been a significant demonstrated reduction
in consumptive water use after this program has gone on.
It turns out that the early estimates of water use
by tamarisks were probably way too high.
And this beetle has now spread to other riparian basins
and it's causing some environmental damage.
We really need to have better information about water use
at the scale of individual land use patches, and again,
critical for Landsat to be able to resolve that riparian vegetation
from the surrounding landscape.
Okay, I'm going to shift gears just a little bit.
We've got these time sequences of maps of water use or ET.
We can start to mine these data cubes in space and time,
try to identify places and times
where the water use was anomalously low indicating drought.
These ET maps can also serve as a kind of unique
and very valuable drought monitoring technique
and crop condition monitoring technique.
I mean, I've seen maps like this before.
This is the U.S. Drought Monitor, and it's the major record
of drought events over the United States.
It's updated every week, and if you want
to see what the latest conditions are,
you can look at this URL or drought.gov.
The drought monitor is being used extensively for many,
many different types of applications,
and many of these applications are not commensurate
with the spatial scale of the data that's provided in this monitor.
Those contours are very coarse, but they're being used
to determine how compensation is distributed, you know,
due to yield loss at the end of the year.
The Risk Management Agency, which is the arm of the USDA
that operates the Federal Crop Insurance Program, is always looking
for higher and higher spatial resolution drought information
so that they can verify that these payments, these insurance payments,
are going to the people who were really impacted.
And this is another area
where I think satellite-derived ET can play a major role.
In our group, we've developed an ET-based drought index.
And this is showing a map made with these GOES satellites.
The red areas are areas of anomalously low water use, or ET.
And indeed, the southeast
of the U.S. was experiencing an exceptional drought during 2007.
And this is a picture of Lake Lanier,
which is the major reservoir supplying water
to the city of Atlanta.
By October of 2007, that reservoir was down to a three-month supply
of water for Atlanta, and people were really starting to panic.
There were also threats of water war outbreaks between the states
of Georgia and Alabama and Florida, or some disputes
about water allocation and just some shared river basins.
Things were kind of bad in 2007 in the Southeast.
The satellite-derived ET map gives us much higher spatial resolution
information about how that drought was distributed than some
of the standard drought indicators that are currently used
to create these drought monitor reports.
So over here, this is a map of soil moisture anomalies
that were computed based on ground base measurements of rainfall
from the Doppler radar and from the ground-based Rain Gauge Network.
The reliance on ground data, which has been kind of standard so far,
has limited the resolution of these drought mapping products.
With the satellites, we're not constrained
by the ground measurements.
We can get much higher spatial detail.
It's reproducing those patterns and the rainfall deficits pretty well,
but we're not using the precip data.
That also makes these maps much more portable to other parts of world
where there aren't, you know,
Doppler radars every couple hundred of kilometers.
Here's another satellite-derived drought index.
This is soil moisture data retrieved using microwave remote sensing,
which is even longer wavelengths than the thermal band.
And you can see that that, in that wavelength,
the drought signal is completely lost in 2007.
And this is because, this is an area of very dense vegetation.
There's a lot of forest cover down there.
But water contained in the forest canopy absorbs the microwaves.
This is why your microwave oven works.
Water absorbs microwaves.
So the signal is being lost in this dense vegetative canopy
in the microwave wavelengths, but we're seeing it really strongly
in the thermal wavebands.
We're getting a different picture depending
on what wavelength we look at.
This is because, you know, the crops are stressed,
the forest is stressed.
They are shutting down their transpiration.
The reduced evaporative cooling due
to transpiration is causing the leaf temperatures to elevate,
and we're detecting that elevated temperature signal using our


So we're working colleagues at NOAA to put
out these ET-based drought maps in real time.
We have a web-based delivery system
that will be going live within a month or so.
We hope to expand this very quickly to include all of North America
and Central America, as well.
We're also working on techniques to combine this kind
of coarser scale information that we get
from the geostationary satellites about drought with finer scale data
from some of these polar orbiting systems, like MODIS and Landsat,
to try to get down to stress conditions at a farm-to-farm level.
And here's another example
when we can resolve riparian ET from the background.
That's the kind of information that's really needed right now
by the action agencies.
And all these types of monitoring techniques
that I've been talking about, we've tested them very well
over the United States,
can be applied globally wherever we have this thermal imagery.
And, in fact, it probably has more value in some
of these other countries where they don't have the dense ground-based
monitoring infrastructure that we have here
in the U.S. We've recently imported all these tools to Europe
and Africa using geostationary satellites
that are operated by the Europeans.
The meteo satellite, operational satellite,
sits at 0 degrees longitude over the Equator,
and it gives us a really nice full face on view
of the African continent.
One of the projects we're working on is
to use these Meteosat land surface products, again,
mostly land surface temperature -- you can see a little movie here --
to do a kind of detailed water balance analysis
over the Nile River Basin.   And this is a very interesting case.

It's a trans-border, large river basin
where multiple countries are dependent
on water coming out of the same source.
You've got the White Nile that's originating
in the Equatorial Lakes Region.
You have the Blue Nile that's draining much
of the Ethiopian Highlands and feeding into the main stem Nile.
But all along the course of these different segments of the Nile,
water is being diverted to support irrigated agriculture,
and we need to know how much water is being diverted
to do the water budget analyses.
For example, in the Nile Delta where the Nile in Egypt runs
into the Mediterranean Sea, there's a gigantic irrigated zone there
where they're using a lot of the water.
Also in Sudan, there is a big irrigated area, the Gezira Scheme.
You can see it in detail using Landsat.
You can see these intricate, elongated networks of fields
that have developed along these irrigation canals.
We need to be able to quantify this water use, this diversion.
Water use by upstream countries is going to affect the water supply
for the downstream countries.
So really, the only objective source of information about what's going
where may be coming from these satellite products.
This is a map of seasonal water use over the Nile Basin,
and this is the index that we use in the drought monitoring.
And in both cases, you can see there is again a strong gradient
in moisture conditions as you move from the Sahel, the grasslands
of the Sahel, up into the Sahara Desert.
You can also see the additional rainfall
that the Ethiopian Highlands receive relative to the surrounding areas.
A couple of other features of note again, the enhanced ET we see
over at the Nile River Delta.
There's also a big wetland in the White Nile called the Sudd,
and it's a major sink of water along the Nile, very important
to quantify how much water is being lost there.
Difficult to do, but I think these satellite tools are going
to give us a means for estimating water loss there.
So here's just a quick movie on changing water use patterns
over this -- whoops -- there we go, water, this Nile Delta.
You see as the rains come this wave of green vegetation moves northward,
and, you know, the highlands of Ethiopia are greening up
and then it also recedes in the fall as the rains go away.
You can also see cycles of water use in the Nile Delta related
to the cropping cycles and the supply of water from the Nile.
Very useful information, temporal and spatial.
We've applied our drought monitoring tools over this region
and we see a very clear hotspot and anomalously low water use,
water availability, coinciding with the hotspot of the famine in Somalia
and the Horn of Africa of 2011.
The hope is that with the increased spatial resolution that we can get
from these satellite products and the timeliness,
because we can create these products almost, you know,
that night for the previous day, this may help in the future
to better target relief efforts, mobilize efforts earlier.

We've been kind of test driving some of these monitoring techniques
in a case study over the Blue Nile Highlands, the Abay Highlands.
And this is in collaboration with Ethiopian scientists
and U.S. scientists in an effort to promote better climate resilience
in some of the very vulnerable agroecosystems that exist
in the Blue Nile Highlands.
So they get a lot of rainfall there, but they're still,
these rural subsistence farming communities are faced
with a lot of challenges.   The soils are highly erosive.

You can see right here some of the critical erosion that's going on.
There's also significant land use pressures causing them to farm
on unsuitably steep hill slopes, and oftentimes, look at this.
They're farming right up the edge of a gully.
This is exacerbating the erosion problem.
The erosion carries away the rich, good topsoil,
and what's left behind has much lower water-holding capacity.
It makes these soils much more susceptible to flood and drought.
So it's a big problem.   Then all that sediment gets carried

into the Blue Nile, which isn't very blue.
It gets transported downstream.
It's silting up downstream dams and downstream nations causing,
you know, significant costs for dredging annually, another aspect
of transporter water management problems.
So the goal of this case study is to try to develop a suite
of information products that is really tailored
to this Blue Nile Highlands area that might provide some information
for better land management, better adaptation
to changing climate conditions.
If we have soils mapped and maps of hill slope,
or digital elevation models, along with kind
of natural moisture availability maps,
we can map out erosion potential.
These types of maps can be then used by the communities
to better plan their land use.   Where should we grow crops,

where shouldn't we grow crops to minimize erosion.
What diversity of crops may be best suited for different parts
of the landscape that may be more resilient to the range
in climate conditions that they're being subjected to now.
And then we also have flood
and drought monitoring capabilities with the satellites.
[Inaudible] had very, very high spatial resolution allowing these
communities to take some precautionary measures.
So these communities are remote, but there are a lot of cellphones
out there, and there's good political infrastructure,
good peer networking, so there's a very good kind of route
for this information to get disseminated to the actual farmers.
It's a participatory project, so we're hoping
that by engaging the communities, we have a better chance
of making products that are relevant to their needs.
And again, the spatial scale of the information is going to be critical.
You need high spatial resolution to match the variability
that they're seeing in the landscape and match the scale
at which these decisions are being made.
Again, that's the scale of the Landsat data.
Okay, future prospects, we'll wind up here.
I tried to make the case that satellite imaging is useful
for monitoring water resources at different spatial scales.
What are our prospects for water resource monitoring
from space moving forward.
Now, at the coarse spatial scale, I think the prospects are pretty good.
There's geostationary coverage of most of the landmass around globe,
and these weather satellite programs tend
to be relatively, consistently funded.
We always want to know what the weather is.
We are developing global ET
and drought monitoring products using this international suite
of geostationary satellites, and we hope to have this automated
and broadcast on the web in a routine fashion.
On the other end of the spectrum, at the very fine scale end
of the spectrum with Landsat,
our prospects are not quite as clear at present.
We have the Landsat Data Continuity Mission,
which is due to launch in January of 2013.
And so that's the next Landsat, Landsat 8,
and we're looking forward to that.
We hope that all goes off as planned.
We have a bit of a data gap right now.
Landsat 5 and 7 both are kind of in a failure mode,
so this growing season is a little touchy.
We really want to get something back up there
so we can start monitoring it again.
After LDCM, however, it's not clear, at present, what's happening
with the Landsat program.
There has been an ongoing effort to shift management of Landsat
from NASA, which is more of a short-term mission,
research mission agency, to USGS for longer term operational management
with NASA still involved in the satellite development,
but this transfer has not been going smoothly;
it's kind of stalled out present.
And then we top this off with our budgetary issues.
The funding for the Landsat program is going right
down from an FY-12 budget request of 48 million for development
of the follow-on Landsats, of which 2 million was actually allocated,
and that for studying kind of the range of alternatives to Landsat
that might exist out there.
And to FY-13 with a request of only $250,000, again,
for just assessing Landsat alternatives.
None of this bodes very well, given that the typical launch window,
from funding to launch of a Landsat satellite is
about six to seven years.
So even if the program for Landsat 9 was fully funded next year
by some miracle, we're not going to get a launch until well
into the 2020s, and this is well beyond the design life of LDCM.
So we're looking at another potential extensive data gap.
So there's some [inaudible] growing
in the water resource monitoring community based
on this uncertainty in the Landsat program.
And this is a cartoon that was drawn by the son of a colleague of mine
out in New Mexico, so he's got, you know, Earth is sick in bed
and he's saying, you know, "Dr. NASA, can you take my temperature,
I'm feeling kind of sick."
And Dr. NASA says, "Sure, I'll take your temperature
with my Landsat," which is right here.
And he says, "Okay, can you take my temperature."
"Well, you know, maybe we'll take your temperature in 2020,
maybe 2025, we'll see when we can get around to this."
So, I thought that was pretty good.   Given the kind of broad array

of water resource monitoring applications and other types
of resource monitoring applications that have developed
around this Landsat data archive, there is some urgency
in finding this Landsat program a good home pretty quickly
and starting work immediately on developing the follow-on Landsat
so we can maintain this continuity of data collection
that we've had over the past years.
Again, I'll end with my message, we can't manage what we can't measure.
With that, I'll stop, and if there's time for questions,
I'd be glad to answer them.   [ Applause ]

>> Please, just repeat the questions after they've asked
so that they can be captured fully, and go ahead.

>> Okay, right here.
>> You mentioned that, I think Landsat 5 and 7,
were kind of in a failure mode and Landsat 8 hasn't launched yet.
>> Yes.
>> What time period do we have data [inaudible] 2007
to 2010, that's coming in, or --   >> Yes, yes.

So it was kind of the end, and Jeannie can correct me
on the exact date, but it was the end of last year
where Landsat 5 stopped collecting routine data, and they either have
or will turn on another instrument, the MSS, right?
>> Yeah, but we hasten to say that Landsat 5 was designed
for a three-year lifetime launch in 1984.
>> Yeah, it's well, it's done its time.
Yeah, yeah.
>> So if you're interested in, say, California drought data
from 2007 to 2010 [inaudible].   >> Yes, yes.

Landsat 5 was operating during that time.
Landsat 7 is still operating, and it was,
it's been up there for a long time, too.
But it has an issue, it's got the scan line corrector issue
where it collects full data in the center of its swath,
but there are these little Venetian blind gaps off to the edges.
So you just kind of have to hope that your site is
in the center of the swath.
And you can piece together information from Landsat 7,
but it's not a very satisfactory mode of operation.
>> And again, that one was launched in 1999 with a five-year design.
>> Yeah, those instruments have really lasted.
We cannot necessarily count on that for LDCM, that would be great,
but we can't count on it.
>> What other inputs do you use for your model
for determining evapotranspiration
for land surface temperature this time of year?
>> That's a very good question.
So the question is what other information do we need
to compute ET other than just land surface temperature.
Anything that affects the exchange of water between the land
and the atmosphere, and the primary inputs are amount of vegetation,
because that kind of determines the transpiration component of ET.
We also need to know the solar radiation, how much, you know,
what the solar inputs, and we can also derive solar radiation very
accurately using geostationary satellites.
We need to know wind speed, because wind determines the turbulence,
and the turbulence is what drives the eddies
that bring the moisture off the land into the atmosphere.
We need to know surface roughness.   Rougher surfaces transfer water

and heat more efficiently than smoother surfaces.
So something like a map of vegetation height would be useful.
Those are the major other inputs.
The other, the meteorological inputs we can derive
with pretty good accuracy just from some kind
of weather analyses frameworks.
We don't need to necessarily have very detailed measurements
of those properties, and so land surface temperature we need
with very high spatial resolution and good accuracy.
>> Are the details published in [inaudible].
>> Absolutely.
And I can direct you to some papers, if you're interested.
Very good.   Any other questions?

One more question right here.   [ Inaudible question ]

That's absolutely right.
So the question is what if there are some natural disasters that are kind
of bringing debris into the atmosphere that are kind of,
maybe the primary impact would be to reduce the solar radiation load
on the land surface, so we would have to accurately know how
to model those impacts on the solar radiation inputs.
Whether we can do that accurately, there are models to do
that to monitor aerosols and volcanic ash and compensate
in the radiated transfer models for those aspects.
It's something we would really need to test what are the impacts
on our accuracy under those kind of very unique events.
But I think the main impact would be on the solar radiation load.

Back here?
>> [Inaudible] the size of some of these systems [inaudible] elsewhere
in the world they're often much,
much smaller than in the Western United States, for example.
How do you deal with that when you're looking at [inaudible].
>> That's a good question.
So the question is, there are fields in other parts of world that tend
to be much smaller size than the fields we have in the United States.
We're actually relatively large.
So Landsat at 100 meters resolution is very adept as resolving many
of our fields in the United States.
It's going to be too coarse for a lot of the cropping systems
in Europe and in these,
the Ethiopian Highlands images that I showed you.
We do have techniques for sharpening up the thermal imagery,
and the person who's doing that in our lab is sitting in the second
to the last row back there, Dr. Feng Gao.
We can use the shortwave bands that are collected
on the same instruments,
so on Landsat it's collecting thermal imagery at about 60
to 100 meters resolution,
but the shortwave data at 30 meters resolution.
We can use that shortwave data, which is related more
to vegetation amount and other factors,
to sharpen up the thermal imagery to 30 meters.
That gets us a little closer to the target scale we need.
Ideally, we would have thermal sensors
that had 10 meters resolution,
and we're collecting daily all over the globe.
This would be ideal.   We have to do what we can.

We're not going to resolve every water use feature that we want to.
Right here.
>> You talked about the design life of Landsat programs,
but are there other alternatives that are being discussed with regard
to completely private satellites and/or relying on information
from other nations, are those being considered.
Is that partly why this funding has dropped off?
>> So the question is, besides Landsat,
are there other alternatives
for collecting this high resolution thermal imagery,
including private companies running the satellites in other countries.
So Landsat kind of went through a period of private operation,
and it wasn't a good time, I don't think, in the history.
>> That was, it was attempted.
Landsat 6 was designed by a private company, and it was the only one
in the series that failed to achieve orbit, as they say.
And there was a study done before Landsat 8,
the Landsat Data Continuity Mission on whether
or not private companies could manage this, and it was determined
over 12 years that no, they couldn't.
So this is not an easy question for us to answer
in the position we're in.   >> Right, right.

One of the issues is that, you know,
for a while the Landsat data were being -- there was a charge,
a fee for the Landsat data, and it was quite hefty.
And in fact, it was too hefty for most researchers to do any research,
a lot of detailed research.
We certainly couldn't build up these dense time data cubes if we had
to pay $6,000 for every scene we acquired.
The research was not going to get done.
The science was not going to get advanced
under that mode of operation.
Since USGS opened up the Landsat Data Archive,
the research has just exploded,
and all sorts of new phenomena have been detected and new types
of monitoring tools have developed.
So from a researcher's standpoint, I'm very happy
to have this open access to the data.
Whether another country will put up a satellite of the quality
of the Landsat, I mean, hopefully, that would be one way
to increase our temporal resolution.
If we had Landsat, and there was another country
that flew another similar satellite in a different orbit,
we could improve our data collection frequency.
That would be ideal.
I have not heard of Landsat-type data in the near term
from other countries that would be freely available to us
in the U.S. That's another issue.   And would be consistent [inaudible].

Right.
>> Does it, do private companies, does it exist, though,
privately, for a fee for others?
>> So the question is, are thermal data available by fee
through companies, and you know, I think there is a bit
of an industry developing in the western states,
people with aircraft flying over fields
and collecting thermal imagery
because the growers are recognizing the value of these data
for managing their crops.
It's really expensive for them, you know.
I think Gallo, you know, has thought about this, too,
but it's really prohibitively expensive.
So if we can develop some routine techniques
at somewhat lower spatial resolution,
but something that can be processed automatically as soon
as the imagery comes down,
this could be a service that's could be provided broadly
over the whole western states.
But yes, there are companies that are providing this data
at very fine spatial resolution from airborne, yeah.
And there may be some plans out there
for some commercial thermal satellites, as well.
>> I'm so sorry.
I encourage you to come down and speak in person.
But we're going to wrap up this session for today,
and I want to again thank Dr. Anderson for being with us.
>> This has been a presentation of the Library of Congress.