“Earth Science Vision”
Remarks of NASA Associate Administrator
Dr. Ghassem R. Asrar
International Geoscience and Remote Sensing Symposium
July 24, 2000


Good morning.  It is a pleasure to address this group today.

I would like to commend this organization and its members for the 
role they have played in the growth and change of remote sensing.  

In the twenty years since this organization had its first symposium, 
remote sensing technology and applications has evolved from the 
Nimbus 7 spacecraft that collected kilobits of data to multi 
spacecraft systems collecting and processing gigabits of scientific 
data.  We are now deploying the Earth Observing System and a 
series of small exploratory missions to characterize all the major 
interactions among components of the Earth system.

This is Step 1 in our great endeavor to characterize, understand and 
predict Earth system change. We want to know how the Earth 
system is changing, and what are the consequences for life on 
Earth.  Our vision and goal is to enable reliable prediction of 
climate, weather and natural hazards, and to see these predictions 
broadly used in commerce and in society.

This vision, and NASA’s plans to get there, is the subject of my 
remarks today. And we are dependent upon the scientific, 
technological, and operational services expertise of the IGARSS 
member communities to make it happen.  

First, let me illustrate with a few examples what I mean by Earth 
system prediction in service to society.  Currently, our national 
capability to predict hurricane landfall is +/-400Km 2-3 days out. 
Plus or minus 400Km is a span from Jacksonville, FL almost to 
Norfolk, VA…that’s a huge uncertainty. For Hurricane Dennis, we 
missed badly on the call for evacuation, costing the US economy 
$100 million.  And that’s just one hurricane!  But experiments that 
combine ocean winds with precipitation data indicate that we can 
do a lot better. We think that the observing and modeling 
capabilities we are planning for the next decade can reduce that 
uncertainty from 400Km to 100Km; from Jacksonville to 
Savannah, GA.  In 25 years, we’d like to be at 30Km with 5 days’ 
notice.  

On a larger scale, consider the potential of short-term climate 
prediction. El Nino-generated weather events are of particular 
interest to both scientists and citizens worldwide. The costs 
associated with the eight-month period of weather disturbances 
attributed to the most recent El Nino include almost $3.0 billion in 
property losses, $400 million in federal government relief, and 
over $650 million in agricultural losses.   As part of the on-going 
Earth Science program we are looking at climate conditions for 
several years prior to the last El Nino and entering those conditions 
into a model that combines ocean height, temperature and surface 
wind data.  We’ve already shown that the signs were there 15 
months in advance.  That system could help us to develop a tool to 
predict the timing and intensity of future El Nino events.  Just 
imagine what that would do for farmers, fishermen, or trans-
oceanic shippers.

Improved prediction of climate changes will have profound impact 
on crop forecasts and energy usage predictions.  For example, 
farmers could choose drought resistant seed instead of mildew 
resistant seed, and cities could base their snow response budgets on 
accurate seasonal forecasts.  Improved climate predictions could 
save upwards of $6 to 10 billion annually.


In the area of natural hazards emanating from inside the Earth, 
earthquakes and volcanoes have proven extraordinarily hard to 
predict.  But even there, progress is possible.  We can now 
accurately characterize volcanic activity, which is the first step 
toward a predictive capability.  In 20 years’ time, we might be in a 
position to see reliable forecasts of eruptions 5 days in advance.  
Our network of GPS receivers have documented the rate at which 
Los Angeles is migrating northward (no reason for them to run out 
and buy parkas just yet!).  And our experimental models are 
showing some limited success at 20 year timescales for regional 
earthquake prediction.  In 20 years’ time, we could see 1 to 5 year 
earthquake forecasts at the neighborhood level.  No promises on 
this one.  But we won’t make progress without a vision and a goal 
to work toward.  Our ultimate vision for natural hazards is to be 
able to predict surface deformation events with the same 
confidence that we can predict this week’s weather today.

That is what I mean by Earth system prediction in service to 
society—the ability to provide information to people, to cities, to 
businesses that they can act on to save lives and time and money.

How will we get there?  By the means I described earlier; the 
pathway of characterizing, understanding, and predicting.  We are 
doing all three of these all the time, acting on the knowledge we 
have at any point in time.  But the vision I have illustrated above –
to know how the Earth is changing and what are the consequences 
for life on Earth -- requires characterizing, understanding and 
predicting on a 20 year scale.  

Again, an example should serve to illustrate what I mean.  Today, 
reliable weather prediction extends 3 to 5 days.  This is enabled by 
weather satellites literally watching weather systems move and 
evolve, and using those data to initialize relatively crude numerical 
models.  In the EOS era, we will be able to demonstrate a much 
refined characterization of weather and climate-related parameters.  
We will make global temperature and humidity measurements 
within 1º at 1km resolution.  We have begun providing ocean 
surface winds measurements at 25km resolution globally, enabling 
NOAA to correlate storm energy vectors to storm movement, and 
enabling us to view storms in their climate context.  TRMM is 
showing us how energy is distributed and released in the 3-
dimensional structure of a tropical storm.  In the next few years, 
we hope to be able to make the first space-based measurements of 
the vertical profile of winds in the troposphere, and the 3-D 
structure of clouds.  

Combining all these data, we will greatly deepen our 
understanding of how the atmosphere works. Not only will we 
have better data to initialize numerical models, we will have better 
models.  We will quantitatively understand the air-sea interactions 
and deduce from measurements the amount of water evaporated 
and thus be able to predict the amount of water to be precipitated 
down wind.  Better data and better models, that is, better 
characterization and better understanding, mean more accurate and 
more reliable prediction.  It is this combination of tools that will 
yield the 10 year goal of  7-10 day weather forecasting, and the 25 
goal of 14 day forecasting.

All three steps of characterizing, understanding and predicting are 
powered by geospatial, computing and communications 
technology.  And so the Earth Science vision is also a technology 
vision.  At NASA, we hope to spark a revolution in geospatial 
technology, and to leverage industry’s revolutionary advances in 
computing and communications.  

In the realm of geospatial technology, the revolution we want to 
forment at NASA is focused on sensors and networks of sensors.  
Most of the sensors we fly today are passive; they take in reflected 
light through lenses and spectral filters.  And we’re learning a 
tremendous amount from them about the total composition of the 
atmosphere and the surface. But the next step in understanding the 
Earth system requires the ability for 3-dimensional view – to see 
the vertical profile of the atmosphere and the shape and 
deformation of the surface.  And that means active remote sensing, 
with radars and lidars.

The great promise of new radar technologies was demonstrated 
this past Spring by the Shuttle Radar Topography Mission. In just 
ten days a spaceborne radar interferometer provided us with a 
nearly complete map of the Earth’s land surface. A task which 
millenia of explorers and surveyors could not complete with the 
efficiency and accuracy of a single spaceborne mission. 
Spaceborne Interferometric Synthetic Aperture Radar is now 
capable of providing us with a day by day view of the slow, minute 
but so important accumulation of strain during the earthquake 
cycle or previously imperceptible inflation of volcanoes due to the 
flow of magma deep within the Earth. We can now observe in a 
single INSAR image detailed images of the deflation of 
underground aquifers beneath cities such as LasVegas, Los 
Angeles, or Houston.  This information will certainly lead to a 
more informed use of our precious water resources. Thanks to 
these new technologies and those to follow, the next generation 
may be able to prepare for the occurrence of an earthquake or 
volcanic eruption with the same certainty that we now anticipate 
equally destructive hurricanes. 

Beyond advances in individual sensors, we will need networks of 
sensors, a concept we refer to as “sensorwebs”.  These can 
accomplish at least two goals.  One is to achieve affordably the 
robustness that an operational system requires by allowing several 
small, separate platforms working together to accomplish what was 
formerly thought to require large, multi-instrument platforms.  The 
other is to achieve affordably  the high temporal and spatial 
coverage needed to monitor rapidly evolving phenomena such as 
tornadoes and volcanic eruptions.  
We will demonstrate this concept in the EOS era. The EOS 
platforms, as well as Landsat, Cloudsat, and Picasso will be flown 
in concert to create an “observatorium” if you will.  NASA and its 
partners within the US and internationally will operate each 
mission independently but the collective data will be correlated 
and processed as if the formation were a “superinstrument”.  In the 
future, sensorwebs will consist of a reconfigurable, smart  
constellation of  microsatellites  that will adjust their tasking based 
on emerging events on Earth.  

Another dimension of this geospatial revolution is altitude, for lack 
of a better term.  We want to migrate sensors from low Earth orbit 
to geostationary and higher orbits to achieve greater temporal and 
diurnal coverage.   Our New Millennium Program EO-3 mission is 
intended to demonstrate the capability to do EOS-type atmospheric 
observations from geostationary orbit.  This will be a major step 
toward achieving our weather prediction improvement goals.  The 
Triana mission will be the first Earth observing mission to L1, 
allowing us to understand the sunrise to sunset changes in ozone 
and other atmospheric variables.  Such a mission to L2 to view the 
night side of the Earth would be an ideal capability to measure 
global average night time temperatures, where suspected global 
warming will be most evident.

The computing and communications requirements for such a 
system will be enormous, growing from today’s terabytes to 
tomorrow’s giga and petabytes of data per day.  Industry will 
provide the advances in computing; NASA’s job will be to put 
these capabilities in space to enable on-board data processing and 
data compression.  NASA will also need to do the software design 
that will allow these high performance computing machines to run 
the coupled models that make prediction possible.

Computing and communications capability is not only required to 
get large volumes data from space to ground.  Inherent in the 
creation of a sensorweb is the ability to perform information 
synthesis.  Technology investment areas include adaptive 
information processing with distributed, reconfigurable computing 
and intelligent agents.  Information synthesis also includes real-
time integration of observations and models.  

Further, advanced computing and communications are also need to  
enable broad access to knowledge.  Remember, the goal is to 
enable Earth system prediction that is broadly available in society.   
In our space-based observing context, it could mean on-board data 
fusion to allow transmission of  tailored information products 
directly to a user’s desktop at no more that the cost of today’s 
international telephone call.

A host of technology advances are required to make the 
communications end of this vision possible.  And many will come 
from industry.  NASA’s role will focus on those aspects with 
particular utility to Earth science, such as knowledge presentation 
via new visualization techniques like immersive environments; 
knowledge generation via data mining, 3D and 4D dynamic data 
fusion; and development of Geo-reference standards for distributed 
multi-source data fusion.

The technical challenges to the vision I’ve just outlined are enough 
to start any engineer salivating…I see many hungry looks out there 
in the audience.  I’m glad that so many are eager to work on these 
exciting technologies.  But let’s keep in mind the larger context, 
the true motivation for this work.  Let’s finish by revisit the place 
we started—with proactive Earth system prediction.

The economic implications of improved weather forecasting and 
severe storm warning, not to mention human health and social 
factors, are compelling.  Between 1991 and 1995, the average 
annual weather-related damage in the US was in excess of $17 
billion per year, and more than 500 deaths per year were 
considered weather-related. The ability to forecast and respond 
proactively to natural hazards such as disease, volcanoes, and 
earthquakes has worldwide implications with costs estimated as 
high as $100 billion.

The Earth observation, research and modeling that NASA and its 
partners provide will forge the tools the nation needs to reduce its 
vulnerability to weather, climate change and natural hazards.  We 
look forward to working with you to forge those tools.  And we 
look forward to working with NOAA, USGS, USDA, FEMA, 
State and local governments and commercial enterprises to ensure 
these new predictive capabilities are commonly available in 
society.

We want to extend the frontiers of human knowledge in the 
understanding of the Earth’s environment, and the links between 
our environment and human activity.

Once we can do that, we will be ready to answer -- accurately -- 
questions like:

When will the next El Nino be?

What will the weather be in two weeks?

What temperature and precipitation pattern can we expect 
this growing season?

What impact will climate change have on my region of the 
country over the next 25 years?

NASA is committed to studying the Planet Earth in pursuit of 
answers to questions such as these.  Aggressive science-driven 
global monitoring, unique data applications and revolutionary 
advances in technology will enable us to truly  predict the future of 
the Earth system.   

Thank you.


7

1