Hydraulic engineer David
Goodrich examines cottonwood
leaves for signs of water
stress in the San Pedro
It's been called "a ribbon of life," "a miracle of the
desert," and "one of the last great places."
The Upper San Pedro River Basin of southeastern Arizona is perhaps
best known for its astounding variety of bird life. Some are seldom-seen
species like the azure bluebird, green kingfisher, and American bittern.
Others are more familiar, like the red-tailed hawk. Ornithologists estimate
that anywhere from 1 to 4 million birds visit the basin on their way
to or from their summer breeding and nesting grounds.
The cottonwood forests and willow thickets that line the banks of the
San Pedro make a perfect place for these birds to feed and rest. Too,
this unique region also hosts an impressive array of other wildlife,
including blacktail prairie dogs, gray foxes, and piglike javelinas.
Roughly 130 miles long, the San Pedro River extends from the state
of Sonora in northern Mexico into Arizona, finally disappearing into
the Gila River about 50 miles north of Tucson, Arizona. Along the way,
the river cuts through oak savannas, desert scrub, and grassland expanses.
In the Upper San Pedro Basin, the river is fed in part by a vast aquifer.
Experts estimate that this porous underground layer of sand, clay, and
fractured rock holds some 13 trillion gallons of water.
Rick Edwards (left), University
of Arizona graduate student,
and David Goodrich download
data from a meteorology/flux
system, which measures the
amount of water and carbon
exchanged among vegetation,
soil, and atmosphere.
| How Much Water Do Riverbank
Trees Really Drink?
The intent of SALSA is to produce science-based ways to predict the
effect that various natural or human changes might have on the water
balance and the plant and animal life in the basin. Plants play a key
role in cycling water through the basin, taking moisture up through
their roots, then giving it off as water vapor through leaves, stems,
"Everyone knows that plants like cottonwoods take up and use a
lot of water," says Goodrich. "Even so, it would help if we
knew how much." Until now, the best anyone could come up with was
a very rough estimate of the monthly or yearly water needs of plants
like cottonwood or willow.
To solve this problem, Goodrich and co-investigators created some high-tech
approaches to measure and predict the amount of water riverside plants
require. They used scanner laser beams to measure the amount of water
the plants gave off, sap-flow sensors to detect the quantity traveling
through tree trunks, and naturally occurring atomic tracers to identify
the trees' water sources.
In addition, they relied on devices called scintillometers to track
the amount of heat radiated from the ground and vegetation. They also
used satellite-mounted sensors called radiometers to determine the amount,
type, and temperature of the vegetation.
"As far as we know," says Goodrich, "these are the first
direct estimates of how much water these riparian plants really use."
The team's preliminary results indicate that plants draw up about 15
percent less water than was generally thought to be the case. That figure
is significant enough "to make a real difference in choosing among
management strategies," Goodrich says.
Grasslands Also Garner Attention
About one-third of the San Pedro Basin is covered with hardy grasses
like blue grama, sideoats grama, and sacaton grass. These rugged plants
make ideal cover for mice, rabbits, and other creatures that are, in
turn, prey for the region's raptorsowls, hawks, falcons, and eagles.
On both sides of the U.S.-Mexico border, beef cattle graze this rangeland
mix of grasses and desert shrubs throughout the year. Ranchers have
to be careful not to put too many cattle on these ranges because overgrazing
can lead to erosion. What's more, the loss of plant cover can hasten
the rush of floodwaters, which, in turn, can damage river channels and
lessen the quality of river water.
Perhaps surprisingly, the size and condition of the grasses' roots
is one of the most telling indicators of the health of the rangelandand
a good predictor of the plants' future growth. ARS research leader and
hydrologist M. Susan Moran at Tucson and Yann Nouvellon, formerly a
NASA-sponsored postdoctoral researcher at Moran's laboratory, have developed
a mathematical model that will, among other things, give ranchers a
good indication of whether a stand of grass has strong, healthy roots.
"It's somewhat like knowing if your front lawn has a good root
system," Moran explains. "If it does, you know it is more
likely to green up after a rain."
Similarly, if they have good information about the size and condition
of roots, ranchers should be able to use the model to estimate how Southwest
desert grasses will respond to the region's typical summer monsoon and
SEHEM for Savvy Management
Moran and Nouvellon have dubbed their model SEHEM, short for Spatially
Explicit Hydro-Ecological Model. It's based on mathematical formulas
that can be run on an ordinary desktop computer. When data about the
grasses and the weather are plugged into the formulas, the computer
can generate customized color maps that ranchers can use to estimate
where, and how much, forage might be available for hungry cattle to
graze in the coming months.
To create a model like this basically required taking known facts about
how these wildland grasses grow, as well as data about weather, and
then converting that information into a numerical form that a computer
can process and analyze. Computer models that simulate weather or how
a plant might grow above and below ground, given certain typical weather
patterns, aren't new. So what's new about SEHEM? Likely the broad expanse
of land for which SEHEM has so far proven accurate and reliable.
Moran and Nouvellon tested the model using weather data from a solitary,
self-operating meteorological station. They meshed it with data from
the huge swaths viewed by satellite-mounted sensors. The information
from the meteorological station was very accurate for a small, localized
part of the basin where the station was located. This 10-foot-high tower,
loaded with weather instruments, is situated at the ARS Walnut Gulch
Experimental Watershed, near Tombstone, about 175 miles southeast of
"We have 12 years' worth of detailed weather information from
the station," Moran notes. "But it's very localized. If you
were to try to extrapolate it to predict grass growth over a much broader
area, like someone's 10,000-acre ranch, you might lose a significant
amount of accuracyunless you had the help of SEHEM.
"The satellite imagery from NASA's LANDSAT Thematic Mapper and
LANDSAT Enhanced TM Plus," she says, "covers the exact same
site as our Walnut Gulch meteorological station and more than 12,000
square miles, as well. But this satellite imagery of plants, soil, and
other natural features is at a much different scale and a much less-frequent
time intervaljust 29 images for an entire 10-year period."
Next Target: Great Basin
"The SEHEM model we developed allows us to use both of these important
data sets," Moran explains. "With its scope and specificity,
SEHEM should be appropriate for broad areas." Moran says that puts
the researchers in the ballpark for having the scale and degree of accuracy
to make the model a practical, workable tool for the region's ranchers,
conservationists, and others interested in safeguarding the desert grasslands.
Moran and Nouvellon validated the model by having it project plant
growth for specific periodsin the pastfor which they already
had accurate information about the vegetation. Says Moran, "We
already knew about aboveground growth, but we wanted to see whether
the model could accurately predict the same thing. The correlation was
quite good, and we've been able to make adjustments in the equations
to fine-tune or calibrate them."
This procedure made SEHEM one of the first validated models for predicting
growth of rangeland grasses over a very broad area. The scientists are
using the same approachteaming up single-point, very-long-term
meteorological data with infrequently obtained, high-resolution satellite
imageryfor other regions. In particular, they're doing that to
develop a new model specifically for predicting growth of wildland grasses
in the Great Basin ranges of Utah, Nevada, and parts of Idaho. In the
meantime, Moran expects SEHEM to "go a long way toward helping
prevent the overgrazing that occurred on some San Pedro Basin grasslands
in the past."
For their studies, Goodrich, Moran, and 67 other SALSA team members
have won a USDA honor award for environmental protection. SALSA scientists
have reported their findings in Agricultural and Forest Meteorology,
Remote Sensing of Environment, Water Resources Research,
and other scientific publications. Two CDs, available from Goodrich
at no charge while supplies last, tell more about SALSA. They are "Miracle
of a Desert River" and "San Pedro River Basin Spatial Data
Wood, Agricultural Research Service Information Staff.
This research is part of Water Quality and Management, an ARS National
Program (#201) described on the World Wide Web at http://www.nps.ars.usda.gov.
David C. Goodrich
and M. Susan Moran are
with the USDA-ARS Southwest
Watershed Research Center, 2000 E. Allen Rd., Tucson, AZ 85719;
phone (520) 670-6380, fax (520) 670-5550.