Taming the Tamarisk

A new GIS weapon helps a county fight a dangerous shrub.

by / April 27, 2005
A high-tech war is beginning to unfold across the American landscape -- a quiet war far removed from terrorist threats and CNN news coverage.

On one side, high above the earth, is the QuickBird satellite, a commercial spacecraft able to offer submeter resolution imagery of Earth's surface.

On the other side is a pretty, hardwood shrub -- the tamarisk, also known as the salt cedar.

Growing from 5 to 20 feet tall, the tamarisk was originally introduced into the United States from Eurasia in the early 1800s as an ornamental plant. Because of its dense, deep root system, settlers in Southwestern states planted salt cedar along streambeds to prevent erosion from flooding. Over the years, tamarisk spread from the Colorado River basin to New Mexico, and as far east as Texas.

This proved disastrous for ecosystems in affected states, including Kansas. Kearny County, Kan., is fighting the spread of tamarisk along the Arkansas River with a combination of satellite imagery and GIS.

Targeting the Tamarisk
"We needed to find out the population of salt cedar on our stretch of the river and see if we could more accurately determine the full use of water by the salt cedar," explained Shannon McCormick, a Kearny County commissioner. "People don't really understand it is more than just a weed problem. But if you relate the use of the water by salt cedar to the acreage of corn you could water with that, then you can accurately estimate the economic impact. Everybody understands dollars."

Rough estimates reveal that tamarisk along parts of the Arkansas River in Kansas consume enough water to supply 20 million people or irrigation of more than 1 million acres of land.

Past efforts to estimate the extent of salt cedar infestation involved either extensively flying over infested areas or hiring people to scout plant growth on foot -- a long and laborious task.

"Flying, you just get too much air and you still only end up with a guesstimate on how many acres," McCormick said. "With digital imagery, on the other hand, and using the spectrum of the plants, you can get a very accurate number. You don't just end up with some abstract type of graph. You get actual maps that are really animated and friendly to use. We were happy with the results."

The tamarisk is now labeled an invasive species because it displaces native plants through its aggressive growth.

"In short, salt cedar is just a nasty plant," said Jason San Souci, director of remote sensing applications at the Native Communities Development Corp. (NCDC), which specializes in using high-resolution imagery for a wide range of natural resource applications, such as wildfire risk assessment, forest composition analysis and invasive species tracking. "It steals water. It's bad for fire. It's bad for wildlife. It's bad for soil. Based on the latest projections, there are probably 1.5 million acres affected in the western United States."

The tamarisk consumes far more water than other indigenous plants -- a single tree can use as much as 200 gallons per day. As a result, tamarisk infestations lower water tables and destroy wildlife habitat by sucking millions of gallons of water from streams, rivers and lakes. The plant also exudes salt from its leaves, increasing the salinity of soil, which further inhibits native plant life. And its high-density growth is a serious factor in helping spread forest fires.

"As each plant gives off 600,000 microscopic seeds every season, there is no way we are going to get rid of this problem," added San Souci. "But we can certainly control it, and we can start to restore some of the areas that have been affected. That's where we come in."

Managing the tamarisk consists of four components, he explained, control, revegetation, monitoring and maintenance.

"Without all four components, it is unlikely that tamarisk-control projects will be successful over the long term," he said. "Tamarisk can be managed using successional weed management techniques, including chemical, mechanical and biological techniques."

San Souci said restoring vegetation in previously tamarisk-dominated areas can present numerous technical challenges caused by the negative effects of the aggressive plant. Restoring vegetation after tamarisk removal, however, is essential to prevent reinfestation, and it's possible, especially if flooding is allowed.

In some cases, he explained, successful restoration of riverbank habitat will require remediation of conditions that led to tamarisk infestation in the first place, such as ecosystem changes through land and water alterations.

Focusing the Satellite
What Kearny County ended up with -- not just printed maps, but also a GIS application showing the salt cedar infestation -- involved a complex series of technical steps. As good as the satellite imagery is, the information requires expertise to use and interpret.

Coming up with a solution to a problem such as tamarisk invasion typically requires DigitalGlobe to work with another vendor because end-users aren't experts in remote sensing imagery.

"The satellite essentially takes two images simultaneously within microseconds of each other through two different sensors. One image is black and white with a 2-foot resolution and the other is multispectral -- red, green, blue and near-infrared -- with about an 8-foot resolution," explained Chuck Herring, director of marketing communications for DigitalGlobe. Those infrared images were key to addressing Kearny County's tamarisk problem.

"If a city has been shot in near-infrared, for instance, all the vegetation areas are bright red, and everything else that is not vegetation should be dark or black," said Herring. "However, the detail -- what species or type of plants these are -- is mushy. That is where someone like the NCDC comes in."

San Souci said the NCDC turns the pixels into GIS data points that can be used to solve whatever problem is at hand. In this particular case, the NCDC started by taking the two images, and through a process called pan-sharpening, essentially fused the two. This gives the resolution of the black-and-white, or panchromatic images, and adds the spectral information of the color images.

"The panchromatic has all the texture we're looking for in the spatial context, while the four-band color, and in particular, the near-infrared, give us the sense of vegetative health and vegetation in general," said San Souci. "Additionally we have a set of proprietary algorithms that we're still refining to calibrate the imagery and enhance it to pull out the features we're most interested in."

At this point, the images go into a commercial software package -- Feature Analyst from Visual Learning Systems -- that uses machine learning technology to classify object-specific geographic features specified by the user.

"We actually go into the field with a submeter GPS unit, and we gather training samples and accuracy assessment samples for salt cedar," he said. "Using these, we can train the computer to recognize salt cedar as opposed to other types of vegetation. And then finally, running those algorithms, it will spit out an ESRI Shapefile for the area that can be used in most GIS applications.

"Without these enhancement stages, you can't pull out salt cedar," he continued. "We tried, but it didn't work very well. There is a lot of confusion with all the other vegetation classes out there. But using these enhancements specific to salt cedar allows us to pull it out very precisely."

Call to Action
The results in Kearny County are gaining wide attention, according to McCormick.

"We've just sent the information to the state, and statewide there have been more meetings," he said. "They are now using the maps we supplied to talk about it as a statewide problem, which is how it has to be addressed. If salt cedar isn't handled upstream from us, then anything we do will largely be futile in the long run."

"Hopefully we will get some state and federal funding," McCormick added. "That was what we were hoping our $24,000 investment would generate. Plus the rest of the state could see how well the new technology worked for us. We weren't just dealing with the guesstimates of salt cedar infestation that we have been living with up to now."

Over the past two years, six separate bills have been introduced in Congress specifically addressing the tamarisk problem, said San Souci, noting that during last year's session of Congress, two bills, S. 1516 and H.R. 2707, consolidated previous language from several other bills, and both passed but were tabled at the conclusion of the session.

San Souci said S. 177 was introduced by Sen. Pete Domenici during this year's session of Congress and is currently scheduled for debate.

"The 2005 bill provides $20 million in fiscal 2006 and $15 million each subsequent fiscal year for on-the-ground demonstrations of tamarisk and Russian olive control and revegetation, with monitoring and research to identify changes to the environment," he explained. "Demonstrations could occur in any of the Western states and most of the Plains states having the problem."

Under the terms of the bill, he said, $4 million would be authorized in fiscal 2006 for assessment of existing conditions and techniques for control and restoration with an additional $2 million authorized to develop long-term management and funding strategies.

If Kearny County's approach to managing the tamarisk catches Uncle Sam's eye, Kansas could grab a share of the federal funding being considered. San Souci said funding would be 75 percent federal with the local share made up of funds and in-kind contributions, including state-agency-provided services.
Blake Harris Contributing Editor