Using infrared aerial photography and computer enhancement, this West Texas rancher is waging a high-tech war on rangeland brush.

Every ranch battles a troublesome weed or bush that encroaches on its land and chokes off its cattle forage. Each region has its own invader. In the mountain states, the culprit is sagebrush, while in the Southwest the problem is juniper trees and prickly pear cactus.

On Bert Wallace's Peace Pipe Ranch in West Texas, the war is waged on mesquite. Wallace spends a significant portion of his operating budget each year on chemicals and aerial spray services to reduce mesquite and open up more of the 66,000 acres he oversees near Jayton to the grasses his cow-calf operation depends on.

To maximize payback, Wallace has devised a high-tech - and highly successful - method of brush control that may earn the nickname, "Precision Ranching."

For a decade, he's been fine-tuning it and using it to find the most dense mesquite on his land in order to determine which invaded area holds the best forage growth potential. This information helps plan a cost-effective campaign of pinpoint aerial spraying.

A Unique Approach Wallace's approach is unique. He contracts an aerial firm to photograph his property every few years. But, instead of simply examining the photos with the naked eye, he enhances and analyzes the photos digitally with a commercial image processing software he taught himself to run on his desktop computer.

"I can create a map of all the mesquite on my property and then prioritize which areas should be sprayed based on their value for growing forage grasses," said Wallace. "I'm controlling the mesquite and seeing more grass as a result."

The Peace Pipe approach may be novel on a cattle ranch now, but aerial surveying, image processing and related geographic information technologies are becoming increasingly accessible to individual ranchers. Significant enhancements in affordability and user friendliness of digital image processing software are driving this trend.

The most important advantage of the precision ranching technique is that it's not limited to West Texas or mesquite control, says Jim Ansley, an associate professor at Texas A&M Research and Extension Center in Vernon.

"You can use these technologies to differentiate a variety of shrubs and trees from grasses," says Ansley, adding that "ranchers will use (these techniques) for many other applications in the general management of their lands."

Waging the Battle Mesquite competes with forage grasses across West Texas. It grows no higher than a bush in dry areas, but can reach 60 ft. in height with a 4-ft. -diameter trunk near water. Grass can't survive in the shade cast by its canopy, and the thicket it weaves hinders livestock handling.

Wallace remembers his father contracting to have the entire ranch sprayed for the princely sum of $30,000, or $1/acre, in the early 1960s. That wiped out most primary growth, butpockets have since returned, usually in wet areas and on the best soil.

"Spraying the entire ranch is simply too expensive, and it would be impractical because secondary mesquite doesn't grow uniformly," Wallace says. "The only practical solution is precision aerial spraying."

But to decide where to spray, a rancher must know exactly where the brush grows. Haphazard growth patterns make it impossible to map mesquite from the ground, so Wallace opted to go the extra mile and survey his property from the sky.

After researching aerial photography, he concluded color infrared film was best for mapping mesquite because near infrared energy is strongly reflected by chlorophyll-rich green vegetation like mesquite. This would make the invading growth stand out strongly as bright red in the photos. He also correctly theorized that taking the pictures in July would further highlight mesquite in full foliage while surrounding grasses were dormant.

"Finding an aerial survey company that could do color infrared photography was more difficult than I expected because this film can be quite hard to work with," Wallace says.

He was finally satisfied with the work of Krawietz Aerial near San Antonio. In its most recent survey, they charged $5,600 to take a total of 80 photos at 12,000 feet covering the entire Peace Pipe Ranch and surrounding lands he manages. This included printing 9x9-in. transparencies at 1:40,000 scale.

Going Digital Wallace realized he could analyze the photos more quickly, and map mesquite automatically, with a computer. So he purchased a Sharp scanner to convert the photos to digital files.

He scans each print at a resolution of 600 dots/in. (dpi). This means the scanner measures the color brightness at 600 individual points, or pixels, within each inch of the photo and assigns a number value to each pixel based on its brightness.

In an infrared photo, color brightness corresponds directly with reflectance of near infrared (IR) energy. Healthy, chlorophyll-rich vegetation strongly reflects IR energy and appears as a bright red. After scanning is completed, this brightness or high degree of reflectance is represented by a high number. Surfaces reflecting less IR light, such as dormant grass, appear less bright and as other colors in the photo and are assigned lower reflectance numbers by the scanner.

This concept of reflectance values may seem complex, but it's critical to understanding how Wallace differentiates mesquite from other land cover in air photos of his lands.

"My decision to go digital proved crucial in the development of the technique," he says.

A digital mapping expert soon introduced Wallace to digital image processing software. This is written specifically for enhancing, analyzing and extracting land information from air photos, satellite images and digital maps. Commonly used by geologists, foresters, urban planners and environmental scientists to help manage large land areas, Wallace knew he could accomplish more with it than with a computer-aided drafting and design (CADD) package.

In 1987, he purchased TNTmips image processing software developed by MicroImages of Lincoln, NE, because it was easy to use and designed to integrate several different types of digital geographic data. He runs the current version of the same software today (now with integrated geographic information system (GIS) capabilities and selling for $3,000-5,000) on a standard Pentium computer in his office.

"It's exceeded all of my expectations for the brush control project and gets easier to use all the time," Wallace says.

Processing the Digital Photos Wallace loads the 70 megabyte digital photos individually into his system for processing. The software has several enhancement routines that allow him to improve the photo quality with a click of the mouse. He usually applies a basic algorithm to adjust the contrast so the vegetation stands out.

The image is then ready for automated classification, one of the most powerful capabilities of computerized image processing. Wallace first applies a classification routine called Adaptive Resonance, which automatically clusters the pixels into groups of similar reflectance values. It also compensates for differences in reflectance numbers resulting from shadows or other distortions on the photo.

"This routine begins the process of locating ground features based on reflectance," said Wallace.

In other words, each cluster contains pixels representing similar ground features. Grasses would be in one cluster, and mesquite would be in a different one while bare soil might be in yet another. But because this first round of classification is somewhat arbitrary, two ground features with similar reflectance might end up in the same cluster.

"Cottonwoods and mesquite can look pretty similar in infrared photos," said Wallace.

The next phase of processing, called Feature Mapping in TNTmips, refines this classification and uses operator input to actually identify clusters - something the computer can't do on its own.

Wallace displays the classified photo on his computer screen with each cluster represented in a different color. He then uses the mouse to point and click on specific clusters he knows from ground observations are mesquite, cottonwood, hackberry, grass or soil.

This procedure is called Training because it trains the software to reclassify the clusters based on input, which means it finds all other pixels in the image with spectral values similar to the ones specified. Instead of just having anonymous pixel clusters on screen, he has land cover features clearly differentiated and identified as mesquite, cottonwood and so forth.

"The beauty of Feature Mapping is that I can keep inputting ground truth information to further subdivide the classes," says Wallace.

For instance, once he identifies mesquite stands, he measures their canopy density on screen. He then enters this as additional training information and reclassifies the already identified mesquite classes as high, medium or low density.

This is very important because Wallace figures that grass is inhibited from growing under mesquite trees once canopy closure exceeds 45%, which is when spraying becomes cost effective.

Turning Data into Information The Peace Pipe Ranch targets chemical spraying on mesquite stands with the densest canopies because eradicating them provides the best return on investment. But TNTmips enables Wallace to go another step and make a more sophisticated calculation of investment payoff by taking into account soil type and its potential to grow grasses.

"I get soil maps from the Natural Resources Conservation Service (formerly Soil Conservation Service)," he explains. "They have calculated the growth potential for each soil type."

Wallace scans the soil maps into the image processing system. An option within the Feature Mapping function enables him to digitally overlay the classified air photo onto the soil map. The software then analyzes and correlates the two data sets and derives statistics from them.

"It gives me an output file that tells me a lot about the soil and mesquite cover, such as the number of acres of each soil type and the percent of each soil type covered by mesquite," he says.

The TNTmips output file is loaded into an Excel spreadsheet to determine the cost/benefit ratio of applying chemical to any given area. Wallace uses a financial program developed by Jim McGrann at Texas A&M to determine the productive value in dollars for each acre. His overall goal is to clear dense mesquite from soils with the highest growth potential.

The spreadsheet multiplies the soil productive value, percent canopy cover, number of acres, and expected root kill percent to calculate the dollar benefit of spraying any particular area. That dollar amount is compared against Wallace's $26/acre cost to apply Dow Reclaim control chemical.

"I can figure out my exact return on investment so I know whether it is better to spray or put my money in a CD," he jokes.

In some cases, Wallace adds weight to certain land areas in the spreadsheet calculations based on ranch operating factors. For instance, areasaround fences and water bodies are considered prime targets for spraying because they are important to handling the cattle.

Another benefit of performing this technique in a totally digital environment is that coordinates of spray areas can be downloaded directly from the image processing system to a disk and input into the GPS guidance system in the spray plane. This improves efficiency of the chemical application, which translates into saved money for Wallace.

Growing More Grass Wallace hasn't quantified the impact of the technique in dollar terms. He budgets a certain amount for brush control and assumes he'll spend it. "The bottom line for this technique is not to save money but to utilize what I do spend in the best possible way with the highest payout down the road," he says.

The next time he has the Peace Pipe surveyed from the air he plans to run what is called a Change Detection routine in TNTmips to compare new and old photos and quantify exactly how many acres of mesquite have been replaced with new grass.

Judging from his ground observations while working cattle, Wallace knows he's beating back the invader. Now, he's turning his computer power to other chores.

"I'm using the image processing system and the photos for a lot of other management operations such as designing fence layouts, placing water tubs and planning grazing strategies," he says.

This article was provided by Kevin Corbley on behalf of his client MicroImages, Lincoln, NE.