Agricultural Education & Outreach
Water Resources Planning
GIS Applications
Web Design Services

Home  |  About Us  |  Services  |  Contact

MAPPING OF POTENTIAL
GLASSY-WINGED SHARPSHOOTER HABITATS: A PILOT PROJECT

Project Team:

  • Bay Area Shared Information Consortium (BASIC)
  • Spatial-Information Visualization & Analysis (SIVA) Resources Center of California State University at Monterey Bay (CSUMB)
  • Monterey County Agricultural Commissioner's Office (MCACO)
During the year 2000, there were approximately 46,200 acres of wine grapes in Monterey County. According to Gomgerg, Fredrikson & Associates, Wine Industry Consultants (April 2001), total (grape and wine) industry revenues topped \$384 million in the same year. Since the glassy-winged sharpshooter (GWSS) was found to be a vector spreading the deadly Pierce's disease (PD) in Temecula in southern California, Monterey County with has taken the threat of GWSS invasion into the County very seriously. As a result, in addition to the survey, trapping and intensive inspection of nurseries, and other quarantine activities, the Monterey County Agricultural Commissioner's Office (MCACO) has taken an aggressive, proactive action - developing a (digital) map showing the locations of potential habitats of GWSS. On the first day of March 2001, the MCACO entered into an agreement with the Bay Area Shared Information Consortium (BASIC) which agreed to support the mapping of potential GWSS habitats in a pilot project area within the Salinas Valley. BASIC is a California registered 501 (c) (3) corporation (a non-profit organization).

The mapping effort was funded by BASIC under the ongoing National Aeronautics and Space Administration's (NASA) Earth Science Information Partner (ESIP) program, cooperative agreement NCC5-317, in conjunction with the ongoing NASA-VINTAGE project. The Spatial-Information Visualization & Analysis (SIVA) Resources Center of California State University at Monterey Bay (CSUMB) provided technical assistance in developing a monitoring and planning tool using remote sensing (RS), Geographic Information Systems (GIS), and image processing (IP) technologies. Various in-kind services were provided by the MCACO. Local communities such as Monterey County Vintners and Growers Association, the Monterey County Farm Bureau, and the Central Coast Pierce's Disease Task Force supported this effort.


Pilot Project Area

The northern part of mid-Salinas Valley was selected for the site of the pilot project area. It extends the valley floor and the foothills from south of Chualar down (south) to Gloria-Camphora Road located south of the Soledad State Correctional Facility. The City of Gonzales and the Soledad State Correctional Facility lie within the pilot project area. It also is the northern most area where vineyards are planted in the Salinas Valley.

Technically, the pilot project boundary was demarcated along the boundaries of the digital color infrared (DCIR) image frame provided by the NASA Ames. It was discovered that the ground coverage of some vineyards has extended beyond the western boundary of the pilot project area reaching the ridgeline of Sierra de Salinas ranges.


Overview of GWSS and the spread of Pierce's Disease (PD)

 

The GWSS is a leafhopper insect that has a stylus - like a little drill - that bores into the xylem of the plant. With this unique apparatus, the insect can transmit lethal diseases into the wood portion of plants and trees. With this stylus, it also sucks life-giving water out of all the plants it feeds on. Each adult GWSS sucks out 200 to 300 times its body weight in water every day. This is the equivalent of an adult human drinking 4,300 gallons of water per day. According to the University of California (UC) Agriculture and Natural Resources Department, GWSS (Homalodisca coagulata), was accidentally introduced into California over a decade ago. It was first collected near Irvine in 1989, but positively identified only in 1994. GWSS is the carrier of a bacterium, X. fastidiosa, causing Pierce's disease to grapevines and many other food crops and ornamentals. X. fastidiosa is responsible for a new oak leaf scorch disease being found in the East Coast. GWSS has robust mouthparts and can feed on older wood. When it feeds on wood at the base of canes and on older stems the rate of spread of X. fastidiosa appears to increase from vine to vine. Systemic infection is further promoted by placing the bacterium in closer proximity to the xylem.

Since the early 1990s, the glassy-winged sharpshooter has been seen in high numbers in citrus orchards along the coast of southern California. It spread further inland, and locally abundant in Riverside and San Diego Counties. In 1997, the outbreak of PD in Temecula, Riverside County, was first detected. The following year the symptoms of the disease were visible in only a few localized areas. By summer 1999 hundreds of acres of vines were dead. In 1998 and 1999, high populations on citrus and adjacent vineyards were seen in southern Kern County. As of today, the GWSS infested areas include the entire counties of Los Angeles, Orange, Riverside, San Bernardino, San Diego, Ventura, and portions of Butte, Contra Costa, Fresno, Kern, Imperial, Sacramento, Santa Barbara, and Tulare Counties. Viable GWSS life stages were also found in nursery stock shipped into many counties, including neighboring Santa Clara, Santa Cruz and San Luis Obispo Counties. Santa Clara County is eradicating an infestation in the southern part of the county. Monterey County is still free from any GWSS infestation. Viable egg masses have been found in shipments of nursery stock into Monterey County.

Hewitt et al. (1946) documented that grasses in pasture or weedy alfalfa harbored the PD vectors in San Joaquin Valley. Purcell A.H. (1975) confirmed that the riparian areas were the habitats of primary PD vector -- blue-green sharpshooter, in California northern coastal region, however, vineyards located 400-600 away from riparian areas were found not affected by PD.

GWSS is a serious threat to California vineyards because it disperses faster and travels longer distances than other sharpshooters. GWSS lives on a wide variety of plants in Southern California including commercial crops (e.g., citrus, avocado and macadamia), ornamentals (e.g., eucalyptus, crape myrtle and ash), and natural vegetation, including oak, sycamore and sumac. The UC scientists noted that GWSS' ability to feed on dormant trees and vines during winter months is unprecedented for California sharpshooters, and it likely can reproduce on an even wider variety of cultivated plants and natural vegetation than is currently documented. The GWSS host list includes more than 100 species of plants including commercial crops such as almonds, citrus, peaches, plums, alfalfa and many ornamental plants produced by the state's commercial nursery industry. (See Appendix 1: Host Plants) The scientists also believe GWSS has the potential to increase both the incidence and severity of Pierce's disease in California because it can utilize more breeding habitats and plant hosts than native PD vectors (opportunistic behavior).

GWSS has been found in citrus and avocado groves, and on woody ornamentals. Oleander leaf scorch has become a serious problem for the California Department of Transportation (CalTrans). The sharpshooter threatens native plants, shrubs and trees. It has been documented GWSS feeding on over 70 species of plants in 35 different plant families and is active throughout the year. Favorite residential host plants for the sharpshooter include ash, bougainvillea, camellia, citrus, crape myrtle, eucalyptus, loquat, magnolia, mulberry, oak, privet, umbrella tree, apple, most stone fruits, blackberry, Chinese elm, oleander, philodendron, pine, sycamore, trumpet flower, wisteria, willow and yucca. It reproduces on Eucalyptus and coast live oaks in southern California.

A study conducted by Perring, TM, et al. (2000) has proved that proximity to citrus has influenced the incidence and severity of PD. In other words, higher disease severity is found closest to the citrus, and lower severity further from citrus. In the Temecula Valley, vineyards located within 1,000' (38%) and 2,000' (72%) of citrus groves allowed the rapid dissemination of PD throughout the Valley. They suggested that "Grape growers throughout the state should be particularly watchful of vineyards near citrus groves. Similarly, other areas containing non-citrus GWSS host plants should be monitored for GWSS, and vineyard borders near these areas should be surveyed for PD."

The effort to map the potential habitat areas for glassy-winged sharpshooters in the Pilot Project Area aims to produce a monitoring and planning tool for the MCACO that would apply remote sensing (RS), Geographic Information Systems (GIS), and image processing (IP) technologies. The pilot project effort has established methodologies to achieve the objectives of identifying and mapping existing vineyards, and high risk areas for being invaded by GWSS. As the objectives are achieved, the effort has reached the state of readiness to conduct a (Salinas) valley-wide mapping of vineyards and potential GWSS habitat areas in the near future.


Methodology

This pilot project focused on the feasibility of using remotely-sensed image data, image processing, and geospatial information system (GIS) technology to map potential GWSS habitats. In general, GWSS potential habitats are perennial vegetation such as shrubs, trees, and grapevines. Thus, the success of any remote-sensing approach to mapping GWSS habitats depends on the spatial resolution and spectral characteristics of the associated imagery. Since shrubs, trees, and grapevines often have a distinct spectral reflectance signatures compared to bare soil, herbaceous vegetation (crops, grass, pastures), open water and a host of manmade materials (roads, roofs, buildings), multispectral remote sensing can be a successful way to map GWSS potential habitats. Row crops and herbaceous crops of low to medium cover may have spectral signatures similar to that of shrubs, trees, and grapevines. If remotely-sensed data are available for a number of dates during a growing season, the spectral confusion may be overcome by using changes from date to date. GWSS potential habitats have spectral signatures that change much shower than the spectral signatures of herbaceous crops, especially short-season crops or crops that are mowed often during the growing season. Longer season herbaceous crops or herbaceous crops that are perennial have spectral signatures that are distinctly different from shrubs, trees, and grapevines. In addition to spectral and temporal patterns, a GWSS potential habitat mapper can take advantage of spatial properties such as size (of agricultural fields and individual plants such as trees), patterns (row crops versus random placement of plants), shape (agricultural shapes or natural shapes), texture (smooth herbaceous crops versus "rough looking" shrubs and trees), and context (e.g., urban environments contain numerous shrubs and trees that may not be identifiable in spectral data due to mixtures with non-vegetative materials). All of these considerations were the subject of this pilot project.


Data Collection and Image Manipulation

A variety of remotely sensed data were used in this effort.

One was a set of digital color-infrared (DCIR) imagery of the pilot project area that was received from the NASA Ames Research Center (ARC). ARC produced the DCIR by scanning aerial 9-inch color-infrared (CIR) photos taken from the NASA ER-2 high-altitude aircraft (about 65,000 ft). The ground spacing distance (GSD) of the DCIR is 1-meter between neighboring pixels. While these DCIR data were high resolution, they lacked spectral purity due to the use of CIR film, which blurs the differences among green light (GL), red light (RL), and near infrared (NIR).

SIVA also purchased four natural-color (NC) aerial photos covering the northwestern section of the pilot project area. These NC photos were taken from an altitude of 12,000 feet. After being scanned, the is 0.4-meter per pixel.




To analyze the temporal changes in the land cover in the pilot project area, the sponsor, BASIC, purchased a scene of Landsat 7 ETM+ (Enhanced Thematic Mapper Plus) dated June 19, 2001, which is relatively comparable to the same scene of Landsat 7 ETM+ taken on June 30, 1999. The GSD of Landsat images are 30-meter and 28.5-meter respectively. The Instantaneous Field of View (IFOV) for ETM+ data is 30-m by 30-m (except for the panchromatic Band 8, which has an IFOV of 15-m by 15-m).




USGS 7.5-minute (1:24,000) Digital Elevation Model (DEM) datasets were also obtained having a GSD of 30 meters.

For the cross-referencing purposes in the interpretation of the remotely sensed imageries, a few GIS datasets were obtained from the County of Monterey. They include:

 

  • Roads & High ways line networks (source: Monterey County Water Resources)
  • Ranch map boundary polygons (source: Monterey County Agricultural Commissioner's Office)
  • City limits boundary polygons (source: Monterey County Water Resources)
  • 1999 Land use polygons (source: Monterey County Water Resources)

Determining Potential GWSS Host Plants for Identification

As a group effort, the MCACO biologists and staff, a UCCE agent, who specialized in PD and GWSS, and the SIVA staff collectively selected the following GWSS host plants and their locations as features to be identified by using remote sensing methods: vineyards, riparian area, citrus orchards, oaks, eucalyptus, avocados, cactus, and ornamental vegetation.


Image Processing, Analysis and Ground-truthing

SIVA used ESRI's ArcView GIS program to access available GIS coverages maintained by the County of Monterey. These GIS coverages were then imported into MicroImages' TNTmips program for geospatial analysis. TNTmips was also used for handling all remotely-sensed image data.


 



Using the TNTmips Feature Mapping technique, areas above and beyond the cultivated areas were masked to exclude them from the analysis of the image. Feature Mapping is a supervised classification approach (based on simple range tests within each spectral band) that the analyst could override by what he or she sees on the display screen (i.e., based on spatial properties of size, shape, pattern, texture, and context). The Landsat 7 ETM+ (June 19, 2001) image was enhanced calibrated and used in Feature Mapping to locate the vineyards. Due to the low resolution (30-meter) with the Landsat 7 ETM+ image, it was required to crosscheck the ETM+ results with the results from the DCIR image, aerial NC photos and land cover GIS data from the County of Monterey.

A different approach to spectral analysis was unsupervised classification using a clustering algorithm such as ISODATA (in the TNTmips software). Using only the spectral signatures, unsupervised classification was conducted to produce 150 spectrally-different classes. Following this, the TNTmips Feature Mapping process, which is heavily supervised by the analyst, was used to identify spectral clusters/classes associated with vineyards and other potential GWSS host plants.

Due to the low ETM+ resolution, it is difficult to distinguish between early-stage vineyards (having grassy ground cover) and grasslands based off of the ETM+ image. The same situation occurred between mature (full-foliated) vineyards and other mature crops such as cactus, citrus orchards and row crops. The DCIR data with their high-resolution 1-meter resolution allowed the analyst to distinguish among different types of crop (plants) by visually identifying row patterns, and width of the plant-beds. Cross-referencing with other data such as the County of Monterey Land Cover was useful in the verification of types of crop (plant). Citrus and avocado allowed detecting different spectral signatures. The DCIR, County Roads, County Land Cover, and Ranch GIS information were used in verification and cross-checking of the sites.


Results

As a result, ground-truthing was essential for the verification of what was identified to be on the image and what was really on the ground. Assisted by the MCACO field staff, ground-truthing work was conducted a few times during the summer of 2001. At the beginning of the project, a few points (locations) of the features (host plants) were mapped by a Trimble Global Positioning System (GPS) unit. These GPS locations were displayed over the imagery under spectral analysis with the aid of the TNTmips GIS program. The points indicate the locations of different types of plants. This approach was helpful in identifying and distinguishing the spectral signatures of one type of plant to another within the same crop or land-cover class. After preliminary identification of the locations of the host plants was completed, another round of ground-truthing was conducted to validate the GWSS potential habitat mapping units. Ground-truthing helped in differentiating the vineyards in early stage, as well as vineyards in initial stage which spectral signature was similar to those of bare soils. Ground-truthing also has allowed us to distinguish among tree features such as eucalyptus, oak and other trees.

Areas of riparian vegetation are located within the river channel, ditches, and ponds. For this mapping project, the riparian zone as defined by the County of Monterey was assumed to be valid. Ornamental plants, another GWSS potential habitat, are mostly used for the landscaping in the urbanized areas. The locations of ornamental plants were assumed to be within general city-limits (urban-boundaries). The City of Gonzales, the property Soledad State Prison, and a property owned by the City of Soledad are the areas where ornamental host plants are considered to be located.

The map above shows the spatial distribution of vineyards and other major potential sharpshooter habitats. The proximity of vineyards to the other potential sharpshooter habits can be visually identified. In other words, it is easy to identify which vineyards are most vulnerable should there be a sharpshooter invasion. For the monitoring and planning purpose, the map allows one to select strategic sites where the (GWSS insect) traps should be placed.

By overlaying the "Ranch Map" GIS data (layer) the name of each ranch can be identified. It is important to note that the geopositional accuracy of ranch boundaries in the available GIS data is very low, however, it is useful in identifying the name of the ranch and approximate location of the ranch, and plants that have been cultivated on it.

© 2000-2004. Winners' Circle Consulting

703 visitors have visited this page.