Lee Johnson’s research while affiliated with California State University System and other places

NASA's Ames Research Center is performing an AVIRIS demonstration project in conjunction with the U. S. Environmental Protection Agency (Region 3). NASA and EPA scientists have jointly defined a Study Area in eastern Virginia to include portions of the Chesapeake Bay, southern Delmarva Peninsula, and the mouths of the York and James Rivers. Several environmental issues have been identified for study. These include, by priority: 1) water constituent analysis in the Chesapeake Bay, 2) mapping of submerged aquatic vegetation in the Bay, 3) detection of vegetation stress related to Superfund sites at the Yorktown Naval Weapons Station, and 4) wetland species analysis in the York River vicinity. In support of this project, three lines of AVIRIS data were collected during the Wallops Island deployment on 17 August 1997. The remote sensing payload included AVIRIS, MODIS Airborne Simulator and an RC-10 color infrared film camera. The AVIRIS data were delivered to Ames from the JPL AVIRIS Data Facility, on 29 September 1997. Quicklook images indicate nominal data acquisition, and at the current time an atmospheric correction is being applied. Water constituent analysis of the Bay is our highest priority based on EPA interest and available collateral data, both from the surface and from other remote sensing instruments. Constituents of interest include suspended sediments, chlorophyll-a and accessory pigments, Analysis steps will include: verification of data quality, location of study sites in imagery, incorporation of relevant field data from EPA and other Chesapeake Bay cooperators, processing of imagery to show phenomenon of interest, verification of results with cooperators. By 1st quarter CY98 we plan to circulate initial results to NASA and EPA management for review. In the longer term we will finalize documentation, prepare results for publication, and complete any needed technology transfer to EPA remote sensing personnel.

Ames Research Center, together with the Robert Mondavi Winery (Oakville, California) and Terra Spase Vineyard Consulting (Napa, California), is evaluating the use of geospatial technology (remote- sensing, geographic information systems) in a precision agriculture context. In the Canopy Remote-Sensing for Uniformly Segmented Harvest (CRUSH) project, high-spatial resolution multispectral images were collected at midseason 1997 with an airborne ADAR-5500 digital camera system. In a technology demonstration, an image of a processed to a Normalized Difference Vegetation Index to improve sensitivity to grapevine canopy density. The Index values were then stratified and color-coded for visual discrimination. A georegistered output image was delivered to the winery for input to the grower's geographic information system. NASA and winery researchers field-sampled vines within the study block for canopy density (light interception), vine physiology (leaf-water potential, chlorophyll concentration), fruit characteristics (maturity, potential quality), and yield. The grower used a laptop computer with image display and onboard Gloha! Positioning System to physically subdivide (with flagging tape) the study block for harvest based on vine vigor (high, medium, and low). Grapes from each field segment were fermented separately and the resulting wines were formally evaluated by the winery.

High spatial resolution airborne imagery was acquired in California's Napa Valley in 1993 and 1994 as part of the Grapevine Remote sensing Analysis of Phylloxera Early Stress (GRAPES) project. Investigators from NASA, the University of California, the California State University, and Robert Mondavi Winery examined the application of airborne digital imaging technology to vineyard management, with emphasis on detecting the phylloxera infestation in California vineyards. Because the root louse causes vine stress that leads to grapevine death in three to five years, the infested areas must be replanted with resistant rootstock. Early detection of infestation and changing cultural practices can compensate for vine damage. Vineyard managers need improved information to decide where and when to replant fields or sections of fields to minimize crop financial losses. Annual relative changes in leaf area due to phylloxera infestation were determined by using information obtained from computing Normalized Difference Vegetation Index (NDVI) images. Two other methods of monitoring vineyards through imagery were also investigated: optical sensing of the Red Edge Inflection Point (REIP), and thermal sensing. These did not convey the stress patterns as well as the NDVI imagery and require specialized sensor configurations. NDVI-derived products are recommended for monitoring phylloxera infestations.

During the 1993 and 1994 growing seasons, airborne digital sensors were used to collect visible and near-infrared images of phylloxera-infested vineyards near Oakville in Napa County. Computerized processing enhanced the information content of the images with respect to leaf area of the canopy. Processed image values were strongly related to ground measurements of vine pruning weight and leaf area made within a 12-acre study site. The images were useful for mapping patterns of leaf area throughout the site and in surrounding vineyards, and for assessing year-to-year changes in canopy. The vineyard manager found the imagery valuable in planning for replacement of phylloxera-infested fields, managing for crop uniformity and segregating grapes of differing quality during harvest. This tool was particularly useful in evaluating and managing newly acquired property.

Grape Phylloxera damage grapevines by feeding on roots. Vines with damaged roots may be less able to replenish their leaves with water than vines with healthy roots. If these vines with damaged roots continue to transpire at the rate of healthy vines, their leaves should exhibit lowered water potentials.

Grape Phylloxera will cause the California wine industry to lose over one billion dollars by the year 2000. Grape growers with grape phylloxera- infested soils graft scion varieties onto what are popularly termed 'resistant rootstocks.' Rootstocks, however, differ in their suppression of phylloxera: some do not support the insect at all, while other support low populations. In addition, phylloxera biotypes vary in their growth on different rootstocks. In California's Napa and Sonoma county vineyards, about 75% of the vines have AXR#1 rootstock that tolerates phylloxera biotype A. In the early 1980's biotype B emerged there. It so devastates AXR#1 that the vineyards must be replanted with rootstocks resistant to biotypes A and B. Timing replanting is difficult because vineyards do not decline uniformly. A patchwork of uninfested vines, infested but asymptomatic vines, declining but productive vines', and unproductive vines typifies most vineyards. The grower must determine the proportion of vines in each category and estimate the yield loss the stressed vines will suffer. During 1993, 1994 and 1995 the NASA-Ames GRAPES study used remotely sensed leaf reflectance, temperature, and canopy size data and geographic information system (GIS) technology to study infestations in Napa County vineyards. As part of this study a vineyard with a range of phylloxera induced stress and accompanying symptoms -- reduced growth, less chlorophyll, and lower reflectance of near infrared:red light -- was investigated to determine the degree to which stress measurements predict the current and following season's yields from stressed vines relative to healthy vines. Such yield estimates could enable a grower -- before obtaining actual yields -- to calculate the economics of replanting. A grower who decided to replant would have 2-14 months additional lead time to plan and prepare.

Leaf/canopy model simulations and measured data were used to derive information on the form and strength of the nitrogen (N) "signal" in near-infrared (1100-2500 nanometer (nm)) spectra of fresh leaves. Simulations across multiple species indicated that in total, protein absorption decreased near-infrared reflectance and transmittance by up to 1.8% and 3.7% respectively, and all other inputs held constant. Associated changes in spectral slope were generally in the range of plus or minus 0.02% per nanometer. Spectral effects were about an order of magnitude more subtle for a smaller, though potentially ecologically significant, change in N concentration of 0.5% over measured. Nitrogen influence on spectral slope was fairly consistent across four empirical data sets as judged by wavelength dependence of N correlation. The observed and simulated data showed similar trends in sensitivity to N variation. Further, these trends were in reasonable agreement with locations of absorption by protein-related organic molecules. improved understanding of the form and strength of the N signal under differing conditions may allow development of reasonably robust spectral measurement and analysis techniques for "direct" (based strictly upon N-related absorption features) N estimation in fresh leaves. A pragmatic approach for remote sensing might additionally consider surrogate measures such as chlorophyll concentration or canopy biophysical properties.