This paper presents data on utilizing remote sensing technology for characterizing a riparian zone in southern Texas. Radiometric ground reflectance measurements, color-infrared aerial photography, and computer image processing techniques were conducted for this study. Reflectance measurements were made on 8 dominant vegetation types, soil, and water. Spectral measurements were made in the visible green (0.52-0.60 pLm), visible red (0.63-0.69 [pm), and near-infrared (0.76-0.90 pm) wavelengths. Reflectance values differed significantly (P = 0.05) among the vegetation, soil, and water parameters at all 3 wavelengths. Differences in reflectance among the vegetation cover types were attributed to variable foliage coloration and vegetative density. A color-infrared photograph of the study area showed that many of the ecological surface types could be readily distinguished. An unsupervised computer classification of the photograph identified 8 ground classes. An accuracy assessment performed on the classification showed an overall accuracy of 88%. RESUMEN-Se presenta infomaci6n sobre c6mo utilizar la tecnologia de la recepci6n remota para caracterizar una zona riparia en el sur de Texas. Medidas radiom&tricas de reflectancia del suelo, fotograffia aerea infrarroja a color, y t&cnicas de procesamiento digital de imaigenes se realizaron para este estudio. Medidas de reflectancia espectral se hicieron en ocho tipos dominantes de vegetaci6n, suelo y agua. Medidas espectrales se efectuaron en longitudes de onda correspondientes al intervalo visible verde (0.52-0.60 Rm), visible rojo (0.63-0.69 [Lm), y del infrarrojo cercano (0.76-0.90 [Lm). Los valores de reflectancia variaron significativamente (P = 0.05) entre los parametros de vegetaci6n, suelo y de agua en las tres longitudes de onda. Diferencias de reflectancia entre los tipos de cobertura vegetal se atribuyeron a la coloraci6n variada del follaje y a la densidad de la vegetati6n. Fotografia aerea infrarroja del drea bajo estudio mostr6 que muchos de los tipos surperficiales ecol6gicos pueden ser ficilmente identificados. Una clasificaci6n no supervisada con computadora identific6 ocho clases de capas superficiales. Una evaluaci6n de la clasificaci6n demostr6 una precisi6n total del 88%. Riparian zones and other wildland areas are often too large and inaccessible to determine their characteristics by ground surveys. Remote sensing techniques offer potentially timely, cost effective means of obtaining reliable data for these areas (Tueller, 1982). Color-infrared (CIR) aerial photography has been used extensively to inventory and classify riparian and wildland areas (Bonner, 1981; Tueller, 1982; Carneggie et al., 1983; Everitt and Deloach, 1990; Lonard et al., 1998). Field spectroradiometric light reflectance measurements have been used to distinguish among weed, wetland, and rangeland species (Best et al., 1981; Gausman et al., 1981; Everitt et al., 1986). Reflectance measurements have also been used to distinguish woody plant species (Gausman et al., 1977; Everitt et al., 1989) and related to their tonal responses on CIR aerial photographs. Aerial photographs can be digitized and subjected to computer analysis to quantify ecological cover types within these images (Everitt et al., 1990). The objectives of this study were to: 1) describe reflectance characteristics of major vegetation types, soil, and water in a riparian zone in southern Texas; 2) evaluate large scale CIR aerial photography for distinguishing among This content downloaded from 158.135.136.72 on Mon, 29 Sep 2014 10:50:43 AM All use subject to JSTOR Terms and Conditions 434 The Southwestern Naturalist vol. 47, no. 3 these cover types; and 3) determine potential of computer image processing for automated classification of the area. MATERIALS AND METHODS-This study was conducted near Tilden, in southern Texas. Tilden is located approximately 100 km south of San Antonio in the Rio Grande Plain vegetational region of Texas (Hatch et al., 1990). The study site was a riparian area located on the Frio River in Choke Canyon State Park. The adjacent area is dominated by Tamaulipan thorn shrublands on clayey calcareous soils (Correll and Johnston, 1970). Aerial photography, radiometric reflectance data, computer image analysis, and ground truth observations were conducted for this study. Reflectance measurements were collected to determine the spectral characteristics of plant, soil, and water variables within the study site and to help interpret aerial photography, and ground truth observations were collected to verify interpretations of the aerial imagery. Radiometric measurements were collected from 10 randomly selected plant canopies of each plant species or species mixture, soil, and water surface with a Barnes modular multispectral radiometer (Robinson et al., 1979). Reflectance measurements were obtained on black willow (Salix nigra Marsh.), dryland willow (Baccharis neglecta Britt.), bermudagrass [Cynodon dactylon (L.) Pers.], mixed herbaceous species (MHS)-lush, MHS-moderately stressed, MHS-severely stressed, hydrilla [Hydrilla verticilata (L. F.) Royle], algae (Spirogyra), bare soil, and water. These 10 vegetation, soil, and water surface types were the dominant land cover types in the study area. The area was experiencing drought conditions during this study and consequently, the 3 classes of MHS were observed. Lush MHS were found on the floodplain zone adjacent to the river, and moderately stressed and severely stressed MHS were located on the well drained slopes. Lush MHS were comprised of grasses, sedges, and broad-leaved herbs, whereas moderately stressed and severely stressed MHS were comprised of grasses and broad-leaved herbs. Black willow and dryland willow were the dominant woody plants in the area, and bermudagrass was a dominant herbaceous species. Hydrilla and algae were dominant aquatic plant species in the area. Measurements were obtained from the visible green, visible red, and near-infrared (NIR) spectral bands with a sensor that had a 15-degree fieldof-view placed 1.0 to 1.5 m above each plant, soil, and water target. The area within the sensor fieldof-view ranged from 0.26 to 0.39 m. Reflectance measurements were obtained on 24 June 1998 under sunny conditions between 1100 and 1500 h CST, with some measurements made from a step ladder. Radiometric measurements were corrected to reflectance using a barium sulfate standard. Overhead vertical photographs were obtained of plant canopies, soil, and water measured with the radiometer to help interpret reflectance data. Kodak Aerochrome CIR (0.50 to 0.90 VLm) type 2443 film was used for aerial photographs. Colorinfrared film is sensitive in the visible green (0.50 to 0.60 [Lm), visible red (0.60 to 0.70 [Lm), and NIR (0.70 to 0.90 ptm) spectral regions. Photographs were obtained with a Fairchild type K-37 large-format (23 cm by 23 cm) mapping camera. The camera had an aperture setting of f8 at 1/250 sec and a 305mm lens equipped with a Wratten 15 orange (minus blue) filter. Photographs were obtained on 17 June 1998 at an altitude above ground level of 600 m (scale 1:2,000). Photos (nadir) were acquired under sunny conditions with a Cessna 404 aircraft between 1130 and 1200 h CST. The camera was mounted vertically in a camera port in the floor of the aircraft. A CIR photographic transparency of the study site was scanned and subsequently digitized to perform a computer classification of land cover and an accuracy assessment. A Trimble global positioning system (GPS) Pathfinder Pro XRS system was used in the field to establish control points on the digitized photographic transparency for georeferencing. All image processing was performed using Erdas Imagine software (Version 8.3). The image was subjected to an unsupervised classification using ISODATA (Erdas, Inc., 1997). The ISODATA technique uses minimum spectral distance to assign a cluster for each selected pixel. With a specified number of arbitrary cluster means, the technique repetitively processes them where new means shift toward the means of the clusters in the data. Initially the unsupervised classification created 16 classes at the 99% convergence threshold. Some of the initial classes were combined resulting in 8 final classes. These classes were assigned to major land cover types based on field observations. Classes consisted of black willow, dryland willow, green grass (lush MHS, moderately stressed MHS, and bermudagrass), stressed grass (severely stressed MHS), hydrilla, algae, water, and bare soil/roads. For accuracy assessment of the classification, 150 points were assigned to the 8 classes in a stratified random pattern. The geographic coordinates of these points were determined and the GPS unit was used to navigate to these points in ground truthing. Both a producer's and user's accuracy were calculated. The producer's accuracy (measure of omission error) is the total number of correct points in a class divided by the total number of points of that class as derived from the reference data (ground truthing). The user's accuracy (measure of commission error) is the total number of correct points in a class divided by the total number of points of that class as derived from the classification data (map data). Green, red, and NIR reflectance data were anaThis content downloaded from 158.135.136.72 on Mon, 29 Sep 2014 10:50:43 AM All use subject to JSTOR Terms and Conditions September 2002 Everitt et al.-Remote sensing of a riparian zone 435 TABLE 1-Field light reflectance measurements of 5 plant species, 3 mixtures of plant species, soil, and water at the green, red, and near-infrared wavelengths. Measurements were made in a riparian area along the Frio River near Tilden, Texas in June 1998. Light reflectance (%) for 3 wavelengths Plant species/mixtures, Nearsoil, and water Green Red infrared Black willow 5.6 d' 2.7 f 38.4 b Dryland willow 4.6 e 3.6 e 21.6 de Bermudagrass 6.2

Video imaging systems are not radiometrically calibrated, thus it is difficult to obtain quantitative remotely-sensed imagery for natural resource applications. Video camera automatic gain controls (AGC) present potential problems in calibrating video systems for quantitative analysis because they compensate for changing solar illumination conditions. In this experiment aerial video calibrations to ground reflectance standards were compared for AGC turned on and off. The calibrated aerial video was evaluated for guinea grass (Panicum maximum L.) biomass treatments on two dates. Results showed that there was more atmospheric light scattering in the red than in the NIR video band. Light scattering affects could be detected only when the AGC was off because when AGC was on light scattering affects were masked. These results also showed that video imagery produced significant correlations with guinea grass biomass that were comparable to ground reflectance measurements.

The Rio Grande contributes approximately 97% of the water used in Texas' Lower Rio Grande Valley for consumption, crop irrigation, industry, support of riparian vegetation and wildlife and for recreation. Anthropogenic impacts on this international river have significantly degraded both the quantity and quality of its water. The objectives of this project were to assess the utility of multispectral aerial videography for assessing a point‐source wastewater discharge to the Rio Grande and to track the plume downstream to assess distribution and dilution of the contaminants over distance. An airborne 12‐band imaging system [11 visible and 1 near‐infrared (NIR) band] was used to collect videography of the study area while in situ physicochemical data were collected along three transects established across the river. Transect 1 was located at the discharge source, while transects 2 and 3 were located approximately one and two km downstream, respectively. Visual observations were used to divide each horizontal transect into plume, mixing and non‐plume zones. Each of the three zones was sampled along each transect. With the exception of the 820 nm NIR image band, all image bands recorded a higher light intensity (luminance) in the plume zone than in the mixing or non‐plume zones for all transects. Visible band results were associated with higher abiotic and biotic suspension concentrations in the effluent (plume) than in the river water. Video data correlated best with total chlorophyllous aggregates [ΣCA, (chlorophyll a + pheophytin a)], followed by total suspended solids (TSS), pheophytin a, total dissolved solids (TDS) and chlorophyll a concentrations. The 720 nm band demonstrated the best overall relationship with the selected parameters, followed by the 560 nm band, and then by the 700 mn band. No significant correlations were noted with any of the bands for total organic carbon (TOC) concentration. Correlations conducted for all transects indicated a significant relationship between plume luminosity and ΣCA, TSS, TDS and pheophytin a concentrations. Results indicated a significant potential for the use of multispectral narrowband visible imagery in conjunction with ground measurements (water physicochemistry) for rapid monitoring of environmental water chemistry. Development of imaging techniques to qualitatively estimate concentrations of certain constituents in impacted water columns without the need for in situ sampling may be possible.

CMOS-based (Complementary metal-oxide semiconductor) imager technology offers the opportunity for large scale (1 megapixel and greater) image capture using substantially lower power consumption than conventional imaging sensors such as CCD's (Charge Coupled Device). In addition, CMOS-based image sensors provide the advantages of greater on-chip integration of features and reduced system cost. Such devices are excellent candidates for many consumer imaging applications including digital still cameras, video cameras, and cellular phones with imagers. Typically, any final product using a CMOS-based image sensor will possess myriad additional components each of which may produce heating of the entire system. Increased system temperature arising from dense circuitry can lead to degradation of image quality as the sensor itself heats up. An investigation of the thermal profile of a CMOS image sensor will provide useful information for device design and optimization. In this paper, we present thermographs of CMOS imager sensors produced by a technique using a nematic liquid crystal.

Conventional color (CC) and color‐infrared (CIR) photography have been used in a wide‐range of remote sensing applications. Many users, however, do not fully understand or know how to interpret it. Published plant photography/reflectance studies make little effort to help the lay‐reader understand the relationship of plant spectral reflectance with their color tonal response on color photographic film. The objective of this paper is to describe the relationship of how spectral reflectances of plant species are registered on photographic film and how the resultant color tonal responses of the plants are produced in relation to their spectral reflectance. Illustrations are presented which simplify the interpretation of the interaction between plant canopy reflectance and CC/CIR photographic film emulsion layers, including the resultant color tones of the plants. Research studies addressing both plant photography and reflectance data can be more meaningful and better appreciated if one has a general understanding of these relationships.

Phytophthora foot rot, caused by Phytophthora parasitica (Dast), can result in economic losses for the citrus industry in the Lower Rio Grande Valley of Texas. Therefore, locating foot rot-infected trees in citrus groves is important to citrus growers. Symptoms of the infection include leaf yellowing, canopy defoliation, twig dieback, and short growth flushes. This study evaluated the use of the latest remote sensing technology, that of airborne digital imagery, for the detection of citrus trees exhibiting mild symptoms of foot rot infection. Airborne color-infrared (CIR) digital imagery of two citrus orchards having problems with foot rot infection was acquired. In addition to the aerial digital imagery, ground spectroradiometric measurements were conducted to determine the visible and near-infrared (NIR) spectral reflectance differences between healthy and infected trees. These measurements were also used to help interpret the color tonal renditions between the trees. The CIR digital imagery distinguished infected from noninfected trees. The noninfected trees had a bright red-magenta color rendition, while the infected trees had a dull grayish red tonal response. The NIR spectroradiometric and digital readings were significantly lower for infected trees than for healthy ones ( P ≤ 0.05), whereas the visible reflectance and digital data revealed no significant differences between the trees. The infected trees' dull tonal response in the CIR image was attributed to their lower NIR light intensity. These results indicated that digital imagery has potential for detecting foot rot-infected trees in citrus groves based upon NIR spectral differences. The advantage of airborne digital imagery is its real-time survey for quick field assessment.

Airborne digital imagery in conjunction with ground reflectance and plant physical data was used to evaluate simulated boll weevil damage in a cotton field in 2000. Five different levels of artificial square damage (control, 10, 20, 40, and 60%) with three replications were assigned across 15 experimental plots in a randomized complete block design. The artificial square damage was performed on 25 May when punctured squares by boll weevils were found. Airborne color-infrared (CIR) digital images were obtained from the field on 20, 28 June and 11 July. Ground reflectance and plant physical data, including plant height, number of leaves, number of squares, and chlorophyll, were collected on 5 July. The 20 and 28 June images revealed that plants with high artificial square damage levels showed higher spectral response in the near-infrared (NIR) band and lower spectral response in the red and green bands than those with lower damage levels. Plant height and number of leaves were significantly higher for plants with high damage levels than for those with low damage levels. However, there were no significant differences in either spectral response or plant physical characteristics between some of the damage levels. These preliminary results indicate that airborne imagery has potential for assessing boll weevil infestations in cotton fields, but more experiments are needed.

This paper reports on the application of an airborne digital video imaging system with visible red (R) (0.625‐0.635 μm), near‐infrared (NIR) (0.845‐0.857 μm), and mid‐infrared (MIR) (1.631‐1.676 μm) spectral sensitivity for distinguishing among a diversity of ecological ground conditions in a rangeland area in southern Texas. The system produces false color imaging similar to that of the Landsat Thematic Mapper (TM) satellite bands 5, 4, 3. Imagery from this system was useful for differentiating among a diversity of cover types including brushlands, grasslands, and wetlands. Computer analysis of the imagery showed that many of the ecological variables could be quantified. An accuracy assessment performed on the classified image showed an overall accuracy of 85%. The imagery was also useful for interpreting the coarser resolution TM. Ground radiometric reflectance measurements were made on various ecological ground types to assist in interpreting the imagery.

The objective of this paper is to present multispectral imagery acquired with a twelve‐band airborne digital video imagery system (ADVIS) to demonstrate its potential use as a research tool for ascertaining spectral bands and/or band combinations to better characterize and assess natural resources. The ADVIS is capable of sequentially generating four real‐time TlF‐formatted digital false color composite images that can be readily displayed immediately after the flight mission in order to evaluate which composite image(s) provides better differences among land‐use cover types in scenes of interest. The imagery, however, needs to be registered for image processing and analysis. Emphasis is addressed to image comparisons of a visible false green color composite to a conventional color composite. The false green color imagery showed better differences among land use cover types than the conventional color imagery. The image comparison results showed that there are other visible spectral bands or band combinations superior to the typical band combination of conventional color imagery for enhancing differences among land cover features of scenes. The ADVIS has potential use for ascertaining optimal band(s) or band combination(s) for distinguishing, characterizing and/or detecting problems of natural resources.

This paper reviews the application of aerial photography, airborne videography, and satellite imagery for distinguishing brush and weed species on rangeland and other wildland areas in the western United States. Ground reflectance measurements are used to determine the spectral characteristics of plant species. Season is an important variable for detecting many species because their reflectance varies at different times of the year and many species are distinguishable only when in a specific phenological stage. Computer image analyses are used to quantify weed and brush infestations, thus providing area estimates of noxious plant populations on rangelands. The integration of videography with global positioning system and geographic information system technologies is demonstrated. Plant species addressed include silverleaf sunflower (Helianthus arqophyllus). Texas lantana (Lantana horrida), false broomweed (Ericameria austrotexana). broom snakeweed (Gutierrezia sarothrae). spiny aster (Aster spinosus). blackbrush (Acacia rigidula). huisache (Acacia farnesiana). Mexican palo‐verde (Parkinsonia aculeata). common goldenweed (Isocoma coronopifolia). Drummond goldenweed (Isocoma drummondii). Chinese tamarisk (Tamarix chinensis), pricklypear (Opuntia lindheimeri). leafy spurge (Euphorbia esula). Big Bend locoweed (Astragalus mollissimus), Wooton locoweed (Astragalus wootoni). sand sagebrush (Artemisia filifolia). and shin oak (Ouercus havardii).

Leafy spurge is a troublesome weed on the northern Great Plains of the United States that chemicals and grazing management have not controlled. Remote sensing and geographic information system (GIS) technology have been used to detect and monitor numerous grassland related problems. The objectives of this study were to use both technologies to map and quantify the extent of leafy spurge within Theodore Roosevelt National Park and to provide information for managing the infestation. Analysis of the data indicated that 702 ha of the 18,680 ha park were infested by leafy spurge; however, leafy spurge populations occurring under dense woody canopies, in deep stream channels, and on steep slopes were not always detected. Infestations were especially dense in the western and southeast portions of the park. Most infestations were restricted to riparian zones and smaller drainage channels. Leafy spurge infestations decreased exponentially as distance from stream channels increased (r2=0.98). The significant association of leafy spurge with drainage channels suggests that the weed might be effectively managed on a watershed sub‐basin level. The joint use of GIS and remote sensing proved to be a powerful combination of tools which provided previously unavailable information about the extent and spatial dynamics of leafy spurge within the park. The results of this study will contribute to the development of a comprehensive leafy spurge management plan for Theodore Roosevelt National Park (South Unit).

Ground reflectance measurements, aerial video, and SPOT satellite data were compared for assessing herbaceous phytomass production and cover measurements on a south Texas rangeland area. Aerial video and ground reflectance data were acquired in April 1990 and in May and October 1991. For the October date, SPOT data was also obtained. Vegetation vigor was generally low in April 1990 due to insufficient precipitation and past drought, whereas more vigorous growing conditions prevailed in May and October 1991. The near‐infrared(NIR)/red ratio and normalized difference vegetation index (NDVI) were computed from the respective reflectance, video, and SPOT data. Regression analysis was used to relate the remote sensing data to phytomass and cover. Neither video or reflectance data were related to phytomass in April 1990, but they were significantly (p = 0.01) related to cover. In May and October 1991, video and reflectance data were significantly (p = 0.01) related to both phytomass and cover. SPOT data was also related to phytomass and cover in October. The better relationships of the remote sensing data to the biophysical variables in May and October 1991 was primarily attributed to the more vigorous growing conditions on these dates. These results showed that reflectance, aerial video, and SPOT satellite indices data may have potential for assessing phytomass production and foliar cover on rangelands.

Abstract. Multispectral video data were evaluated to determine their sensitivity to detect environmental conditions associated with hydrocarbon microseepage at the Pollard oil field. Biogeochemical analysis of soil and tree leaves indicates that plants are extracting available heavy metals from the soil environment. Anomalous concentrations of Mn were found in plant tissue. Correlation analysis reveals statistically significant relationships between Mn concentrations and vegetation growth around active producing wellsites. Although single‐band video data were generally not sensitive to biochemical variations, findings indicate that transformed video data may have the potential to detect environmental conditions associated with hydrocarbon microseepage.

We evaluated the effectiveness of airborne video imagery in distinguishing and mapping oyster (Crassostrea virginica) reefs in Aransas Bay on the southern Texas gulf coast. Imagery obtained using a high resolution multispectral videosystem was compared with color infrared (CIR) aerial photography of the same oyster reef. The reef was clearly visible in the video imagery, but the video image lacked the detailed resolution of the CIR photograph. And, the mean deviation from ground measures was significantly smaller for measurements taken from the CIR photography. Videography may be most useful when employed in conjunction with a global positioning system to locate exactly the positions of oyster reefs (and their approximate sizes) in a large area. Because video imagery can be acquired more cheaply than aerial photographs, the cost of the survey would be less. Subsequently, photography may be used to map the sizes of specific reefs where the accuracy of size estimates is crucial.

The expanse and inaccessibility of rangelands make them difficult to assess with conventional ground surveys. This paper illustrates the application of a relatively new technology, that of airborne video systems, for managing rangelands. Video imagery has been used to detect or distinguish among a variety of rangeland variables such as plant communities and species, drought stress, burned areas, grazing intensity, phytomass production, and rodent and insect infestations. Presently video does not have the detailed resolution of film, but it can provide range managers with immediately available, inexpensive remote sensing data that can be useful in managing their resource.

Airborne videographic remote sensing is less than a decade old. Its recent growth at the USDA-ARS at Weslaco, Texas and selected universities is evidence of a plethora of research and applications in this field. Video imagery, when interpreted by visual or computer techniques, has provided insights into a variety of Earth features. Much research has been done on detecting and assessing variables relating to cropland, rangeland, tree crops, and soils. In addition, videographic research and its applications have been conducted in forestry, water quality, wetlands, land-cover inventory, and urban land use. The poorer resolution of video compared with film has limited some applications, but the numerous advantages of videography have promoted its increasing use. In the future, advances in video technology will result in improved spatial resolution suitable for expanded applications.

The design, operation, and testing of a high resolution multispectral video system (HRMVS) is described. The system uses state‐of‐the‐art video technology. It incorporates three black‐and‐white (B & W) visible/near‐infrared (NIR) (0.4–1.1 μm) light sensitive solid‐state cameras equipped with band‐pass filters and provides two kinds of simultaneously synchronized video images: (1) color‐infrared (CIR) composite imagery and (2) its three‐band B & W image components. Only CIR composite imagery is presented here with its B & W components (yellow‐green, red, and NIR bands), but any false color combination can be generated by the encoder. Images are recorded on high resolution (400 horizontal lines) Super (S)‐VHS recorders. An independent solid‐state conventional color (0.4–0.7 μm) camcorder (S‐VHS) was optional to this system. It was set up to acquire imagery at approximately the same field‐of‐view as the three‐camera synchronized system. Examples of imagery of various natural resource characteristics are given. Color‐infrared composite imagery had similar color tonal renditions to that of CIR film. The high resolution multispectral Black & White B & W images showed that some terrain features could be discriminated better in certain bands. For example, the yellow‐green (543–0.552 μm) band was best for distinguishing chlorosis in grain sorghum (Sorghum bicolor Moench), whereas the NIR band was optimum for separating biomass levels in alfalfa (Medicago sativa L.). The color and Black & White B & W multispectral image results showed this system to be a valuable and versatile tool for a variety of remote sensing applications.