Optical remote sensing applications in viticulture - a review

Transcription

1 Optical remote sensing applications in viticulture - a review 1 A HALL 1,2,3, DW LAMB 1,2, B HOLZAPFEL 1,2 and J LOUIS 1,2 Abridged Title: Optical remote sensing applications in viticulture 1 Cooperative Research Centre for Viticulture, PO Box 154, Glen Osmond, SA National Wine and Grape Industry Centre, Charles Sturt University, Locked Bag 588, Wagga Wagga, NSW Corresponding author: Andrew Hall, anhall@csu.edu.au, Fax: Abstract The emergence of precision agriculture technologies and an increasing demand for higher quality grape products has led to a growing interest in the practice of precision viticulture; monitoring and managing spatial variations in productivity-related variables within single vineyards. Potentially, one of the most powerful tools in precision viticulture is the use of remote sensing through its ability to rapidly provide a synoptic view of grapevine shape, size and vigour over entire vineyards. Its potential for improving viticultural practice is evident by the relationships that are known to exist between these canopy descriptors and grape quality and yield. This paper introduces the reader to remote sensing and reviews its recent, and potential, applications in viticulture. Abbreviations EM electromagnetic; GPS global positioning system; GIS geographical information system; NDVI normalised difference vegetation index Key words: remote sensing, precision viticulture, multispectral imaging, grapevine, vegetative vigour

2 2 Introduction Grapevine (Vitis vinifera) health and productivity are influenced by numerous physical, biological and chemical factors, including spatial variations in topography, physical and chemical characteristics of soils and the incidence of pests and diseases. The spatial variation in these factors effects a spatial variation in grape quality and yield within vineyards leading to an overall reduction in wine quality and volume. With the likelihood of increased differentiation in pricing between grapes based on measured quality attributes (Winemakers Federation 1996), vineyard management decisions must account for spatial variability in quality and yield in order to produce a higher-quality higher-value product. However, these decisions rely on the availability of accurate and reliable data that describe spatial variability in relevant vine descriptors. The emergence of global positioning systems (GPS) technology means traditional on-site measurements of physical, chemical and biological parameters associated with vine productivity can now be linked to specific locations within vineyards. This information, when used in conjunction with computer-based geographical information systems (GIS), provides viticulturists with the capability to process and map spatial relationships between attributes and make management decisions based on numerous layers of information (Taylor, 2000). The process of modulating cultural practices as a function of spatial and temporal variation within agricultural fields is known as precision agriculture (Cook and Bramley 1998, Moran et al. 1997). In the context of the grape and wine industry, precision viticulture may be defined as monitoring and managing spatial variation in productivity-related variables (yield and quality) within single vineyards (Lamb and Bramley 2001). In recent years, yield maps produced by grape-yield monitors in Australia have shown up to eight-fold differences in yield can occur within a single vineyard block (Bramley and Proffitt 1999). Furthermore, there are considerable spatial variations in quality indicators such as colour and baume (Bramley and Proffitt 2000). Relationships between yield and quality

3 3 indicators are often inferred, however these relationships do vary significantly between vineyards (Holzapfel et al. 1999, Holzapfel et al. 2000), and possibly within vineyards. Moreover, preliminary data suggest regions of high and low-yielding vines in a vineyard tend to remain stable in time, inferring that soils play a significant role in such variability (Bramley et al. 2000). The accurate characterisation of spatial variations in those parameters that influence vineyard productivity requires a considerable amount of data. Traditional methods of generating such data are generally time consuming and expensive. For example, measuring basic fruit quality and yield parameters of sixty sample sites in a one hectare block requires more than thirty work-hours. The move toward on-the-go sensing of yield and quality parameters by combining the latest sensor technology with GPS-equipped vehicles is slow and currently limited to grape yield. However, rapid sensing techniques such as measurement of baume using near infrared (NIR) spectroscopy (Williams 2000) and grapephenolic composition using visible-nir spectroscopy (Celotti et al. 2001) are potential candidates for on-the-go sensing. The use of rapid electromagnetic induction or EM-survey techniques to accurately characterise soil structure is also becoming more widely used in the grape and wine industry (Lamb and Bramley 2001). The use of remote sensing as a means of monitoring crop growth and development is attracting interest from researchers and commercial organisations alike. This interest is primarily driven by the opportunities for cost-effective generation of spatial data amenable to support precision agriculture activities (Lamb 2000). To date, limited use is being made of this technology in the grape and wine industry, either for research support or as a commercial monitoring tool. This paper presents some of the key principles of remote sensing, reviews the current status of remote sensing in viticulture, and discusses the potential of remote sensing as part of an integrated management tool for vineyards.

4 4 How does remote sensing work? Remote sensing involves measuring features on the earth's surface using remote satellite or aircraft-mounted sensors. In terms of optical remote sensing, sensors detect and record sunlight reflected from the surface of objects on the ground. The ability of a sensor to detect these objects is quantified in terms of the sensor's spatial, radiometric, spectral and temporal resolution. Spatial resolution is a measure of the smallest object detectable on the ground. The number of available image-forming pixels in the sensor itself, and its distance from the ground, contribute to determining the pixel-size on the ground and the overall image footprint. For example, the American Landsat satellite, orbiting at a height of 705 km above the Earth s surface is capable of recording images with a 30 m x 30 m pixel size (referred to as a 30 m pixel), and a footprint of 185 km x 185 km. The French SPOT satellite orbits 832 km above the earth' s surface, generating full scenes of 60 km x 60 km and a 20 m pixel. This means the smallest object that can be directly detected by the sensor is 30 m (Landsat) or 20 m (SPOT) in each dimension (Barret and Curtis 1999) (Figure 1). More recently, highresolution satellites such as IKONOS, which provides 4-m resolution multispectral imagery, have come on line, however, the cost of such data remains a significant impediment to its widespread use (Lamb et al. 2001b). Airborne mounted sensors such as airborne digital cameras or video systems, which are flown up to 3 km above the ground, generally have 1- to 2-m pixels and corresponding image footprints of the order of 100 Ha (Figure 2) (eg Lamb 2000). Figures 1 and 2 illustrate that while Landsat and SPOT satellite imagery, with spatial resolution of the order of tens-of-metres, is suitable for applications requiring regional coverage, the pixel size precludes its use in the investigation of targets of the size of typical vineyard blocks, and of features that may vary within vineyards. Radiometric resolution specifies the number of discrete radiometric levels available to individual pixels to record the intensity of measured radiation from a target in a given

5 5 waveband. For example, 8-bit radiometric resolution means there are 2 8 = 256 levels available (0 = darkest, 255 = brightest) while 10-bit sensors have 2 10 = 1024 levels available to each image pixel. In practise, however, n-bit systems tend to only have (n-2)-bits of information in image pixels as usually the lowest 2-bits of data carries the system noise, including dark-current and thermal noise (King 1992, Louis et al. 1995). Temporal resolution or, more simply, revisit-frequency is an important attribute of any sensor when used for commercial monitoring or management purposes. Typical commercial satellites like the American Landsat and French SPOT satellites have revisit intervals of 16 and 26 days, respectively. In the case of SPOT imagery, a target-pointing capability during different overpasses could reduce this interval to as low as 2 days (Barrett and Curtis 1999). Aircraft mounted sensors, on the other hand, are more amenable to user-defined visitations, and have the added advantage of being able to operate under a high-cloud base (Figure 3). The spectral resolution is the number of wavebands of data that can be simultaneously recorded at each pixel. The amount of sunlight reflected off a target is described in terms of the target's reflectance profile. The spectral reflectance profiles for Cabernet Sauvignon vines, underlying covercrop (chick-peas) and bare soil are given in Figure 4. These profiles indicate the amount of sunlight these targets reflect as a function of the wavelength (or colour). All photosynthesising plants, including vine canopies and covercrops, do not reflect much light in blue or red wavelengths because chlorophylls (and related pigments) absorb much of the incident energy in these wavelengths for the process of photosynthesis. However, these targets reflect a higher proportion of light in the green wavelengths, again due to chlorophylls and related pigments, and this is why such targets appear green when viewed by the human eye. However, in the near infrared wavelengths (wavelengths greater than about 700 nm) photosynthesising plants reflect large proportions of the incident sunlight (in excess of 65%). These wavelengths, to which the human eye is insensitive, can be detected by appropriate instruments. The amount of sunlight reflected in these wavelengths is very sensitive to leaf cell structure and this is influenced by water content (Campbell 1996,

6 6 pp ). Figures 5(a) and (b) show the reflectance profile of a typical vegetated target. Superimposed on these profiles are a set of wavebands corresponding to the sensitivity of a hypothetical instrument and the reflectance profile that would be inferred from the response of that instrument to the ground target. In Figure 5(a), the hypothetical instrument measures the spectral signature of the target in four wavebands. While an accurate measure of the target reflectance would be extracted at the four specified wavebands, the shape of the reflectance profile of the vegetated target is only poorly described. Using thirteen closely spaced wavebands (Figure 5(b)), the reflectance of the target is recorded for each waveband and the shape of the entire spectral profile is more accurately described. In an application where fine detail in the shape of the spectral profile is required, the higher spectral-resolution instrument (Figure 5(b)) would be appropriate. A consequence of the upper limit on the amount of data that can be processed and stored in real-time by any remote sensing system is the compromise between spatial, radiometric and spectral resolution. In general, this equates to a trade-off between spatial and spectral resolution. The terms multispectral and hyperspectral are often interchanged, although they usually define instruments according to the number of wavebands of information that is recorded for each image pixel. The more general adjective multispectral is used to describe instruments that record information in only a small number of wavebands; typically Hyperspectral instruments record information in a large number of wavebands, typically greater than 10. Spectral vegetation indices reduce the multiple-waveband data at each image pixel to a single numerical value (index), and many have been developed to highlight changes in vegetation condition (eg Wiegand et al. 1991, Price and Bausch 1995). Vegetation indices utilise the significant differences in reflectance of vegetation at green, red and near infrared wavelengths. For example, Normalised Difference Vegetation Index (NDVI) images are created by transforming each multi-waveband image pixel according to the relation:

7 7 ( near infrared) - ( red) NDVI = (Equation 1) ( near infrared) + ( red) where near infrared and red are respectively the reflectances in each band (Rouse et al. 1973). The NDVI, a number between 1 and +1, quantifies the relative difference between the near infrared reflectance peak and red reflectance trough in the spectral signature (refer to Figure 4 for an example). This index is the most widely used indicator of plant vigour or relative biomass. For highly vegetated targets, the NDVI value will be close to unity, while for non-vegetated targets the NDVI will be close to zero. Negative values of NDVI rarely occur in natural targets. One important advantage of ratio indices such as the NDVI is that the intensity of the total light reflected from a target does not influence the calculation. An object under shadow will reflect light reduced by approximately the same amount across the entire spectrum. Although there is a reduction in the precision of NDVI for areas in shadow, because of a reduction in the total range of reflectance levels, the ratio of two spectrally similar features should be the same. Through the use of vegetation indices, shadows, which may otherwise be a significant problem in imaging a vineyard with closely spaced rows, are effectively removed. Airborne imaging systems The use of airborne colour and colour-infrared photography for monitoring crops in Australia was established in the early 1970 s (eg Harris and Haney 1973). These techniques were later extended to detect weeds in crops and pastures (Barrett and Leggett 1979, Arnold et al. 1985). However, limitations of aerial photography for crop monitoring include the absence of a quantitative data acquisition capability, the high cost and availability of colour infrared film and processing, and the requirement for manual scanning or digitising. The intrinsic analogue nature of the imagery results in significant additional processing and delay prior to incorporating the imagery into a GIS.

8 8 Airborne imaging systems, incorporating in-flight or post-flight image digitisation can provide sub-metre resolution images of crops at any revisit frequency and in a timely and cost-effective manner. Multispectral imaging systems provide user-selectable spectral bands, some as narrow as 10 nm bandwidth. These bands are commonly available in the visible and near infrared (NIR) (eg Manzer and Cooper 1982, Louis et al. 1995, Anderson and Yang 1996, Sun et al. 1997), and the mid-infrared bands (SWIR) (Everitt et al. 1986, Everitt et al. 1987). Hyperspectral imaging systems also provide user-selectable wavebands. Systems such as the Compact Airborne Spectrographic Imager (CASI-2) offers up to 288 wavebands with approximately a 2.2 nm bandwidth in the range of nm (ITRES 2001). By comparison, Hymap imagery offers up to 200 wavebands in the visible, NIR, SWIR and thermal infrared (TIR) (Intspec 2001). However, due to power and stability requirements, airborne hyperspectral imaging systems are confined to operation in larger twin-engine aircraft and this makes them significantly more expensive to operate than multispectral imaging systems which can be deployed in single-engine aircraft. Airborne multispectral and hyperspectral systems are ideal for quantification of crop growth in agricultural research applications. These systems have spectral bands in the visible green ( nm) and red ( nm) wavelengths, and in the near-infrared ( nm) wavelengths, and provide the high temporal and spatial resolution needed for agricultural research plot evaluation (Clevers 1986, Clevers 1988a, Clevers 1988b, Lamb 2000). Insights provided by such research and the increasing affordability of multispectral imaging systems have resulted in them becoming more widely used over a wide variety of Australian crops (Lamb 2000). Remote sensing as a tool for precision viticulture Remote sensing of soils Along with climate and topography, soil is a key factor influencing vineyard productivity (Jackson 2000). At the regional or between-vineyard scale, soil has been described as having the least significant effect on grape wine and quality (eg Rankine et al. 1971, Wahl 1988).

9 9 However, at the scale of individual vineyards, soil and topography are often strongly connected (Jackson 2000, Yule et al. 2001, Taylor and McBratney 2001), as are topography and climate, in particular microclimate (Percival et al. 1994, Hutchinson 2001). On-ground physical measurements of soil structure and condition in vineyards have demonstrated significant variations can exist within single vineyards. For example, using EM-surveying to measure soil electrical conductivity in two contrasting Australian vineyards demonstrated up to three-fold differences in conductivity existed within each. In the case of the 7 hectare Coonawarra vineyard used in this study, conductivity was highly correlated to soil depth and the latter varied by a factor of two (Bramley et al. 2000). Similarly, large variations in petiole nutrient levels in the same vineyard suggested large spatial differences in soil mineral content (Bramley 2001a). Spatially referenced grape yield maps acquired from a number of Australian vineyards over the past three years suggest regions of high- and low-yielding vines tend to remain stable in time. This suggests that soils, and their association with topography and microclimate, play an important role in characterising the spatial characteristics of within-vineyard variability. Therefore, it is no surprise that considerable effort in precision viticulture research is targeting measuring and mapping spatial variability in soils at the single vineyard scale. Often, different soils, because of differences in intrinsic colour, moisture levels and organic and mineral constituents, have different optical reflectance characteristics (Condit 1970, Colwell 1983, Escadafal et al. 1989). However, care must be taken when using optical remote sensing for mapping soil structure on the basis of surface reflectance as visible and near infrared radiation penetrates only to within a few millimetres of the soil surface (Lee 1978). Numerous studies have reported varying levels of agreement in comparing bare-soil images with other on-ground soil data such as EM survey (Pitcher-Campbell et al. 2001) and traditional soil sampling (Grierson and Bolt 1995, Ryan and Lewis 2001). In a situation where the ground has been ploughed, as in the preparation of a new vineyard site, the soil surface may more accurately reflect soil variations in the vicinity of the vine root-zone.

10 10 Imagery, highlighting differences in the vigour of the pre-existing pasture or crop, may also be useful in identifying different soil zones in a potential vineyard site before cultivation, (for example, see figure 2 in Lamb 1999). Recent work, involving the Hymap hyperspectral sensor demonstrated the enormous potential of high-order image processing of many wavebands of spectral information (Ryan and Lewis 2001). Ryan and Lewis contended that using 128 spectral wavebands of Hymap allowed them to discriminate numerous soil zones underneath mature vines. However, the extent to which these soil analyses relied on direct soil spectral information versus indirect measurements of the subtle variations in vine vigour was not established. By identifying regions of similar soils and matching suitable varieties and clones to the particular soil types, remotely sensed images can be a valuable tool at the planning stage of vineyard development. Soil-related effects in a given field will vary from season to season, and may completely reverse under different rainfall conditions (eg Lamb 2000). However, positioning varietal blocks so that they are contained within only one soil type with its own irrigation system allows easy management and a more consistent product (Grierson and Bolt 1995). Although physical soil sampling will remain an essential requirement of groundtruthing, the major advantage of remote sensing is in reducing the amount of soil sampling required to adequately characterise and delineate soil zones. A single imaging mission, with a view to segregating a site into homogeneous blocks has potential to characterise variability and increase overall quality and productivity of a vineyard development. The economic benefits of planning in this way are considerable, as the cost of imaging at this stage can be inexpensive (Grierson and Bolt 1995) yet beneficial to a vineyard over its entire lifetime. Remote sensing of vines Despite its increasing level of application usage in the analysis of broadacre agriculture crops, airborne imaging is yet to be fully evaluated over established vineyards with respect to quantifying attributes of the vines themselves. Two distinct functions of imaging established

11 11 vineyards have so far been identified. The first is the general mapping of vines to accurately establish numbers of different varietals within vineyards, and the second is the mapping of levels of relative vigour to establish spatial differences in vine performance within single varietal blocks. In terms of general mapping, accurate information concerning the location and size of blocks containing different varietals allows for more accurate forecasting of regional productivity and the allocation of resources for subsequent winemaking (Bramley 2001b). Subtle differences in leaf spectral signature and phenology, and vine shape/size, suggest that it may be possible to discriminate and map different varietals using remote sensing. However, such differences may be quite small and would likely require a sensor with a combination of metre-resolution imagery and a large number of spectral wavebands. To date, only hyperspectral instruments such as CASI have been successfully used to discriminate different varietals within vineyards and to identify mis-planting of one variety within a block containing another (eg Bradey and Wiley 2000). Information regarding relative vigour levels has many applications for improving management at the precision scale, such as the early detection of certain vine diseases or the identification of discrete management zones. Spatial differences in environmental factors result in significant spatial variations in vigour throughout a vineyard. Vine vigour is reported to have a considerable effect on fruit yield and quality (Dry 2000, Haselgrove et al. 2000, Petrie et al. 2000, Tisseyre et al. 1999, Iland et al. 1994). For example, in a single block of Cabernet Sauvignon, researchers demonstrated that yields of vigorous vines were nearly double that of stressed vines (Clingeleffer and Sommer 1995). In the same study, considerable delays in fruit maturation were also associated with the more-vigorous and higher-yielding vines. Three levels of vine vigour used in the study produced significant differences in juice and wine parameters; higher-yielding vines produced grapes of lower quality. Based on such observations, vine vigour could be used as surrogate indicators of vine yield and grape quality.

12 12 In addition to vine vigour, links between canopy shape and vine physiology have also been reported by several studies. For example, Intrieri et al. (1997) describe significant differences in total vine assimilation of CO 2 by vines before and after various canopy shape and thickness manipulations. Similarly, Smithyman et al. (1997) report on the influence of three different canopy configurations on vegetative development, yield and composition of grapevines. Furthermore, several studies have shown a link between fruit exposure on the vine and some of its characteristics at harvest. This has led to practices such as basal leaf removal; a late season trimming technique where leaves are removed from around fruit clusters to improve ripening conditions. This method of leaf removal has been associated with increased evaporation potential, wind speed, higher temperature and improved light exposure around the fruit (Thomas et al. 1988). As well as leaf removal increasing fruit exposure to light and air movement, the resulting improvement in access of chemical sprays to the fruit produce a less favourable environment for the development of fungal infections. In uniform-cover crops like wheat and canola, different levels of plant vigour often appear as differences in the crop density against a background of underlying soil. Generally, a region of healthy crop has a high plant density. Such a region would appear as all crop plants, typically a deep green as viewed by the eye. Conversely, poor crops with a lower plant density would appear as a mixture of soil and crop. These crops would appear to look greenbrown as viewed by the eye (Lamb 2000). Grapevines, however, express vigour not only in terms of the density of the canopy, but also in the spatial extent of the canopy itself. Therefore, the relationship between spatial variations in vine vigour, as perceived by a remote sensing instrument, and spatial variations in vine productivity (yield and quality) may be complex. Identifying the most appropriate means of quantifying vine vigour in remotely sensed imagery is currently the subject of research worldwide. Conceptually, and from a computational point of view, the most convenient approach to quantifying vine vigour is in blending canopy spectral signature, a combination of single leaf spectral characteristics and canopy density, with canopy size/shape. This is achieved by

13 13 using remotely sensed imagery with a spatial resolution comparable to the inter-row spacing of the target vines (Figure 6). Assuming the background covercrop is uniform, or at the least shaded by vines, this process produces image pixels that are a local average of vine and inter-row space (non-vine) spectral signatures. Changes in leaf spectral signature or the proportion of vine and non-vine area within single image pixels will change the average pixel value (Figure 7). Johnson et al. (1996) have successfully used this technique to identify broad areas of vineyard infested with the highly damaging vine aphid, phylloxera (Daktulosphaira vitifoliae). This was achieved by relating the level of phylloxera incidence to the level of vine vigour. The level of vine vigour on the ground was quantified in terms of pruning weight where the largest or most dense vines yielded the greatest weight of vegetation during subsequent pruning. Correlations established between the NDVI values extracted from imagery and canopy pruning weights were used to indicate areas subject to phylloxera infestation. Significant correlations have also been achieved between NDVI and canopy leaf area index (m 2 leaf area per m 2 of ground) and leaf area per vine (m 2 per vine). These correlations have been established over multiple vineyards using 4 metre-resolution IKONOS satellite imagery (Johnson et al. 2001). A consequence of the link between canopy vigour and grape yield is that significant correlations between image-derived NDVI values and subsequent grape yield is possible (Baldy et al. 1996, Lamb et al. 2001a). These relationships remain valid regardless of whether the driving influence behind the spatial variation is water and nutrient status (eg Clingeleffer and Sommer 1995) or pests and diseases (eg Baldy et al. 1996, Munkvold et al. 1994). Similarly, studies involving assessment of the effect of canopy morphology on fruit characteristics have suggested some qualities of the fruit may also be inferable from vine size/shape or vigour. Where it can be established from remotely sensed imagery that vines within a block have, for example a more open canopy, it could be expected that the fruit character and other biophysical properties of the vine are being influenced. Numerous

14 14 researchers worldwide have indicated that links between remotely sensed imagery and grape quality indices are being investigated (eg Vintage 2001, CRCV 2001). However, outcomes have yet to be reported in the scientific literature. Separation of leaf spectral signature from vine size/shape characteristics in remotely sensed imagery can only be achieved through using more complex data-extraction procedures. Furthermore, extraction of vine size/shape descriptors requires images of spatial resolution of tens of centimetres, as large numbers of image pixels must be covered by individual vines. Hall et al. (2001) have reported developing a "vinecrawler" algorithm for extracting both spectral-signature and canopy dimension information from ultra-high (25-cm) resolution multispectral images of vines. This process first involves the classification of vine and nonvine pixels. The inter-row space, which would otherwise confound vine shape/size measurement, can be eliminated from the analysis (Figure 8). The vinecrawler algorithm progressively moves along the centre of the classified vine rows and records spectral signature parameters as well as size/shape descriptors such as the width of the canopy crosssection (number of pixels), skew and kurtosis (Figure 9). Importantly, this technique has allowed the identification of individual vines, which allows the generation of a row-vine coordinate system from vineyard imagery. This has an immediate application in terms of directing on-ground field visitations to regions identified from remotely sensed imagery (Hall et al. 2001). Work on linking these complex vine descriptors with grape quality indices is reported to be in progress. The way ahead Although it is undergoing rapid growth on the heels of precision viticulture, the application of remote sensing in viticulture is in its infancy. With the proliferation of newer, more advanced remote sensing technologies, growers are being tempted by the promise of valueadded products such as yield and quality maps. Scientific investigations are only now in progress to evaluate the capability and utility of remote sensing to directly estimate yield and

15 15 quality parameters. In the meantime, research has demonstrated the ability of remote sensing for simply monitoring and mapping vine-canopy vigour within vineyards. The link between remotely sensed imagery and simple canopy spectral signature and size/shape descriptors is more clearly understood. The ability of remote sensing to provide a synoptic snapshot of vineyard variability could be used for directing in-vineyard sampling to ascertain causes of variability, or as a means of detecting changes in spatial variations of vine vigour during or between seasons. It is recommended that such information be used as part of a greater management strategy. Acknowledgments This work is supported by the Commonwealth Cooperative Research Centres Program and is conducted by the CRC for Viticulture. The authors appreciate ongoing support provided by Charles Sturt University s Spatial Analysis Unit (CSU-SPAN). References Anderson, G.L. and Yang, C. (1996) Multispectral videography and geographical information systems for site-specific farm management. Proceedings of the 3 rd International Conference on Precision Agriculture. Eds. R.C. Robert, R.H. Rust, W.E. Larsen (ASA, CSSA, SSSA: Madison, WI, USA). pp Arnold, G.W., Ozanne, P.G., Galbraith, K.A. and Dandridge, F. (1985) The capeweed content of pastures in south-west Western Australia. Australian Journal of Experimental Agriculture 25, Baldy, R., DeBenedictis, J., Johnson, L., Weber E., Baldy M., Osborn, B. and Burleigh, J. (1996) Leaf colour and vine size are related to yield in a phylloxera-infested vineyard. Vitis 35, Barrett, E.C. and Curtis, L.F. (1999) Introduction to environmental remote sensing (Stanley Thornes: Cheltenham).

16 16 Barrett, M.W. and Leggett, E.K. (1979) The development of aerial infrared photography to detect Echinochloaspecies in rice. Proceedings of the 7 th Asia-Pacific Weed Science Society Conference. (APWSS: Melbourne, Australia). pp Brady, J. and Wiley, S. (2000) Taking AIMS amongst the vines. The Australian Grapegrower and Winemaker 441, Bramley, R. (2001a) Progress in the development of Precision Viticulture - Variation in Yield, Quality and Soil Properties in Contrasting Australian Vineyards (3.3M). In: Precision tools for improving land management. Occasional report No. 14. Eds L.D. Currie and P. Loganathan. (Fertilizer and Lime Research Centre: Massey University, Palmerston North, unpublished). In press. Bramley, R.G.B (2001b). Vineyard sampling for more precise targeted management Proceedings 1st National Conference on Geospatial Information & Agriculture, Sydney (Causal Productions: Sydney) pp Bramley R., and Proffitt, A.P.B. (1999) Managing variability in viticultural production. The Australian Grapegrower and Winemaker 427, Bramley, R.G.V. and Proffitt, A.P.B. (2000) Variation in grape yield and quality in a Coonawarra vineyard. Proceedings of the 5th International Symposium on Cool Climate Viticulture & Oenology, Melbourne, Australia. In press. Bramley, R.G.V., Proffitt, A.P.B., Corner, R.J. and Evans, T.D. (2000) Variation in grape yield and soil depth in two contrasting Australian vineyards. Soil New Horizons for a New Century, Australian & New Zealand 2 nd Joint Soils Conference, 2: Oral Papers. Eds J.A. Adams and A.K. Metherell, (NZSSS) pp Campbell, J.B. (1996) Introduction to Remote Sensing (The Guildford Press: New York, London).

17 17 Celotti, E., De Prati, G.C. and Cantoni, S. (2001) Rapid evaluation of the phenolic potential of red grapes at winery delivery: application to mechanical harvesting. Australian Grapegrower & Winemaker 449a, Clevers, J.G.P.W. (1986). The Use of Multispectral Photography in Agricultural Research. Proceedings 7th ISPRS Commission Symposium, Remote Sensing for Resources Development and Environmental Management, Enschede, 1, Clevers, J.G.P.W. (1988a). Multispectral Aerial Photography as a New Method in Agricultural Field Trial Analysis. International Journal of Remote Sensing 9, Clevers, J.G.P.W. (1988b). Multispectral Aerial Photography as a Supplemental Technique in Agricultural Research. Netherlands Journal of Agricultural Science 36, Clingeleffer, P.R. and Sommer, K.J. (1995) Vine Development and Vigour Control. Proceedings ASVO Viticulture Seminar (Australian Society of Viticulture and Oenology) (Winetitles: Adelaide, Australia). Colwell, R.N. (1983) 'Manual of remote sensing. Vol. II. Interpretation and applications' (Sheridan Press: Virginia, USA). Condit, H.R. (1970) The spectral reflectance of American soils. Photogrammetric Engineering 36, 955. Cook, S. E., and Bramley, R. G. V. (1998). Precision agriculture - opportunities, benefits and pitfalls of site-specific crop management in Australia. Australian Journal of Experimental Agriculture 38, CRCV (2001) Cooperative Research Centre for Viticulture, Project Precision viticulture - Investigating the utility of precision agriculture technologies for monitoring and managing variability in vineyards.

18 Dry, P.R. (2000) Canopy management for fruitfulness. Australian Journal of Grape and Wine Research 6, Escadafal, R., Girard, M-C, Couralt, D. (1989) Munsel soil color and soil reflectance in the visible spectral bands of Landsat MSS and TM data. Remote Sensing of Environment 27, Everitt, J. H., Escobar, D. E., Blazquez, C. H., Hussey, M. A, and Nixon, P. R. (1986). Evaluation of the mid-infrared ( µm) with a black-and-white infrared video camera. Photogrammetric Engineering and Remote Sensing 52, Everitt, J. H., Escobar, D. E., Alaniz, M. A., and Davis, M. R. (1987). Using airborne middle-infrared ( µm) video imagery for distinguishing plant species and soil conditions. Remote Sensing of Environment 22, Grierson, I. and Bolt, S. (1995) Aerial video: A new effective method for vineyard planning. The Australian Grapegrower & Winemaker May 1995, 8-9. Hall, A. Louis J.P. and Lamb, D.W. (2001). A method for extracting detailed information from high resolution multispectral images of vineyards. Proceedings of the 6th International Conference on Geocomputation, University of Queensland, Brisbane. In Press. Harris, J. R., and Haney, T. G. (1973). Techniques of oblique aerial photography of agricultural field trials. Division of Soils Technical Paper 19 (CSIRO: Melbourne, Australia). Haselgrove, L., Botting, D., van Heeswijck, R., Høj, P.B., Dry, P.R., Ford, C. and Iland, P.G. (2000) Canopy microclimate and berry composition: The effect of bunch exposure on the phenolic composition of Vitis vinifera L. cv. Shiraz grape berries. Australian Journal of Grape and Wine Research 6,

19 19 Holzapfel, B., Rogiers, S., Degaris, K. and Small, G. (1999). Ripening grapes to specification: effect of yield on colour development of Shiraz grapes in the Riverina. The Australian Grapegrower & Winemaker 428, Holzapfel, B., Rogiers, S., Degaris, K. and Small, G. (2000). Identifying factors effecting grape berry ripening and berry colour development. Proceedings 5th International Symposium on Cool Climate Viticulture & Oenology, Melbourne, Australia. In Press. Hutchinson, G.K. (2001) Getting mud on the boots (and the lap-top!) - The topoclimate process and providing credible resource information for farmers. Proceedings 1st National Conference on Geospatial Information & Agriculture, Sydney (Causal Productions: Sydney) pp Iland, P.G., Botting, D.G., Dry, P.R., Giddings, J. and Gawel, R. (1994) Grapevine canopy performance. Proceedings ASVO Viticulture Seminar: Canopy Management (Winetitles: Adelaide). Intrieri, C., Poni, S., Rebucci, B. and Magnanini, E. (1997) Effects of canopy manipulations on whole-vine photosynthesis: Results from pot and field experiments. Vitis 36, Intspec (2001) Hymap- airborne hyperspectral scanner. ITRES (2001) Hyperspectral CASI mode. Jackson, R.S. (2000) Wine Science: Principles, Practice, Perception (Academic Press: San Diago). Johnson, L., Lobitz, B., Armstrong, R., Baldy, R., Weber, E., DeBenedictis, J. and Bosch, D. (1996) Airborne imaging aids vineyard canopy evaluation. California Agriculture 50, 14-8.

20 20 Johnson, L., Roczen, D. and Youkhana, S. (2001) Vineyard canopy density mapping with IKONOS satellite imagery. Proceedings 3rd International Conference on Geospatial Information in Agriculture and Forestry, Denver, Clorado (ERIM International Inc.: Ann Arbor, MI, USA). In Press. King, D. (1992) Evaluation of radiometric quality, statistical characteristics and spatial resolution of multispectral videography. Journal of Imaging Science & Technology 36, Lamb, D.W. (1999) Monitoring vineyard variability from the air. Australian Viticulture 3, Lamb, D.W. (2000) The use of qualitative airborne multispectral imaging for managing agricultural crops a case study in south-eastern Australia. Australian Journal of Experimental Agriculture 40, Lamb, D.W. and Bramley, R.G.V. (2001) Managing and monitoring spatial variability in vineyard productivity. Natural Resource Management 4, Lamb, D., Hall, A. and Louis, J. (2001a) Airborne remote sensing of vines for canopy variability and productivity. Australian Grapegrower & Winemaker 449a, Lamb, D.W., Hall, A. and Louis, J.P. (2001b) Airborne/spaceborne remote sensing for the grape and wine industry Proceedings 1st National Conference on Geospatial Information & Agriculture, Sydney (Causal Productions: Sydney) pp Lee, R. (1978) Forest microclimatology (Columbia University Press: New York). Louis, J., Lamb, D. W., McKenzie, G., Chapman, G., Edirisinghe, A., McCloud I. and Pratley, J. (1995). Operational use and calibration of airborne video imagery for agricultural and environmental land management. Proceedings 15th Biennial American Workshop on

21 Colour Photography and Videography in Resource Assessment. Ed P.W. Mausel (ASPRS: Bethesda, MD, USA). pp Manzer, F.E. and Cooper, G.R. (1982) Use of portable videotaping for aerial infrared detection of potato diseases. Plant Disease 66, Moran, M. S., Vidal, A., Troufleau, D., Qi, J., Clarke, T. R., Pinter, P. J., Mitchel, T. A., Inoue, Y., and Neale, C. M. U. (1997) Combining multifrequency microwave and optical data for crop management. Remote Sensing of Environment 61, Munkvold, G.P., Duthie, J.A. and Marios, J.J. (1994) Reductions in yield and vegetative growth of grapevines due to Eutypa dieback. Phytopathology 84, Percival, D.C., Fisher, K.H. and Sullivan, J.A. (1994) Use of fruit zone leaf removal with Vitis vinifera L. cv. Riesling. American Journal of Enology and Viticulture 45, Petrie, R.P., Trought, M.C.T. and Howell, G.S. (2000), Fruit composition and ripening of Pinot Noir (Vitis vinifera L.) in relation to leaf area. Australian Journal of Grape and Wine Research 6, Pitcher-Campbell, S., Tuohy, M. and Yule, I.J. (2001) The application of remote sensing and GIS for improving vineyard management. Proceedings 1st National Conference on Geospatial Information & Agriculture, Sydney (Causal Productions: Sydney) pp Price, J.C. and Bausch, W.C. (1995) Leaf area index estimation from visible and nearinfrared reflectance data. Remote Sensing of Environment 52, Rankine, B.C., Fornachon, J.C.M., Boehm, E.W. and Cellier, K.M. (1971) Influence of grape variety, climate and soil on grape composition and quality of table wines. Vitis 10,

22 22 Rouse, J. W. Jr., Haas, R. H., Schell, J. A., and Deering, D. W. (1973). Monitoring vegetation systems in the great plains with ERTS, Proceedings of the 3rd ERTS Symposium, NASA SP-351 1, (U.S. Government Printing Office: Washington DC.) pp Ryan, S. and Lewis, M. (2001) Mapping soils using high resolution airborne imagery, Barossa Valley, SA. Proceedings 1st National Conference on Geospatial Information & Agriculture, Sydney (Causal Productions: Sydney) pp Smithyman, R.P., Howell, G.S. and Miller, D.P. (1997) Influence of canopy configuration on vegetative development, yield and fruit composition of Seyval blanc grapevines. American Journal of Enology and Viticulture 48, Sun, X., Baker, J., and Hordon, R. (1997). Computerized airborne multicamera imaging system (CAMIS) and its 4-camera applications. Proceedings of the 3rd International Airborne Remote Sensing Conference and Exhibition, Copenhagen, Denmark (ERIM International Inc.: Ann Arbor, MI, USA) pp Taylor, J. (2000) Geographic information systems a step into the information age. The Australian Grapegrower and Winemaker 435, Taylor, J.A. and McBratney, A.B. (2001) Environmental management of a viticultural irrigation district. A "top-down bottom-up" model. Proceedings 1st National Conference on Geospatial Information & Agriculture, Sydney (Causal Productions: Sydney) pp Thomas, C.S., Marios, J.J. and English, J.T. (1988) The effects of wind speed, temperature, and relative humidity on development of aerial mycelium and conidia of Botrytis cinera on grape. Phytopathology 78,

23 23 Tisseyre, B., Ardoin, N. and Sevila, F. (1999) Precision viticulture: Precise location and vigour mapping aspects. Proceedings 2nd European Conference on Precision Agriculture (Sheffield Academic Press, Sheffield, UK). pp Vintage (2001) Viticultural Integration of NASA Technologies for Assessment of the Grapevine Environment. Wahl, K. (1988) Climate and soil effects on grapevine and wine: The situation on the northern border of viticulture - the example Franconia. Proceedings Second International Cool Climate Viticulture and Oenology Symposium, Auckland (New Zealand Society for Viticulture and Oenology: Auckland) pp Wiegend, C.L., Richardson, A.J., Escobar, D.E., Gerbermann, A.H. (1991) Vegetation indices in crop assessments. Remote Sensing of Environment 35, Williams, B. (2000) GPS Applications in Viticulture. Proceeding 5th International Symposium on Cool Climate Viticulture & Oenology, Melbourne, Australia. In Press. Winemaker s Federation of Australia (1996) Strategy The Australian Wine Industry (Magill: South Australia). Yule, I.J., Woodgyer, W.R. and Murray, R. (2001) RTK DGPS and Autodesk Land Development Desktop; a powerful combination for rapid accurate surveying and land development planning. Proceedings 1st National Conference on Geospatial Information & Agriculture, Sydney (Causal Productions: Sydney) pp

24 24

25 25 (a) (b) (c) (d)

26 26 (a) (b)

27 27 Relative reflectance 0.8 Cabernet Sauvignon 0.7 Covercrop 0.6 Soil Wavelength (nm)

28 28 % Reflectance Wavelength (nm) actual spectrum selected wavebands inferred spectrum (a) % Reflectance Wavelength (nm) actual spectrum selected wavebands inferred spectrum (b)

29 29 (a) (b) (c)

30 30

31 31

32 32

33 33 Figure 1 Satellite image (French SPOT) of the Wagga Wagga region, SE NSW, acquired on October Pixel size = 20 m. (a) 60 km x 60 km Full scene, (b) magnified, 13 km x 13 km, sub-scene of Wagga Wagga, (c) magnified, 3 km x 3 km, sub-scene of Charles Sturt University Wagga Wagga Campus, (d) magnified, 400 m x 600 m (24 Ha), sub-scene of Charles Sturt University Vineyard Stages IIA & IV. Figure 2 Multispectral airborne image (false-colour) of Charles Sturt University Wagga Wagga Vineyard acquired January, (a) Altitude = 2.25 km, pixel size = 1.5 m, area coverage = 110 Ha, (b) Altitude = 1.5 km, pixel size = 1.0 m, area coverage = 49 Ha, (c) Altitude = 750 m, pixel size = 50 cm, area coverage = 12 Ha (d) Altitude = 300 m, pixel size = 20 cm, area coverage = 2 Ha. Figure 3 Multispectral airborne image (false-colour) of Charles Sturt University Wagga Wagga Vineyard acquired late February, 2001 under (a) full cloud-cover with cloudbase at 2.4 km, imaging altitude = 1.5 km, pixel size = 1.0 m, (b) clear skies, imaging altitude = 1.5 km, pixel size = 1.0 m. Note the absence of shadows in 3(a). Figure 4 Spectral reflectance profiles for Cabernet Sauvignon, covercrop (chick-peas) and exposed red-brown soil. (Percentage of reflected sunlight = 100 x Relative reflectance). Data acquired from Charles Sturt University's vineyard in Wagga Wagga, NSW. Figure 5 Comparison between an actual vegetation reflectance profile and an inferred reflectance profile using (a) 4 wavebands (multispectral), and (b) 13 wavebands (hyperspectral).

34 34 Figure 6 NDVI images of a Cabernet Sauvignon block with different spatial resolutions. (a) 20 cm, (b) 1 m, and (c) 3m. Vine row spacing = 3 m. Extracted from Lamb et al (2001). Figure 7 Synthetic NDVI image of a block of vines having the same spatial characteristics and vigour. Pixels with dimensions equal to the vine-row spacing will give the same combined signature regardless of where they lie relative to vines or inter-row space. Modified from Lamb et al (2001). Figure 8 Pseudo-colour NDVI image of CSU vineyard with fully developed canopy, January With the inter-row space eliminated from this image, a good indication of vine size as well as overall vigour is conveyed. Figure 9 Grey-scale representation of a single vine-canopy unit extracted from highresolution (25 cm) imagery. Light grey areas represent a high NDVI and dark grey represent low NDVI.