The Okanagan and neighbouring Similkameen valleys in south-central British Columbia (Fig. 1) are home to a young, expanding premium-wine industry. Premium-wine production in the two valleys has risen from 0.8 to 4.7 million litres in the last 12 years (British Columbia Wine Institute, 2004). Although Vitis vinifera varieties have been cultivated in the area for decades, prior to 1990, most of the vineyards were of French hybrid and V. labrusca varieties, which were found to be more productive, hardy and disease resistant than the enologically superior V. vinifera (Association of British Columbia Grape Growers, 1984). In the late 1980s, it became clear that development of an internationally competitive wine industry in the region must be based on premiumV. vinifera wines. This spurred a government-supported vine pullout and replant program that resulted in the conversion of nearly all plantings to V. vinifera. In addition, the higher profit potential compared to some tree fruit production, and the opening of previously uncultivated vineyard-suitable lands, have resulted in further expansion of the vineyard area to its highest ever, totalling more than 2,000 ha. However, sites suitable for V. vinifera production are greatly limited by climate, terrain and the scarcity of arable soil. Thus individual vineyards are small, averaging ca. 10 ha.

The bedrock, soils and terrain in the two valleys are highly diverse owing to the complex geological history of the region. East of Okanagan Valley, the Okanagan Highlands are composed of the Monashee Gneiss, representing the ancient western margin of North America, which was uplifted during the Cretaceous (Roed and Greenough, 2004). West of the Okanagan Valley, the bedrock geology is dominated by Jurassic granites and Eocene volcanic sedimentary rocks, which slid downward and westward off the rising crustal massif of the Highlands during the Eocene. The separation plane between these two bedrock terrains crops out as a complex network of faults in and around the margins of the lake, collectively referred to as the Okanagan Fault (Templeman-Kluit and Parkinson, 1986). Eocene volcanics of the Marron Group are of andesitic composition and were extruded in association with large calderas, the remnants of which are still present at Vernon, Kelowna, Summerland and Penticton (Roed and Greenough, 2004). Eocene sedimentary rocks of the White Lake Formation, including conglomerate, sandstone, siltstone and minor coal, were deposited as fluvial sediments in the landscape surrounding the volcanic edifices; these bedrock units underlie vineyards in the Okanagan Valley.

The valleys have endured many glacial advances, the last occurring in the late Pleistocene (Nasmith, 1962). River valleys were clogged with large volumes of sand and gravel during glacial advances. Although glacial processes eroded bedrock locally, they mainly redistributed pre-existing unconsolidated materials. Thus the materials lining the Okanagan and Similkameen valleys are heterogeneous mixtures derived from many different bedrock sources (see detailed maps in Nasmith, 1962, and simplified map in Roed and Greenough, 2004)

The late glacial history of the Okanagan Valley has been described by Nasmith (1962). Most valley bottom landforms were created during deglaciation. The most influential activities were the melting of stagnant ice lobes in the valley, the periodic creation and breaching of ice dams south of Skaha Lake that created glacial Lake Penticton, and the deposition of thick glaciolacustrine silts which filled the valley. The draining of this lake and the erosional cutting led to benches and bluffs consisting of glacial lake sediments along the shores of Okanagan and Skaha lakes (Fig. 2), and fluvioglacial materials that gave rise to silty, sandy, gravelly and stony soils throughout the area but particularly south of Skaha Lake and to the east of Osoyoos Lake. The presence of ice lobes that filled the valley bottoms also allowed for the formation of alluvial fans that sit above the current valley floor.

Vineyards in the Okanagan and Similkameen valleys (Fig. 1) are established mostly on glaciofluvial and fluvial terraces (benches) and fans, glacial lake deposits, and slope deposit complexes that are reworked glacial deposits (Fulton, 2003). Soil parent materials are composed primarily of unconsolidated glacial deposits having a lithology reflecting regional geological inputs and producing a largely uniform, mixed mineralogy in most soils. What is variable is soil texture and thickness, characteristics that influence rooting depth and the water and nutrient holding capacities. Bedrock is rarely prominent in vineyards developed to date.

The geological events and processes, which led to the diverse bedrock, landforms and soils in the region, and the Okanagan Valley's wide latitude span (49° N to 50°30' N) create a wide range of vineyard edaphic and climatic conditions. Recognizing the challenge grape-growers face in selecting vineyard sites and varieties to cultivate in the valleys, government specialists in the early 1980s put significant effort into identifying and rating the viticultural suitability of agricultural lands. Lands were rated according to the suitability of their soil, climate and sun exposure, and maps showing these ratings were published in the widely used "Atlas of Suitable Grape Growing Locations" (Association of B.C. Grape Growers, 1984). At the time the atlas was published, French hybrid varieties were recommended, and dominated in plantings. Winter survival and high yields, rather than premium winemaking quality, were the main goals of viticulture in the Okanagan and Similkameen valleys. Since that publication, the conversion to V. vinifera and emphasis on quality for winemaking have created needs for more detailed ratings of vineyard site characteristics, and a better understanding of how these characteristics interact with management practices to affect the performance of V. vinifera varieties. While much has been learned from other areas about the response of V. vinifera varieties to climate, soils and management practices, information derived from local experience can be more relevant in developing strategies to improve crop performance while minimizing risk and cost. Indeed, within the Okanagan and Similkameen valleys the combinations of soils, landforms and climates - those characteristics that define terroir (Haynes, 1999; Wilson, 2001) - are unique.

Our approach in constructing our viticultural GIS application for the Okanagan and Similkameen valleys has been to focus on site-specific vineyard attributes. The GIS contains information on the conditions and activities that have resulted in the range of success growers have attained in producing high quality wine grapes from their sites and varieties planted. Its purpose is to provide a data model that growers and researchers can use to study relationships among the site characteristics, including physical, soil and climatic attributes, varieties grown, and viticultural practices that determine grape and wine quality. A similar model for France was described by Morlat (2001) in which a chain of factors contribute to the quality of the wine produced, beginning with site conditions, then influenced by planting material, annual climate, and human practices, including viticultural and wine-making methods.

Any grower or researcher who keeps and accesses records of field conditions, management practices and crop performance by block or site has, in effect, already created and is using a GIS. For large sets of location-based records, modern GIS software (i.e., ArcGIS, ESRI, Canada) facilitates data handling and analysis, and can be used to create maps displaying spatial relationships. In recent years the adoption of GIS analysis has become widespread in agriculture, both by single-farm operators and by regional commodity associations. For viticulture, GIS has been used successfully in the precision management of nutrients and irrigation (e.g., Bramley and Williams, 2001; Lamb et al., 2004) and for modelling vineyard site suitability (e.g., Tesic et al., 2002a, b; Tesic, 2004; Wolf and Boyer, 2003). Application of GIS tools may be particularly useful in the Okanagan and Similkameen valleys where growers are challenged in managing vineyards sited on a limited land base characterized by variable bedrock, soils, climate and terrain.

Although V. vinifera culture in the Okanagan and Similkameen valleys is relatively recent, the potential for world-class wine production has already been verified by the increasing number of press commendations and gold medals received (British Columbia Wine Institute, 2004). Obviously some growers and wineries are already making good management choices and others may be able to learn from them. While France took centuries to develop its viti-culture for each appellation, Okanagan and Similkameen growers have quickly adopted and modified techniques developed elsewhere (including France), and may be able to further fine-tune their varietal choices and management techniques for the specific conditions of their site. Since no two wine regions have the same growing conditions, the GIS should be able to provide growers with the most relevant information on what leads to the finest wines under the unique conditions in British Columbia's interior.

There are currently ca. 250 commercial vineyards in the Okanagan and Similkameen valleys. Vineyards operated by wineries average ca. 29 ha, whereas those operated by independent growers average ca. 4 ha. The average area of individual grape variety blocks is ca. 1 ha. As most vineyard managers keep records for cultural practices and performance on a variety block basis, information in our GIS is collected on the same basis and is organized in three modules: Site Conditions, Management Practices, and Vineyard Performance (Table 1). The Site Conditions module has information on soils, terrain and climate. The Management Practices module includes varieties planted, planting systems designs and management practices. The Vineyard Performance module contains phenological dates, crop yield, basic juice composition, winemaking quality ratings and the vineyard sources of medal-winning wines. An outline of the procedures used for collecting data for each module follows.

To help evaluate the effects of differences in regional conditions on viticultural performance, we have divided the Okanagan and Similkameen grape-growing area into six viticultural regions based primarily on landform and climate (Fig. 7). Differences in the complement of grape varieties planted among regions (Fig. 8) likely reflect general knowledge of the suitability of V. vinifera to climatic conditions, and to published climatic suitability ratings for grape-growing (Association of B.C. Grape Growers, 1984) available to growers when a high number of vineyards were planted in the early 1990s. White varieties dominated plantings in the northern regions and red varieties were planted more in the south. Experience from these V. vinifera plantings, and the value of varietal wines on the market, has likely affected varietal choices in more recent plantings as these contain relatively fewer varieties and are dominated by high-value noble varieties, including Chardonnay, Merlot and Cabernet Sauvignon. A general description of the site conditions and varieties grown in each region follow.

Gauging the suitability of grape varieties to growing regions can be difficult, as differences in viticultural practices and winemaking styles may confound the effects of site conditions on fruit and wine quality. Unfortunately, chemical analyses used routinely to evaluate the composition of grapes do not give clear indications of quality, particularly those for aroma and flavour components, and it is rare to find sites that have had rigorous fruit analyses conducted routinely. A further complication in a young industry is the evolution of grape and wine quality resulting from adjustments made in grape-growing and winemaking methods while the industry gains experience. During that time sufficient evaluation of wines across vintages by experienced judges can be difficult to accomplish and may take several years. Although it may not be feasible to conduct a rigorous analysis of varietal suitability in a small and evolving industry, studies that provide early indications of success in grape varietal choices can be valuable to the industry. Thus we have developed a method for evaluating varietal suitability to regions in the Okanagan and Similkameen valleys based on outcomes of an annual contest for medal awards, the Fall Okanagan Wine Festival, in which wines from all regions are entered and judged blindly by well-recognized and internationally experienced judges. The analysis reported here includes judging outcomes from four years (2000 through 2003), but future results from additional years and a more mature industry may be different and more reliable. Thus our present findings on varietal performance should be considered as a preliminary evaluation.

One of the most variable characteristics of soils in the Okanagan and Similkameen vineyards is texture. Soil texture is determined by the size distribution of component mineral particles (Wittneben, 1986). Coarse-textured soils have sand-sized particles as their main component, whereas fine-textured soils consist mostly of silt- and clay-sized particles. Soil texture affects permeability for drainage, water holding capacity, and the tension with which water is held to soil particles. It can thus have substantial impact on vine growth and development including vegetative vigour and berry composition, which determines wine-making quality. To examine whether soil texture has a major effect on the quality of wines produced in the Okanagan and Similkameen valleys, the distribution among soil textural classes of all grape-variety blocks was compared with that of blocks from which medal-winning wines were produced (Fig. 12). If soil texture has no effect on wine quality the distributions are expected to be similar. Chi-square goodness-of-fit tests were used to detect differences in the distributions.

Although most of the vineyards in the Okanagan and Similkameen valleys are sited on coarse (i.e., loamy sand) or moderately coarse (i.e., sandy loam) soils, we found that medal-winning wines have been made from grapes originating from a range of soil textures, from loamy sands and sandy loams to finer textured clay loams and silty clay loams (Fig. 12). Nonetheless, based on the chi-square analyses, the distributions of all blocks and medal winning blocks among textural classes were different in Kelowna, Penticton, Golden Mile and Similkameen. Kelowna is the only region with significant grape production on moderately fine-textured soils (i.e., silty clay loams; Fig. 12), and these soils have produced a disproportionately high number of medal-winning wines. In Penticton (2) where a discontinuous eolian veneer overlies the glacial lacustrine deposits, medals have favoured grape production from the sandy loam soils associated with this veneer. In Golden Mile (4), where the soils originated from fluvial fan deposits, vineyards located on sandy loam soils have also performed exceptionally well. In Similkameen (6), medals also favour grapes produced on sandy loams. This suggests that loamy soils, which are neither extremely coarse nor fine, are especially suitable for producing quality wine grapes, perhaps because they are adequately drained yet hold sufficient water and nutrients to meet the requirements of vines for optimum performance.

Further study is needed to elucidate the mechanisms by which soil textural properties, as well as the mineralogical and chemical composition of soils, affect wine quality. As noted by Wilson (2001) and others, such affects may constitute an important aspect of terroir. The question of whether wine quality characteristics, such as aroma and flavour, are correlated with specific chemical or physical properties of the soils in the Okanagan and Similkameen valleys is being explored in our current work.