Assessing a site's physical terroir aspects is arguably the single most important process that any potential grape grower will encounter when starting out (Jones and Hellman, 2003). Combined with matching the site to a grape variety, this decision will ultimately affect the vineyard's yield, the quality of the wine produced, and the vineyard's long-term profitability (Wolf, 1997). In addition, site selection normally involves compromises, in that few sites will possess ideal landscape and climate characteristics in every respect. Terroir assessment (site assessment) represents a complex suite of issues that must be factored into any plan to establish a successful vineyard operation. Numerous overviews exist that detail site selection in general (e.g., Dry and Smart, 1988; Gladstones, 1992) or for specific regions (e.g., Shaulis and Dethier, 1970; Wolf, 1997; Jones and Hellman, 2003) and focus mostly on climate, topography, and soil factors.

Of all site factors, climate exerts the most profound effect on the ability of a region or site to produce quality grapes. Over time the average climate structure of an area has proven to determine to a large degree the defining wine style, with variations in wine production and quality being chiefly controlled by husbandry decisions and short-term climate variability. Climate factors include solar radiation, heat accumulation, temperature extremes (including winter freezes and spring and fall frosts), precipitation during the principal growth stages, wind, and extreme weather events such as hail. Each climate factor should be assessed from the overriding macroscale down to the site or toposcale and even into the canopy (microscale).

The topography and soils of a site play important roles in grapevine growth and quality, and have interactive effects with climatic elements. Topographic factors that exert the greatest influence on a site's climate include elevation, slope, aspect, hill isolation and how it affects air drainage, and proximity to bodies of water. A marginal climate can be mitigated to some degree by locating the vineyard on an ideal site. From a soil standpoint, high-quality wines are made from grapes grown in many different types of soils with no single type considered ideal, however each soil imparts its own unique taste and mouth-feel to a given variety (Wilson, 1998). Grapevines will tolerate a wide range of soils, with nutrient imbalances and moisture issues being the principal factors to consider. Soil drainage is the most important physical property. Waterlogged soils will lead to a reduction in vine health and added difficulties in vineyard management. Adequate soil depth is next in importance. A minimum of 30-40 inches of permeable soil and no impeding layers (shallow bedrock, chemical or physical hardpans) is needed for optimum vine growth (Dry and Smart, 1988). Finally, soil fertility should be moderate for grapevines. Soil that is too fertile along with excessive vegetative growth can be problematic; in contrast, impoverished soils can require economically intensive applications of nutrients. Soil pH is a good indicator of the availability of nutrient cations and therefore fertility.

Oregon currently ranks as the fourth largest wine producer in the United States, growing grapes in six appellations (Federally designated grape growing and wine making regions, called American Viticultural Areas -AVAs): Applegate Valley, Columbia Valley, Rogue Valley, Umpqua Valley, Walla Walla Valley, and Willamette Valley (Fig. 2). In 2002 there were 582 vineyards and 176 wineries growing more than 25 different varieties on over 10,000 harvestable acres with an economic benefit of over 250 million dollars to the state (Oregon Vineyard and Winery Reports, 2002). While Oregon is widely known for its Pinot Noir, other regions in Oregon are finding that warmer climate varieties can be ripened to produce quality wines (e.g., Merlot, Tempranillo, Syrah, and Malbec).

The Umpqua Valley American Viticultural Area (AVA) is located entirely within Douglas County (Fig. 3) (Code of Federal Regulations, 2000). Grape growing and wine making started in the Umpqua Valley during the 1880s, but because of the region's distance to a larger market (i.e., Portland, San Francisco) it never established itself as a dominant industry in the region. In 1961 the region became the home for some of the earliest plantings of post-prohibition winegrapes in Oregon and in 1984 the Umpqua Valley AVA was established. Within the region, there is a diverse array of landscapes and climates that offer the various conditions needed to produce both cool and warm climate grape varieties. Today, the Umpqua Valley AVA has over 650,000 acres zoned for agriculture and, as of 2002, was planted with over 1,100 acres of vines on nearly 60 vineyards (Jones, 2003a). Production totaled over 1200 tons of grapes crushed at twelve wineries in 2002 (Oregon Vineyard and Winery Reports, 2002) with the dominant varieties grown being Pinot Noir, Chardonnay, Riesling, Cabernet Sauvignon, Pinot Gris, Merlot, and Gewurztraminer. Although the varieties listed above make up over 80 percent of the crop in the Umpqua Valley, the region is proving to be well suited for other varieties such as Baco Noir, Maréchal Foch, Tempranillo, Syrah, Malbec, and Dolcetto.

In comparison to the other grape growing regions in Oregon, the Umpqua Valley AVA has the most beneficial climate structure in the state, with the longest growing season, the lowest risk of both spring and fall frosts, and low-ripening period rainfall and temperature extremes (Jones, 2003b). The region is warmer than the Willamette Valley to the north, but cooler than the Rogue Valley to the south (Fig. 2). Heat accumulation varies north to south in the AVA with growing degree days averaging 2200-2400 in the north and 2400-2700 in the south. Precipitation is also quite variable from north to south, averaging 30-60 inches annually (Oregon Climate Service, 2002). The growing season averages 200 days across the AVA, varying from 180 to over 215 days depending on elevation, while the median last and first frost (32°F) occurs on April 10 and October 31, respectively.

The topography of the region is extremely diverse, being derived from the joining of three mountain ranges of varying ages and structure: the Klamath Mountains, the Coastal Range, and the Cascades. The Klamath Mountains extend into the southwestern portion of the AVA and consist of complex folded and faulted igneous and metamorphic rocks that are the oldest in the region. The Cascade Mountains to the east consist of the younger High Cascades and the older, more deeply eroded Western Cascades that make up the eastern boundary of the AVA. The region is protected from the ocean largely by the Coastal Mountains, which are composed of mostly oceanic sedimentary rocks and volcanic islands that were accreted to the landscape over the last 50 million years. Often called the "Hundred Valleys of the Umpqua," the region is drained by the North and South Umpqua rivers, which join near the center of the AVA and empty into the Pacific Ocean. The landscape of the Umpqua Valley AVA is mostly valley lowlands with some isolated hills, stream terraces or benches, and footslopes of alluvial fans scattered by isolated hilltops and ridges. From this diverse geology comes a widely varying mix of metamorphic, sedimentary, and volcanic derived soils. The lower elevations of the valley are mostly deep alluvial material or heavy clays while the hillside and bench locations have mixed alluvial, silt, or clay structures. Complex faulting, especially in the southern part of the AVA, can produce large variations in soil types over areas the size of a vineyard. Drainage and moisture-holding capacity vary greatly by soil type, and while most soils in the region do retain water into the growing season, available water for irrigation during mid to late summer growth is generally needed. Soil fertility varies greatly over the region with issues generally related to either imbalances of nitrogen, calcium, potassium, phosphorous, magnesium, boron, or zinc. Soil pH also varies from region to region (roughly from 4.5 to 7.0) and is mostly due to differences in climate and parent rock material. In general, the soils in the northern portion of the AVA are slightly more acidic than those of the south as a result of more rainfall and greater leaching potential.

Even with over 40 years of post-prohibition grape growing experience, deciding to grow grapes in the Umpqua Valley can be a complex issue owing to the diverse geography and climate. For any existing and potential grape grower, vineyard site selection, or choosing the best terroir, is the single most important decision to be made. Therefore, the research described herein attempts to clarify the locational factors important to grape growing in the Umpqua Valley AVA through a multi-factor terroir-related analysis. The goal of the research is to model the best grape growing landscapes in the Umpqua Valley AVA, and to provide an inventory of sites to facilitate greater success in the grape growing to wine quality continuum.

The topographic landscape was analyzed through the use of over 120 United States Geological Survey 10 meter digital elevation models (DEM). Each 1:24,000 scale quadrangle for Douglas County, Oregon was mosaiked into a complete county DEM grid (raster data). The entire landscape in Douglas County and the Umpqua Valley AVA was then categorized for the most advantageous elevations, slopes, and aspects for growing grapes. The categorization was constructed as a multilayer, topographically driven potential site analysis using ArcGIS (ESRI, 2003) with class rankings given each grid based on its potential. Elevations were categorized from 0-1200 feet into six classes (although grapes are grown at elevations greater than 1200 feet in the region, anything above that value has a much greater frost/freeze risk). Elevations from 400-799 feet were given the highest value and those greater than 1200 feet given a -1 or not viable (Table 1). Slopes were categorized into six classes from less than 1% (or basically flat with poor cold air drainage) to those over 30% being classed -1 or not viable (increasing slopes cause problems using vineyard equipment). The best slopes are considered to be those in the 5-15% range (Table 1). The aspect of the landscape was categorized into five classes with those from south-southeast to south-southwest given the highest ranking because of greater solar radiation receipt and ripening potential (Table 1). The three separate grids were then added together to produce a single topographical suitability grid with values ranging from not viable (anything with a negative score, coming from either being too high in elevation or on extremely steep slopes) to least suitable through most suitable landscapes. Owing to the spatial heterogeneity of most landscapes and realizing that potential sites need to be of adequate size to be profitable, the final grid was resampled using a nine acre neighborhood mean calculation. This size was used as smaller neighborhoods did not adequately produce homogeneous regions while larger neighborhoods created unrealistically smoothed regions.

To analyze the Umpqua Valley AVA's soils, data were obtained from the State Soil Geographic (STATSGO) Data Base for Douglas County Area, Oregon (NRCS, 1997). Site suitability relative to soils was analyzed in a similar manner to Margary et al. (1998) for New York, using the Soils Suitability Extension (SSE) with adjustments made for soils found in Oregon (Stulz, 2001). Four soil properties were used in the categorization of suitability: drainage, depth to bedrock, available water-holding capacity, and pH.

Drainage is thought to be the most important soil factor in establishing and maintaining a vineyard (Cass, 1999) and is influenced by many structural issues such as texture, depth, slope, and aspect. To assess soil drainage, the STATSGO database was analyzed by individual Hydrologic Soil Groups by map unit in the database. The four groups represent variations in drainage from good to poor (groups A to D). In dry growing regions like Oregon, depth to bedrock gives an indication of how well vines can cope with dry periods, with a minimum of 30-40 inches generally needed (Jordan et al., 1980; Dry and Smart, 1988). Mean depth to bedrock was calculated using the STATSGO database from the low to high bedrock depths for each component, and a weighted average was obtained for each map unit. While drainage is extremely important in vineyards, a soil's available water-holding capacity (AWC) is important as those soils with adequate water-holding capacity are at an advantage, giving vines the greatest ability to tolerate periods of moderate drought (Cass, 1999). AWC was calculated from the STATSGO database by computing the mean value for each soil layer, summed over the layers for each component, and then weighted by the percentage of each component per map unit (Margary etal., 1998). Soil pH gives an indication of fertility and nutrient balance with most ideal vineyard soils being found between 5.5-8.0. Outside this range, nutrients may become out of balance, with deficiencies or toxic levels affecting vine uptake or beneficial relationships with microorganisms. Soil pH was computed from the STATSGO database by computing the mean value for each soil component and then weighting by the percentage of each component per map unit.

The spatial polygon data was then converted to grids at the same nine acre resolution to match the landscape suitability grid. Each soil factor was then grouped into classes based upon their individual values; drainage from poor to excessive, available water holding capacity from 0.10 to 0.30 inches of water per inch of soil; depth to bedrock from 25 to 65 inches; and pH from values less than 5.0 to values over 6.0 (Table 2). All classes in each grid were then scaled and weighted with drainage given the greatest weight (40%) and each of the other factors weighted 20%. A final grid of soil suitability was then constructed from the weighting of the four soil factors.

To incorporate land-use issues relative to agricultural development, a statewide generalized zoning coverage was used in the analysis (OGDC, 2003). The zoning data used was initially digitized from data collected from 1983 through 1986 by the Oregon Department of Land Conservation and Development. Since its production, limited zoning changes have occurred and the dataset represents the best statewide zoning coverage available for the state of Oregon. For this analysis only lands zoned agriculture, farm/forest transition, and rural residential are considered as viable parcels on which to develop a vineyard. The farm/forest transition zoning was included, as this type of land tends to be on slopes that are currently used by vineyards and is easily converted to new vineyard sites. The rural residential zoning is set for both two and five acre parcels and is included because many vineyards in southern Oregon can be found on sites in this zoning. The coverage was converted to a grid with all agriculture and farm/forest transition zones given a value of 2, rural residential given a value of 1, and all others a -1 representing a zoning limit to viability.

To assess the climate suitability of the Umpqua Valley AVA, a composite climate model was constructed based upon earlier work conducted for Oregon by Aney (1974) and Mastin (1983). These early analyses of climate suitability used available climate station data to create crude isolines of suitable areas for either cool or warm climate grape varieties. Climate factors used included winter minimum temperatures, growing season length, growing season potential evapotranspiration, heat summations, sunshine hours, and a precipitation index. To better depict the spatial variation in climate factors, this analysis uses the PRISM (Parameter-elevation Regressions on Independent Slopes Model) model, which is derived from a combination of point data, a digital elevation model, and other spatial data sets to create estimates of monthly and annual climate variables that are gridded at roughly 1.45 miles resolution (Daly et al., 2001). The composite climate model was developed for the western Oregon AVAs (Jones et al., 2002) using the most agriculturally and economically important variables.

Absolute climate viability was determined from extreme winter temperatures where areas that experience values lower than 5°F were deemed not viable for grape growing. The model then proceeds to build climate/maturity groupings for cool, intermediate, warm, and hot climates (Fig. 4) based initially on growing degree days (GrDDs; from April to October using a base of 50°F) using practical experience in Oregon. Each climate maturity group is classed to a 25 point range that represents a growing degree day potential within that grouping. The initial GrDD grids are then adjusted for the relative impact that each of the other climate variables has on production viability (frost-free period, last spring frost, first fall frost, rainfall frequency and amount during flowering and harvest, and ripening period temperature variability). For example, the length of the growing season, defined as the days between the first and last frost, is most advantageous when it is at least 180 days or greater (Dry and Smart, 1988). In the modeling, the frost free period was weighted as +2 if over 200 days; +1 if 180-200 days; 0 if 160-180 days; -1 if 140-160 days; and -2 if less than 140 days. The resulting grid was then added to the initial GrDD grid to adjust the grouping relative to how equitable the growing season length is to grape maturation potential. Each of the other climate variables are weighted individually, then added or subtracted to/from the initial GrDD grid. Frost timing is included by using the median last spring and first fall dates (based on 32°F) in reference to the known timing of phenological events for grape varieties grown in Oregon (Jones et al., 2002). Precipitation effects were assessed by weighting climate grids on the frequency and amount of rainfall that occurs during the critical growth stages of flowering and maturation. Finally, a ripening period temperature variability factor (extremes of both daily and monthly temperature), as defined by Gladstones (1992), was included. The temperature variability index accounts for how relatively constant, intermediate temperatures during the ripening period enhance the biochemical developmental processes of colour, flavour, and aroma that lead to high-quality wine. Grids are scored relative to a given ideal range as defined by Gladstones (1992). The final grid represents an adjusted climate/maturity value relative to the sum effects of multiple climate factors that attempt to define maturity potential by minimizing climate-related risks.

The three site factors – topography, soil, and land-use grids – after being internally scaled relative to their individual grape-growing influences (as defined above), were then combined to produce a composite suitability grid taking into account the combined landscape/land-use effect. The composite grid was then masked with the climate maturity group grid to remove all non-climatically viable areas and then overlaid with the climate maturity group grid to produce a spatial depiction of the best vineyard sites in each maturity group in the Douglas County and Umpqua Valley AVA areas.