Vous êtes sur la page 1sur 6

Digital Soil Assessments and Beyond Minasny, Malone & McBratney (eds) 2012 Taylor & Francis Group,

p, London, ISBN 978-0-415-62155-7

Development of terrons for the Lower Hunter Valley wine-growing region


Philip Hughes, Alex. B. McBratney, Brendan P. Malone & Budiman Minasny
Faculty of Agriculture and Environment, The University of Sydney, Australia

ABSTRACT: This study developed terrons for the Lower Hunter Valley wine-growing region. Terrons, combined soil and landscape classes, are a quantitative and practical approach to developing terroirs. From a survey of some 1500 data points covering some 22,077 ha, and eight individual soil and seven landscape attributes 12 distinct terrons were created. By utilising terrons it is possible to summarise large disparate soil characteristics in a manner that was easy to represent graphically. It may be possible to use this model to simplify the manner in which consumers and growers look at soil and as such may be able to add value to local wine production. 1 INTRODUCTION iii. To evaluate the geographic distribution and utilitiy of the terron classes. 1.1 Study area

Terroir is an accepted method of valuing the attributes of a wine-producing (or other food) soil. It comprises more than soil however. The term terroir is hard to quantify. Described as an area whose soil and climate create distinctive properties to food products (Barham 2003), it is nonetheless subjective and determined by perception (Cadot et al., 2011). It is usually specifically linked with a particular region of the earth e.g., Burgundy France, it is a useful method of determining the regionality and some aspects of a wines quality. Carr and McBratney (2005) developed the concept of terron. Soil and terrain attributes from a landscape are analysed by an objective classification procedure and from this the attributes of terron prototypes are established. This study concentrates on the creation of terrons in a small but historically significant winegrowing region of Australia, the lower Hunter Valley, NSW. No terrons nor terroirs previously have been recognised in this region. The aims of this study are: i. To harmonise and use the soil information obtained from a series of soil surveys conducted in the Hunter Valley wine region to create and verify a map of important soil features for wine growing such as marl occurrence, soil class, pH and drainage. ii. To incorporate these soil data along with terrain attributes derived from a digital elevation model to create a set of terron classes.

This study is a part of the lower Hunter Valley on the edge of the Sydney basin, around latitude 32.8S and longitude 151.3E. Originally cultivated for staple foods, emphasis rapidly shifted to luxury products (Moore 1970). In order to accommodate the increase in wine production, irrigation was installed, creating the Hunter Wine Country Private Irrigation District (HWCPID). To the West of the HWCPID is Brokenback Mountain, a part of the Brokenback Range which constitutes a small portion of the Great Dividing Range. The HWCPID has a varied array of geological features. The older units include the remnant of an ancient sea bed, which was deposited during the early Permean and Triassic period (290250 MA) and carboniferous material from this time was compressed. (Gale 1992). This unit contains calcareous deposits as well as coal. On top of this layer is Nowra sandstone followed by more late Permean coal. These layers have folded and in some cases turned nearly 90 degrees. It can therefore be expected that substrate minerals for soil types can change with geographic location. The mix of coal and calcareous material has resulted in a combination of wine and mining as the predominant industries. Landuse in 2005 was reported as 2.8% built up areas, 10.9% vineyards, 38% woodland and 46% pasture. (Manandhar et al., 2009).

31

Figure 1. Interpolation of predicted soil observations in the HWCPID byusing direct soil observations and statistical methods.

2 2.1

MATERIALS AND METHODS Surveys and data

An ongoing soil survey has been conducted since 2000. Sampling was done at different times according to various designs, particularly the Latin hypercube design (McBratney and Minasny, 2006), random toposequence sampling (Odgers et al., 2008) and stratified simple random sampling (Malone 2011). The amalgamated 1543 sites are shown in Fig. 1. At each site the soil was cored to a depth of 1 metre and a full soil description made. From the soil descriptions we extracted the data which we perceived to be important. Soil data was interpolated onto a 25 m by 25 m grid using the methods outlined in (Figure 2). 2.2 Creation of terrons

The predicted soil properties discussed in Figure 2 were combined with the landscape attributes mentioned in Figure 3 via a hierarchical fuzzy k means based on the presence or absence of marl to create terrons. The process was performed twice; initially with the P < 0.5 marl values to create ten non-marl terrons, then with the P 0.5 marl values to create two marl terrons. This split was done because presence of carbonates is seen as very important for vine husbandry. 3 3.1 RESULTS Soil properties

Figure 2. The huntervalley wine country principal irrigation district (HWCPID).

3.1.1 Marl Figure 4 indicates a large concentration of marl to the southwest and this mirrors the pH. The map demonstrates that the vineyards to the south of the HWCPID and a small portion of vineyards to the centre of the map have the necessary substrate for terra rosa soils. There are private

vineyards either directly or partially over these prime areas. The first terrons to be produced were of the most accurate marl prediction. This marl map was split by fuzzy two means into two terrons; one dominating the north, the other dominating the south. The two terrons were named LHVT1 and LHVT2. Terron identification Each terron was differentiated by differences in attributes.

32

Figure 3. Production of a Terron map in the HWCPID by comparing marl data with soil properties and terrain attributes via fuzzy K means.

usually be found in the north. Although the altitudes and slopes are not as extreme as the LHVT2 terron, this terron is still excellent for wine owing to the presence of marl. Soils typical of this terron will have red hues, high pH and excellent drainage. The predominant soil class is red dermosol (60%). 10% of the soil can be classed as calcarosols. LHVT2 This is a marl terron that is identifies itself by extreme altitudes. Sitting at around 170 m it dominates some of the highest areas in the Hunter region. As well as extreme altitude there are larger slope gradients. It is possible that this is the most preferable terron in the Hunter Valley. Vineyards situated in this terron can expect to see red, well drained clays dominating their soil profiles. The high altitude also ensures that these soils are less prone to waterlogging. This soil has fewer calcarosols (9%), and 51% red dermosols. Browner soils are increased with 26% brown dermosols in this terron class. 3.2.2 Nonmarl terrons Predicted from a fuzzy 10 means analysis of the areas in which marl was not indicated.

Figure 4. The presence of marl as indicated by a binomial regression of covariates in the hunter valley.

3.2 3.2.1

TERRONS The marl terrons

LHVT1 This is a marl terron. It is distinguished by its lower altitude. It sits at around 130 m and can

33

LHVT3 Distinguished by having an extreme elevation (over 100 m) but contains no marl. The drainage, wetness index and other such attributes make it otherwise unremarkable. It has a high MRVBF which indicates it is an erosional area. Despite this the altitude, drainage and occasional influence from the higher marls make it a good wine soil. This terron is 61% brown dermosols. LHVT4 This particular terron is in a class of its own. Separated from the previous 4 by RAW MRVBF and TWI and distinguished from the next four by drainage and elevation, and is a terron that has a relatively high membership of wineries. It has high elevation. Older vines can be found here as the poor drainage was used in earlier times as a safeguard against drought. The majority of the soils are an even mix of brown dermosols and brown chromosols. LHVT5 This terron Shares the same basic traits as LHVT7 but is differentiated by a high MRVBF. It has a high spatial distribution to the North West of the HWCPID. Soils in this region can be expected to have variable characteristics dependent on the attributes of upstream areas. As many of the upstream areas contain calcareous material, this terron can be productive for vines. The soils here are an even mix of red dermosols and brown chromosols. LHVT6 LHVT6 shares the same basic traits as LHVT7 and LHVT5 but is separated by elevation. It sits extremely low on the landscape. As a consequence, the TWI is also much higher. The combination of high TWI and high MRVBF would mean this area is prone to erosion. The terron follows the length of Black Creek. The low altitude and high TWI indicates that this area is where most of the regions water terminates. The soils in this terron are an even split between brown chromosols, red dermosols and brown dermosols. LHVT7 Distinguished by a low MRVBF, it is also separated from the rest of the terrons by RAW, a second level of MRVB, TWI and drainage. This Terron is fairly ubiquitous but seems to run nearly half the length of McDonalds Rd. Notable in that it seems to enclose any of the river systems. The higher than normal drainage properties suggest the presence of calcareous material. A relatively low MRVBF of 1.1 suggests that on occasions that are not extreme this area would erode or deposit

dependent on the severity of the rain. This terron is mostly composed of red dermosols (67%) with the majority of the other soil composing brown dermosols. LHVT8 In the same group as LHVT7, LHVT5 and LHVT6 but different in slope and MRVBF. This terron has a much more extreme slope than the other four. It has two large clusters, one near the trig station, the other one on the eastern part of Palmers lane. Its MRVBF of 0.9 makes it the most severe erosion risk of all the terrons in this particular cluster. This terron has the highest proportion of a single soil sub-order, which is 78% red dermosol. LHVT9 LHVT9 has low elevation, high MRVBF, high TWI, low slope, medium insolation, medium to low drainage, medium RAW, moderate average pH and moderate to high pH subsoil. It is similar to LHVT10, but is different by its higher elevation and lower RAW. Although the drainage would need to be managed carefully. It is possible to grow wine on some of these soils, having an even split of red dermosols and brown dermosols as the dominant soil sub-orders. LHVT10 Having the lowest elevation of the terrons, combined with low slope, high TWI mid to low drainage and average RAW. This means that although the soils here are constantly waterlogged, there is evidence to say that the excess water is rapidly cleared. A vigneron using this terron would have to plan carefully to ensure the quality of their harvest. The soils would be in a riparian setting, the colours would be more often brown to black, reflecting this state of affairs. It can be expected that the soils here would be extremely mottled in their B horizons. This terron is composed of just under 50% brown chromosols with a large but smaller proportion of brown dermosols. LHVT11 LHVT11 is separate from the previous terron by drainage and from the next by elevation this is a part of two classes that have differing slopes and TWIs. The major distinguishing factor of this terron is a high TWI. This land class exists almost solely in Werekata national park. The lack of agricultural allocation suggests that this terron is not well suited to crops further evidenced by the fact that this terron almost exactly overlays the remaining natural vegetation. There are areas where there are vineyards in the north east portion of this land class. The straight borders imply that human development is a dominant

34

their own position in the landscape and are desired by vignerons for different reasons. Marl affected terrons are desirable because of the influence lime has on the soil. Altitude affected terrons are desirable because of the superior characteristics that slope brings to air temperatures. Deposition affected terrons are highly prized by viticulturalists as soils that benefit from attributes found all over the region. Water affected terrons need to be managed well, but can also be productive because of the dissolved carbonates and sediment transport. 4.1 Map of terrons overlaid with cadastre

The relative abundance of soil classes within individual terron types is presented in (Table 3).

5
Figure 5. A map of terrons in the HWCPID.

CONCLUSIONS AND FUTURE WORK

factor in the distribution of this terron. This terron has large numbers of Kurosols. There is an even split between red (31%) and brown (22%) kurosols. LHVT12 LHVT12 is another terron which is distinguished by TWI. This one has a lower TWI than LHVT11, its other features are similar. The lower TWI means that it is less prone to water logging. As with LHVT11, almost all of the area delineated by this terron is dominated by bushland. This terron is composed of 59% red dermosols and is the only other terron other than LHVT1 and LHVT2 which has a proportion of calcarosols that is identifiable. At 0.1% this proportion is small. Composite Map of Terrons Combining all of the terrons in the Hunter valley makes a composite map that can display information about soils with regard to their physical location (Figure 5). Locations of wineries can be determined and added as a GIS layer. This has not been done at this stage for commercial reasons, but the presence of vineyards in terron types is clear.

1. From the analysis of the Hunter Valley it was determined that landscape data could be combined with samples collected in the field and meaningful predictions of marl, soil pH, RAW, drainage and soil class could be obtained. 2. When viewed in map form, these attributes demonstrated common features which related to known structures in the Hunter Valley. 3. This information could then be used to statistically create a terron map. The terrons, viewed by local experts, corresponded with what was considered to be the natural state of the soil. This demonstrates that the notion of terron has validity. In the future work should be made to highlight the value of terron and incorporate it into the concept of terroir. 4. There are many different avenues that can be explored from this. There needs to be an entirely different random sample of data taken in the hunter valley to strengthen the validation of the terroir map. A map has been prepared with 200 coordinates for this purpose. All that needs to be done is to travel to each of these locations, survey the area, analyse cores and confirm the absence or presence of marl. The soil data needs to be correlated with the output of the vineyards. Wine quality needs to be assessed in an objective manner. This project would require the cooperation of several vineyards and winemakers.

DISCUSSION REFERENCES
Abdi, H. and Williams, L.J. (2010). Principal component analysis. Wiley Interdisciplinary reviews: Computational Statistics 2(4): 433459.

The terrons described can be divided into categories; Terrons affected by marl, terrons affected by altitude, terrons affected by deposition and terrons affected by water. Each of these tend to have

35

Barham, E. (2003). Translating terroir: the global challenge of French AOC labeling. Journal of Rural Studies 19(1): 127138. Cadot, Y., Caill, S. et al. (2012). Characterisa tion of typicality for wines related to terroir by conceptual and by perceptual representations. An application to red wines from the Loire Valley. Food Quality and Preference 24(1): 4858. Carr, F. and McBratney, A.B. (2005). Digital terron mapping. Geoderma 128(34): 340353. Gale, S.J. (1992). Long term landscape evolution in Australia. Earth Surface. Processes and Landforms 17(4): 323343. Malone, B., McBratney, A. et al. (2011). Empirical estimates of uncertainty for mapping continuous depth functions of soil attributes. Geoderma 160(3): 614626.

Manandhar, R., Odeh, I. et al. (2009). Improving the Accuracy of Land Use and Land Cover Classification of Landsat Data Using Post-Classification Enhancement. Remote Sensing 1(3): 330344. Minasny, B. and McBratney, A.B. (2006). A conditioned Latin hypercube method for sampling in the presence of ancillary in formation. Computers & Geosciences 32(9): 13781388. Moore, D.R. (1970) Results of an Archaeological Survey of the Hunter River Valley, New South Wales, Australia. (Australian Museum). Odgers, N.P. and McBratney, A. (2008) Generation of kth-order random toposequences. Computers & Geosciences 34, 479490.

36

Vous aimerez peut-être aussi