EFFECT OF SNOW LAYER ON GROUND SURFACE TEMPERATURE IN COOL WINE GROWING REGIONS

Juha Karvonen
Department of Agricultural Sciences, Faculty of Agriculture and Forestry, University of Helsinki, Finland

ABSTRACT

This study set out to find how snow affects ground surface temperatures and to evaluate what kind of influence this has on grapevine growing. Temperatures were measured with digital maximum-minimum thermometers that stored the readings. The research shows that in winters when there is little snow, the correlation between ground surface temperature and air temperature is strong (R² = 0.93–0.99) because there is not enough snow to act as an insulator. In winters when there is plenty of snow, no such correlation exists (R² = 0.0001–0.3) because the thick layer of snow acts as an insulator and controls changes in ground surface temperature and air temperature. In very cold weather (-25.0 to -27.4°C), a thick layer of snow (61–67 cm) kept the ground surface minimum temperature at -0.5 to -1.3°C. A thin snow layer (18–32 cm) did not stop the ground surface temperature from falling; the ground surface temperature fell to -10.8°C when air temperature stayed within the same temperature range. In north Europe, central Europe and other cool wine-growing regions, a snowy winter prevents the ground from freezing and thus protects the vines from very cold temperatures.

Key words: snow and viticulture, ground temperature, snow insulation, freezing, Nordic viticulture.

INTRODUCTION

The north European climate cycle traditionally entails a long and snowy winter that central and southern Europeans think prevents successful wine-growing. These days, however, winters in the southern parts of north Europe are rather similar to winters in the northern parts of central Europe [Rötzer and Chmielewski 2001]. In recent years, snow depth and winter temperatures have varied considerably in southern Finland [Pirinen et al. 2012]. This results partly from the climate changes in the Baltic Sea [HELCOM 2007] and partly from the changing currents and winds in the North Atlantic Ocean [Kerr 2004]. In the 2006–2007, 2007–2008 and 2008–2009 winters in Finland, there was little snow and temperatures were mild [Finnish Meteorological Institute 2007]. In contrast, the 2009–2010 and 2010–2011 winters were long, cold and snowy [Finnish Meteorological Institute 2012].

In the Nordic countries (60–65°N and 5–32°E), a permanent snow cover typically falls in November or December. Snow that falls in early winter before severe freezing temperatures contains air, making it a good insulator, keeping the ground from freezing, protecting the base of vines and preventing frost from forming on the roots. Even in severe temperatures, a thick snow cover keeps the ground surface temperature above -5°C; a temperature of -5°C is considered to be destructive to vine roots [Ahmedullah and Kawakami 1986].

In central Europe, vineyards at higher elevations have a thicker snow cover than vineyards at lower elevations, and snow also stays longer on the ground. In Heidadorf, Switzerland, vineyards at an elevation of 1,150 m receive a snowfall of 100–110 cm annually and snow covers the ground from November to March [Meteoswiss 1961–1990]. At lower elevations, snow cover tends to be thinner and last for a shorter period of time. This means that snow provides little protection to the trunks and roots; when air temperature drops to -20°C, vines are at risk of being damaged.

Some research has been conducted on the effect of snow depth on ground surface temperature in China and Mongolia [Guo et al. 1987, Zhang et al. 2007, Xiaoyan et al. 2012], but the topic has received little attention in European textbooks on viticulture. This might be because short-lasting snow cover has not been considered a problem in the mild winters of central and south Europe.

This study set out to examine how snow depth affects ground surface temperatures during freezing weather in the southern parts of north Europe, in the Helsinki-Vantaa region in Finland (Tuusula, 60°24´10´´N; 25°25´45´´E), by measuring simultaneously snow depth as well as air and ground surface temperatures. These observations were then used to evaluate, in regard to previous studies, what opportunities and threats winters with plenty of snow or little snow pose to wine-growing in this region.

MATERIALS AND METHODS

This study examines how two consecutive winters with little snow and two consecutive winters with plenty of snow affect ground surface temperatures (Table 1).

Snow depth was measured manually once a week [Kaukoranta and Niinimäki 2010] and reported as monthly mean thickness in centimetres. Air temperatures were measured at 150 cm and ground surface temperatures at 0 cm with digital thermometers that record maximum and minimum temperatures (Waterproof In-Out Door Max-Min Thermometer with Hygrometer, Shenzhen Hong Tong Yuan Technology Ltd, China, Shenzhen). The thermometers were protected from moisture but not from sunlight. The results of the measurements were recorded daily from 1 November until 30 April, but only monthly absolute minimum and maximum air and ground surface temperatures were taken into account, because the intention was to examine how low ground surface temperature may fall under a snow layer during extreme cold spells, and how the insulating effect of snow reduces variation of frost. Temperatures were measured using the centigrade scale [°C].

The correlation between snow depth and surface ground temperature, and snow depth and air temperature were reported using the Pearson correlation coefficient R2. Its difference to zero was tested with test quantity t and calculated with the formula  when the degrees of freedom (df) were n – 2. The P-value corresponding to the observed test quantity was extracted from the t distribution. Two-tailed t-tests (Student’s t-tests) were used to estimate statistical significance (P = 0.05).

RESULTS AND DISCUSSION

In mid-winter, snow depth remains more or less unchanged in the course of one month and the maximum and minimum temperatures remain similar for several days. Tables 2 and 3 illustrate how a thin snow cover affects ground temperature. In the Helsinki-Vantaa area, there was no statistically significant difference in snow depth (t = 1.2727, P > 0.1) in winters with little snow (2006–2007 and 2007–2008), thus their influence on ground temperature can be regarded as comparable. Air temperatures were slightly colder in winter 2007–2008 than in winter 2006–2007 but the difference was not statistically significant.

In the winter of 2006–2007, the Helsinki-Vantaa area had a very thin snow cover Table 2. The R2 between air and ground surface minimum temperatures was 0.93, and the R2 between air and ground surface maximum temperature was 0.99, meaning there was a very strong correlation between air and ground surface temperatures (Fig. 1 and 2). The minimum ground surface temperature dropped to -10.8°C during March, the coldest month of the 2006–2007 winter. At the time, the average snow depth was 3 cm and the minimum air temperature was -16.1°C (Table 2). This indicates that a thin snow cover does not protect the ground from frost or fluctuating temperatures.

When comparing the snow and climate conditions of the Helsinki-Vantaa area in 2006–2007 [Pirinen et al. 2012] with the climate conditions of vineyards in the central European towns of Heidadorf and Sion (in the Canton of Valais, Switzerland), it is evident that the differences are fairly minimal (Table 2). In Sion, the coldest spell is in January (-12 to -13°C), which is two months before the coldest spell in the Helsinki-Vantaa area. In Heidadorf (1,150 m above sea level) and Sion (790 m above sea level), snow covers the ground from December to March (Statistique de la Ville de Sion, 2010) – almost as long as in the Helsinki-Vantaa area.

In the Hungarian and Austrian wine-growing areas of central Europe, snow falls in late December and a thin layer of snow can still remain on the ground in February. Temperatures are lowest in January when they can fall below -20ºC. In southern Finland, too, snow cover can be thin (Tables 2 and 3) and only stay for a few weeks at a time in mild winters, periodically melting away. These conditions resemble winters in Denmark and the highlands of central Europe.

There is no statistically significant difference (t = 0.7867, P > 0.1) between snow depths in the snowy winters of 2009–2010 and 2010–2011, nor between the lowest air temperatures in these winters (t = 0.9311, P > 0.1) (Tables 4 and 5). This means that the snow and frost conditions of snowy winters are also comparable

In the 2009–2010 and 2010–2011 winters, the lowest ground surface temperature was measured in March when the snow cover was still very thin. Tables 4 and 5 show how the thick snow cover of the 2009–2010 and 2010–2011 winters affects ground surface temperature. In both snowy winters, snow cover was at its thickest around February and March [Solantie 2010]. In the winter of 2009–2010, even though the temperature dropped to -30°C in March, the ground surface temperature remained at -0.5°C under the 60 cm snow coat. The thick snow coat thus protected vegetation very well at the ground surface level.

In southern Finland, snow depth generally varies between 20 cm and 80 cm. In 2010 to 2012, the coldest spell fell at the turn of February and March when snow depth was also the greatest [Finnish Meteorological Institute 2012]. In general, severe temperatures and snowy winters go together. This was demonstrated in the winter of 2009–2010 (Table 4). Pearson’s correlation coefficient (R²) 0.74 between snow depth and the lowest air temperature in Fig.3 demonstrates that the correlation is statistically significant (t = 3.3653 and P < 0.025). The correlation (R²) between air and ground surface minimum temperatures is 0.1 and the correlation between air and ground surface maximum temperatures is 0.2, meaning that these two are not interdependent due to the insulating effects of snow.

In the winter of 2010–2011, the correlation between snow depth and lowest air temperature was almost as great (R² = 0.58) as in the previous winter (Tables 4 and 5 and Fig.4). The correlation is statistically significant (t = 2.3515 and P < 0.05). This observation, made over consecutive snowy winters, confirms the assumption that severe temperatures and snowy winters do in fact go together.

Also in the winter of 2010–2011, the correlation between maximum and minimum air temperatures and maximum and minimum ground surface temperatures was insignificant (R² = 0.0001–0.096) due to the insulating effect of snow. As demonstrated by the results of the measurements presented in Tables 4 and 5 and the correlations and statistical significances calculated from those results, snow acts as an effective insulator and evens out changes between ground surface temperatures and air temperatures.

Previous studies have examined the correlation between snow depth and ground temperatures [Karvonen 2008]. In southern Finland in January–April of the winter of 2005–2006, when the air temperature was -25.4°C and there was a 6–32 cm layer of snow, the lowest ground temperature was 0.0°C, measured at 25 cm below ground; when the air temperature was -33.6°C and there was a 10–55 cm layer of snow, the lowest ground temperature was 0.1°C, measured at 40 cm below ground. When the ground was covered thickly in snow before severe cold temperatures hit, the ground temperature stayed higher and thus the ground did not freeze.

In the fringes of wine-growing regions, such as Quebec in Canada or Siberia in Russia, vines make it through -40°C winters with the help of deep snow or mulch. Thierny et al. [2001] kept the ground free of snow as an experiment and noticed that a ground temperature of -4°C measured at 30 cm underground causes fine root system mortality in the slender roots of maples and birches – although the roots revive to some extent in the following spring.

Zhang et al. [2007] covered vine roots in EVA film and buried the roots under layers of soil. They found that using soil and EVA film together protects side roots better than a simple layer of soil. In the Helsinki-Vantaa area, the ground surface temperature remained at -0.5 to -1.3°C under a thick snow cover, which indicates that a thick layer of snow alone is enough to protect the roots, even without the EVA film or soil layers (Tables 4 and 5).

Snow is used in viticulture in different parts of the world. In Canada, the growth period lasts six months and in winter temperatures can fall to -30°C, the ground freezes and several feet of snow cover the ground [Bells 2013]. In Canada, the snow cover is raised to 1.5 metres by installing a snow fence in the middle of the vineyard to cumulate snow on both sides. The same method is used in Siberia [Yaschenko 2006]. In the Helsinki-Vantaa area, there is no need to cover the vines in both snow and soil as the temperature in southern Finland does not drop to -40°C as it does in Quebec, Mongolia, China or Siberia.

CONCLUSIONS

1. Contrary to common belief, a thick snow layer protects crops and wine from  frost damage during hard winter months.
2. In the Nordic countries and in emerging wine-growing regions, a thick snow coat is highly important for vines in winter.
3. The most harmful conditions are cold winters with little snow: this is when the thin coat of snow is not enough to protect the base and roots of the vines.

ACKNOWLEDGEMENTS

I thank M.Sc. Jyrki Ollikainen of the School of Information Sciences of the University of Tampere for helpful comments on statistical treatment of the results.

REFERENCES

Accepted for print: 12.09.2013

Responses to this article, comments are invited and should be submitted within three months of the publication of the article. If accepted for publication, they will be published in the chapter headed 'Discussions' and hyperlinked to the article.