17550238, 2011, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1755-0238.2010.00118.x by Cochrane Romania, Wiley Online Library on [21/12/2022].

See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
52 Effects of climate change on grape phenology Australian Journal of Grape and Wine Research 17, 52–61, 2011

Projecting the impacts of climate change on the phenology of

grapevine in a mountain area _118 52..61

A. CAFFARRA and E. ECCEL

IASMA Research and Innovation Centre – Fondazione Edmund Mach Environment and Natural Resources Area Via
Mach, 1 – 38010 San Michele all’Adige (I), Italy
Corresponding author: Dr Emanuele Eccel, fax +39 0461 650956, email [email protected]

Abstract
Background and Aims: The strong link between climate and grapevine phenology suggests a potentially stronger
impact of climate change on viticulture in climate-limited areas, including mountain zones. The aim of this work was
to evaluate the potential effects of climate change on grapevine phenology and viticultural suitability in a mountain
region.
Methods and Results: Climatic projections were applied to phenological models to determine the effect on stages
of budburst, flowering and veraison for Vitis vinifera cv. Chardonnay. Calibration and validation of the models had
been previously carried out in the same alpine region. The output of the general-circulation climatic model HadCM3,
run with two different Intergovernmental Panel on Climate Change emission scenarios, was statistically downscaled
to 10 locations in different agricultural sites in Trentino, Italian Alps, some of which are presently unfit for viticulture
due to climatic limitations. Results yielded a trend of phenological advance, with interesting differences among
phases and sites. Simulated advance was more pronounced at higher elevations, and larger for veraison than for
spring phenophases.
Conclusions: As a consequence of the considerable warming projected by future climate scenarios, some mountain
sites at about 1000 m were expected to fall within areas climatically suitable for viticulture before the end of this
century. Nevertheless, noticeable differences from present conditions are not expected within a short timescale.
Significance of the Study: These projections, suggesting a more pronounced phenological response at higher
elevations, may support the development of adaptation strategies aimed at maintaining the profitability of mountain
viticulture even in times of global change.

Abbreviations
AOGCM Atmosphere-Ocean Global Circulation Model; DOC Denominazione d’Origine Controllata (Italian
‘controlled denomination of origin’); DOY day of the year; GHG green-house gases; IPCC Intergovernmental
Panel on Climate Change; SRES Special Report on Emission Scenarios

Keywords: budburst, climate projection, flowering, mountain area, phenology, veraison, Vitis vinifera

Introduction carried out (Mariani et al. 2009), as well as in many other parts
Much has been written on the several issues of the impact of of the world (Jones and Davis 2000, Chuine et al. 2004, Meier
climatic change on agriculture (for reviews, see: Olesen and Bindi et al. 2007, Schleip et al. 2008). In the last 20 years, in the
2002, Fuhrer 2003, Maracchi et al. 2005, Grünberg et al. 2007). southern-alpine region of Trentino, the dates of attainment of
According to the Intergovernmental Panel on Climate Change spring and summer phenophases have significantly advanced at
(IPCC) Forth Assessment Report, global warming could substan- both valley floor (San Michele) and mountain (Cembra) loca-
tially alter agricultural systems (Easterling et al. 2007). In par- tions, as shown in Figure 1.
ticular, climate change is expected to have a significant impact on Alpine zones are climate-limited and may be particularly
crop growth, which is determined by the weather conditions sensitive to changes in temperature (Thuiller et al. 2005). For this
during the growing season. Models linking plant growth with reason, mountain agro-ecosystems have been listed as particu-
climate patterns are particularly useful as they can project the larly vulnerable to the impacts of climate change by the Food and
effects of climate change on future yields. In particular, the study Agriculture Organization. The strong link between climate and
of the climate control of crop phenology is fundamental for a grapevine physiology suggests that impacts on viticulture might
better formulation of adaptation policies (Duchene and be more significant in mountain areas. Changes in temperature
Schneider 2005, Wolfe et al. 2005, Moriondo and Bindi 2007). are likely to affect the duration of the growing season and the
Important effects on agriculture may stem from a changed quality of harvested grape, which determine viticultural profit-
water balance regime; nevertheless, temperature plays a major ability. It is known that grape quality depends on meteorological
role in regulating plant phenology (Battey 2000). The effects of features during the ripening season, and particularly during the
climate warming are already evident on the phenological timing period immediately preceding vintage (Jones et al. 2005, White
of grapevine in the southern Alps, where the present study was et al. 2006, Moriondo and Bindi 2007).

doi: 10.1111/j.1755-0238.2010.00118.x
© 2011 Australian Society of Viticulture and Oenology Inc.
 17550238, 2011, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1755-0238.2010.00118.x by Cochrane Romania, Wiley Online Library on [21/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Caffarra and Eccel Effects of climate change on grape phenology 53

Ukraine. While the above study is certainly interesting for

assessing the general impacts of climate change and direct Euro-
pean policies, smaller-scale studies are also necessary to support
local viticulture. In particular, the results from large-scale
studies (e.g. Webb et al. 2007) are not applicable to mountain
regions, where large climate variability and environmental het-
erogeneity exist within short distances.
The aim of this work was to evaluate the impacts of climate
change on the phenology of grapevine in an alpine region with
a high viticultural value, and assess potential changes in the
viticultural suitability of different locations, sited at different
elevations. The analysis of potential impacts of climate change is
important to develop adaptation strategies that will help to
maintain the competitiveness of the wine industry.

Materials and methods

Geographic and climatic features of the target areas
The target area covers several climatic and agricultural sites in
Trentino, a region in the central-eastern Italian Alps (Figure 2).
This mountainous area consists of a system of minor valleys
converging to the larger Adige River Valley. The latter is deep,
with steep flanks and a flat bottom formed by alluvia from the
Adige River. It gently slopes down towards the Po plain, with a
decrease in elevation of about 50 m in 65 km.
In general, Trentino has a humid, temperate, oceanic
climate (maritime air masses have influence all over the year),
particularly in the pre-alpine areas. Inner valleys show a cooler
and drier, more continental-alpine climate. Precipitation
amounts are distributed over two maxima, in autumn (main)
and in spring (secondary), although in some mountain areas,
rainfall peaks in summer (Eccel and Saibanti 2007). Viticulture
occupies a set of climatically diverse areas (Figure 2), from the
sub-Mediterranean area north of Lake Garda, at about 70 m
a.s.l., to the continental-alpine area of the highest commercial
vineyards at about 700 m in the northern part of the region.
Figure 1. Observed dates (day of the year) of budburst (䊉), flow- Chardonnay is the most widely grown variety, covering about
ering (䉱) and veraison (䉬) for Chardonnay in San Michele and one-third of the total viticultural area. Thus, the interest in
Cembra (budburst series for Cembra not available). All regression phenological projections for this variety is particularly high.
lines are significant at P < 0.05, except for budburst in San Michele. For the simulations, 10 sites were chosen in Trentino.
Among these, five are located in valley floors, for a simulation of
Another likely consequence of increasing temperatures is impacts on low-elevation viticultural areas, and the other five
an expansion of viticulture-suitable areas into more elevated are located on mountains at more than 500 m of elevation, to
areas, where climate is presently too cold for grapevines. gain insight into the impacts of climate change on mountain
However, the suitability and potential profitability of candidate viticulture and its potential upward expansion. The projections
viticultural areas have to be considered very carefully. In tem- of future viticultural suitability were based on future tempera-
perate zones, temperature variations will affect grapevine ture projections, while other variables affecting the establish-
growth depending on local climatic constraints, the variety ment of viticulture (i.e. site morphology, cultural and
considered, growing practices and the temperature control of economical features) were not considered. The selected sites are
pathogen life cycles. described in Table 1 and shown in Figure 2.
Phenological models provide a powerful tool for predicting
the timing of key physiological events during the annual cycle of Creation of temperature climatic series for
grapevine, making up for the absence of direct observations, as phenological projections
in the case of simulation of future scenarios. They are largely For the simulation of future climate, a statistical downscaling
applied in viticulture: for planning cultural practices (Williams was carried out on the output of the Hadley Centre’s
et al. 1985, Mariani et al. 2006), for the forecast of plant vul- Atmosphere-Ocean General Circulation Model (AOGCM)
nerability to pest attacks (Thomson et al. 2009) and for the HadCM3 (Pope et al. 2000). Two projected temperature series
simulation of the fitness of cultivars to given climatic conditions were available for Trentino from the statistically downscaled
(Gladstones 1992, Orlandini et al. 1993, Bindi et al. 1996, output of HadCM3 model, one representative of the large, low
Oliveira 1998). At a very large scale, Harrison and Butterfield valley floor environment (San Michele, 210 m), the other of a
(2000) have considered the potential expansion of viticultural mountain valley context (Cles, 650 m). The model HadCM3
areas in Europe. Their projections suggest that a temperature was run with a resolution of 2.5° latitude and 3.75° longitude.
increase of 4°C will result in a general shortening of the growing The downscaling algorithm was a daily-resolution ‘transfer
season and an eastward and northward shift of viticultural function method’ (Eccel et al. 2009). For each site, tempera-
areas, to include England, Poland, Rumania, Belarus and ture series were available for the A2 and B2 scenario families,

© 2011 Australian Society of Viticulture and Oenology Inc.

 17550238, 2011, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1755-0238.2010.00118.x by Cochrane Romania, Wiley Online Library on [21/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
54 Effects of climate change on grape phenology Australian Journal of Grape and Wine Research 17, 52–61, 2011

Figure 2. Sites under study

and grape growing areas in
Trentino (shaded).

Table 1. Stations used for the simulations, their elevation above sea level, morphology, mean observed temperature
and land use.

Station Elevation (m) Site morphology Mean temperature Land use

2000–2009 (°C)

San Michele all’Adige 228 Valley floor 12.3 Viticulture

Cembra 600 Mountain slopes 11.1 Viticulture
Avio 131 Valley floor 13.8 Viticulture
Borgo Valsugana 386 Valley floor 11.0 Fruit-growing/forage
Rovereto 204 Valley floor 12.2 Viticulture
Storo 409 Valley floor 11.0 Fruit-growing/forage
Fondo 982 Mountain slopes 9.5 Fruit-growing/forage
Lomaso 500 Mountain slopes 9.6 Fruit-growing/forage
Sant’Orsola 1022 Mountain slopes 9.8 Forage/fruit-growing/woodland
Zortea 1028 Mountain slopes 8.4 Forage/woodland

chosen from IPCC’s ‘Special Report on Emission Scenarios’ – scenario) projection is in the average of five benchmark
SRES (IPCC 2000) as representing two intermediate hypo- models from 2000 to the years 2040s, and slightly raises above
theses on the increase in concentration of greenhouse gases group average after the first half of the century.
(GHG). A2 represents a planetary development of emissions Sub-downscaling to more sites was performed by translating
without any self-regulation, whereas B2 assumes the adoption the two downscaled series to the other sites by applying a
of self-imposed standards of reduction in the emissions. Com- normalization of daily series. For each of the 10 sites, a reference
parisons with other AOGCMs, including IPCC 4th Assessment (‘master’) series was chosen, either San Michele (for valley floor
Report, point out that HadCM3 projections for the selected locations) or Cles (for hill side or mountain locations). Normal-
atmospheric scenarios yield intermediate global temperature ized master series were obtained by subtracting monthly aver-
increase rates (Meehl et al. 2007). Comparison for the Medi- ages and dividing by the relevant standard deviation. For each
terranean area (including the northern Italian region) shows one of the remaining 10 sites, the normalized series derived
that ‘HadCM3 clearly exhibits an average response for A2 from the master series were reverted to absolute, site-specific
and a conservative one for B2, with respect to the time of values by applying the inverse transfer function (multiplying by
2°C global warming between 2049 and 2060’ (Giannakopou- the standard deviation of each series and finally by adding to
los et al. 2009). Also, the comparison of the temperature each series their monthly means). This approach allows consid-
model outputs from the Coupled Model Intercomparison eration of both temperature offsets from the ‘master’ series and
Project (http://cmip-pcmdi.lhl.gov) reveals that HadCM3 (A2 their differences in scattering around mean values.

© 2011 Australian Society of Viticulture and Oenology Inc.

 17550238, 2011, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1755-0238.2010.00118.x by Cochrane Romania, Wiley Online Library on [21/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Caffarra and Eccel Effects of climate change on grape phenology 55

Phenological model slope environment) are represented in Figure 3. A moderate

The chosen model (called ‘FENOVITIS’ from now onwards) was increase in temperature can be seen from the present to the
recently developed for the simulation of spring and summer middle of the century, with little or no differences between the
phenological stages (budburst, flowering and veraison) of Char- two scenarios. In contrast, the second half of the century shows
donnay (Caffarra and Eccel 2010). It was calibrated using a a stronger trend, with A2 warming abruptly and B2 maintaining
long-term phenological dataset from Veneto (Northern Italy) a constant but moderate temperature increase. At the end of the
with mean annual temperature 13.1°C, and validated on a set of simulation period, mean annual temperatures are projected to
four stations (three of which were located in Trentino) with exceed present values by about 2 and 4.5°C, in the B2 and A2
mean annual temperatures ranging from 10.4 to 13.4°C. The scenarios, respectively. No major changes are expected for the
FENOVITIS model considers the action of cool (chilling) tem- next 20 years, but a slight temperature increase is predicted
peratures for dormancy release, and the action of warm according to scenario A2.
(forcing) temperatures for subsequent bud growth. It describes Emission scenarios A2 and B2 become different already in
phenological progression in terms of developmental units, cal- the early years of the present century; nevertheless, differences
culated through fitted or experimentally measured relation- in atmospheric composition and, consequently, in temperature
ships: chilling units are accumulated up to a critical chilling increases between the two scenarios become evident from about
threshold simulating dormancy release, which is followed by the middle of the centennial simulation period. This is the result
the accumulation of forcing units up to a critical forcing thresh- of the mechanism of transfer of the increase in GHG emissions
old, simulating budburst. The inclusion of chilling and the use of into their atmospheric concentration, which involves a large
an experimentally established relationship for quantifying the time lag due to oceanic uptake, scavenging by the natural envi-
action of warm temperatures on growth make the FENOVITIS ronment and the mechanisms of GHG release in response to
model comparatively more process-based than phenological temperature increases. One consequence can be seen from the
models based on degree-days (such as Winkler degrees) or other years in which simulated mean annual temperatures of scenario
bioclimatic indices. When applied to the validation (external) B2 exceed those of scenario A2. This behaviour strengthens the
data series, it yielded smaller prediction errors than the Winkler- confidence in simulations of climatic evolution in the next few
degree model, and showed a better performance on warm years, decades, which are a little responsive to the emission rates
including extreme years such as 2003, suggesting its reliability hypothesized by SRES.
for the application to climate change scenarios. The projected dates of budburst, flowering and veraison
During the simulation, we considered the 31st of August during the period 2000–2098 are shown in Figure 4, for three
as the latest possible date of veraison for quality Chardonnay sites representative of different climatic areas: valley floor (San
growing, based on common experience in this alpine area. In Michele), present ‘mountain’ growing context (Cembra) and
fact, to ensure optimal ripening, veraison should occur in warm mountain agricultural environment, presently unsuitable for
conditions with time ahead for sugar accumulation (Huglin and grape growing (Zortea). There is a clear advance in the dates of
Schneider 1998). At the upper limits of the viticultural area of phenophases at all stations. Flowering and veraison are most
Chardonnay in this region (around 700 m, depending on slope influenced by warming, while budburst does not show large
and aspect), veraison usually occurs before the end of August advances. Nonetheless, the trend towards earlier occurrence is
Thus, when the occurrence of veraison was simulated later than negative for all phenophases, at all sites (Table 2). Significance
31st of August, the year was considered too cold for quality- levels vary according to the climate scenario and the location.
wine production. For example, the advance of budburst is significant at all stations
according to scenario A2, but only at Borgo Valsugana, Lomaso
Analysis and Zortea according to scenario B2. These differences are due
The projected series of budburst, flowering and veraison were to the less dramatic increase in temperature projected by sce-
visually evaluated using scatter plots and bar plots. Bar plots nario B2, but they also depend on the winter climate at each
were created comparing mean decadal dates of veraison over station. The budburst sub-model predicts that in the warmer,
two different periods: 2020–2029 and 2070–2079 for the A2 and valley-bottom stations, like San Michele and Avio, the faster
B2 scenarios. This comparison enabled evaluation of the differ- accumulation of forcing units during late winter months will be
ences between the beginning and the end of the projected counterbalanced by the delayed chilling fulfilment due to tem-
period. Regression lines were fitted on the projected phenologi- perature increase. Indeed, the trends of the simulated dates of
cal series to analyse temporal trends, for each of the 10 stations chilling fulfilment are positive for all stations with particularly
considered. Finally, in order to investigate differences in the late dates (first week of January) for warmer sites like Avio
modelled phenological trends among sites, their regression co- (results not shown). The final effect at some sites is a non-
efficients were plotted against site elevation and the mean significant trend in the timing of budburst. The less marked
2000–2098 temperature. advance of budburst compared to flowering and veraison, and
the faster phenological development due to higher tempera-
Results tures, will shorten the time intervals between these phe-
The comparison between simulated and observed dates of verai- nophases (Figure 5).
son during the periods overlapping with the scenarios revealed On the other hand, flowering and veraison show significant
that simulated phenology was reasonably close to reality, in advances at all stations but Zortea, due to the short duration of
particular in scenario A2. In San Michele (1990–2000), the the available series (Table 2). The slopes of the regression lines
observation mean was 221 (day of the year) and the prediction (regression coefficients) for the projected dates of veraison are
mean was 223 (A2) or 224 (B2). In Cembra (1991–1994; 2002– particularly pronounced at mountain stations (i.e. Sant’Orsola
2007, 2009), the observation mean was 222, and the prediction and Fondo) at the upper limit of the considered range. The
mean was 222 (A2) or 226 (B2). climate of these sites is currently too cold for grape ripening, but
The downscaled monthly climatic projections (according to the projected increase in temperature would make them suit-
scenarios A2 and B2) for the ‘master’ stations of San Michele able for the commercial growth of Chardonnay by the end of the
(low valley floor environment) and Cembra (elevated, valley century. In Zortea, projections simulate a climate warm enough

© 2011 Australian Society of Viticulture and Oenology Inc.

 17550238, 2011, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1755-0238.2010.00118.x by Cochrane Romania, Wiley Online Library on [21/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
56 Effects of climate change on grape phenology Australian Journal of Grape and Wine Research 17, 52–61, 2011

Figure 3. Projected yearly

temperatures over the period
1990–2098 at San Michele (a)
and Cembra (b) for scenarios
A2 and B2. Lines: 10-year
moving average (2000–2098)
for scenario A2 (solid line),
and scenario B2 (broken line).

to induce veraison before the beginning of September only from between the 20s and the 70s of the century (averages of the
the 2060s; thus, the simulated veraison series include only a few periods 2021–2029 and 2070–2079, Figure 7). The advance is
years, at the end of the period under study. Thus, for this site, more pronounced for scenario A2 than B2, 21 versus 15 days
the trend in the timing of veraison is not significant. (difference between the two periods, averaged over all sites in
The higher climate sensitivity of grape vines grown at which viticulture is practicable in both periods). For the two
mountain stations is shown by the significant correlations of the sample stations of Cembra (mountain slope) and San Michele
phenological advance rates (regression coefficients of the phe- (valley floor), we also calculated the mean veraison date on
nological trends, shown in Table 2) with the mean projected the hindcast simulation of the period 2001–2010, and the
temperature and with site elevation. The projected advances of phenological advance occurring by 2020s and by 2070s. This
flowering and veraison are more pronounced for stations with a advance, averaged over the two A2 and B2 scenarios, would
lower mean temperature, or at higher elevation (Figure 6). be of approximately 4 and 5 days by the 2020s and of 15 and
Magnitude and significance of these trends are shown in 19 days by the 2070s, for San Michele and Cembra, respec-
Table 3. The linear relationship between elevation and the pro- tively. While at present veraison occurs on average in the first
jected rates of advance of flowering and veraison implies that (San Michele) and second (Cembra) week of August, our pro-
the same increase in temperature will result in a larger pheno- jections show that by the 2070s, it would occur around the
logical advance at higher (and colder) sites. same time for both sites, i.e. on the third week of July (mean
The comparison between phenological simulations for sce- dates are 20 and 21 July for Cembra and San Michele, respec-
narios A2 and B2 shows that they differ noticeably only from tively). By the end of the century (2090s), the simulations
the second half of the 21st century. This seems in good accor- show an advance in veraison of 21–30 days compared to
dance with the projected temperature series. Bar plots of the present, depending on scenario (A2 or B2) and site (San
mean dates of veraison show a clear phenological advance Michele or Cembra).

© 2011 Australian Society of Viticulture and Oenology Inc.

 17550238, 2011, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1755-0238.2010.00118.x by Cochrane Romania, Wiley Online Library on [21/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Caffarra and Eccel Effects of climate change on grape phenology 57

Figure 4. Model simulations of future dates of budburst (x), flowering (䉱), and veraison (䊉) for Chardonnay, according to scenario A2
(upper row) and B2 (lower row) for three of the sites under study: San Michele (low elevation), Cembra (mid altitudinal range) and Zortea (high
elevation). Interpolation LOWESS (LOcally WEighted Scatterplot Smoothing) curves (solid lines) are superimposed to the phenological series.

Table 2. Rates of shift in phenophase occurrence (d/yr) for each station and climate scenario.

Station Budburst Flowering Veraison

A2 B2 A2 B2 A2 B2

San Michele all’Adige -0.07 -0.04 -0.20 -0.14 -0.31 -0.21

Avio -0.06 -0.03 -0.18 -0.12 -0.27 -0.18
Borgo Valsugana -0.10 -0.05 -0.26 -0.17 -0.39 -0.25
Rovereto -0.07 -0.04 -0.21 -0.15 -0.33 -0.22
Storo -0.08 -0.05 -0.26 -0.17 -0.39 -0.25
Cembra -0.07 -0.04 -0.20 -0.16 -0.36 -0.23
Fondo -0.10 -0.04 -0.33 -0.22 -0.42 -0.26
Lomaso -0.12 -0.05 -0.28 -0.19 -0.38 -0.19
Sant’Orsola -0.10 -0.05 -0.30 -0.21 -0.42 -0.20
Zortea -0.13 -0.06 -0.40 -0.27 -0.36 -0.04

Figures in bold indicate a significant advance over the period 2000–2098 (P < 0.05).

Discussion The different timing of phenophases and their shorter

At valley floor stations, the projected phenological advances duration is likely to affect vintage quality. The simulated phe-
would shift veraison from early–middle August to the middle of nological advances will shift the ripening period from late
July by the 2070s. Thus, future phenology of Chardonnay in summer/early autumn to the middle of summer, implying that
cool Italian areas, like Trentino, might resemble current phenol- ripening will occur in warmer conditions and in a shorter time
ogy in warmer regions where Chardonnay is grown, such as interval. This will likely lead to a compression of vintage times
southern Italy. Phenological observations from Ragusa (south- (Webb et al. 2007) and might influence grape quality, which
ern Sicily) show that Chardonnay attains veraison 2 to 4 weeks has been shown to depend on temperature during the ripen-
earlier than in Trentino. ing period (Duchene and Schneider 2005, White et al. 2006,

© 2011 Australian Society of Viticulture and Oenology Inc.

 17550238, 2011, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1755-0238.2010.00118.x by Cochrane Romania, Wiley Online Library on [21/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
58 Effects of climate change on grape phenology Australian Journal of Grape and Wine Research 17, 52–61, 2011

for high vintage quality. In the coming decades, most likely from
the second half of the 21st century, present DOC areas may no
longer feature the optimal climate for their reference varieties,
and it may be necessary to shift them towards higher or cooler
areas. Our simulations show that the climate suitable for Char-
donnay may expand upward to sites currently too cold for this
variety. This might occur in the next few decades for sites where
viticulture is now at its beginnings (e.g. Lomaso area), and
during the second half of the 21st century for cooler, higher sites
in the mountain belt at around 1000 m.
Nevertheless, other issues have to be considered when pro-
jecting the expansion of viticulture-suitable areas. Pre-existing
land uses, including other crops, possible difficulties in exploit-
ing mountain environments, terroir features and vulnerability to
pathogens, are all matters that have to be taken into account
when assessing the suitability of a new area for viticulture.
The bar plots of the mean dates of veraison for the decades
2020–2029 and 2070–2079 show that for the mountain sites of
Fondo, Sant’Orsola and Zortea, the phenological requirements
for grape growing may be met only from the second half of the
century, while for Lomaso, the heat needs may be satisfied
already in the next decades. In the mountain sites considered –
where the term ‘mountain’ means locations at about 1000 m or
more a.s.l. – simulated veraison projected for years 2070–2079
occurs at dates perfectly comparable with the present ones for
sites like Cembra (today at the upper limit of the viticultural
area in Trentino), or even earlier. Thus, given the present upper
limit of viticulture in Trentino (about 700 m), climate change
would result in an upward shift of the climate suitable for
Chardonnay of 300–400 m by the end of the century.
The more pronounced advance of veraison at higher eleva-
tions suggests a higher phenological sensitivity of mountain sites
to climate change. This response is typical of climate-limited
zones, and is analogous to the phenological patterns found by
other studies (Defila and Clot 2001, Schleip et al. 2006), report-
ing stronger trends of advance at higher elevations. Indeed,
ecotones between different ecoclimatic regions have been
indicated as particularly sensitive to global warming by the
Figure 5. Time intervals between simulated phenophases for (a) IPCC (Barker et al. 2007). While this indication was originally
San Michele, and (b) Cembra. referred to natural ecosystems in transition areas, it can also be
reasonably applied to transition agro-ecosystems, like viticul-
ture at its upper altitudinal limit. It is reasonable to expect that
Moriondo and Bindi 2007). A long-term study (1952–1997) an increase in temperature will lead to more pronounced
on vineyard observations in Bordeaux showed an earlier growth changes in areas with a sub-optimal growing tempera-
occurrence of phenological events, which was related to better ture, than in areas whose climate is already close to the
vintage quality, and a higher sugar/acid ratio in Merlot and optimum. In fact growth rate of plants increases at higher tem-
Cabernet Sauvignon (Jones and Davis 2000). However, in peratures, but this increase is lower close to the optimal growing
warmer regions like Tuscany, climate change could result in a temperature. This response was represented in the model
higher frequency of extreme events with temperature above FENOVITIS through the use of a temperature-dependent sig-
35°C, leading to heat stress in grapevine (Moriondo and Bindi moidal function for determining daily growth rates, whose
2007). Heat stress during berry ripening is associated with optimum was reached above 25°C. A consequence of this
lower sugar accumulation (Sepulveda and Kliewer 1986), approach is that the phenological response to increasing tem-
lower yield (Kliewer 1977) and reduced berry colouring perature depends on the initial climatic conditions at the site.
(Kliewer and Torres 1972). However, some caution should be used when considering these
Thus, the final effect of temperature increase on vintage results, as the model was calibrated and validated in areas where
quality will depend on the present climate of viticultural regions soil moisture is not a limiting factor. In alpine regions, future
and on the degree of adaptation of varieties to local climates warming could lead to drier conditions and water stress, and
(Webb et al. 2008). In areas where varieties are grown at their possibly impact phenology. Other meteorological agents may
climatic optimum or at higher temperatures, an increase in affect the phenological development of grapevine. For instance,
temperature will inevitably lead to a decrease in vintage ratings Failla et al. (2004) found that potential photosynthetically
(Jones et al. 2005). In Italy, varieties are grown in optimal active radiation is a driver for the zoning of viticultural suitabil-
climatic conditions in their Denominazione d’Origine Control- ity of a red grapevine variety (Nebbiolo) in an alpine region
lata (DOC) areas, which are determined on the basis of experi- similar to the one of this study. In addition, as the sigmoidal
ence as those areas where optimal soil, terrain and climate function used by the FENOVITIS model was derived from an
features are combined to produce the best growing conditions experimental study investigating the effect of mean daily tem-

© 2011 Australian Society of Viticulture and Oenology Inc.

 17550238, 2011, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1755-0238.2010.00118.x by Cochrane Romania, Wiley Online Library on [21/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Caffarra and Eccel Effects of climate change on grape phenology 59

Figure 6. Relationship between the rates of advance (d/yr) of flowering (䊊) and veraison (䊉) and the mean annual temperature (°C) over
the period 2000–2098 (upper row), and elevation (m a.s.l.) (lower row) at each site, for scenarios A2 and B2. The relationships are all
significant at P < 0.05, except for the rate of advance of veraison versus elevation in scenario B2.

Table 3. Regression coefficients between rates of pheno- heat episodes are predicted to become more frequent by climate
logical shift of flowering and veraison (d/yr) versus mean change scenarios, so it might be worthwhile to integrate this
temperature (°C) and elevation (m) at all sites, for sce- effect into the model, through further experiments.
narios A2 and B2. While the projections for veraison and flowering offer a
good picture of possible phenological changes for Chardonnay,
given current responses, projections of the date of budburst
IPCC SRES Phenophase Independent variable
have to be considered with more caution. The sub-model of
Scenario budburst was calibrated and validated using data from stations
Mean temperature Elevation
(d/yr/°C) (d/yr/m) experiencing cold winters and these conditions might not have
provided enough information for a realistic fit of the end of
A2 Flowering 0.040 -0.0002 endodormancy (chilling fulfilment). In addition, the beginning
Veraison 0.036 -0.0001 of chilling unit accumulation is arbitrarily set as 1 September,
which might not reflect year-to-year variations in the beginning
B2 Flowering 0.026 -0.0001
of dormancy. Schleip et al. (2008), for example, postulate the
Veraison 0.027 -0.0001 existence of an effect of summer temperatures of the previous
year on spring phenophases of grapevine. Similar effects were
Figures in bold indicate a significant relationship (P < 0.05). See Figure 6. IPCC,
reported for Betula pubescens by Heide (2003). In addition, pro-
Intergovernmental Panel on Climate Change; SRES, Special Report on Emission
jections based on the application of phenological models should
Scenarios.
be taken with caution, as they do not take into account pheno-
logical plasticity, leading to a phenotypic adaptation. These
peratures up to 25°C (Pouget 1963), the model does not take mechanisms seem able to reduce the sensitivity of phenological
into consideration a possible decrease in growth rate at very response of plants to warming, as already assessed by previous
high temperatures (>30°C). While days with a mean tempera- works carried out in the alpine region (Rutishauser et al. 2007,
ture above 30°C are currently exceptional in Trentino, extreme 2008, Eccel et al. 2009).

© 2011 Australian Society of Viticulture and Oenology Inc.

 17550238, 2011, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1755-0238.2010.00118.x by Cochrane Romania, Wiley Online Library on [21/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
60 Effects of climate change on grape phenology Australian Journal of Grape and Wine Research 17, 52–61, 2011

Figure 7. Bar plots of the mean dates of veraison for the decades 2020s and 2070s for all stations. Grey bars: scenario A2, black bars:
scenario B2. Missing bars indicate that veraison did not occur before the 1st of September in most of the years considered.

Conclusions Mountain viticulture is already a vulnerable sector. The

This study produced a pheno-climatic simulation of budburst, small size of farms and the morphological constraints of terrain
flowering and veraison of grapevine cv. Chardonnay in a viti- limit crop mechanisation, and the steep slopes and rocky terrain
cultural area in the southern Alps, extending over a wide alti- decrease soil fertility. The future climatic changes might further
tudinal range. The simulations for the next century show an increase this vulnerability, adding new challenges to this already
advance for all phenophases, whose magnitude varies according fragile sector, whose intrinsic value is not purely economical but
to the applied climate change scenario, the site and the phe- includes cultural, environmental and landscape conservation
nophase considered. The largest advance is for veraison in sce- aspects. Further projections of changes in the phenology of
nario A2. Our phenological simulations applied to the HadCM3 grape varieties and in the climatic suitability of mountain areas
model projections show that the climatically suitable areas for to viticulture may support the development of adaptation strat-
Chardonnay are likely to expand to more elevated sites cur- egies, like changes in cultural practices or variety renewal,
rently too cold for this variety. This might occur as early as the assuring the profitability of mountain viticulture even in times
next few decades for sites where viticulture is today in its begin- of global change.
nings, and during the second half of the present century for
cooler, higher sites in the mountain belt at around 1000 m.
Current viticultural sites will experience an advance in the dates Acknowledgements
of phenophases, more pronounced for flowering and veraison Project FENOVITIS has been funded by a grant provided by
than for budburst. This advance would shift the dates of verai- CARITRO Foundation, Trento (I). We Thank F. Zottele (IASMA)
son from early–middle August to the early–middle July. Conse- and A. Crisci (CNR-IBIMET, Florence) for data processing and
quently, according to the simulations, the time intervals Mauro Varner (cantina di Mezzocorona) for data supply.
between budburst and flowering and flowering and veraison
will possibly become shorter, and ripening will occur in warmer
References
conditions, with possible effects on vintage quality. A compari-
Barker, T., Bashmakov, I., Bernstein, L., Bogner, J.E., Bosch, P., Dave, R.,
son between the rates of phenological advance at different sites Davidson, O., Fisher, B.S., Gupta, S., Halsnaes, K., Heij, B.J., Kahn
highlights a higher sensitivity of mountain sites to climate Ribeiro, S., Kobayashi, S., Levine, M.D., Martino, D.L., Masera, O., Metz,
change, with more consistent advances of flowering and verai- B., Meyer, L., Nabuurs, G.J., Najam, A., Nakicenovic, N., Rogner, H.H.,
son at higher elevations. Roy, J., Sathaye, J., Schock, R., Shukla, P., Sims, R.E.H., Smith, P.,
Tirpak, D.A., Urge-Vorsatz, D. and Dadi, Z. (2007) Technical summary.
Even with some caveats, the simulations proposed in this In: Climate change 2007: mitigation. Contribution of Working group III to
work, based on current phenological responses, suggest a the Fourth Assessment Report of the Intergovernmental Panel on Climate
picture of possible changes of viticultural areas in a mountain Change. Contribution of Working group III to the Fourth Assessment
region for a widely grown variety of grapevine under two dif- Report of the Intergovernmental Panel on Climate Change. Eds. B. Metz,
ferent climate change scenarios. As these scenarios evolve as O.R. Davidson, P.R. Bosch, R. Dave and L.A. Meyer (Cambridge Univer-
sity Press: Cambridge, United Kingdom and New York, NY).
pathways into the future, electing to follow a higher emissions Battey, N.H. (2000) Aspects of seasonality. Journal of Experimental Botany
pathway now will have consequences for the future, even if in 51, 1769–1780.
the short term the impacts seem the same. Indeed, according Bindi, M., Fibbi, L., Gozzini, B., Orlandini, S. and Miglietta, F. (1996)
to the projections, major changes in the phenology of grape- Modelling the impact of future climate scenarios on yield and yield vari-
ability of grapevine. Climate Research 7, 213–224.
vine are to be expected only from the second half of the 21st Caffarra, A. and Eccel, E. (2010) Increasing the robustness of phenological
century, while in the immediate future, only minor changes models for Vitis vinifera cv. Chardonnay. International Journal of Bio-
will occur. meteorology 54, 255–267.

© 2011 Australian Society of Viticulture and Oenology Inc.

 17550238, 2011, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1755-0238.2010.00118.x by Cochrane Romania, Wiley Online Library on [21/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Caffarra and Eccel Effects of climate change on grape phenology 61

Chuine, I., Yiou, P., Viovy, N., Seguin, B., Daux, V. and Ladurie, E.L. climate projections. In: Climate change 2007: the physical science basis.
(2004) Historical phenology: grape ripening as a past climate indicator. Contribution of Working Group I to the Fourth Assessment Report of the
Nature 432, 289–290. Intergovernmental Panel on Climate Change. Eds. S. Solomon, D. Qin, M.
Defila, C. and Clot, B. (2001) Phytophenological trends in Switzerland. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller
International Journal of Biometeorology 45, 203–207. (Cambridge University Press: Cambridge, United Kingdom and New York,
Duchene, E. and Schneider, C. (2005) Grapevine and climatic changes: a NY).
glance at the situation in Alsace. Agronomy and Sustainable Development Meier, N., Rutishauser, T., Pfister, C., Wanner, H. and Luterbacher, J.
25, 93–99. (2007) Grape harvest dates as a proxy for Swiss April to August tempera-
Easterling, W.E., Aggarwal, P.K., Batima, P., Brander, K.M., Erda, L., ture reconstructions back to AD 1480. Geophysical Research Letters 34,
Howden, S.M., Kirilenko, A., Morton, J., Soussana, J.-F., Schmidhuber, L20705.
J. and Tubiello, F.N. (2007) Food, fibre and forest products. In: Climate Moriondo, M. and Bindi, M. (2007) Impact of climate change of typical
change 2007: impacts, adaptation and vulnerability. Contribution of Mediterranean crops. Italian Journal of Agrometeorology 12, 5–12.
Working group II to the Fourth Assessment Report of the Intergovern- Olesen, J.E. and Bindi, M. (2002) Consequences of climate change for
mental Panel on Climate Change. Eds. M.L. Parry, O.F. Canziani, J.P. European agricultural productivity, land use and policy. European Journal
Palutikof, P.J. van der Linden and C.E. Hanson (Cambridge University of Agronomy 16, 239–262.
Press: Cambridge) pp. 273–313. Oliveira, M. (1998) Calculation of budbreak and flowering base tempera-
Eccel, E. and Saibanti, S. (2007) Inquadramento climatico dell’Altopiano di tures for Vitis vinifera cv. Touriga Francesa in the Douro Region of Portugal.
Lavarone-Vezzena nel contesto generale trentino. Studi Trentini di Scienze American Journal of Enology and Viticulture 49, 74–78.
Naturali. Acta Biologica 82, 111–121. Orlandini, S., Bindi, M. and Gozzini, B. (1993) Effect of CO2-induced
Eccel, E., Rea, R., Caffarra, A. and Crisci, A. (2009) Risk of spring frost to climatic change on viticultural environments of cool climate regions of
apple production under future climate scenarios: the role of phenological Italy. Viticultural and Enological Sciences 48, 81–85.
acclimation. International Journal of Biometeorology 53, 273–286. Pope, V., Gallani, M.L., Rowntree, P.R. and Stratton, R.A. (2000) The
Failla, O., Mariani, L., Brancadoro, L., Minelli, R., Scienza, A., Murada, G. impact of new physical parameterizations in the Hadley Centre climate
and Mancini, S. (2004) Spatial distribution of solar radiation and its model: HadAM3. Climate Dynamics 16, 123–146.
effects on vine phenololgy and grape ripening in an alpine environment. Pouget, R. (1963) Recherches physiologiques sur le repos végétatif de la
American Journal of Enology and Viticulture 55, 128–138. vigne (Vitis vinifera L.): la dormance des bourgeons et le mécanisme de sa
Fuhrer, J. (2003) Agroecosystem responses to combinations of elevated disparition. PhD Thesis, Bordeaux University & Annales Amelioration des
CO2, ozone, and global climate change. Agriculture, Ecosystems and Envi- Plantes 13, Special Issue.
ronment 97, 1–20. Rutishauser, T., Luterbacher, J., Jeanneret, F., Pfister, C. and Wanner, H.
Giannakopoulos, C., Le Sager, P., Bindi, M., Moriondo, M., Kostopoulou, E. (2007) A phenology-based reconstruction of interannual changes in past
and Goodess, C. (2009) Climatic changes and associated impacts in the spring seasons. Journal of Geophysical Research – Biogeoscience 112,
Mediterranean resulting from a 2°C global warming. Global and Planetary G04016.
Change 68, 209–224. Rutishauser, T., Luterbacher, J., Defila, C., Frank, D. and Wanner, H.
Gladstones, J. (1992) Viticulture and environment (WineTitles: Adelaide) (2008) Swiss spring plant phenology 2007: extremes, a multi-century
p. 130. perspective, and changes in temperature sensitivity. Geophysical Research
Grünberg, K.U., Beltrano, C., Perini, L., Rossi, F., dall’Olio, N., Tripaldi, V., Letters 35, L05703.
Toller, G., Eccel, E., Reitschläger, J.D., Petrakis, M., Verbič, J., Mateescu, Schleip, C., Menzel, A., Estrella, N. and Dose, V. (2006) The use of
E., Turcu, V. and Cortez-Sack, S. (2007) Agriculture and climate change Bayesian analysis to detect recent changes in phenological events
in Europe. Regional facts and challenges. Project ACCRETE, INTERREG III throughout the year. Agricultural and Forest Meteorology 141, 179–191.
B CADSES (Univ. of Rostock: 82). Schleip, C., Rutishauser, T., Luterbacher, J. and Menzel, A. (2008) Time
Harrison, P.A. and Butterfield, R.E. (2000) Modelling climate change series modelling and central European temperature impact assessment
impacts on wheat, potato and grapevine in Europe. In: Climate change, of phenological records over the last 250 years. Journal of Geophysical
climate variability and agriculture in Europe: an integrated assessment. Research 113, G04026.
Research Report 21. Eds. T.E. Downing, P.A. Harrison, R.E. Butterfield Sepulveda, G. and Kliewer, W.M. (1986) Effect of high temperature on
and K.G. Lonsdale (Environmental Change Unit, University of Oxford: grapevines (Vitis vinifera L.). II. Distribution of soluble sugars. American
Oxford) pp. 367–390. Journal of Enology and Viticulture 37, 20–25.
Heide, O.M. (2003) High autumn temperature delays spring bud burst Thomson, L.J., Macfadyen, S. and Hoffmann, A.A. (2009) Predicting the
in boreal trees, counterbalancing the effect of climatic warming. Tree effects of climate change on natural enemies of agricultural pests. Biologi-
Physiology 23, 931–936. cal Control 52, 296–306.
Huglin, P. and Schneider, C. (1998) Biologie et écologie de la vigne 2e Thuiller, W., Lavorel, S., Araujo, M.B., Sykes, M.T. and Prentice, C. (2005)
Édition (Lavoisier TEC & DOC: Paris). Climate change threats to plant diversity in Europe. Proceedings of the
IPCC (2000). Special report on emission scenarios. Eds. N. Nakicenovic and National Academy of Sciences of the United States of America 102, 8245–
R. Swart (Cambridge University Press: Cambridge, UK) p. 570. 8250.
Jones, G.V. and Davis, R.E. (2000) Climate influences on grapevine phe- Webb, L.B., Whetton, P.H. and Barlow, E.W.R. (2007) Modelled impact of
nology, grape composition, and wine production and quality for Bor- future climate change on the phenology of winegrapes in Australia. Aus-
deaux, France. American Journal of Enology and Viticulture 51, 249–261. tralian Journal of Grape and Wine Research 13, 165–175.
Jones, G.V., White, M.A., Cooper, O.R. and Storchmann, K. (2005) Webb, L., Whetton, P. and Barlow, E.W.R. (2008) Climate change and wine
Climate change and global wine quality. Climatic Change 73, 319–343. grape quality in Australia. Climate Research 36, 99–111.
Kliewer, W.M. (1977) Influence of temperature, solar radiation, and nitro- White, M.A., Diffenbaugh, N.S., Jones, G.V., Pal, J.S. and Giorgi, F. (2006)
gen on coloration and composition of Emperor grapes. American Journal Extreme heat reduces and shifts United States premium wine production
of Enology and Viticulture 28, 96–103. in the 21st century. Proceedings of the National Academy of Sciences of
Kliewer, W.M. and Torres, R.E. (1972) Effect of controlled day and night the United States of America 103, 11217–11222.
temperature on grape coloration. American Journal of Enology and Viti- Williams, D.W., Andris, H.L., Beede, R.H., Luvisi, D.A., Norton, M.V.K. and
culture 23, 71–76. Williams, L.E. (1985) Validation of a model for the growth and develop-
Maracchi, G., Sirotenko, O. and Bindi, M. (2005) Impacts of present and ment of the Thompsom Seedless grapevine. II. Phenology. American
future climate variability on agriculture and forestry in the temperature Journal of Enology and Viticulture 36, 283–289.
regions: Europe. Climatic Change 70, 117–135. Wolfe, D.W., Schwartz, M.D., Lakso, A.N., Otsuki, Y., Pool, R.M. and
Mariani, L., Failla, O., Del Monte, G. and Facchinetti, D. (2006) Iphen: a Shaulis, N.J. (2005) Climate change and shifts in spring phenology of
model for real time production of grapevine phenological maps. In: Con- three horticultural woody perennials in northeastern USA. International
gress on climate and viticulture. Ed. M.A. Mainar (Consejero de Agricul- Journal of Biometeorology 49, 303–309.
tura y Alimentation de Gobierno del Aragona: Zaragoza) pp. 272–278.
Mariani, L., Parisi, S., Failla, O., Cola, G., Zoia, G. and Bonardi, L. (2009)
Tirano (1624–1930): a long time series of harvest dates for grapevine.
Italian Journal of Agrometeorology 14, 7–16. Manuscript received: 10 February 2010
Meehl, G.A., Stocker, T.F., Collins, W.D., Friedlingstein, P., Gaye, A.T.,
Gregory, J.M., Kitoh, A., Knutti, R., Murphy, J.M., Noda, A., Raper,
Revised manuscript received: 16 August 2010
S.C.B., Watterson, I.G., Weaver, A.J. and Zhao, Z.-C. (2007) Global Accepted: 6 September 2010

© 2011 Australian Society of Viticulture and Oenology Inc.