Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Insensitivity of Tree-Ring Growth to Temperature and Precipitation Sharpens the Puzzle of Enhanced Pre-Eruption NDVI on Mt. Etna (Italy)

  • Ruedi Seiler ,

    Contributed equally to this work with: Ruedi Seiler, James W. Kirchner, Paolo Cherubini

    ruedi.seiler@wsl.ch

    Affiliations Swiss Federal Institute for Forest, Snow and Landscape Research WSL, Birmensdorf, Switzerland, Department of Geography, University of Zurich, Zürich, Switzerland

  • James W. Kirchner ,

    Contributed equally to this work with: Ruedi Seiler, James W. Kirchner, Paolo Cherubini

    Affiliations Swiss Federal Institute for Forest, Snow and Landscape Research WSL, Birmensdorf, Switzerland, Department of Environmental Systems Science, ETH Zurich, Zürich, Switzerland

  • Paul J. Krusic ,

    ‡ These authors also contributed equally to this work.

    Affiliations Navarino Environmental Observatory, Messinia, Greece, Institutionen för Naturgeografi, Stockholm University, Sweden

  • Roberto Tognetti ,

    ‡ These authors also contributed equally to this work.

    Affiliation Dipartimento di Bioscienze e territorio, Università del Molise, Contrada Fonte Lappone, Pesche, Italy

  • Nicolas Houlié ,

    ‡ These authors also contributed equally to this work.

    Affiliation Department of Earth Sciences, ETH Zurich, Zürich, Switzerland

  • Daniele Andronico ,

    ‡ These authors also contributed equally to this work.

    Affiliation Osservatorio Etneo, INGV, Sezione di Catania, Italy

  • Sebastiano Cullotta †,

    † Deceased.

    ‡ These authors also contributed equally to this work.

    Affiliation Università di Palermo, Palermo, Italy

  • Markus Egli ,

    ‡ These authors also contributed equally to this work.

    Affiliation Department of Geography, University of Zurich, Zürich, Switzerland

  • Rosanne D'Arrigo ,

    ‡ These authors also contributed equally to this work.

    Affiliation Lamont-Doherty Earth Observatory, Palisades, New York, United States of America

  • Paolo Cherubini

    Contributed equally to this work with: Ruedi Seiler, James W. Kirchner, Paolo Cherubini

    Affiliations Swiss Federal Institute for Forest, Snow and Landscape Research WSL, Birmensdorf, Switzerland, Department of Geography, University of Zurich, Zürich, Switzerland, Lamont-Doherty Earth Observatory, Palisades, New York, United States of America

Insensitivity of Tree-Ring Growth to Temperature and Precipitation Sharpens the Puzzle of Enhanced Pre-Eruption NDVI on Mt. Etna (Italy)

  • Ruedi Seiler, 
  • James W. Kirchner, 
  • Paul J. Krusic, 
  • Roberto Tognetti, 
  • Nicolas Houlié, 
  • Daniele Andronico, 
  • Sebastiano Cullotta, 
  • Markus Egli, 
  • Rosanne D'Arrigo, 
  • Paolo Cherubini
PLOS
x

Abstract

On Mt. Etna (Italy), an enhanced Normalized Difference in Vegetation Index (NDVI) signature was detected in the summers of 2001 and 2002 along a distinct line where, in November 2002, a flank eruption subsequently occurred. These observations suggest that pre-eruptive volcanic activity may have enhanced photosynthesis along the future eruptive fissure. If a direct relation between NDVI and future volcanic eruptions could be established, it would provide a straightforward and low-cost method for early detection of upcoming eruptions. However, it is unclear if, or to what extent, the observed enhancement of NDVI can be attributed to volcanic activity prior to the subsequent eruption. We consequently aimed at determining whether an increase in ambient temperature or additional water availability owing to the rise of magma and degassing of water vapour prior to the eruption could have increased photosynthesis of Mt. Etna's trees. Using dendro-climatic analyses we quantified the sensitivity of tree ring widths to temperature and precipitation at high elevation stands on Mt. Etna. Our findings suggest that tree growth at high elevation on Mt. Etna is weakly influenced by climate, and that neither an increase in water availability nor an increase in temperature induced by pre-eruptive activity is a plausible mechanism for enhanced photosynthesis before the 2002/2003 flank eruption. Our findings thus imply that other, yet unknown, factors must be sought as causes of the pre-eruption enhancement of NDVI on Mt. Etna.

Introduction

Early detection of precursors to volcanic eruptions is important in preventing major damage and loss of life. To date, these precursors have mainly included seismic, geochemical, petrographic, ground deformation and gravimetric changes that are used to assess volcanic activity shortly before eruptions (e.g., [16]). Surface deformation, small earthquakes, and release of volcanic gases are typically triggered by the ascent of magma in volcanoes [7]. Volcanic monitoring by remote sensing includes the acquisition of different geochemical and geophysical parameters which record volcanic processes, such as gas emissions and hydrological variations, over time (e.g., [810]). Using remote sensing data, an increased Normalized Difference Vegetation Index (NDVI) was observed along subsequent eruptive fissures on Mt. Nyiragongo (Congo) and Mt. Etna (Italy), as early as two years prior to eruptions of both volcanoes [11]. NDVI is closely associated with the amount of photosynthetically active radiation intercepted by vegetation, and thus with both the spatial coverage of green biomass and the chlorophyll content in leaves [12,13]. On Mt. Etna, the enhanced NDVI signal [11] was detected along a narrow line on the northeastern flank; this line later developed into an eruptive fissure during the 2002/2003 flank eruption, suggesting that enhanced photosynthesis may be related to a coming volcanic eruption.

The observed NDVI signal raises the question of whether, and how, eruptive precursor activity could influence photosynthetic rates. A comparison of tree growth with environmental parameters is necessary to estimate their influence on tree growth and to assess the potential contribution from volcanic activity. Increased photosynthesis may be induced by a number of environmental factors that are affected by pre-eruptive volcanic processes, but the most probable are an increase in heat or water availability associated with volcanic degassing (e.g., [1416]).

Trees in temperate climates form annual growth rings, and variations in their tree-ring characteristics (width and density) reflect changes in the environmental conditions in which they grow. At high elevations and high latitudes, where the limiting factor is summer air temperature, tree growth is typically enhanced during warm summers (e.g., [17,18]). Conversely, in semi-arid ecosystems, such as in Mediterranean lowlands, growth is primarily regulated by precipitation and is enhanced during wet years [19,20]. Consequently, tree rings, often used as indicators of photosynthetic rates [21], also serve as useful proxies for climate [22,23].

The Mediterranean region is characterized by hot and dry summers and mild, humid winters [24,25]. Maximum rainfall occurs predominantly in autumn and sometimes during winter [26]. Precipitation minima and temperature maxima coincide with the period of most intense solar radiation, limiting water availability during the summer season (e.g., [27]). High rainfall variability over the year greatly affects drought severity and hampers growth [28]. Rainfall during spring is the most important factor influencing tree growth and vegetation activity of Mediterranen forests, particularly at more xeric, low-elevation sites, as also shown by remote-sensing-based model simulations and tree-ring-based growth analyses [29]. In more temperate high-elevation conditions, drought often has a minor impact on tree growth because precipitation is less limiting. In the high-elevation forests on Mt. Etna where the increased NDVI prior to the eruption was detected [11], tree growth might be enhanced by increased air temperature during the vegetation period or, given the southern latitude, by increased water availability.

Here we analyse the relationships between ring-width indices of Pinus nigra J.F. Arnold and monthly precipitation and air temperature, and compare our ring-width series from Mt. Etna with series from trees growing at similar elevations in Calabria, a region located at a similar latitude on the Italian peninsula without the direct influence of volcanic activity. Our hypotheses are that i) ascending magma led to an increase in local ambient air and soil temperature which positively influenced photosynthesis rates and tree growth, and that ii) water vapour from volcanic degassing locally provided additional humidity/moisture/water which became available to trees influencing photosynthesis rates and tree growth. To address these issues we assess to what extent tree-ring growth at the highest elevations on Mt. Etna is influenced by climate, i.e. air temperature and precipitation, to indirectly determine i) whether an increase in temperature caused by an incipient volcanic eruption (e.g., [30,14]) would likely induce higher photosynthetic productivity, and ii) whether, at specific locations close to rift zones, additional water availability induced by degassing of water vapour associated with the rise of magma prior to an eruption (water is the most abundant component in volcanic gas, [16]) would likely increase the photosynthetic capacity of Mt. Etna's trees (see [11]).

Materials and Methods

Study area

Mt. Etna is a stratovolcano situated in the northeastern part of Sicily. With an area of approximately 1600 km2 and a summit elevation of roughly 3330 m a.s.l., Mt. Etna is an isolated high mountain exposed to air masses from the Mediterranean Sea. The climate on Mt. Etna is strongly maritime on the eastern flank [31,32], with drier conditions on its western flank [33]. The slopes are characterized by lava flows of different ages [34]. Most of the lower elevations, being especially fertile, have been settled and used for agriculture for thousands of years. The higher elevations, from 1000–1600 m a.s.l., are dominated by European beech (Fagus sylvatica L.), and from 1600 m to treeline (~2000 m a.s.l.) by Corsican black pine (P. nigra). Though the treeline climatically determined at such latitudes would otherwise be higher [35,36], eruption-induced wildfires and the lack of soils on lava flows hinder its uphill development [3739].

Besides volcanic eruptions and lava flows on Mt. Etna, other volcanic processes, such as degassing through small vents, are also present but difficult to quantify [40]. The soils of Mt. Etna are primarily classified as Regosols, Eutric or Dystric Cambisols and (Mollic) Andosols. The characteristics of these soils predominantly depend on the surface age of the lava flow and volcanic deposits from which they have developed [38,41,42]. In general, soils at intermediate to high elevation on Mt. Etna (i.e. above 800 m a.s.l.) are mostly described as Andisols with a sandy loam texture, vitric characteristics, an udic moisture regime [39] and good water holding capacity [43]. However, less mature, young soils on fresh lava flows may be less developed resulting in lower water holding capacity.

The forests around the flanks of Mt. Etna are greatly affected by both natural and anthropogenic disturbances, such as wildfires, lava flows, avalanches and logging. At the lowest elevations, from the plains up to 900 m a.s.l., agricultural crops and orchards, e.g., orange, lemon, almond, pistachio, and chestnut, are found. Only a few forest stands, mainly at the highest elevations on the northern or northeastern side of the mountain, are undisturbed [31]. Meteorological station data from Linguaglossa (530 m a.s.l., 15°08'42" E, 37°50'27" N, timespan: 1893–2004) based on daily temperatures and precipitation measurements give an average annual temperature of 18°C and a total annual precipitation of 1400 mm. Additionally, monthly temperature averages in winter are above zero at at all stations.

Sampling

In total, we sampled 143 trees (P. nigra), with permission issued by the local forest authorities (Corpo Forestale della Regione Siciliana, Distaccamento di Bronte, Piazza Cadorna 11, I-95034 Bronte, Catania, Italy), at four high-elevation (1500 to 1900 m a.s.l.) forest sites on the northeastern and western slopes of Mt. Etna (Fig 1): 52 trees growing close to the 2002/2003 eruptive fissure (Group 1), 27 trees growing close to the 1928 eruptive fissure (Group 2), 38 trees growing in the same elevation band but far from any obvious fissures (Group 3), and 26 trees growing close to the 1974 eruptive fissure (Group 4). All sites are dominated by P.nigra and Fagus silvatica and are located at a comparable elevation and slope with NNE aspect except for group 4, which is located on the western flank on a western-aspect slope. Apart from that, there were no evident differences between the four sites in terms of forest-stand density, composition of tree species, topography and slope (about 14%). From each tree, two 0.5 cm diameter cores were taken orthogonally with respect to each other using a corer with a three-threaded auger by Haglof (Haglof Inc., Sweden), wrapped in paper and transported to the laboratory. All samples were mounted on wooden supports and cut using a microtome at an angle of roughly 30° to the radial axis of the tree to prevent core breakage [44]. For later comparisons, three ring-width chronologies, located close to Mt. Etna, derived from coniferous trees uninfluenced by Mt. Etna growing at similarly high elevations in Calabria (Gambarie, Monte Pollino and Sierra da Crispo) were retrieved from the International Tree-Ring Databank (ITRDB, NOAA, U.S.A.)

thumbnail
Fig 1. Map of sample locations.

Mt. Etna sample sites (Group 1–4) on the northeastern and western slopes at an elevation range from 1600 to 1850 m a.s.l. indicating the location where samples were taken and the location of the meteorological stations. The NDVI anomaly overlays with the 2002 fissure line. A more detailed map can be found in Houlié et al. (2006). The map of Italy was created using the program R (Version 3.1.3; URL: http://www.R-project.org/) [45], the topographic map showing Sicily was created using Generic Mapping Tools (Version 5.2.1; URL: http://gmt.soest.hawaii.edu/) [46] and the basis map was taken from Egli et al. (2007) [47].

https://doi.org/10.1371/journal.pone.0169297.g001

Ring-width measurements

All ring widths were measured to the nearest 0.01 mm using a Leica Wild M32 binocular microscope (Leica, Germany) with 25-50x magnification, coupled to a LINTAB measuring table and computer with TSAPwin (Time Series Analysis Program) software (RinnTech, Heidelberg, Germany). Core measurements were visually crossdated against each other and any inconsistencies, if found, were eliminated. Subsequent crossdating of the single-tree chronologies with their respective mean site chronology by visual and statistical measures was performed using TSAPwin and COFECHA (50-year segments with a 25-year overlap) [48,49]. Since the sampling dates of all trees were known, crossdating was primarily used to ensure that prominent tree-ring patterns were not shifted between trees and no rings were missing.

Meteorological data

We used monthly precipitation and air temperature data recorded at three meteorological stations on Mt. Etna: Floresta (1275 m a.s.l., 14°54'31" E, 37°59'15" N, Timespan: 1924–2004), Linguaglossa (530 m a.s.l., 15°08'42" E, 37°50'27" N, Timespan: 1893–2004) and Taormina (248 m a.s.l., 15°17'34" E, 37°17'34" N, Timespan: 1906–2004). Although longer records are available for Linguaglossa and Taormina, we only used data from 1924–2004 at all three sites so that they could be compared over a common period. In addition, interpolated monthly temperature, precipitation, cloud cover and Palmer Drought Severity Index (PDSI) data for the Mt. Etna region and Calabria from the Climatic Research Unit, University of East Anglia, Norwich, U.K. (CRU) [50] were compared to the above-mentioned station data and the tree-ring data [51,52]. Opposed to temperature and precipitation data, the PDSI incorporates both temperature and precipitation, representing long-term drought taking prior months' condition into account. On the other hand, delayed water runoff from snow during spring is not accounted for in the index (e.g., [53]). Correlations between the data recorded at the meteorological stations and the interpolated datasets were calculated, to assess whether the interpolated data could be used to further analyse relationships between climate and tree growth.

Data analysis

All raw measurement series were standardized using the program ARSTAN (http://www.ldeo.columbia.edu/tree-ring-laboratory) by applying 30-year spline detrending combined with a variance stabilization, to remove the age trend and produce detrended ring-width indices [48,54]. All chronologies were used individually to analyse the relationships between climate and growth at different sampling sites.

We performed correlation analysis and response function modelling to quantify the influence of climate on tree growth [55]. We tested the statistical significance of temperature, precipitation, cloud cover and Palmer Drought Severity Index (PDSI; e.g., [56,57]) in different months and seasonal combinations of monthly values, including prior-year values, using Spearman rank correlation. Linear regression, as described by [58], was used to remove long-term trends in the meteorological data and avoid artificially inflating correlation values. Based on results from simple Spearman rank correlations between all the Mt. Etna chronologies and monthly meteorological data, we built "Visual Regression Models" (VRM) which were defined as standard multiple linear regression models including statistically significant (p < 0.05) monthly variables or monthly groupings.

In addition to VRM, Stepwise Linear Regression Modelling (SLRM), based on the Akaike Information Criterion (AIC) and using a forward-backward approach [59], was used to identify those climate variables that explained the greatest variance in each chronology on Mt. Etna. The SLRMs were based on monthly variables and groupings that defined the spring and summer seasons. For all models we only used variables that did not overlap in time (e.g., precipitation in May and precipitation in spring would not be used together because both contain May precipitation values).

The models (SLRM) were tested for collinearity among explanatory variables using variance inflation factor (VIF) analysis [60]. VIF values were all lower than 4, implying a lack of strong collinearity among explanatory variables [61]. This led to selecting the final climate models based on the AIC.

We compared how well the two model types (VRM and SLRM) explained the climate—ring width relationship:

  1. Qualitative differences between the two model types included differences in monthly parameters used in the models.
  2. Quantitative differences, comparing adjusted-R2, show the increase/decrease of explanatory power from one model type to another.

To test the reliability of the models, as well as the degree of overfitting due to including too many variables during the stepwise model selection process, we divided the timespan covered by both meteorological and ring-width measurements at Mt. Etna (1924–2004) and Calabria (1924–1980) into two segments [62,63]. The models were run on both segments to compare differences over time. In addition, the two segments were used for both forward- and backward validation to quantify model robustness over time [53].

On Mt. Etna, forward validation used a model based on the time segment 1925–1964 to predict tree growth during 1965–2004, and backward validation used a model based on the latter time segment to reconstruct 1925–1964 tree growth. For the Calabria chronologies we used the same validation procedure based on the two time segments from 1925–1951 and 1952–1980. AIC was used to test for overfitting.

We calculated how the explanatory power (R2) varied through time by applying our SLRMs to a 15-year moving window with 14 years overlap to identify periods of higher and lower correlations between ring-width and our climate models [52].

Results

The Mt. Etna chronologies show higher variability in their ring-width patterns than the chronologies from Calabria (Fig 2). The four raw ring-width chronologies (Group1-4) from Mt. Etna display growth patterns that are strongly influenced by the establishment of new generations of trees. Young trees, most germinating after 1950, greatly increase the mean ring-width, producing a clear age trend (Fig 2). Tree age ranges from 55 to 122 years on Mt. Etna, and from 128 to 299 years in Calabria. Descriptive chronology statistics are given in Table 1.

thumbnail
Fig 2. Chronologies and sample replication.

Average chronologies (black lines) and sample replication of all samples (grey area) of Mt. Etna (Group 1–4) and Calabria (Gambarie, Monte Pollino and Sierra da Crispo.

https://doi.org/10.1371/journal.pone.0169297.g002

thumbnail
Table 1. Sample overview information.

Descriptive statistics of sample chronologies (Group 1–4) from Mt. Etna, and the chronologies from Calabria (Gambarie, Monte Pollino and Sierra da Crispo) displaying number of series (core-series), total length (years) of group chronologies, series intercorrelation (measure of common growth signal in the chronology), mean sample length (years), elevation (m a.s.l.) and species.

https://doi.org/10.1371/journal.pone.0169297.t001

The monthly meteorological data show stronger inter-station correlations for air temperature (Spearman correlation coefficients up to r = 0.86; p < 0.01) than for precipitation (up to r = 0.83; p < 0.01), as usually reported in the literature (e.g., [51,64]). In addition, we find that instrumental station data also correlates significantly with the interpolated CRU data (highest Spearman r = 0.67 to 0.87 for temperature; r = 0.67 to 0.78 for precipitation; p < 0.01). Given these correlations, comparable to those found in previous studies [51], we used the interpolated data to assess the climate influence on tree growth.

In general, total summer precipitation produces similarly high correlation values with tree growth as single monthly variables in the same season. Average spring and summer temperature and total spring precipitation produce generally lower correlations than single months during the same season (Table 2). Prior-year precipitation and temperature were also considered but results are not reported because the current year's correlations are higher. Cloud coverage is not significantly correlated with tree growth. In contrast to all other sites, tree growth at Group 3 exhibits a significant negative correlation with PDSI from April to December (results not shown).

thumbnail
Table 2. Climate-ring width correlation statistics.

Spearman rank correlations between climate variables and detrended ring width from Mt. Etna chronologies (Group 1–4) and from Calabria chronologies (Gambarie, Monte Pollino and Sierra da Crispo), where P = precipitation, T = temperature, tot. = total amount of precipitation, avg. = average temperature, and prior = prior year. Values printed in bold are statistically significant with (* = p < 0.05, ** = p < 0.01, two-tailed). The significance threshold at Mt. Etna is lower than in Calabria (r = 0.222 vs. r = 0.271, respectively) because the climate and tree ring records overlap for longer on Mt. Etna than in Calabria (81 vs. 57 years, respectively).

https://doi.org/10.1371/journal.pone.0169297.t002

In Calabria the Gambarie chronology is primarily sensitive to summer temperature and summer precipitation (Table 2). The Monte Pollino chronology responds more to spring temperatures as well as spring and summer precipitation, and the chronology from Sierra da Crispo is not significantly correlated with any climate variables. The Mt. Etna chronologies correlate with spring and summer temperatures (Groups 1–3), as well as spring and summer precipitation (Groups 1 and 3). Group 4 has no significant correlations with climate. Over all months and seasons the Mt. Etna (Groups 1–4) chronologies are, on average, less strongly correlated with temperature and precipitation than the Calabria chronologies are. The difference in the (absolute value) strength of the correlations between Mt. Etna and Calabria was not, however, statistically significant (Student t-test).

Overall we observe similar responses in the chronologies from Mt. Etna and in those from Calabria (Table 2 and Fig 3). The raw numbers of significant correlations between temperature or precipitation and ring width are broadly similar (20 out of 88, or 23% at Mt. Etna, and 14 out of 66, or 21%, in Calabria).

thumbnail
Fig 3. Climate—ring-width correlations.

Spearman rank correlation data points of single months and seasonal groupings (red dots) of chronologies from Mt. Etna (top panel) and Calabria (bottom panel) showing the correlation range of each monthly- or seasonal variable with the different group-chronologies; where T = temperature, P = precipitation and p = prior year. Boxplots show the median and lower and upper quartiles, and the whiskers display the minimum and maximum values.

https://doi.org/10.1371/journal.pone.0169297.g003

We used the sign test to evaluate whether the correlations between climate and ring width were positive or negative more often than expected by chance. If we pool both the Mt. Etna and Calabria correlations (Table 2), we see that ring widths are positively correlated with spring temperature (22 of 28 month and site combinations; p < 0.01 by the two-tailed sign test), negatively correlated with summer temperature (22 of 28 correlations; p < 0.01 by the two-tailed sign test), and positively correlated with summer precipitation (18 of 21 correlations; p < 0.05 by the two-tailed sign test), but not significantly correlated with spring precipitation (17 negative correlations out of 28 month and site combinations; p > 0.05 by the two-tailed sign test). Thus tree growth in these mountain environments tends to be favoured by warm springs (suggesting that water is not limiting in the springtime) and cool and relatively wet summers.

Comparing the VRM and SLRM modelling results (Table 3), we distinguished between models explaining ring width on Mt. Etna and models explaining ring width in Calabria. SLRM's yielded average R2 values of 20% (8% to 33%) on Mt. Etna and 26% (13% to 39%) in Calabria, demonstrating that precipitation and temperature are not strongly correlated with tree growth compared to other climatic regions. The regression models include both precipitation and temperature variables from spring and summer. There are, however slight differences between the models (Table 4).

thumbnail
Table 3. Statistics of climate models.

Overview of model R2 and adjusted R2 statistics of the Mt. Etna and Calabria chronology models. Visual Regression Models (VRM) are shown in the left panel, Stepwise Linear Regression Models (SLRM) in the middle panel and the percentage of "model-improvement" from VRM to SLRM is shown in the right panel.

https://doi.org/10.1371/journal.pone.0169297.t003

thumbnail
Table 4. Model variables used by climate models.

Climate variables (single months and seasonal groupings) used in VRM and SLRM models. Monthly variables included in the models are designated as X. Due to time overlaps between single months and seasons, variables that were excluded from the models are designated as e.

https://doi.org/10.1371/journal.pone.0169297.t004

We calculated an average improvement in adjusted R2 from VRM to SLRM of 5% on Mt. Etna and 8% in Calabria (Table 3; Fig 4).

thumbnail
Fig 4. Climate model comparison.

Comparisons between individual VRMs (visual regression models) and SLRMs (stepwise linear regression models) revealed only one case (Group 1) where the VRM and the SLRM were both significant (p < 0.05) and where a statistically significant model improvement (p < 0.05) was calculated. On Mt. Etna, the Group 1 model significantly improved from adjusted R2 = 0.19 (VRM) to adjusted R2 = 0.24 (SLRM). Details are summarized in Table 3.

https://doi.org/10.1371/journal.pone.0169297.g004

When comparing VRM and SLRM which were statistically significant (p < 0.05), a significant model-improvement (p < 0.05) was only obtained with Mt. Etna's Group 1 models. These results show that on Mt. Etna even complex models such as our SLRM are not able to explain tree-growth variability much better, demonstrating that tree growth is further influenced by parameters other than climate which induce additional noise to our ring-width data.

When comparing differences over time by running the SLRMs on both time segments separately, on Mt. Etna only two out of eight model-runs on the two time segments led to significant R2 values (p < 0.05), whereas in Calabria the R2 in four out of six runs was significant (results not shown).

Mt. Etna model validations demonstrate that the only forward verification resulting in a significant R2 value (p < 0.01) was obtained with the Group 3 model. Forward validation of the Group 1, Group 2 and Group 4 models resulted in statistically non-significant R2 values. Further, backward validations show that all models calibrated on the second segment show a decrease to non-significant R2 values when run on the first segment. Out of eight validation runs (forward and backward) on Mt. Etna only one retained significant R2 values (p < 0.01). Including all validation runs (statistically significant and non-significant), forward validation lost 4% (from an average R2 of 0.18 in the first period to 0.14 in the second period) and backward validation lost 37% (from an average R2 of 0.48 in the second period to 0.11 in the first period) of explained ring-width variance on Mt. Etna.

The same calculations for Calabria showed that forward validation (earlier to later time-segment) lost 32% of the explained variance (average R2 changed from 0.53 to 0.21), while backward validation lost 11% (average R2 changed from 0.55 to 0.44). We calculated that of all six validation runs (forward and backward) in Calabria, only two backward validations retained significant R2 values (p < 0.05).

These results show that our models (SLRMs) of tree-ring data on Mt. Etna and in Calabria do not withstand the cross-validation test and demonstrate the importance of validating tree-ring based climate models, especially in regions such as the Mediterranean, where climatic factors are not strongly limiting tree growth.

The explanatory power of all models is generally low. The changes in the explanatory power of all models over time are shown in Fig 5. The Mt. Etna models (Groups 1–4) display a wider range of performance with some visual suggestion of a trend of increasing explanatory power over time, while the Calabria models showed less temporal variability.

thumbnail
Fig 5. Model strength over time.

Change of SLRM model goodness-of-fit statistics for 15-year moving windows over time, showing a visual suggestion of higher temporal consistency among the Calabria models.

https://doi.org/10.1371/journal.pone.0169297.g005

Discussion

In the Mediterranean region, seasonal drought conditions can persist at low elevations for up to five months and have a strong negative impact on tree growth [19,65], whereas at higher elevations precipitation is usually not a limiting factor [e.g., 29]. At the treeline in temperate regions, air temperature is generally the major driver of tree growth, being positively correlated with ring width (e.g., [6670]). Similar correlations have been observed in the Mediterranean mountains as well [71,72]. At lower elevations, higher temperatures reduce tree growth by increasing evaporative demand and drought stress [73,74]. Our study showed that tree growth at high elevation on Mt. Etna is not much limited by climatic conditions: the correlations between tree-ring width and meteorological data (monthly precipitation, air temperature, PDSI and cloud coverage) were rather weak, suggesting that an increase in local moisture or temperature caused by pre-eruptive volcanic activity was unlikely to have affected tree growth. On Mt. Etna, the correlations between ring width and meteorological data were weaker than in Calabria, suggesting that the tree-ring/climate relationship on Mt. Etna might be affected by other factors. Stepwise linear regression models (SLRM) explained an average adjusted R2 of 16% of tree growth on Mt. Etna and 22% in Calabria. By comparison, linear regression models using spring to summer temperature at similar elevations in south-western Anatolia explained up to 51% of ring-width variance [75]. Furthermore, climate correlation values on Mt. Etna and in Calabria are lower than those found using response function analyses in the Aegean [76]. Tree-growth response to temperature across the Mediterranean region is complex and strongly affected by the change of environmental factors over longitudinal, lattitudinal and elevational gradients [77] and the age of the trees studied [78]. The dependence of tree growth on spring temperature may be reduced at sites that, like Mt. Etna and Calabria, are very close to the sea.

In general, the ring-width chronologies on Mt. Etna showed higher inter-annual variability than those from Calabria, even though the Mt. Etna chronologies were based on greater numbers of trees with a correspondingly greater averaging of random inter-tree variations. Our analyses suggest that non-climatic factors may give rise to a greater variability in tree-ring growth on Mt. Etna than in Calabria. This argument is supported by the larger differences on Mt. Etna between the regression models fitted to the two time periods.

On Mt. Etna, the correlation with summer precipitation (r = 0.27 to 0.35) is lower than in previous studies that found that water is the limiting factor in the Mediterranean region [63,75,76,7981]. Near treeline on Mt. Etna, which is lower than at other sites at similar latitudes, because not determined by climatic conditions but rather by volcanic activities, other factors, such as higher air humidity with increasing elevation, seem to reduce the influence of summer precipitation on tree growth.

The low correlation values between PDSI and ring width on Mt. Etna (maximum r = 0.29) confirm that moisture availability is not strongly limiting tree growth. Similar results have been reported for Pinus halepensis Mill. in Tunisia [82] and for Pinus sylvestris L. in south-eastern France [83]. It is notable that P. nigra exhibits a drought-avoidance strategy characterized by efficient stomatal control of transpirational water loss [84]. At the same latitudes in Spain, the drought response of P. nigra varied along an aridity gradient with the strongest response at the most xeric sites [85]. At comparable latitudes and elevations in Calabria, studies on F. sylvatica found no correlations between either water use efficiency or basal area increment and an estimated drought index, suggesting a minimal effect of climate on tree growth during the last century [86], especially when considering the rather udic soil moisture regime at intermediate to high elevations on Mt. Etna [39]. No significant differences in soil properties, such as nitrate and ammonium or phosphorus from soils adjacent to eruptive fissures and soils away from such fissures [87] revealed homogenity of soil characteristics on Mt.Etna and demonstrated that nitrogen (nitrate and ammonium) is unlikely to have induced stronger tree growth.

Due to their high elevation and the subsequent cold, snowy, foggy and humid environmental conditions, the Mt. Etna trees do not appear to be strongly affected by Mediterranean summer drought. The combination of proximity to the sea and high elevation favours persistent seasonal fog and shading by clouds, thus limiting the vapour pressure deficit and reducing water use efficiency [88]. To survive summer droughts, vegetation uses water coming from spring precipitation or, at higher elevation, melting snow, which in some regions can be half of the annual precipitation amount [89]. At the highest elevations on Mt. Etna, winter precipitation is mainly snowfall and, based on our analyses, is not strongly correlated with tree growth (Table 2). The high porosity of the volcanic soils allows rapid infiltration, making a large fraction of the annual precipitation unavailable to the trees, and making it unlikely that winter or spring precipitation could be stored long enough to significantly alleviate summer drought stress.

Significant positive correlations between ring width and spring temperatures were found, except for Group 4. These results confirm those of previous studies on silver fir in southern Italy [52], and in southwestern Anatolia [90]. Based on the positive correlations between March temperatures and ring width on Mt. Etna, high spring temperatures appear to promote an early start of the growing season. Based on the above zero average temperatures during winter measured at all meteorological stations, possible heat discharge from the volcanic fissure in the years before the eruption (e.g., [91]) could not have caused such an early start of the growing season. In contrast, pines are able to photosynthesize during winter; thus mild temperatures in late winter may enhance the availability of reserves (non-structural carbohydrates) for allocation to cambial growth in spring. Mild conditions in spring may stimulate cambial dynamics or induce early cambial reactivation, increasing production of early-wood.

We found negative correlations between ring width and summer temperature, as previously described in Turkey [51,76,92,93], as well as in north-eastern Greece and the Spanish Pyrenees [94]. These correlations may be related to heat waves and the negative effect of drought stress on tree growth. An increase in frequency in hot and dry summers under climate change during the recent past [95] may also increase drought stress indirectly in autumn and further constrain tree growth.

Low climate sensitivity of the studied trees suggests that other factors must have caused the increased pre-eruption NDVI signal. One of the factors which could have induced fertilization is an increased carbon dioxide (CO2) concentration, which is commonly found in the surrounding of volcanic fissures (e.g., [96,97]).

Pre-eruption CO2-degassing has been observed on Mt. Etna [98] supporting the possible effects on trees analysed in this study. Even though it has been shown that the deposition of volcanic trace elements in tree rings appears to be unrelated to the occurrence of volcanic events [99] trees have been found to grow faster upon elevated concentration of CO2 gas in the surrounding atmosphere (e.g., [100,101]), as may be induced by emissions from the volcanic system. Analyses of tree rings have shown that trees close to natural CO2 springs did take up fossil CO2 but did not grow faster [102104]. Depending on the amount of increased CO2 concentration the effects on tree growth can either be positive [105], negative [106] or even lethal, such as at Mammoth Mountain in California [107]. However, a number of studies have shown that adult, mature trees do not grow faster under elevated CO2 concentrations [102,108,109]. Therefore, it is not clear if an enhanced CO2 concentration would be a suitable candidate to explain the increased NDVI.

Conclusions

We conclude that tree growth at the highest elevations on Mt. Etna is not significantly limited by climate. Our samples were taken near Mt Etna's upper treeline, but this treeline is not climatically determined; instead it is defined by volcanic activity and related disturbances such as wildfires. Consequently, climatic influences on tree growth are weaker than would be expected in trees growing at a climatically induced treeline where temperature is the limiting factor [29]. At the same time, the Mt. Etna trees are growing at an elevation that is too high to be strongly affected by summer drought as in Mediterranean lowlands [19]. The intermediate elevation between the two extremes (high elevation where temperature is limiting and low elevation where summer drought is limiting) makes it difficult to explain the tree-growth variability using meteorological data. The low sensitivity of tree growth to climate suggests that neither i) an increase of surrounding air temperature caused by heating from magma at shallow depths, nor ii) an increase in water availability induced by pre-eruptive subsurface pressures and water vapour, is likely to have enhanced photosynthesis before the 2002/2003 flank eruption. Thus to explain the NDVI signal previously observed [11], one must search for factors (volcanic or not) other than additional water or heat induced by volcanic activity.

Acknowledgments

We thank our colleagues from the WSL, Dr. Ulf Büntgen, Dr. David Frank and Richard L. Peters, for their help with ring-width and climate analyses, Dr. Alexander Bast for helping with programming in R, and Anne Verstege for her continuous support in the dendro laboratory. Also we wish to thank Vincenzo Crimi (Corpo Forestale, Bronte) for his logistical support during our fieldwork on Mt. Etna, as well as the Corpo Forestale della Regione Siciliana (Distaccamento di Bronte, Catania, Italy) for their permission to take samples.

Sadly, Sebastiano Cullotta passed away before the submission of the final version of this manuscript. Ruedi Seiler accepts responsibility for the integrity and validity of the data collected and analyzed. We celebrate the memory of our colleague and friend who tragically passed away while this paper was in preparation.

Author Contributions

  1. Conceptualization: RS JWK PJK RT NH DA SC ME RD PC.
  2. Data curation: RS JWK PJK.
  3. Formal analysis: RS JWK PJK PC.
  4. Funding acquisition: NH ME PC.
  5. Investigation: RS JWK PJK RT DA ME PC.
  6. Methodology: RS JWK PJK RT NH DA SC RD PC.
  7. Project administration: RS RT SC PC.
  8. Resources: JWK PJK RT SC ME RD PC.
  9. Software: RS JWK PJK.
  10. Supervision: ME PC.
  11. Validation: RS JWK PJK NH SC ME PC.
  12. Visualization: RS JWK PC NH.
  13. Writing – original draft: RS JWK PJK RT DA SC ME RD PC.
  14. Writing – review & editing: RS JWK PJK RT NH DA SC ME RD PC.

References

  1. 1. McNutt SR. Seismic monitoring and eruption forecasting of volcanoes: A review of the state-of-the-art and case histories. Berlin Heidelberg: Springer-Verlag; 1996.
  2. 2. Battaglia M, Roberts C, Segall P. Magma intrusion beneath long valley caldera confirmed by temporal changes in gravity. Science 1999;285:2119–2122. pmid:10497128
  3. 3. Williams-Jones G, Rymer H. Detecting volcanic eruption precursors: a new method using gravity and deformation measurements. J Volcanol Geotherm Res 2002;113:379–389.
  4. 4. Sherburn S, Scott B.J, Olsen J, Miller C. Monitoring seismic precursors to an eruption from the Auckland volcanic field, New Zealand. New Zeal J Geol Geop 2007;50(1):1–11.
  5. 5. Hooper A, Prata F, and Sigmundson F. Remote sensing of volcanic hazards and their precursors. Proceedings of the IEEE 2012;100(10):2908–2930.
  6. 6. Sicali S, Barberi G, Cocina O, Musumeci C, Patanè D. Volcanic unrest leading to the July-August 2001 lateral eruption at Mt. Etna: Seismological constraints. J Volcanol Geotherm Res 2015;304:11–23.
  7. 7. Sparks RSJ, Biggs J, Neuberg JW. Monitoring Volcanoes. Science 2012;335:1310–1311. pmid:22422969
  8. 8. Jong SM. Imaging spectrometry for monitoring tree damage caused by volcanic activity in the Long Valley caldera, California. ITC Journal 1998;1–10.
  9. 9. McNutt SR, Rymer H, Stix J. Synthesis of volcano monitoring. In: Encyclopedia of volcanoes. Sigurdsson H, et al., editors. San Diego: Academic Press; 2000. pp. 1167–1185.
  10. 10. Andronico D, Scollo S, Cristaldi A, Ferrari F. Monitoring the ash emission episodes at Mt. Etna: the 16 November 2006 case study. J Volcanol Geotherm Res 2009;180:123–134.
  11. 11. Houlié N, Komorowski JC, de Michele M, Kasereka M, Ciraba H. Early detection of eruptive dykes revealed by normalized difference vegetation index (NDVI) on Mt. Etna and Mt. Nyiragongo. Earth Planet Sci Lett 2006;246:231–240.
  12. 12. Gamon JA, Field CB, Goulden ML, Friffin KL, Hartley AE, Joel G, et al. Relationships between NDVI, canopy structure, and photosynthesis in three Californian vegetation types. Ecological Application 1995;5:28–41.
  13. 13. Goetz SJ, Bunn AG, Fiske GJ, Houghton RA. Satellite-observed photosynthetic trends across boreal North America associated with climate and fire disturbance. Proc Natl Acad Sci USA 2005;102:13521–13525. pmid:16174745
  14. 14. Andronico D, Branca S, Calvari S, Burton M, Caltabiano T, Corsaro RA, et al. A multi-disciplinary study of the 2002–03 Etna eruption: insights into a complex plumbing system. Bull Volcanol 2005;67:314–330.
  15. 15. Aiuppa A, Moretti R, Federico C, Giudice G, Gurrieri S, Liuzzo M, et al. Forecasting Etna eruptions by real-time observation of volcanic gas composition. Geology 2007;35:1115–1118.
  16. 16. Shinohara H. Excess degassing from volcanoes and its role on eruptive and intrusive activity. Rev Geophys 2008;46:RG4005.
  17. 17. Büntgen U, Frank DC, Schmidhalter M, Neuwirth B, Seifert M, Esper J. Growth/climate response shift in a long subalpine spruce chronology. Trees 2006;20:99–110.
  18. 18. Briffa K, Melvin TM, Osborn TJ, Hantemirov RM, Kirdyanov AV, Mazepa VS, et al. Reassessing the evidence for tree-growth and inferred temperature change during the Common Era in Yamalia, northwest Siberia. Quat Sci Rev 2013;72:83–107.
  19. 19. Cherubini P, Gartner BL, Tognetti R, Bräker OU, Schoch W, Innes J. Identification, measurement and interpretation of tree rings in woody species from Mediterranean climates. Biol Rev Camb Philos Soc 2003;78:119–148. pmid:12620063
  20. 20. Gea-Izquierdo G, Cherubini P, Cañellas . Tree-rings reflect the impact of climate change on Quercus ilex L. along a temperature gradient in Spain over the last 100 years. For Ecol Manage 2011;262(9):1807–1816.
  21. 21. Dobbertin M. Tree growth as indicator of tree vitality and of tree reaction to environmental stress: a review. Eur J For Res 2005;124:319–333.
  22. 22. Fritts HC. Tree Rings and Climate. London: Academic Press Inc. Ltd; 1976.
  23. 23. Begum S, Nakaba S, Yamagishi Y, Oribe Y, Funada R. Regulation of cambial activity in relation to environmental conditions: understanding the role of temperature in wood formation of trees. Physiol Plant 2013;147:46–54. pmid:22680337
  24. 24. Köppen W. Die Klimate der Erde. Grundriss der Klimakunde. Berlin: Walter de Gruyter; 1923.
  25. 25. Walter H, Lieth H. Klimadiagramm-Weltatlas. Vienna: Gustav Fischer Verlag; 1960.
  26. 26. Bolle HJ, Eckardt M, Koslowsky D, Maselli F, Melia Miralles J, Menenti M, et al. Mediterranean land surface processes assessed from space. Regional Climate Studies; vol. XXVIII. Springer Series; 2006.
  27. 27. Ma S, Baldocchi DD, Xu L, Hehn T. Inter-annual variability in carbon dioxide exchange of an oak/grass savanna and open grassland in California. Agric For Meteorol 2007;147:157–171.
  28. 28. Cotrufo MF, Alberti G, Inglima I, Marjanovic H, LeCain D, Zaldei A, et al. Decreased summer drought affects plant productivity and soil carbon dynamics in a Mediterranean woodland. Biogeosciences 2011;8:2729–2739.
  29. 29. Maselli F, Cherubini P, Chiesi M, Gilabert MA, Lombardi F, Moreno A, et al. Start of the dry season as a main determinant of inter-annual Mediterranean forest production variations. Agric For Meteorol 2014;194:197–206.
  30. 30. Bonneville A, Gouze P. Thermal survey of Mount Etna volcano from space. Geophys Res Lett 1992;19:725–728.
  31. 31. Poli Marchese E, Grillo M. Primary succession on lava flows on Mt.Etna. In: Burga C.A., Klötzli F., Grabherr G. (Eds.), Gebirge der Erde.: Landschaft, Klima, Pflanzenwelt. Stuttgart: Ulmer; 2004. pp. 291–300.
  32. 32. Branca S, Coltelli M, De Beni E, Wijbrans J. Geological evolution of Mount Etna volcano (Italy) from earliest products until the first central volcanism (between 500 and 100 ka ago) inferred from geochronological and stratigraphic data. Int J Earth Sci 2008;97:135–152.
  33. 33. Burga C, Klötzli F. Gebirge der Erde. Landschaft, Klima, Pflanzenwelt. Stuttgart: Eugen Ulmer GmbH & Co.; 2004.
  34. 34. Doglioni C, Innocenti F, Mariotti G. On the geodynamic origin of Mt. Etna. GNGTS—Atti del 17 Convegno Nazionale, 2002.
  35. 35. Hermes K. Die Lage der oberen Waldgrenze in den Gebirgen der Erde und ihr Abstand zur Schneegrenze (Kölner geographische Arbeiten 5). Universität Köln: Geographisches Institut; 1955.
  36. 36. Körner C. A re-assessment of high elevation treeline positions and their explanation. Oecologia 1998;115:445–459.
  37. 37. Certini G, Sanjurjo MJF, Corti G, Ugolini F. The contrasting effect of broom and pine on pedogenic processes in volcanic soils (Mt. Etna, Italy). Geoderma 2001;102:239–254.
  38. 38. Dazzi C. Environmental features and land use of Etna (Sicily—Italy). In: Arnalds O, Bartoli F, Buurman P, Oskarsson H, Stoops G, Garcia-Rodeja E, editors. Soils of Volcanic Regions in Europe. Heidelberg: Springer-Verlag Berlin; 2007.
  39. 39. Egli M, Mastrolonardo G, Seiler R, Raimondi S, Favilli F, Crimi V, et al. Charcoal and stable soil organic matter as indicators of fire frequency, climate and past vegetation in volcanic soils of Mt.Etna, Sicily. Catena 2012;88:14–26.
  40. 40. Allard P. Endogenous magma degassing and storage at Mount Etna. Geophys Res Lett 1997;24:2219–2222.
  41. 41. Lulli L. Italian volcanic soils. In: Arnalds Ò, Bartoli F, Buurman P, Òskarsson H, Stoops G, Garcia-Rodeja E, editors. Soils of volcanic regions in Europe. Berlin Heidelberg: Springer-Verlag; 2007. pp. 51–67.
  42. 42. Egli M, Mirabella A, Nater M, Alioth L, Raimondi S. Clay minerals, oxyhydroxide formation, element leaching and humus development in volcanic soils. Geoderma 2008;143:101–114.
  43. 43. Maeda T, Takenaka H, Warkentin BP. Physical properties of allophane soils. Adv Agron 1977;29:229–264.
  44. 44. Gärtner H, Cherubini P, Fonti P, von Arx G, Schneider L, Nievergelt D, et al. A technical perspective in modern tree-ring research—How to overcome dendroecological and wood anatomical challenges. J Vis Exp 2015;97:e52337.
  45. 45. R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, Austria. URL: http://www.R-project.org/) (2015)
  46. 46. Wessel P. & Smith W.H.F. New, improved version of generic mapping tools released. Eos Trans. Amer. Geophys. Union 79, 579 (1998).
  47. 47. Egli M, Alioth L, Mirabella A, Raimondi S, Nater M, Verel R. Effect of climate and vegetation on soil organic carbon, humus fractions, allophanes, imogolite, kaolinite, and oxhzdroxides in volcanic soils of Etna (Sicily). Soil Science 2007;172:673–691.
  48. 48. Holmes RL. Computer-assisted quality control in tree ring dating and measurement. Tree-ring Bull 1983;43:69–78.
  49. 49. Grissino-Mayer HD. Evaluating crossdating accuracy: A manual and tutorial for the computer program Cofecha. Tree Ring Res 2001;57(2):205–221.
  50. 50. Mitchell TD, Carter TR, Jones PD, Hulme M, New M. A comprehensive set of high-resolution grids of monthly climate for Europe and the globe: the observed record (1901–2000) and 16 scenarios (2001–2100). Working Paper 55. Norwich: Tyndall Centre for Climate Change Research; 2004.
  51. 51. Griggs C, DeGaetano A, Kuniholm P, Newton M. A regional high-frequency reconstruction of May-June precipitation in the north Aegean from oak tree rings, A.D. 1089–1989. Int J Climatol 2007;27:1075–1089.
  52. 52. Carrer M, Nola P, Motta R, Urbinati C. Contrasting tree-ring growth to climate responses of Abies alba toward the southern limit of its distribution area. Oikos 2010;119:1515–1525.
  53. 53. Lüdeke MKB, Ramge PH, Kohlmaier GH. The use of satellite NDVI data for the validation of global vegetation phenology models: application to the Frankfurt Biosphere Model. Ecol Model 1996;91:255–270.
  54. 54. Büntgen U, Frank DC, Nievergelt D, Esper . Summer temperature variations in the European Alps, A.D. 755–2004. J Clim 2005;9:5606–5623.
  55. 55. Cook E, Kariukstis L. Methods of dendrochronology—Applications in the environmental sciences. Dordrecht: Kluwer Academic Pubishers; 1990.
  56. 56. García-Suárez AM, Butler CJ, Baille MG. Climate signal in tree-ring chronologies in a temperate climate: A multi-species approach. Dendrochronologia 2009;27:183–198.
  57. 57. Cai Q, Liu Y, Lei Y, Bao G, Sun . Reconstruction of the March-august PDSI since 1703 AD based on tree rings of Chinese pine (Pinus tabulaeformis Carr.) in the Lingkong Mountain, southeast Chinese loess Plateau. Clim Past 2014;10:509–521.
  58. 58. Tomé AR, Miranda PMA. Piecewise linear fitting and trend changing points of climate parameters. Geophys Res Lett 2004; 31:L02207
  59. 59. Cook ER, Pederson N. Book- Chapter: Uncertainty, emergence and statistics in dendrochronology. In: Dendroclimatology, Developments in Paleoenvironmental Research 2011;11:77–112.
  60. 60. Kutner M, Nachtsheim C, Neter J, Li W. Applied linear statistical models. 5th ed. New York: McGraw-Hill/Irwin; 2005.
  61. 61. O'Brian RM. A caution regarding rules of thumb for variance inflation factors. Quality & Quantity 2007;41:673–690.
  62. 62. Todaro L, Andreu L, D'Alessandro CM, Gutiérrez E, Cherubini P, and Saracino A. Response of Pinus leucodermis to climate
and anthropogenic activity in the National Park of Pollino (Basilicata, Southern Italy). Biol Cons 2007;137:507–519.
  63. 63. Griggs C, Pearson C, Manning SW, Lorentzen B. A 250-year annual precipitation reconstruction and drought assessment for Cyprus from Pinus brutia Ten. tree-rings. Int J Climatol 2014;34:2702–2714.
  64. 64. Graumlich LJ. Precipitation Variation in the Pacific Northwest (1675–1975) as Reconstructed from Tree Rings. Ann Assoc Am Geogr 1987;77:19–29.
  65. 65. Scarascia-Mugnozza G, Oswald H, Piussi P, Radoglou K. Forests of the Mediterranean region: gaps in knowledge and research needs. For Ecol Manage 2000;132:97–109.
  66. 66. Eckstein D, Aniol R. Dendroclimatological reconstruction of the summer temperatures for an alpine region. Mitt Forstl Bundes-Vers.anst Wien 1981;142:391–298.
  67. 67. Schweingruber FH. Flächenhafte dendroklimatische Temperaturrekonstruktionen für Europa. Naturwissenschaften 1987;74:205–212.
  68. 68. Briffa K, Schweingruber FH, Jones PD, Osborn TJ, Harris IC, Shiyatov SG, et al. Trees tell of past climates: but are they speaking less clearly today? Philosophical Transactions of the Royal Society of London B Biological Sciences 1996;353:65–73.
  69. 69. Hughes MK. Dendrochronology in climatology—the state of the art. Dendrochronologia 2002;20(1–2):95–116.
  70. 70. Leonelli G, Pelfini M, Morra di Cella U. Detecting climatic treelines in the Italian Alps: The influence of geomorphological factors and human impacts. Phys Geogr 2009;30(4):338–352.
  71. 71. Serre-Bachet F, Guiot J. Summer temperature changes from tree-rings in the mediterranean area during the last 800 years. In: Abrupt Climatic Change, Evidence and Implications. NATO ASI Series C, Mathematical and Physical Sciences, Vol. 216. The Netherlands: Reidel Publ. Co.; 1987. pp. 89–97.
  72. 72. Büntgen U, Frank DC, Neuenschwander T, Esper . Fading temperature sensitivity of Alpine tree growth at its Mediterranean margin and associated effects on large-scale climate reconstructions. Clim Change 2012;114:651–666.
  73. 73. Campelo F, Nabais C, Freitas H, Gutiérez E. Climatic significance of tree-ring width and intra-annual density fluctuations in Pinus pinea from a dry Mediterranean area in Portugal. Ann For Sci 2006;64:229–238.
  74. 74. Olivar J, Bogino S, Spiecker H, Bravo F. Climate impact on growth dynamic and intra-annual density fluctuations in Aleppo pine (Pinus halepensis) trees of different crown classes. Dendrochronologia 2012;30:35–47.
  75. 75. Touchan R, Akkemik Ü, Malcolm KH, Erkan N. May-June precipiration reconstruction of southwestern Anatolia, Turkey during the last 900 years from tree rings. Quat Res 2007;68:196–202.
  76. 76. Hughes MK. Aegean tree-ring signature years explained, Tree-Ring Res 2001;57:438–450.
  77. 77. Seim A, Büntgen U, Fonti P, Haska H, Herzig F, Tegel W, et al. Climate sensitivity of a millennium-long pine chronology from Albania. Clim Res 2012;51:217–228.
  78. 78. Navarro-Cerrillo RM, Sánchez-Salguero R, Manzanedo RD, Camarero JJ, Fernández Cancio A. Site and age condition the growth responses to climate and drought of relict Pinus nigra subsp. salzmannii populations in Southern Spain. Tree Ring Res 2014;70:145–155.
  79. 79. Maselli F. Monitoring forest conditions in a protected Mediterranean coastal area by the analysis of multiyear NDVI data. Remote Sens Environ 2004;89:423–433.
  80. 80. Akkemik Ü, D'Arrigo R, Cherubini P, Köse N, Jacoby G. Tree-ring reconstructions of precipitation and streamflow for north-western Turkey. Int J Climatol 2008;28:173–183.
  81. 81. Allard V, Ourcival JM, Rambal S, Joffre R, Rocheteau A. Seasonal and annual variation of carbon exchange in an evergreen Mediterranean forest in southern France. Glob Chang Biol 2008;14:714–725.
  82. 82. Aloui A. Recherches dendroclimatologiques en Kromirie (Tunesie). PhD. Thesis, Université d'Aix-Marseille III, France. 1982.
  83. 83. Tessier L. Dendroclimatologie et Ecologie de Pinus sylvestris L. et Quercus pubescens Willd. dans le Sud-Est de la France. PhD. Thesis, Université d'Aix-Marseille III. 1984.
  84. 84. Lebourgeois F, Levy G, Aussenac G, Clerc B, Willm F. Influence of soil drying on leaf water potential, photosynthesis, stomatal conductance and growth in two black pine varieties. Ann For Sci 1998;55:287–299.
  85. 85. Camarero JJ, Manzanedo RD, Sánchez-Salguero R, Navarro-Cerrillo. Growth response to climate and drought change along an aridity gradient in the southernmost Pinus nigra relict forests. Ann For Sci 2013;70:769–780.
  86. 86. Tognetti R, Lombardi F, Lasserre B, Cherubini P, Marchetti M. Tree-ring stable isotopes reveal twentieth-century increases in water-use efficiency of Fagus sylvatica and Nothofagus spp. in Italian and Chilean mountains. PLoS One 2014;9:e113136. pmid:25398040
  87. 87. Reisser M. Soil analysis at Mount Etna to explain faster tree growth preceding a volcanic eruption. Master Thesis. Department of Earth-System Sciences, University of Zurich. Switzerland; 2014.
  88. 88. Limm EB, Simonin KA, Bothman AG, Dawson TE. Foliar water uptake: a common water acquisition strategy for plants of the redwood forest. Oecologia 2009;161:449–459. pmid:19585154
  89. 89. Renault NL, Alvera B, Garcia-Ruiz JM. The snowmelt period in a Mediterranean high mountain catchment: runoff and sediment transport. Cuadernos de Investigacion Geografica 2010;36:99–108.
  90. 90. Touchan R, Anchukaitis KJ, Shishov VV, Fatih S, Attieh J, Ketmen M, et al. Spatial patterns of eastrn Mediterranean climate influence on tree growth. Holocene 2014;24(4):381–392.
  91. 91. Aubert M. Practical evaluation of steady heat discharge from dormant active volcanoes: case study of Vulcarolo fissure (Mount Etna, Italy). J Volcanol Geotherm Res 1999;92:413–429.
  92. 92. Köse N, Akkemik Ü, Dalfes N. Tree-ring reconstructions of May-June precipitation for western Anatolia. Quat Res 2011;75:438–450.
  93. 93. Köse N, Akkemik Ü, Güner HT, Dalfes HN, Frissino-Mayr HD, Özeren MS, et al. an improved reconstruction of May-June precipitation using tree-ring data from western Turkey and its links to volcanic eruptions. Int J Biometeorol 2013;57:691–701. pmid:23015281
  94. 94. Tardif J, Camarero JJ, Ribas M, Gutiérrez E. Spatiotemporal variability in tree growth in the central Pyrenees: Climatic and site influences. Ecol Monogr 2003;73(2):241–257.
  95. 95. Christensen JH, Krishna Kumar K, Aldrian E, An SI, Cavalcani IFA, de Castro M, et al. Climate Phenomena and their Relevance for Future Regional Climate Change. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Interngovernmental Panel on Climate Change (Stocker TF, Qin D, Plattner GK, Tigor M, Allen SK, Boschung J, et al. (Eds.)). Cambridge University Press, Cambridge, United Kingdom and New York, USA. 2013.
  96. 96. Giammanco S, Gurrieri S, Valenza M. Anomalous soil CO2 degassing in relation to faults and eruptive fissures on Mount Etna (Sicily, Italy). Bull Volcanol 1998;60:252.
  97. 97. Liuzzo M, Di Muro A, Giudice G, Michon L, Ferrazzini V, Gurrieri S. New evidence of CO2 soil degassing anomalies on Piton de la Fournaise volcano and the link with volcano tectonic structures. Geochem Geophys Geosyst 2015; 16.
  98. 98. Allard P, Carbonelle J, Dajlevic D, Le Bronec J, Morel P, Robe MC, Maurenas JM, Faivre-Pierret R, Martin D, Sabroux JC, Zettwoog P. Eruptive and diffuse emissions of CO2 from Mount Etna. Nature 1991;351: 387–391.
  99. 99. Watt SFL, Pyle DM, Mather TA, Day JA, Aiuppa A. The use of tree-rings and foliage as an archive of volcanogenic cation deposition. Environ Pollut 2007;184:48–61.
  100. 100. Norby RJ, Wullschleger SD, Gunderson CA, Johnson DW, Ceulemans R. Tree responses to rising CO2 in field experiments: implications for the future forest. Plant Cell Environ 1999;22:683–714.
  101. 101. Ainsworth EA, Long SP. What have we learned from 15 years of free-air CO2 enrichment (FACE)? A meta-analytic review of the responses of photosynthesis, canopy properties and plant production to rising CO2. New Phytol 2005;165:351–372. pmid:15720649
  102. 102. Tognetti R. Cherubini P. Innes JL. Comparative stem-growth rates of Mediterranean trees under background and naturally enhanced ambient CO2 concentrations. New Phytol 2000;146:59–74.
  103. 103. Saurer M, Cherubini P, Bonani G, Siegwolf R. Tracing carbon uptake from a natural CO2 spring into tree rings: an isotope approach. Tree Physiology 2003;23:997–1004. pmid:12952786
  104. 104. Donders TH, Decuyper M, Beabien SE, Van Hoof TB, Cherubini P, Sass-Klaassen U. Tree rings as biosensor to detect leakage of subsurface fossil CO2. Int J Greenh Gas Control 2013;19:387–395.
  105. 105. Knapp PA, Soulé PT, Grissino-Mayer HD. Detecting potential regional effects of increased atmospheric CO2 on growth rates of western juniper. Global Change Biol 2001;7:903–917.
  106. 106. Biondi F, Fessenden JE. Response of Lodgepole Pine growth to CO2 degassing at Mammoth Mountain, California. Ecology 1999;80(7):2420–2426.
  107. 107. Cook AC, Hainsworth LJ, Sorey ML, Evans WC, Southon JR. Radiocarbon studies of plant leaves and tree rings from Mammoth Mountain, CA: a long-term record of magmatic CO2 release. Chem Geol 2001;177:117–131.
  108. 108. Körner C. Carbon limitation in trees. J Ecol 2003;91:4–17.
  109. 109. Klein T, Bader MKF, Leuzinger S, Mildner M, Schleppi P, Siegwolf RTW et al. Growth and carbon relations of mature Picea abies trees under 5 years of free-air CO2 enrichment. J Ecol 2016;104:1720–1733.