but with some characteristic features. First, the negative response to previous
February temperatures became stronger during the late twentieth century in the
declining sites, while a similar response to current July temperatures almost disappeared. Second, the positive response to previous March precipitation and current
June precipitation also turned out to be stronger during the past decades. These
changes coincided with warmer February conditions and the drier September
conditions detected during the last decades.
The maximum growth reduction of silver fir occurred in 1986, which was preceded by the highest water deficit of the available climatic record. Defoliation and
mortality were widespread in western low-elevation sites related to the warming
trend since the 1980s and the outstanding water deficit in 1985, which were higher
there than in the eastern study area. Macias et al. (2006) also suggested that silver fir
is experiencing a greater late-summer drought stress in the Spanish Pyrenees and that
the effects of water shortage on growth were more intense in low-elevation stands
subjected to higher water deficit than elsewhere. The detected increase in drought
stress was not only due to a decrease in precipitation, since similar dry periods
occurred in the 1940s, and suggest a link with the 1980s warming (Vicente–Serrano
et al. 2015) when the water-use efficiency of declining trees also increased sharply
(Linares and Camarero 2012). The most pronounced warming in the western than in
the eastern Aragón Pyrenees indicates that warming-induced drought stress triggered
forest dieback westwards (Fig. 6.2). In addition, climate-growth relationships support a key role of late-summer water deficit controlling the silver-fir growth and
dieback patterns in the Aragón Pyrenees (Figs. 6.3 and 6.7).
These results agree with dendroecological studies which found that silver-fir
growth is very sensitive to water deficit during August and September before ring
formation, responding negatively to high temperatures in those months (Bert 1993;
JFig. 6.7 Climate-growth relationships in Pyrenean silver-fir forests. a The upper graphs show the
significant (P < 0.05) bootstrap correlation coefficients calculated between monthly climatic data
(T mean temperature; P total precipitation) and the residual chronologies for the period 1950–
1999. The strength of the correlation is indicated by the size of the symbol. Open squares indicate
a positive correlation, and solid circles indicate a negative correlation. The months studied go from
previous January to current September (months abbreviated by capital letters correspond to the
year of tree-ring formation). The climatic data were calculated for the two climatic sub-regions
(Western and Eastern Aragón Pyrenees), which are divided by the vertical dashed line. The sites
are arranged from the west (left) to the east (right). The codes of declining sites are underlined.
b Relative positions of silver-fir correlation functions based on the first two components of a
Principal Component Analysis (axes 1 and 2 correspond to the first and second Principal
Components, respectively) calculated on the matrix of climate-growth correlations shown in (a).
Only the most significant climatic variables (arrows) are represented, and they are abbreviated
using a three-letter code. Climatic variables starting with “W” refer to cumulative monthly water
deficit (e.g. WAug, cumulative water deficit from January up to August of the year of growth). The
months studied go from previous January to current September (months abbreviated by lower-/
uppercase letters correspond to the previous/current year of tree-ring formation; e.g. TApr stands
for April temperature of the year of tree-ring formation). The climatic data were calculated for the
two climatic sub-regions (Western and Eastern Aragón Pyrenees). Declining sites are shown as
underlined bold codes
148
J.J. Camarero
February temperatures became stronger during the late twentieth century in the
declining sites, while a similar response to current July temperatures almost disappeared. Second, the positive response to previous March precipitation and current
June precipitation also turned out to be stronger during the past decades. These
changes coincided with warmer February conditions and the drier September
conditions detected during the last decades.
The maximum growth reduction of silver fir occurred in 1986, which was preceded by the highest water deficit of the available climatic record. Defoliation and
mortality were widespread in western low-elevation sites related to the warming
trend since the 1980s and the outstanding water deficit in 1985, which were higher
there than in the eastern study area. Macias et al. (2006) also suggested that silver fir
is experiencing a greater late-summer drought stress in the Spanish Pyrenees and that
the effects of water shortage on growth were more intense in low-elevation stands
subjected to higher water deficit than elsewhere. The detected increase in drought
stress was not only due to a decrease in precipitation, since similar dry periods
occurred in the 1940s, and suggest a link with the 1980s warming (Vicente–Serrano
et al. 2015) when the water-use efficiency of declining trees also increased sharply
(Linares and Camarero 2012). The most pronounced warming in the western than in
the eastern Aragón Pyrenees indicates that warming-induced drought stress triggered
forest dieback westwards (Fig. 6.2). In addition, climate-growth relationships support a key role of late-summer water deficit controlling the silver-fir growth and
dieback patterns in the Aragón Pyrenees (Figs. 6.3 and 6.7).
These results agree with dendroecological studies which found that silver-fir
growth is very sensitive to water deficit during August and September before ring
formation, responding negatively to high temperatures in those months (Bert 1993;
JFig. 6.7 Climate-growth relationships in Pyrenean silver-fir forests. a The upper graphs show the
significant (P < 0.05) bootstrap correlation coefficients calculated between monthly climatic data
(T mean temperature; P total precipitation) and the residual chronologies for the period 1950–
1999. The strength of the correlation is indicated by the size of the symbol. Open squares indicate
a positive correlation, and solid circles indicate a negative correlation. The months studied go from
previous January to current September (months abbreviated by capital letters correspond to the
year of tree-ring formation). The climatic data were calculated for the two climatic sub-regions
(Western and Eastern Aragón Pyrenees), which are divided by the vertical dashed line. The sites
are arranged from the west (left) to the east (right). The codes of declining sites are underlined.
b Relative positions of silver-fir correlation functions based on the first two components of a
Principal Component Analysis (axes 1 and 2 correspond to the first and second Principal
Components, respectively) calculated on the matrix of climate-growth correlations shown in (a).
Only the most significant climatic variables (arrows) are represented, and they are abbreviated
using a three-letter code. Climatic variables starting with “W” refer to cumulative monthly water
deficit (e.g. WAug, cumulative water deficit from January up to August of the year of growth). The
months studied go from previous January to current September (months abbreviated by lower-/
uppercase letters correspond to the previous/current year of tree-ring formation; e.g. TApr stands
for April temperature of the year of tree-ring formation). The climatic data were calculated for the
two climatic sub-regions (Western and Eastern Aragón Pyrenees). Declining sites are shown as
underlined bold codes
148
J.J. Camarero