Temperature and Precipitation Variability and Trends in Northern Spain
267
4.3 Trends and Abrupt Changes
This section deals with identifying the presence of trends and abrupt changes in the
PC time series. For these series, the lag-one correlation is very small (no significant
at 95% level). The right-hand side of Figures 3 - 5 also show the result of the MannKendall test applied to the first principal component series.
All the series except the spring ones show a significant increase. Winter shows
a smooth trend, while autumn presents an abrupt change at the end of 1930s.
Summer series presents and increase, although without significance during the
1970s. For the annual series, the first series shows an abrupt change in 1936 (this
point change is near the significant level and thus perhaps could be described as a
trend).
For the Northern Plateau, the increase appears for the minimum series, with
values that rise more than 2° C between the first and the last years of the records.
Maximum temperatures present a high increase in the last years. (Figure 6).
4.4 Discussion of Temperature Results
The annual Spanish series have been compared with the global and Northern
Hemisphere series (Jones et aI., 1994). The correlations between this first PC annual
series and the global and Northern Hemisphere series are 0.46 and 0.49 respectvely,
(significant at 95% level) In general, the Spanish series show greater variability. For
the Northern Hemisphere series (and also for the global one), the agreement with the
Spanish series appears in the cold period at the end of 19th century, the slight
increase at the beginning of this century and the strong decrease before 1910;
however between 1910 and 1945, these global and hemisphere series do not show
great fluctuations, but rather clearly increasing trends that stabilize during the 1950s
(this last point agrees with Spanish series). The decade of the 1970s was cold
throughout the world, although apparently that it was colder in Spain, which also
had more intense the temperature increase in the 1980s.
With regard to the relationships between temperature and NAO and pressure
conditions, Table 4 shows the correlations between the NAO index proposed by
Jones et al. (1997) (using Gibraltar pressure data as the southern location) and also
with the mean sea level pressure data of Gibraltar as representative of the circulation
conditions over the Iberian Peninsula. It is surprising the zero correlation for winter
between pressure or circulation index, bearing in mind that circulation conditions
have been proposed as an important temperature variability agent during the cold
season (Wallace et aI., 1995). In fact, precipitation over the area is controlled mainly
by the NAO (Esteban-Parra et aI., 1998).This situation may be due to different
effects on maximum and minimum winter temperatures: during high pressure and
high NAO conditions, maximum temperatures increases due to solar heating,
whereas minima decrease due to radiative cooling, and thus, for low NAO and low
pressure conditions, the reverse is true with regard the temperature behaviour. This
267
4.3 Trends and Abrupt Changes
This section deals with identifying the presence of trends and abrupt changes in the
PC time series. For these series, the lag-one correlation is very small (no significant
at 95% level). The right-hand side of Figures 3 - 5 also show the result of the MannKendall test applied to the first principal component series.
All the series except the spring ones show a significant increase. Winter shows
a smooth trend, while autumn presents an abrupt change at the end of 1930s.
Summer series presents and increase, although without significance during the
1970s. For the annual series, the first series shows an abrupt change in 1936 (this
point change is near the significant level and thus perhaps could be described as a
trend).
For the Northern Plateau, the increase appears for the minimum series, with
values that rise more than 2° C between the first and the last years of the records.
Maximum temperatures present a high increase in the last years. (Figure 6).
4.4 Discussion of Temperature Results
The annual Spanish series have been compared with the global and Northern
Hemisphere series (Jones et aI., 1994). The correlations between this first PC annual
series and the global and Northern Hemisphere series are 0.46 and 0.49 respectvely,
(significant at 95% level) In general, the Spanish series show greater variability. For
the Northern Hemisphere series (and also for the global one), the agreement with the
Spanish series appears in the cold period at the end of 19th century, the slight
increase at the beginning of this century and the strong decrease before 1910;
however between 1910 and 1945, these global and hemisphere series do not show
great fluctuations, but rather clearly increasing trends that stabilize during the 1950s
(this last point agrees with Spanish series). The decade of the 1970s was cold
throughout the world, although apparently that it was colder in Spain, which also
had more intense the temperature increase in the 1980s.
With regard to the relationships between temperature and NAO and pressure
conditions, Table 4 shows the correlations between the NAO index proposed by
Jones et al. (1997) (using Gibraltar pressure data as the southern location) and also
with the mean sea level pressure data of Gibraltar as representative of the circulation
conditions over the Iberian Peninsula. It is surprising the zero correlation for winter
between pressure or circulation index, bearing in mind that circulation conditions
have been proposed as an important temperature variability agent during the cold
season (Wallace et aI., 1995). In fact, precipitation over the area is controlled mainly
by the NAO (Esteban-Parra et aI., 1998).This situation may be due to different
effects on maximum and minimum winter temperatures: during high pressure and
high NAO conditions, maximum temperatures increases due to solar heating,
whereas minima decrease due to radiative cooling, and thus, for low NAO and low
pressure conditions, the reverse is true with regard the temperature behaviour. This