The long maximum and minimum temperature series of the Iberian Peninsula are analyzed covering the period 1900-2000. To identify coherent behaviour in these series, the PCA method has been used, resulting in several major noise-filtered time series. For maximum and minimum temperatures two statistically significant EOFs have been obtained with high loading values over the western and eastern parts of the Iberian Peninsula. However, the spatio-temporal variability of the entire region can be described by the first EOF and its time evolution by the corresponding PC series. Minimum temperature series shows a general upward trend, meanwhile maximum one presents a strong increase since the 1970s.
The relationships between extreme temperature and the NAO index were studied on a seasonal basis. For this purpose, we computed the NAO index using Gibraltar pressure data as the southern location for winter and Ponta Delgada (Azores) for the remaining seasons. Only for the winter (DJF) the correlation coefficients are significant (at the 5% level), with values of 0.35 and -0.18 for maximum and minimum temperatures respectively. The opposite signs for these correlation coefficients reflect the fact that, for positive (negative) NAO winter months, the Iberian Peninsula experience anomalously anticyclonic (cyclonic) circulation and reduced (enhanced) precipitation, and these are associated with reduced (enhanced) cloud cover. Thus, when NAO index is positive, the additional daytime incoming solar radiation increases the maximum temperature, while the corresponding night-time minimum temperature decreases due to outgoing long-wave radiation. For the other seasons, no significant correlations are found.
On the other hand, these relationships are non stationary, changing markedly over time. There seems to be an increase on the influence of the NAO pattern on the Iberian temperatures during the last 30 years. In order to explore in more detail this fact, we are currently studying the impact of NAO on cloudiness during the entire period 1901-1995 using the recently developed high resolution (0.5º lat. x 0.5º lon.) database by CRU.