246
E. Piervitali, M. Colacino
i) to analyse long term series of observation data to verify if climatic variability is
already detectable and
ii) to try to forecast the future pattern by statistical models.
With reference to the long term data series different authors agree on a reduction
of rainfall in the Northern Hemisphere for latitudes lower than 50
c : however the
values are in disagreement (Bradley and Groisman, 1989; Diaz et aI., 1989;
Vinnikov et aI., 1990). Also regional studies have been carried out by using different
methodologies and different samples areas (Groisman and Easterling, 1994; BenGai et aI., 1994; Norsallah and Balling, 1996). Since in the Mediterranean region
only few authors give quantitative results (Palutikof et aI., 1994), a systematic
analysis of yearly and monthly precipitation trend has been performed in the
Central-Western Mediterranean basin (Piervitali et aI., 1998).
With reference to the forecast, the statistical approach widely used consists in
trying climate prediction assuming a well defined external forcing. In particular,
many works have been developed connecting the precipitation trend to ENSO and
assessing the future evolution of rainfall by the forecast of the forcing phenomenon
(Price et aI., 1998; Rodo, et aI. 1997).
In the present work the analysis of yearly precipitation over the Central-Western
Mediterranean basin is reported. It has been carried out examining data collected in
the period 1951-1995 in a network of 59 stations lying along the rim of the basin.
It generally indicates a rainfall reduction in the whole region, particularly strong in
the southern part.
Since in a study developed on the teleconnections between ENSO and
precipitation in Central-Western Mediterranean basin (Colocino et aI., 2001) no
correlation has been found, we have analysed the role of the North Atlantic
Oscillation (hereafter NAO) and of the Mediterranean Oscillation (hereafter MO).
The last one seems to be anticorrelated with the precipitation values. On the basis
of the correlation analysis we have assumed the MO as climatic forcing factor. We
have used the statistical model of Winter to forecast the future values of the MO
index and a prediction up to 2030 has been performed. Using, then, the
anticorrelation between the MO index and the precipitation rate, rainfall until 2030
has been predicted for the different stations of the Central-Western Mediterranean
basin.
Since also the evapotranspiration is affected by climatic change, because higher
temperature imply high evapotranspiration and reduced run-off (Henderson-Sellers
and Hansen, 1995) soil moisture has been evaluated, using De Martonne's index,
which depends on precipitation and temperature. Scenario data for 2010 and 2030
show a soil aridity in progress with a possible risk of desertification in some areas.
2 Data Set
Precipitation data used in the present work have been gathered as yearly values for
the 59 stations listed in Table 1. Since these series do not present missing values, no
E. Piervitali, M. Colacino
i) to analyse long term series of observation data to verify if climatic variability is
already detectable and
ii) to try to forecast the future pattern by statistical models.
With reference to the long term data series different authors agree on a reduction
of rainfall in the Northern Hemisphere for latitudes lower than 50
c : however the
values are in disagreement (Bradley and Groisman, 1989; Diaz et aI., 1989;
Vinnikov et aI., 1990). Also regional studies have been carried out by using different
methodologies and different samples areas (Groisman and Easterling, 1994; BenGai et aI., 1994; Norsallah and Balling, 1996). Since in the Mediterranean region
only few authors give quantitative results (Palutikof et aI., 1994), a systematic
analysis of yearly and monthly precipitation trend has been performed in the
Central-Western Mediterranean basin (Piervitali et aI., 1998).
With reference to the forecast, the statistical approach widely used consists in
trying climate prediction assuming a well defined external forcing. In particular,
many works have been developed connecting the precipitation trend to ENSO and
assessing the future evolution of rainfall by the forecast of the forcing phenomenon
(Price et aI., 1998; Rodo, et aI. 1997).
In the present work the analysis of yearly precipitation over the Central-Western
Mediterranean basin is reported. It has been carried out examining data collected in
the period 1951-1995 in a network of 59 stations lying along the rim of the basin.
It generally indicates a rainfall reduction in the whole region, particularly strong in
the southern part.
Since in a study developed on the teleconnections between ENSO and
precipitation in Central-Western Mediterranean basin (Colocino et aI., 2001) no
correlation has been found, we have analysed the role of the North Atlantic
Oscillation (hereafter NAO) and of the Mediterranean Oscillation (hereafter MO).
The last one seems to be anticorrelated with the precipitation values. On the basis
of the correlation analysis we have assumed the MO as climatic forcing factor. We
have used the statistical model of Winter to forecast the future values of the MO
index and a prediction up to 2030 has been performed. Using, then, the
anticorrelation between the MO index and the precipitation rate, rainfall until 2030
has been predicted for the different stations of the Central-Western Mediterranean
basin.
Since also the evapotranspiration is affected by climatic change, because higher
temperature imply high evapotranspiration and reduced run-off (Henderson-Sellers
and Hansen, 1995) soil moisture has been evaluated, using De Martonne's index,
which depends on precipitation and temperature. Scenario data for 2010 and 2030
show a soil aridity in progress with a possible risk of desertification in some areas.
2 Data Set
Precipitation data used in the present work have been gathered as yearly values for
the 59 stations listed in Table 1. Since these series do not present missing values, no