The slight, negligible differences between the maps published in 1994 and those posted on this site are due mainly to the different isometric classes adopted in the two editions. It should also be considered that the evapotranspiration data had been plotted manually in the 1994 study, whereas in the present version the interpolation of the data and the 2D/3D mapping were computerized.

For a detailed description of the rather complicated model originally adopted, see the above-mentioned publication. Suffice it to mention briefly here how the complex orography of the territory made it necessary to use the historical series of data gathered over a period of twenty years on 267 meteorological stations, only 54 of which (the network of stations of the Aeronautica Militare and the station of the Volturno Experimental Farm of the former Southern Italy Development Agency) provided all the parameters necessary for the application of the combined formula chosen. The remaining 213 stations (the network of the Ministry of Public Works) provided only air temperatures and precipitation. To overcome this shortcoming, the following steps were taken:

1 - the Penman-Monteith formula locally calibrated in the energetic term was chosen as the best of the 11 formulae compared for simulating the ETo measured on two agrometeorological areas representative of the coastal climate of Southern Italy. Incidentally, the Hargreaves-Samani formula (coefficient 0.0023) worked well as it had originally been conceived and calibrated for the climate of Southern California which is similar to that of Southern Italy;

2 - ETo was calculated according to Penman-Monteith (Smith, 1991) for the 54 stations which had provided all necessary parameters;

3 - a regression formula was identified which was capable of providing a good estimate of ETo on the 54 stations using only the geographical and temperature data available for the other 213 stations (Ravelli and Rota, 1994):

in which

The following graph shows the good correlation between the ETo according to the regression formula adopted (RRm model) and ETo according to Penman-Monteith.

4 - before calculating ETo, the meteorological data gathered on the stations, which did not have the necessary fetch, were corrected;

5 - the regression formula in step 3 above was used to calculate (on all 267 stations) minimum, average and maximum ETo and ETo at 50% and 75% probability of non-exceedance, for a wind velocity of 150 km/d;

6 - the data obtained were then used to map annual ETo and peak month ETo at probability of non-exceedance of 75%;

7 - a table was prepared to correct ETo for wind velocities other than 150 km/d:

Various procedures were developed to correct the original meteorological data which were conditioned by the manner in which they had been gathered and the location of the stations which did not always meet the standards typical of an agrometeorological station for irrigation purposes. Of particular interest are the algorithms that were developed to make it possible to use the data of some meteorological parameters that had been gathered by the Aeronautica Militare and published by the Istituto Centrale di Statistica (ISTAT, various years) in a form unsuitable for application in the climatic formulae adopted (Ravelli and Rota, 1994):

- for the transformation of wind velocity data:

in which:
Vd₁,…n = average wind velocity in knots from the provenence direction 1, … n
Fd₁,…n = frequency (number) of measurements in the provenance direction 1, … n
F calm = frequency (number) of measurements with wind velocity <2 knots

- for the transformation of cloud data (state of the sky):

in which:

Eeff	=	actual sunshine
Ns, m, c	=	number of sunny, partly cloudy, cloudy days
Eass	=	absolute sunshine
Nmese	=	number of days of the month

ETo Map (probability of non-exceedance 75%)
annual ETo in mm
peak month ETo in mm/d

The inland areas of the peninsula and of the islands above 500 meters a.s.l. are indicated in white.