Sensitivity analysis of different evapotranspiration methods using a new sensitivity coefficient

Abstract

The estimation of evapotranspiration is essential in water resources management. Among a group of
 methods, the Penman–Monteith has been commonly applied to calculate reference
 evapotranspiration as this method has been also recommended by the Food and Agriculture
 Organization of the U.N. (FAO). Other methods widely used are: the FAO 24 Penman, the modified
 Blaney and Criddle, the FAO 24 Makkink, and the Hargreaves.
 Sensitivity analysis is required to gain a better understanding of the meteorological systems;
 particularly to indicate the physical meaning of each meteorological parameter used in the estimation
 of the reference evapotranspiration. Several dimensionless sensitivity coefficients have been
 proposed, based on the partial derivative of the dependent variable (reference evapotranspiration) to
 the independent variables (meteorological variables).
 In this paper, a new sensitivity coefficient is proposed to drive sensitivity analysis of the
 evapotranspiration methods. The new sensitivity coefficient uses the partial derivative and the
 standard deviation of each independent variable. The meteorological variables, whose influence has
 been examined, are all the necessary meteorological parameters for the calculation of reference
 evapotranspiration, such as temperature, solar radiation, wind speed and relative humidity for each
 method. Data from the automatic meteorological station of Aminteo in the Prefecture of Florina,
 Western Macedonia, were used. The sensitivity coefficients were calculated for each month, year
 and irrigation period. The comparison of the sensitivity coefficients is performed for the month of
 water peak demand (July), the irrigation period and the year for each evapotranspiration method.
 Results show that the influence of the variables to evapotranspiration is not the same for each
 period, and also the order that the variables influence evapotranspiration is changing. A comparison
 between the five evapotranspiration methods shows that solar radiation and temperature are the
 main parameters that affect evapotranspiration, while relative humidity and wind speed are not so
 important for the calculation of evapotranspiration.