Table 1.
Summary of the climatic data sets used in this study.
Cumulative rainfall and reference evapotranspiration (ETo), average maximum and minimum temperature, evapotranspiration deficit (rain-ETo) and synthetic agrometeorological indicator (SAM) [52], in the most crucial periods of wheat growth, for the years used in the sensitivity analysis study and for the long term climatic data (30 years).
Fig 1.
Meteorological datasets used in the sensitivity analysis study.
Time trends of daily maximum and minimum air temperature (left), rainfall and reference evapotranspiration (right) for the sites of Xiaotanshan (top), Yangling (centre) and Viterbo (bottom), for the two years considered in each study site.
Fig 2.
Relationship between measured and simulated grain yield in winter wheat with the SAFYE model, across 4 years at Xiaotanghan.
Fig 3.
Values of Morris mean effect (μ*) for the Aquacrop model using climatic scenarios for Xiaotanshan (a), Yangling (b) and Viterbo (c) for two crop growth seasons each.
Bars in red identify drier years, bars in blue wetter years. Parameters abbreviations are given in S1 Table. Parameters with mean μ*values lower than 0.1 t ha-1 have been omitted.
Fig 4.
Values of Morris mean effect (μ*) for the SAFYE model using climatic scenarios for Xiaotanshan (a), Yangling (b) and Viterbo (c) for two crop growth seasons each.
Bars in red identify drier years, bars in blue wetter years. Parameters abbreviations are given in S2 Table. Parameters with μ*values lower than 0.1 t ha-1 have been omitted.
Table 2.
List of Aquacrop factors that were found to have a negligible influence on grain yield, i.e. with Morris total sensitivity index μ* <0.1 t ha-1, in all the climatic scenarios of the present study.
Table 3.
List of SAFYE factors that were found to have a negligible influence on grain yield, i.e. with Morris total sensitivity index μ* <0.1 t ha-1, in all the climatic scenarios of the present study.
Fig 5.
Main sensitivity index (Si) (left panels) and Total Sensitivity index (STi) (right panels), for the three sites, Xiaotanshan (top), Yangling (centre) and Viterbo (bottom), for the two wheat growth seasons examined (wet years bars are blue and dry years red). Abbreviations of parameters and input factors are reported in S1 Table.
Table 4.
List of factor of the Aquacrop model resulting highly influential on the grain yield, according to the main sensitivity index (STi>0.1) from the EFAST analysis, ranked from the most to the least influential.
Fig 6.
First-order and total sensitivity indices estimated by the EFAST method for the Aquacrop model.
The first part of the bars (dark grey) corresponds to the average estimate of the first-order index (Si) over all the sites and scenarios, the full bars indicate average estimates of total index (STi), while the lines indicate extreme estimates of total indices.
Fig 7.
Main sensitivity index (Si) (left panels) and total sensitivity index (STi) (right panels), for the three sites, Xiaotanshan (top), Yangling (centre) and Viterbo (bottom), for the two wheat growth seasons examined (wet years bars are blue and dry years red). Abbreviations of parameters and input factors are reported in S2 Table.
Table 5.
List of factor of the SAFYE model resulting highly influential on the grain yield, according to the main sensitivity index (STi>0.1) from the EFAST analysis, ranked from the most to the least influential.
Fig 8.
First-order and total sensitivity indices estimated by the EFAST method for the SAFYE model.
The first part of the bars (dark grey) corresponds to the average estimate of the first-order index (Si) over all the sites and scenarios, the full bars indicate average estimates of total indices (STi), while the lines indicate extreme estimates of total indices.
Table 6.
Synthesis of the assessment of the plasticity of the SAFYE and Aquacrop models.
The plasticity index L is calculated according to Eq 4 from the normalized synthetic agrometeorological indicator (SAM) (Eq 5) and its standard deviation σSAM and the top-down concordance coefficient (TDCC) [5,3].