عنوان مقاله [English]
Saffron (Crocus sativus L.) is a subtropical plant mostly cultivated in South Khorasan and Khorasan Razavi provinces, Iran. Total saffron production is about 200 ton all over the world, and since it is native to Iran, the country has an important role to produce saffron. In recent years, drought showed a significant effect on water resources in Birjand, South Khorasan. Due to lack of full equipped meteorological stations and sufficient lysimetery studies, over irrigation and deficit irrigation are mostly common among saffron farmers. The cause of saffron yield decline is mostly related to irrigation. In addition, farmers need the simple equation/formula to determine the water demand, so, it is necessary to present a single and simple equation. Regarding to this purpose, this study was conducted to achieve mentioned goals: 1) determination of the best evapotranspiration equation, 2) comparison of saffron water need based on 12 most relevant evapotranspiration equations, and 3) determination a single-parameter model to rapidly and accurately estimate saffron water need based on gamma test (due to lack of all FAO-Penman-Monteith parameters in all Birjand’s meteorological stations).
Materials and methods
This study was conducted using meteorological data collected from Birjand´s synoptic station during 1984-2014. This station is located at longitude 59˚ 21' E and latitude 32˚ 87' N and elevation 1491m. Twelve evapotranspiration equations (Blaney-Criddle, Hargreaves, Turc, Priestley-Taylor, Thornthwaite, Jensen-Haise, Makkink, Modified Jensen-Haise, Irmak (Rn), Irmak (Rs), Lowry-Johnson, Pan-Class A) were evaluate to determine the best accurate one. To compare the 12 evapotranspiration equations results with FAO-Penman-Monteith equation (FAO56-PM), four statistical criteria (R2, MBE, RMSE and EF) were used. In order to comprise saffron water demand to actual water used, amount of irrigation water for saffron cultivation in Birjand plain was measured during 2014. In addition, saffron yield was recorded during this growing season in this region. In order to propose a single parameter and accurate model based on easily accessible meteorological parameters, gamma test was also employed.
Results and discussion
Results of coefficient of determination (R2) were acceptable (>0.9) for all equations except Turk. The results showed that Blaney-Criddle, Modified Jensen-Haise and Hargreaves had better accuracy compared to other equations so that their root mean square errors (RMSE) were 0.572, 0.721 and 0.945 mm.d-1, respectively. In addition, saffron water requirement determined with FAO56-PM was equal to 2350 m3.ha-1.y-1. Hargreaves equation, with differences about -161.23 m3.ha-1.y-1 compared with FAO56-PM, was determined as a good equation. Measured irrigation water by saffron farmers was about 1184.17 m3.ha-1.y-1. According to the results it found that deficit irrigation was applied by saffron farmers. Saffron grow in arid regions, however, drought stress have a negative effect on its plant development and yield (Khorramdel et al., 2014; Khashei Siuki et al., 2015). Despite the lack of doing no test about the effect of deficit irrigation on saffron yield during 2014 in this region, it seems that deficit irrigation may be a reason of low saffron yield (SYASK, 2010) compared to average saffron yield in this region.
Results of gamma test revealed that temperature is a key parameter to develop a single model. In addition, results showed that the single parameter model based on temperature had better accuracy compared to above-twelve mentioned equation.
Hargreaves equation was only used the temperature (T) and extraterrestrial radiation (Ra) for estimating evapotranspiration. In addition, single parameter equation based on the mean temperature was only used one parameter (T) to estimate the evapotranspiration. Since both of the equations had an acceptable accuracy to use in this region, it is recommended to use them to determine saffron irrigation demand. It is due to lack of meteorological equipment for estimating all parameters in the plain.
The authors would like to thank the anonymous reviewers whose valuable comments have helped us to clarify some parts of this paper.