Optimizing the Amount of Irrigation Water and Nitrogen Fertilizer in Ratoon Rice using Response Surface Methodology

Rice (Oryza sativa L.) is one of the most popular cereals in the world and is known as the second most consumed grain in Iran after wheat. For this reason, it is very important to pay attention to the quantity and quality of this agricultural crop. Cultivation of ratoon rice, which has become common in some areas in the north of Iran, enables the re-production of rice in the next cropping season. The lack of water resources in recent years in Iran and the prediction of the severity of the shortage of available water in the coming years have caused the use of methods to reduce water consumption in the agricultural sector, as the largest water consumer in the country, to be considered (Ahmadee et al., 2021). Rice plant has a large share in water consumption per unit area among other agricultural products. For this reason, some methods of reducing water consumption in the cultivation of this product have been suggested. According to the existing restrictions on the use of irrigation water and nitrogen fertilizer, it is necessary to determine the optimal amounts of these factors in the field. But this work requires many field experiments, which require a lot of time and money. To solve this problem, the use of simulation and optimization models has been suggested (Ebrahimipak et al., 2019). Therefore, the aim of this research is to optimize the two factors of irrigation water and nitrogen fertilizer to achieve the most appropriate amount of production and improve the quality of rice using the Response Surface Methodology (RSM).

Considering the effects of two abavementioned factors, in this research the optimization of the amount of irrigation water and nitrogen fertilizer on the quantitative and qualitative characteristics of ratoon rice was applied using the response surface method. The study site was the research farm of the Rice Research Institute located at latitude 37º 16’ N and longitude 49º 63’ E in Rasht, Iran. The studied factors included the amount of irrigation water (with upper and lower levels of 0 and 5 mm of cracks appearance in the soil, respectively) and nitrogen fertilizer (with upper and lower levels of 90 and 0 kg per hectare of pure nitrogen, respectively). As defined by equation (1), the RSM refers to a multivariate function:

(1)



Where:

Y is the response variable and x is the independent variable.

To compare the obtained model results with observed values, from the Root Mean Square Error (RMSE), Normalized Root Mean Square Error (NRMSE), Mean Bias Error (MBE), Efficiency Factor (EF), agreement index (d) and the Coefficient of Determination (R2) was used (Ahmadi et al., 2016).

The results showed that the regression model for predicting the plant height, biomass, gel consistency and yield had an underestimation error (MBE < 0) and for other paramets had an an overestimate error (MBE<0). The value of the NRMSE was less than 0.1, so the quadratic regression model had excellent accuracy for all parameters. The two statistics EF and d, which show the efficiency of the regression model, had acceptable values (d>0.90 and EF>0). After optimization, the yield of seed and biomass was 10.5% and 0.5%, respectively, higher than the average values of these parameters in field conditions. In optimal conditions, the number of tillers, panicle length and harvest index were 1.9%, 1.4%, and 3.6% higher than the observed values, respectively, and plant height and weight of 1000 seeds were almost equal in both optimal and observational conditions. The changes in the quality indicators of gelatinization tempreture, gel consistency, amylase content percentage and grain elongation in optimal conditions compared to the observed values were 1.8 (less), 0.4 (more), zero and 6.2 (more) percent, respectively. The amount of irrigation water and nitrogen fertilizer in optimal conditions was 3.5 mm of cracking ins the soil and 68 kg.ha-1, respectively.

In this research, the effect of two factors, irrigation water and nitrogen fertilizer, was optimized using the response surface method. For this purpose, the minimum and maximum values of these factors were considered with codes of -1 and +1. The results showed that, except yield and biomass, other traits did not have uniform and regular changes with the increase of two factors, irrigation water and nitrogen fertilizer. For this reason, the overlap map of these factors was determined and it was observed that the effect of nitrogen fertilizer on most parameters was greater than that of irrigation water. The optimal range of all factors was in the values of -0.3 to +0.5 nitrogen fertilizer and -1 to +1 irrigation water and +0.5 to +1 nitrogen fertilizer and -1 to +0.7 irrigation water. Of course, a part of the overlapping area of these two ranges also lacked the optimal value for the parameters, which was considered by the surface-response method in the optimization. After optimization, it was observed that the two parameters of gelatinization degree and gel consistency were lower than the average observed values. The parameters of amylase percentage, 1000 seed weight and plant height did not change compared to the field conditions and the value of other parameters increased. To achieve these results, it is necessary to consider the amount of irrigation water in cracks of 3.5 mm and the amount of nitrogen fertilizer as 68 kg/ha.