Modeling solute transport for improved fertiliser use in rice production system

Quantification of water and nutrients and their interactions of a paddy field environment are crucial for the improved utilization of fertilizers for the sustainable rice production. Solutes runoff and leaching are two direct pathways of nutrient pollution from paddy fields to water resources systems. Due to the dynamic nature of paddy fields, solute transport and transformation process are complex and difficult to understand. The past investigation on the water balance components using multifarious parameters did not reflect the true condition of paddy field environments. Quantification of agrochemical losses from paddy fields are generally related to the amount of inflow and outflow water in the paddy field environment which yet to be measured accurately. In order to overcome the challenges, the modern monitoring devices together with sensors and data logging system were installed for intensive field observations in a paddy and developed empirical models to quantify the solute losses through the surface and sub-surface water leaving from a paddy field system for the better utilization of fertilizers (N, P, K). The intensive field investigation was carried out in a paddy plot at Sawah Sempadan compartment of the Tanjung Karang Rice Irrigation Scheme (TAKRIS) for two rice growing seasons (January-April and July-October) in 2017. Firstly, the water balance components in a paddy plot was analysed from the intensive field observations with 1-10 minutes interval of a paddy field. Water balance analysis results revealed that irrigation water accounted 59.6 % of the total water input (irrigation and rainfall) during the January to April (Off Season). However, about 76.2% of total water input during the July to October (Main season). The amount of rainfall contributed to 23.8% and 40.4% of total water input in the main season and off-season, respectively. Drainage flow accounted 37.3% and 43.7% of the total water input during off season and main season, respectively. The daily evapotranspiration accounted 41.7% and 61% of total water input during offseason and main season, respectively. Observed seepage and percolation of 17.1% to 19.2% of total water input accounted during both seasons respectively. The yield of the experimental plot was obtained 2.5 t/ha and 2.7 t/ha for the off season and main season, respectively. Finally, the water productivity index was analyzed 0.72 kg m-3 during off-season and 0.78 kg m-3 during main season, respectively. Based on solute transport analysis, the accumulated total nitrogen (T-N) of 50.3% to 49.7% estimated in the top 40 cm soil layer while 49.7 % to 53 % T-N as leachate obtained below 40 cm soil layer (40-100 cm) during off season and main season, respectively. About 85% of N leaching losses were in the form of NO3-, however there was still a large quantity of NO3- remained below root zone that contributes the groundwater. The total leaching loss of T-N was 34.9 and 27.9 kg/ha during off and main seasons respectively. The estimated loss of total phosphorous during the two rice growing seasons were 3 and 1.7 kg/ha, respectively. The total amount of T-N, TP and K loss through drainage were 27.7 and 18.5, 2.2 and 1.1, 5.9 and 3.5 kg/ha during off-season and main season, respectively. The Hydrus-1D was applied to simulate water and solute movement under different soil depths of 20, 40, 60, 80 and 100 cm in real paddy environment experiments. The simulated and observed water flow and nutrient leaching were in good agreement (R2= 0.98, RMSE = 0.24). Hydrus-1D simulation showed the similar patterns of the water and solute movement under different soil depths during the study period. The observed and simulated N, P and K concentration in paddy was high due to fertilization and other climatic factors. Therefore, reduction of excessive fertilizer rate especially during early rice growing period and adaptation of water saving techniques can reduce the pollutant risks from paddy soil. Regression analyses were performed for the development of the improved fertilizer use models. Multiple linear regression analysis was performed to know the relationships between EC versus solutes (N, P and K) during the both seasons. The polynomial regression analysis was fitted to evaluate whether EC changes has an impact on N, P and K concentrations in paddy field. Finally, empirical models were established to estimate the concentrations of N, P and K using two rice growing season data. MS Excel solver program were used to develop the empirical models. The results obtained a strong agreement between observed and predicted N, P, and K with the determination coefficients (R2) of 0.91 and 0.95 during the both seasons. Therefore, the models could be useful in predicting the solute concentration changes within root zone and below root zone during entire rice growing season for better utilization of fertilizers