<![CDATA[Managing water in the paddy fields, water balance analysis is usually performed to determine the effectiveness of water used. However, with limited advanced instrumentation, time and cost, some water balance components are not measured. This study proposed a novel method, Linear Programming (LP) model, to estimate non-measurable water balance components. The aims of this study were to develop LP model in estimating non-measurable water balance components such as irrigation, runoff and percolation with measured soil moisture data in non-flooded irrigation, and then to evaluate performance of the model by comparing measured and estimated soil moisture. This study was carried out based on two season field experiments of non-flooded irrigation with System of Rice Intensification (SRI) in NOSC, Sukabumi West Java during 20 August to 15 December 2011 (first season) and 22 March to 5 July 2012 (second season). The developed LP model has the objective function by minimizing the differences between total measured and estimated soil moisture. In addition, the LP model has also constraint function and initial condition that were formulated based on actual field conditions. The results showed the LP model estimated non-measurable water balance accurately with the indicators of R2 > 0.85 (p value < 0.01) and percentage error less than 8%. Based on the estimation model results, irrigation only contributed 34-38% of inflow, while crop evapotranspiration and percolation contributed of 40-44% and 11-15%, respectively. Precipitation and runoff were the most contributors of inflow and outflow from the fields. By the current model, water use efficiency and water productivity can be determined with estimated irrigation.]]>
Copyrights © 2019
