A Decision Support System for Rice Cultivation on Acid Sulfate Soils

Acid sulfate soils pose chemical, biological and physical problems for rice cultivation. The proposed amelioration for the study includes: 1) correct water management; 2) applying lime at appropriate rate and time; and 3) adequate fertilizer application. Those procedures were organized into an integrated decision support system (DSS), which used analytical methods and models. Main component for the model was production function as a response of those amelioration processes. The model used that function to simulate yield and in advance step, doing micro economic analysis by calculating profit in order to find the maximum one. In delivering production function, glasshouse experiment has been conducted using soil from acid sulfate soil area in Jelawat Rusa Irrigation Scheme, under Project Kemasin-Semerak (PERKASA), Kelantan. The result of glasshouse experiment showed that yield was significantly affected by the combination of lime and fertilizer (P<0.02). Water management and any of its combination did not affect yield. The highest yield was found on combination of GML at rate 4 t/ha using maximum fertilizer rate. Ameliorative assessment of lime and fertilizer have improved yield shown by increased soil quality and a better plant performance. Liming has increased soil pH, exchangeable Ca and exchangeable Mg, and these have decreased toxicity caused by of Al and Fe. Liming and water management has also improved rice yield by increasing grain weight and decreasing empty spikelet number. Field experiments, which include experimental plot and demonstration plot, have been conducted in that area for 2 seasons with the purpose of validating glasshouse experiment. Statistical analysis showed that yield of plot experiment on the first season was not significantly different (P>0.12). It was primarily because the first season of field experiment was disturbed by flood due to high rainfall and poor drainage system at the location. It may be also because lime is still not stably interacted with the soils to increase soil pH. Crop cutting test (CCT) in the second season of the field experiment showed a significant effect of lime treatment on the yield (P<0.07). The highest yield of 7.52 t/ha was found on treatment 6 using application of GML at 4 t/ha in combination with organic-based fertilizer (JITU). Treatment 5 using application of GML at 8 t/ha resulted on yield of 7.22 t/ha, which was not different from treatment 6 and this become the second highest value. Using field experiment, field adjusting factor (FAF) has been developed. It showed percentage of achieved yield on field trial from the potential yield (glasshouse experiment). The value of FAF was 0.40, which means only 40% of potential yield will be gained on the field. The 60% loss may be because of technical problems,inefficient harvester machines, incorrect fertilizer applications, pest and disease damages. Response curve as production function was formed using TableCurve 3D v4.0. From surface fitting process, an equation has been found (P<0.01). It consists of 4 parameters which were also significant at the same α of 5.00%. The value of r2 of the equation was 0.71, while the curve on initial data and the equation is quite similar. A model, RiCASS, used predicted yield from the equation to calculate the cost of inputs both fixed and variable, and to calculate the profit. It ran simulation under various inputs to find the optimum level that result in maximum profit under 4 different scenarios. Simulation showed that for the first general scenario, maximum profit margin was found to be $ 2,847/ha from predicted yield at 6.15 t/ha. It used lime at the rate of 6.50 t/ha and fertilizer index of 10. The second scenario in which total cost was limited to RM 1,500, maximum profit was found to be RM 2,024. Based on farmer’s cost, the maximum profit was found to be RM 3,624, while under limitation of cost below RM 1,000, the profit was found at RM 3,263. Validation carried out to evaluate the equation which consisted of paired comparison t-test and mean estimation error. Result showed that the equation was able to represent actual yield shown by insignificant difference with field experiment yield (average P>0.20) and small estimation error (2%).

Preview PDF FP_2005_39.pdf Download (145kB)

Actions (login required)

View Item