Modelling the optimal size of silt pits for soil water conservation

Malaysia experiences high total rainfall intensity, which increases soil erosion on steep slopes and causes a reduction in soil fertility, pollution of fresh and groundwater, and the degradation of adjacent lands. Although Malaysia has high rainfall, oil palm may still experience water stress due to high rainfall intensity which results in fast downslope water movement and little time for water to infiltrate into the soil. One of the most effective measures of soil and water conservation in Malaysia is the use of silt pits. The function of a silt pit is to control the runoff, trap and settle down the sediments, increase soil moisture or recharge the groundwater, reduce the effect of slope length and further reduce soil erosion and fertiliser losses. However, what is the optimal size and dimensions of a silt pit to enable the water to reach the farthest roots and empty slowly to release the water over the most extended period? What is the effect of the slope, runoff (volume of water), the volume of the pit, and type of soil of the spatial silt pit size? The study aimed to use the HYDRUS 2D/3D models and to formulate the simulation results as equations to select the optimal size and dimensions of a silt pit depending on the rainfall and soil properties. The treatments used in this study included the following four factors: seven type of soils (sand, sandy loam, loam, silt, sandy clay, silty clay, clay), six surface slopes (0˚, 5˚, 10˚, 15˚, 20˚, and 25˚), three silt pits sizes (3, 4, and 5 m3), each size having three depth levels (50, 75, and 100 cm), and several levels to cater for the volume of water available in the silt pit. Three stages were adopted in this study. The first stage utilised the software HYDRUS 2D/3D models to simulate the soil water content, wetting front, and time-to-empty from a silt pit of various sizes on different soils and slopes. The second stage distinguished the trend and determined the best fit by using statistical methods (Multiple linear regression (MLR) and Artificial neural network (ANN)) to estimate the optimal silt pit size. The last stage applied the fitted model to find the optimum silt pits in some areas of Peninsula Malaysia. From the simulation results, all parameters (distance of wetting front, water content, and time-to-empty) were affected by nearly all the factors (water head in the pits (H), pit width (W), amount of water applied (Vw), pit volume (Vp), and surface slope (Slope)). For instance, increasing the slope will slightly increase the wetting front distance (from 130.54 to 136.45 cm) and soil water content (from 0.374 to 0.375 m3/ m3) with downslope, but decrease significantly the upslope wetting front and soil water content (130.53 to 101.26 cm and 0.373 to 0.333 m3/m3, respectively). However, there was no effect on the time-to-empty. The reduction of unfavourable results of the increasing slope was compensated by manipulating the values of H and Vp. The MLR models did not perform adequately especially for timeto- empty (Mean Squared Error (MSE) = 85.83; R2 = 0.632) compared with the ANN models (MSE = 10.33; R2 = 0.977), mostly due to the non-linear relations between the factors. The results demonstrated that despite requiring the same input data, the ANN models could favourably be used for all parameter predictions. However, processbased numerical models are undoubtedly a better choice for predicting the results with lower uncertainties when the required data are available. The fitted problem was then used to select the optimum sizes of the silt pit in Peninsula Malaysia, based on the soil texture and rainfall intensity. The results show that some types of the soils (sand, loam sand, and sandy loam) which have high values of hydraulic conductivity make the rainwater infiltrate into the soil. So, for these land, there is no need to construct the silt pits. While in the case of soils (sandy clay loam, clay loam, silt clay loam, sandy clay, silty clay, and clay) which have low values of hydraulic conductivity, a large volume of runoff water will be caught in those land compared to the sizes of silt pit used in the experiment. Therefore, length of the pit during construction must be extended to avoid flooding of water.

Text FP 2019 18 -ir.pdf Download (1MB)

Actions (login required)

View Item