Effect of preferential flow dimension on soil moisture distribution under deficit irrigation

Preferential flow is formed through large pores in soil. The flow transmission in the large pores is faster than the regular parts of soil. Therefore, the effect of preferential flows, which are created by factors such as cracks, should be assessed on the distribution of moisture profile. The analytic, empirical and quantitative models can be used to assess the various modeling parameters and water distribution in soil in various environmental conditions, in order to save money and time. Hydrus-2D model is one of the potent models for stimulating the movements of water and salts in soil, based on the Richards' equation numerical solution. Nowadays, the use of deficit irrigation methods is increasing due to the water resources shortage. In addition, the preferential flows creating in farms have an important role in water and contaminant transport. Therefore, the purpose of this study is to investigate the effect of pore geometry (with different width and depth) on soil moisture distribution in soil profiles, at different levels of irrigation. To investigate the effect of preferential flow on soil moisture profile, sandy and sandy loam texture (in the preferred area) was used. In this study, column with a diameter of 160 mm and a length of 350 mm were applied. The soil columns were filled in layers. The method of Wang et al. (2013) was employed to establish the preferential flow. Due to this method, the columns were placed in the water container from the floor and a pipe with a diameter of 16 mm and a height of 350 mm was placed in the center of the column. The soil was poured in layers around a 16 mm pipe and after each layer a little water was sprinkled on the soil surface. After filling the soil column, the 16 mm tube was gently pulled out of the larger column, and then the 16 mm tube was filled with find grain sand (treatment T1). Irrigation levels of 120, 100, 80 and 60% were selected (T120, T100, T80 and T60, respectively). The initial soil moisture of the treatments was the lower limit of readily available water. The T80 and T60 treatments had 20 and 40% less irrigation volume and T120 treatment had 20% more than the T100 treatment, respectively. The RETC software was used to estimate the Van-Genuchten coefficients. To calibrate the Van-Genuchten equation coefficients, soil moisture profiles were used in Hydrus. The model initial conditions were the initial soil water content at different depths. To compare the amount of simulation and observational values, the root mean square error (RMSE), NRMSE and correlation coefficient (r) were used. The Hydrus-2D software was employed to investigate the soil moisture distribution under different geometric shapes of the preferential flow (PF). The PF’s geometries created in V-shaped pores from the soil surface to deep where the top-width was 0.5, 1 and 1.5 cm (B0.5, B1, B1.5) and the total depth was 10, 20 and 30 cm (H10, H20, H30). In this study, the soil moisture was simulated with the Van-Genuchten coefficients. The results showed that the water content with the coefficients was estimated with a good accuracy. The minimum and maximum correlation coefficients were 0.83 and 0.95, respectively. The average percentage of moisture error at depths of 0 and 20 cm in the T60 and T120 was about 2% and the moisture content of the model was slightly different from the observed values. According to the research results, by increasing the depth of cracks, the amount of soil moisture in the soil surface decreases, compared with the non-cracked soils. Thus, the water transfers to the lower depths and consequently increases the soil moisture. In the deficit irrigation condition, the presence of preferential flow causes a more uniform soil moisture distribution in the soil profiles; But, the more irrigation level/ time increases the more deep percolation losses that finally decreases the irrigation efficiency. Some research result showed that PF in most farms has caused increasing deep percolation losses about 71%. Playan and Matthews (2004) evaluated irrigation efficiency by changing irrigation management. Their results showed that with decreasing irrigation time, irrigation water use efficiency increased from 44% to 71%. Also in their study, the soil moisture content at the soil surface in T60, T80, T100 and T120 treatments in the presence of the PF with depth of 30 cm (H30) and a width of 1.5 cm (B1.5), decreased about 11, 9.3, 9.2 and 9.1%, respectively, where the moisture at the depth of 20 cm increased about 24.1, 25, 27 and 21.1%, respectively. In our research, the results for different PF forms in T120 and T80 showed that the presence of a crack with low depth (h=10 cm), for all width scenario (B=0.5, 1 and 1.5 cm), had not a significant effect on the distribution of soil moisture profiles. For the PF with H=20 and H=30 cm, the effect of the crack width (B) on the soil moisture distribution in T80 became more noticeable at a distance of 3.5 cm from the center of the column. The amount of soil moisture in the soil profile did not differ much for the different crack’s width in T120 treatment, but with increasing the depth of the cracks (H20, H30), the moisture distribution at different depths became more uniform. The results of this study showed that with increasing the depth and width of the preferential flow, the moisture distribution is more uniform and the higher the irrigation level, the more uniform distribution is. This process decreases surface soil moisture and causes water flow to the lower depth in comparison with the non- preferential flow condition. For example, in the soil surface in treatments with irrigation levels of 60, 80, 100 and 120%, the amount of soil moisture at the soil surface, under preferential flow conditions with a depth of 30 cm and a width of 1.5 cm, was decreased around 11, 9.3, 9.2 and 9.1 %, respectively and it caused increase in the soil moisture at depth of 20 cm with 24.1, 25, 27 and 21.1 percent, respectively.