Estimation of infiltration rate from readily available soil properties (RASPs) in fallow cultivated land

The information about water transmission rate into the ground is vital concerning the productivity of water system and seepage, advancing the accessibility of water for the plants, enhancing the yield of harvests, limiting degradation of soil and wastage of the water. Infiltration rate can successfully be measured using double ring infiltrometer. However, measurement of infiltration in the field is labour and time consuming and difficult especially in mountainous sites. As an alternative, RASP-based infiltration models can be developed. The study was carried out in the field near the NERIST campus (Nirjuli Complex), Arunachal Pradesh, India. Twenty sites were identified at the grid of 10 m interval, and the field measurement of infiltration was performed. The soil was analysed for properties, namely, soil texture, bulk density (BD), particle density (PD), moisture content (MC), and organic carbon content (OC) for each site. The basic infiltration varied from 1 to 4.84 cm/h. The Scatter plot between RASPs and infiltration rate revealed that there is a positive correlation with OC, PD, and sand, and a negative correlation with BD, MC, silt, and clay. The partial least square regression (PLSR) analysis was carried out to develop predictive models for five different groups of inputs of soil properties. The influential variable projection (VIPs) analysis revealed sand as a highly influential factor, while silt as a reluctant predictor of infiltration characteristics of the study site. It was found that to predict the soil infiltration rate based on RASPs with seven independent variables (Eq. 13) with coefficient of determination ( R 2 ) 0.92, root mean square error (RMSE) 0.378 cm/h, mean absolute error (MAE) 0.143 cm/h, and standard error (SD) 0.398 cm/h is strongly recommended for the prediction of infiltration characteristics.