Estimation of deep percolation in agricultural soils utilizing a weighing lysimeter and soil moisture sensors

Climate change has altered precipitation events over past decades which impacted groundwater recharge in Michigan due to erratic rainfall, sandy soils, runoff fluctuations, seasonal and frozen ground in winter. Deep percolation (DP), critical factor for groundwater balance and hydrological processes, plays a significant role in groundwater sustainability and agricultural water management. Estimation of DP is challenging task and relies on different methods including soil moisture sensors that provides real-time soil water content measurements and weighing lysimeters which offers precise estimation under laboratory conditions. Soil moisture sensors were utilized for DP estimation in common Michigan soils, including Oakville Fine sand, Spinks Sandy loam, Riddles-Hillsdale loamy sand, and Kilmanagh loam soils. The CS655 soil sensors were installed at 15, 30, and 45 cm depths to estimate DP under different irrigation depths (mm) through change in soil moisture contents with time. A multilinear regression analysis was used to model the relationship between irrigation depths (mm), changes in soil water contents (cm³/cm³) and time until stability (min) to predict DP in different soils. Sensors estimated DP accurately with slight underestimation, whereas regression analysis indicated accurate and precise estimation of DP. Across irrigation depths, maximum accuracy for estimating DP through moisture sensors was observed for Oakville Fine sand (99.6 %). Lower accuracy was observed with Spinks Loamy sand (63 %) and Riddles Hillsdale Sandy loam (49.6 %) with the lowest accuracy in Kilmanagh loam soil (29.6 %) due to fine texture. The accuracy of estimating DP was improved with increasing irrigation depth for different soils.