Estimation of ground water content using water balance and regression approach

Abstract

Near-surface water content (NSWC) is a variable in space and time that significantly affect physical processes. The complex interactions between the atmosphere and the surface as well as between the surface and underground, which results in exchange of energy and moisture at quantities varying with time, make it difficult to estimate the NSWC. The traditional techniques of measuring NSWC give point data that does not represent the spatial profile. In this work, attempt was made to estimate the NSWC of an area in Abeokuta using water balance and regression approach. The study was carried out at Akole, Oke-Ata, Abeokuta, Nigeria. NSWC were measured hourly at depths 2 cm, 50 cm 100 cm, 150 cm, and 200 cm between 1 January 2014 and 31 December 2014 using Decagon EM50 data logger. Daily air temperature, solar radiation, relative humidity, and precipitation data for the location were obtained from the Nigerian Metrological Agency. Water balance approach combines empirically only the air temperature and precipitation readings to estimate NSWC. Regression approach employed the knowledge of water content at the surface or near the surface layer to estimate NSWC at other depths. Models were tested by Nash–Sutcliffe (NS) efficiency and coefficient of determination (R2), estimated NSWC using water balance approach was in good agreement with measured data at depths of 2, 50, 100, and 200 cm with R2 of 0.9091, 0.9166, 0.8540, and 0.7139 respectively. NSWC estimated with regression approach was reasonably close to observed values at depths of 50, 100, 150, and 200 cm with R2 of 0.9881, 0.8567, 0.6418, and 0.6278 while the NS were 0.6394, − 0.8132, − 3.5785, and 0.3642 respectively. The NS value of 0.6394 shows that the model performed creditably well at the depth of 50 cm only. The high and consistently positive values of NS efficiency revealed that water balance approach was better than the regression approach in estimating NSWC.