Geonic electromagnetic induction equipment (EM-38) is now commonly used for mapping soil salinity of agricultural fields or landscape due to its rapid and cost-effective in situ measurements. However, its use in quantitative diagnosis of surface (0-15 cm or 0-30 cm) soil salinity is limited notwithstanding its importance to predict establishment and agronomic growth of most agricultural crops. This investigation was thus aimed at evaluating the use of EM-38 to measure surface and sub-surface salinity both in space and time, and relationship of EM-38 readings to observations on growth and establishment of selected perennial pasture species on two different soils underlain with shallow and saline water table. Notwithstanding significant relationships of EC_1:5 with EC_e (r² = 0.98) and EC_P (r² = 0.95), the relationships between EC_e of surface (0-15 cm or 0-30 cm) soil layers and EM-38 readings were not significant. In addition, EC_e predicted from EM-38 calibrations were highly variable and less accurate both in space and time. Hence, the predictability of EC_e from EM-38 readings (EC_a) values seems less promising as the latter is very sensitive to changes in soil moisture profiles both in space and time. In addition, the relationships of EC_a with observations on establishment and growth of perennial pasture species were of low magnitude. However, EC_a values, showed relatively better relationship with EC_e or EC_1:5 of sub-soil layers especially 60-90 cm and 90-120 cm layers. Field soil moisture, clay content of soils and their saturation percentage as measured in laboratory were found to influence EC_a readings significantly. Intensity and magnitude of moisture fluctuations are generally greater in the surface than sub-soil layers under field situations and may probably be responsible for observed poor relationships. These observations construe that although EM-38 is a useful instrument for mapping soil salinity of a landscape or paddocks into different soil salinity categories, its potential to test surface soil salinity for the desired agronomic accuracy is limited.