Diurnal temperature profile impacts on estimating effective soil temperature at L-Band

Near-surface soil moisture is an important parameter in hydrological, meteorological and agricultural applications. Passive microwave observations, both airborne and satellite, can be used to infer near-surface soil moisture. The brightness temperature of the thermal radiation that crosses the soil–air interface is the product of the soil emissivity and the effective temperature. At L-band, the emissivity is sensitive to the moisture content in the top few to several centimetres, while the effective temperature is the weighted average of the emission from all locations within the soil. The effective temperature is used to normalise the observed brightness temperature so that the near-surface soil moisture can be inferred. The effective temperature is a function of the soil temperature and moisture profiles, full information of which is not available for remote sensing applications. The so-called ‘C-parameterisation’ uses two temperatures, one at or near the soil surface and the other at depth, to estimate the effective temperature. Variations in the shape of the temperature profile are not taken into account, so diurnal temperature variations might impact on the effectiveness of this method. Moreover, the best results have been obtained when the temperature just below the surface (e.g., 5 cm) is used. Operational applications thus rely on simulations to provide the necessary sub-surface temperature information. The primary aim of the work presented here is to investigate the influence of diurnal temperature variations on the effectiveness of the C-parameterisation. A new method (the ‘ratio model’) is then introduced by which TEFF may be estimated solely from thermal infrared measurements of the soil skin temperature. The Simultaneous Heat and Water model of Flerchinger et al. (1998) was used to simulate temperature and moisture profiles under bare soil. The daytime profiles were used, with the aim of developing a model for use primarily with airborne observations acquired throughout the day. The C-parameterisation performed best when the temperature at a depth of 5 cm was used, with an RMS error of 0.29 K. When the temperature at the soil surface was used instead, the RMS error increased to 1.7 K and the residuals exhibited an increase from −2 to 3 K between 10:00 and 18:00 hours. This trend is owing to the phase difference between the effective temperature and soil skin temperature. The skin temperature peaks at ∼13:00 hours, while the effective temperature peaks ∼3 hours later. The ratio between the effective temperature and soil skin temperature was found to vary smoothly, decreasing to a minimum near the middle of the day. By fitting a 3-parameter model to the data, the effective temperature could be estimated as a function of the skin temperature and hour of day. With an RMS error of 0.95 K, this model outperformed the C-parameterisation when using the soil skin temperature. The ratio model is primarily expected to be beneficial for airborne operations. A modified form may be required to account for variations in latitude, season, soil and vegetation properties, and meteorological conditions.