A transfer function to predict soil surface reflectance from laboratory soil spectral libraries

Abstract Spectral-based models extracted from laboratory reflectance in the 400–2500 nm spectral range to predict soil attributes may not be applicable to soil spectra acquired in the field. This is because laboratory sampling procedures disturb the natural soil surface's status. We investigated this issue by using the soil surface-dependent property of water-infiltration rate (WIR). We created a dataset with 114 samples collected from six fields with varying textures located in three different Mediterranean countries (Israel, Greece, Italy). Using the field and laboratory spectral datasets, we demonstrated that WIR is better predicted by field vs. laboratory measurements (R2 = 0.92 and 0.56, respectively). We also developed a transfer function (TF) to predict the field spectral measurements from the laboratory spectra. Use of the TF-processed dataset considerably improved the WIR prediction using laboratory information (from R2 = 0.56 to 0.76). It was concluded that soil surface reflectance values can be estimated based on laboratory spectra using a TF. The generated TF enables exploiting soil spectral libraries for remote-sensing views and for assessing surface-related soil properties.