The concentration of soil phosphorus and pH correlated significantly with Landsat TM and Compact Airborne Spectrographic Imager (CASI) data, terrain position and aspect. Band ratios of the infrared and visible wavelengths are particularly significant for phosphorus. Exchangeable soil cations (that is calcium and sodium), as well as electrical conductivity, correlated less strongly with CASI and TM, but the correlations were statistically significant. Magnesium, potassium, and pH were significantly correlated with TM data, but not with the CASI data. pH and phosphorus were significantly correlated with terrain position. Three factors may individually, or in combination, be causing these results: illumination differences caused by terrain, direct reflectance of soil, or differences in vegetation (due to the influence of the soil on vegetation). A review of the literature appears to support the latter as the most likely causal factor. The results are important for establishing whether forest soils may be mapped using only remotely-sensed data, digital terrain data, or a combination.
Skidmore, A. K., Varekamp, C., Wilson, L., Knowles, E., & Delaney, J. (1997). Remote sensing of soils in a eucalypt forest environment. International Journal of Remote Sensing, 18(1), 39-56. https://doi.org/10.1080/014311697219268