Soil moisture estimation from soil spectral characteristics in a precision farming environment

Soil moisture is a critical component of numerous agricultural systems, including crop growth, and precision farming. Consequently, understanding how soil moisture varies across the field is important for planning and management of those systems. However, accurate, detailed maps of soil moisture are difficult to obtain from in situ measurements. This work addresses the variability of soil moisture, the impact of moisture variability on crop yield, and the potential of optical remote sensing for soil moisture monitoring.; Two fields were sampled for surface and near-surface soil moisture and reflectance throughout the growing season. These data were studied to estimate variability in soil moisture, to determine the relationship between moisture and corn yield, and to quantify the relationship between surface reflectance and surface moisture.; Analysis of the spatial structure of soil moisture revealed that the geospatial characteristics of the moisture patterns were similar from one date to another. While there was no strong correlation between moisture patterns and topographic indices, analysis indicated that plan curvature may be important for understanding surface moisture spatial variation. Temporal stability of moisture patterns was studied to identify optimal sampling points for field-average soil moisture. These points tended to be in areas that were neutral in aspect and plan curvature compared to the field average.; Investigation of the relationship between moisture and corn yield indicated that there may be useful information in early (roughly from germination to within two weeks of planting) surface soil moisture measurements for interpreting variation in crop yield.; Spectral reflectance data were analyzed in conjunction with surface moisture data to determine the nature of the relationship between the two. For one of the fields, the strongest relationship between reflectance and surface moisture was between 550 and 620 nm. For the other, the strongest relationship was around 945 nm. Finally, a combination of spectral data and limited moisture data was used to create moisture maps. Use of a cokriging technique generated more detailed soil moisture maps than the limited data alone. This method shows potential for development as part of a data fusion technique to generate moisture maps from a minimum of samples.