CHARACTERIZATION OF WATER AVAILABILITY IN LANDSCAPE USING SEVERAL SOURCES OF DATA.

Authors: Timlin, Dennis; Pachepsky, Yakov; Walthall, Charles

Submitted to: Remote Sensing and Modeling Applications for Natural Resource Management
Publication Type: Proceedings
Publication Acceptance Date: 3/3/2002
Publication Date: 3/23/2002
Citation: Timlin, D.J., Pachepsky, Y.A., Walthall, C.L. 2002. Characterization of water availability in landscape using several sources of data. Remote Sensing and Modeling Applications for Natural Resource Management.

Interpretive Summary:

Technical Abstract: The use of remote sensing and yield monitoring technologies have greatly added to available data on soil and crop properties, and conditions. These data contain much potentially useful information regarding the distribution of important soil properties such as water availability, and can be used to characterize soil management zones. Because these data provide a more or less continuous distribution of measurements, they may be useful to help interpolate values of soil properties that must be manually collected. These properties include soil texture, and soil water holding capacity among others. The data for this study include corn yields from a combine monitor and manually harvested yields on four transects. Sixteen spectral band AISA images were recorded two times during the growing season. In order to obtain a landscape level estimate of soil water holding capacity (WHC), we back calculated soil available water holding capacity from measured yields using a simple water budget model. In order to account for spatial variability, we used conditional simulations. Yields in both years were closely related to landscape position and soil water holding capacity. The estimated soil water holding capacities were similar to measured values and corresponded across years. Soil water holding capacity was also related to surface curvature and slope. We further discuss how these estimated water holding capacities can be tied together with measured values and soil topographic variables such as slope and curvature. We further discuss statistical tools such as Spatial Autoregression that correct for spatial autocorrelation and can be more effective in accounting for local information when quantifying relationships among spatial variables.