Precision Agriculture, 1, 249᎐276 1999
ᮊ 2000 Kluwer Academic Publishers. Manufactured in The Netherlands.
Empirical Modeling of Relationships Between
Sorghum Yield and Soil Properties
T. M. SHATAR
AND A. B. MCBRATNEY email@example.com
Australian Centre for Precision Agriculture, Uni
ersity of Sydney, Australia
Abstract. A crucial part of any site-specific management is the identification of causes of yield
variability and assessment of crop requirements. Therefore, relationships between yield and soil
properties must be identified. In this study, relationships between sorghum yield and soil properties on a
Vertosols within a field located in Moree, in northern NSW, Australia, were examined. Measured soil
properties included pH; available phosphorus; percent clay, silt and sand; gravimetric moisture content
of air-dry soil and at matric potentials corresponding to y1 500 kPa and y33 kPa; percent organic
carbon; CEC and exchangeable calcium, magnesium, sodium and potassium and copper, zinc, man-
ganese and iron contents. The exchangeable sodium percentage ESP and the CarMg ratio were
calculated. We used a number of empirical methods and found that neural networks, projection pursuit
regression, generalized additive models and regression trees are good techniques for modeling yield
response. However, further comparison of these techniques is needed. By modeling yield response to
individual soil properties and using kriging to map yields predicted from these models, it was possible to
identify which soil properties limited production in different areas of the field.
Keywords: empirical modeling, yield response, field-scale, yield variation
Precision Agriculture differs from traditional agricultural management by attempt-
ing to identify and respond to spatial variability in soil and crop requirements at
the within-field scale. Modeling the relationships between soil properties and yield
is necessary in order to identify the causes of yield variation and to determine what
changes to the soil environment management should aim to make.
Soil properties limiting yield may be manageable or unmanageable. Manageable
properties can be altered to prevent them from limiting yield production. Unman-
ageable factors cannot be practically altered and will determine the maximum
potential yield attainable at a site. Since crop requirements vary with yield, the
specific aims of management will change, depending on potential yields, e.g.,
fertilizer requirements may be reduced in lower-yielding areas.
More specifically then, models of yield response are needed to predict yields
yield potentials and to determine crop requirements to reach these yields.
Identifying local relationships is necessary in order to identify factors limiting
production and to more accurately assess crop requirements. For example, tradi-
* Corresponding author: Tamara Shatar, Australian Centre for Precision Agriculture, Department of
Agricultural Chemistry and Soil Science, Ross St Building, A03, Sydney University, 2006 Australia.