Summary of the Presentation
ADVANCES IN ESTIMATING CROP YIELD THROUGH COMBINED
REMOTE SENSING AND GRO~TH MODELING
Stephan J. Maas
USDA-ARS Subtropical Agricultural Research Lab
describes the advances made over the past
two years in a crop yield estimation technique that combines
aspects of satellite remote sensing and crop simulation modeling.
The technique responds to the challenge to meet the goal estab-
lished at the ARS Remote Sensing ~orkshop (20-22 October 1987,
Beltsville, MD) of developing hybrid remote sensingjagroclimatic
models that can estimate foreign yields more accurately and
domestic yields more economically than by current operational
From the start of the project, attention was paid to
developing a technique that would require a minimum of input
data, and the input data would be of a type routinely available
in an operational program.
It was also considered beneficial to
develop a single technique that could be applied to both foreign
and domestic yield estimation.
A number of yield estimation techniques have been proposed
that make use of either remote sensing or growth modeling.
techniques rely on the inherent strengths of these technologies
However, the inherent weaknesses of each technology
have hindered the acceptance of these techniques in operational
yield estimation programs.
The technique described irr~his
combines aspects of remote sensing and growth
modeling such that the strengths of one technology make up for
the weaknesses of the other.
The Objective Yield Survey employed by the National Agricul-
tural Statistics Service (NASS) was used as the starting point
for developing the technique, since it sets the standard for
yield estimation accuracy.
It also demonstrates that observed
data and models can be combined in an operational program.
the Objective Yield Survey, observed data are obtained by ground-
level field sampling.
These data drive the empirical regression
models used to