WSP00527

• File Number: 8271.300
• Description: Colorado River Basin Salinity Control Program - General Information and Publications-Reports
• Basin: Colorado Mainstem
• Water Division: 5
• Date: 9/1/1977
• Title: Optimizing Crop Production Through Control of Water and Salinity Levels in the Soil
• Water Supply Pro - Doc Type: Report/Study

C:J:
"

N
~
to
~

INTRODUCTION

BACKGROUND

The focus of the reported research was to
study and to develop methodology for predicting
crop yields under different soil moisture and
salinity conditions. The work was carried out by
four universities belonging to the Consortium for
International Development (CID) , headquartered
at Logan, Utah. The four experimental sites were
the University of Arizona (Yuma), the University of
. California (Davis), Colorado State University.(Fort
Collins), and Utah State University (Logan). Field
experiments with corn were carried out at each of
these sites in the growing seasons of 1974 and 1975,
under two companion projects funded by the Office
of Water Research and Technology.'

The first project was titled "Water production
functions and predicted irrigation programs for
principal crops as required for water resources
planning and increased water use efficiency." It
was funded under Title II Grant No. 14-'31-0001-
4233, and administered by CID as project number
C-5189. Duration of the project was 30 months,
from April 1. 1974, to September 30, 1976. The
companion project was titled "Optimizing crop
production through control of water and salinity
levels in the soil." It obtained matching funding
under Title I Grant Agreement No. 14-31-0001-
5120, and was administered by the Utah Water
Research Laboratory through the Utah Center for
Water Resources Research as project number
B-121-UT. The remaining funds were provided
from the respective universities. Duration of the
project was 27 months, from July 1, 1974, to
September 30, 1976.

A major objective in organizing the reported
research effort was to join researchers from the
four CID Universities into an effective regional
research unit. This was thoroughly accomplished,
likely beyond any of the participants' real
expectations. The achieved unification of research
procedures was one of the most important and
most gratifYing accomplishments of the entire
effort and was essential to the technical results

achieved. The research led to the L 'lief that it is not
possible to carry out studies of crop responses to
water and salinity at a single site (however well
done), and extend the findings to very different
climates and soils. To work independently on these
questions at several sites, with each set of
researchers following different concepts and using
different techniques. does not and cannot produce
the data base required for full transferability of
knowledge and predictive modeling.

The joint study has produced what is believed
to be an adequate base for both transfer and
prediction. It has also provided conclusive evidence
of a number of basic relationships, knowledge of
which is indispensable to effective modeling for
predictive purposes. One such relation is the linear
nature of the function which describes the decline
of total dry matter production of corn as the
seasonal e1vapotranspiration deficit increases. This
was clearly shown at all four experimental sites in
both cropping seasons despite great diversity in
cropping conditions.

A second basic finding is that the major effect
of salinity in agriculture is to reduce crop water
uptake. i.e., evapotranspiration, and to reduce
production accordingly. Thus water production
functions, and water-salinity production functions,
have been found to be one and the same. rhis
opens the door to development of comparatively
simple models for predicting both the effects of
water shortage and salinity.

Three models for predicting corn dry matter
production (relates to silage) and grain yield as a
function of water deficit have been tested using the
data base developed in the study. They are the
Hanks Model, developed at Logan, the Stewart
Model, developed at Davis, and the Hall-Butcher
Model, published in 1968. All of these models
appear to hold considerable promise for future use
in prediction. In any particular instance selection
among them, and possibly other models as yet
untried. would be largely dependent on the nature
of available input data.

1