United States
Department of
Agriculture
Statistical
Reporting
Service
Statistical
Research
Division _--
StaR Report
No. AGES811214
Washington, D.C.
December 1981
Evaluation of the
CEAS Trend and
Monthly Weather
Data Models for
Spring Wheat Yields
in North Dakota
and Minnesota
JeanneL. Sebaugh
t
EVALUATION OF THE CEAS TREND AND MONTHLY WEATHER DATA MODELS FOR SPRING
WHEAT YIELDS IN NORTH DAKOTA AND MINNESOTA.
By Jeanne L.Sebaugh;
Research Division, Statistical Reporting Service, U. S. Department of
Agriculture, Columbia, Missouri 65201; December 1981. SRS Staff Report
No. AGES 811214
ABSTRACT
The CEAS models evaluated use the basic input variables of year and
monthly average temperature and total precipitation to forecast and
estimate spring wheat yields in North Dakota and Minnesota.
Historic
. trend, meteorological and agroclimaticvariables
are constructed.
Stepwise multiple regression techniques are used to develop state
and crop reporting district regression models based on historic values
of these variables and yield.
Evaluation of yield reliability a~ the
state level indicates that the bias is less than one quintals/hectare.
The Minnesota model is somewhat less reliable than the North Dakota
model.
The models are obje~tive and adequate (in terms of coverage)
"for short-term use in North Dakota and Minnesota.
Consistency with
scientific knowledge could be more thoroughly documented.
Timely
yi'eldforecasts and estimates can be made during the growtngseason
using estimates of climatic division weather data.
The models are'
not costly to operate but the costs of future updates should 'be con-
sidered.
Users can eas i1y understand the.form of the models. and how
to use them. The model standard errors of prediction do',not provide
a useful current measure of modeled yield reliability ...
Key Word s: Model evaluation, crop yield modeling, regression modets,
spring wheat yield models.
****************************************************
~ This paper was prepared for limited distribution ~
~ to the research community