RT Journal Article
ID 11b1c8e2635f68c6
A1 Kogan , Felix
A1 Kussul, Nataliya N.
A1 Adamenko , Tatyana I. 
A1 Skakun, Sergey V.
A1 Kravchenko, Alexey N.
A1 Krivobok , Alexey A.
A1 Shelestov, Andrey Yu.
A1 Kolotii , Andrey V.
A1 Kussul, Olga M.
A1 Lavrenyuk, Alla N.
T1 Winter Wheat Yield Forecasting: a Comparative Analysis of Results of Regression and Biophysical Models
JF Journal of Automation and Information Sciences
JO JAI(S)
YR 2013
FD 2013-07-23
VO 45
IS 6
SP 68
OP 81
K1 winter wheat yield forecasting
K1 linear regression models based on satellite data
K1 nonlinear regression models based on meteorological factors
K1 biophysical models of growth
K1 comparative analysis of results.
AB Relative efficiency of using satellite data to winter wheat yield forecasting in Ukraine at region level is assessed. The efficiency of forecasting on the basis of empirical and biophysical models of agricultural crops is compared. As empirical yields models the linear regression models of yield dependency on 16-day index NDVI composite on the basis of MODIS data with spatial resolution 250 m (MOD 13) are considered as well as nonlinear regression model, in which daily meteorological data of 180 local meteorological stations are used as predictors. The empirical approach to prediction is compared with biophysical which is implemented in the system CGMS, adapted for the Ukraine and based on the WOFOST model. For parameters identification of the yield models the official statistical data is used of winter wheat yield at the regional level for the period of 2000−2009. Validation of models is done on independent data for 2010 and 2011. The obtained results showed that when training models for 2000−2009 and 2000−2010 years and validating for 2010 and 2011 respectively all three approaches show similar accuracy. Average root mean square prediction error is approximately 0.6 c/ha.
PB Begell House
LK https://www.dl.begellhouse.com/journals/2b6239406278e43e,5c31235370ddef1f,11b1c8e2635f68c6.html