Use of meteorological satellite for winter wheat yield estimation and forecasting
2002
Faber, A. (Institute of Soil Science and Plant Cultivation, Pulawy (Poland). Dept. of Agrometeorology and Applied Informatics)
The objective of this paper was to use NOAA/AVHRR indices VCI and TCI calculated according to Kogan for wheat statistical yield estimation and forecasting in regions of Poland. The artificial neural network was fitted 10 years series of VCI and TCI from January and July. The indices were averaged to region. The obtained neural model for 33 regions explained 65.6 percent of relative yield variation with relative root mean square error (RRMSE) equals 5.4 percent. The RRMSE of independent samples from 134 regions in 10 years varied from 6.4 to 7.1 percent. The forecast of winter wheat statistical yields for 49 regions for independent years were: RMSE-2.4 dt per ha and RRMSE-7.1
显示更多 [+] 显示较少 [-]