International Journal of Pure & Applied Bioscience (IJPAB)
Year : 2017, Volume : 5, Issue : 4
First page : (1035) Last page : (1050)
Article doi: http://dx.doi.org/10.18782/2320-7051.5236
Comparison of Simulated and Statistical Model Prediction of Wheat at District Scale Yield under Sub-Temperate Climate of North Western Himalayas
Karan Verma1*, Ranbir Singh Rana2, Shivani Thakur1, and Sahil Salaria1
1, Ph. D Scholar, 2 Principal Scientist
Department of Agronomy, Forages and Grassland Management
CSK Himachal Pradesh Krishi Vishvavidyalaya, Palampur, Himachal Pradesh- 176062 India
*Corresponding Author E-mail: karanverma2123@gmail.com
Received: 17.07.2017 | Revised: 25.07.2017 | Accepted: 26.07.2017
ABSTRACT
Crop growth models were successfully used for forecasting yields of various crops at district and state level. Multiple regression technique was employed to develop models to forecast wheat yield for district Kangra of Himachal Pradesh by using weather and yield data of 25 years (1985 to 2011). Data for two years (2011-12 and 2012-13) was used to calibrate and validate the model. The field experiment were conducted during rabi with three varieties viz., HPW-249, HPW-155 and HPW-42 and four sowing environments viz. 20th October, 10th November, 30th November and 20th December. The statistical model was validated for at mid-season (F2) and pre harvest stage (F3) corresponding to 15th March and 15th April respectively. The sensitive periods of statistical and phenological significance were selected for regression analysis. The GDD decreased with delay in sowing. The crop sown on 20th October accumulated 153-170 degree days to attain complete emergence, 508 to 591 degree days to vegetative, 781 to 853 degree days to heading stage and 1502 to 1688 degree days to physiological maturity. The simulated LAI, dry matter accumulation, grain yield, days to physiology maturity and vegetative stage matched closely with observed values for all sowing environments. The RMSE values for dry matter accumulation varied from 147 to 348 kg/ha in 2011-12 and 234 to 364 kg/ha in 2012-13. It was observed that the differences between simulated and observed biomass were marginal in early and late sown crops in three cultivars. The RMSE values for leaf area index ranged from 0.5 to 0.2. The differences were higher in early and late sown crop. The RMSE values for yield were 258 kg /ha in 2011-12 and 302 kg/ha in 2012-13. The model performance was either underestimated or overestimated but found within acceptable limits. The predicted yields indicated 1.5, 15.2 and 3.7 percent deviation at mid-season (F2) stage during 2009-10, 2010-11 and 2011-12, respectively and 1.3, 10.9 and 2.0 percent errors during 2009-10, 2010-11 and 2011-12, respectively at pre-harvest stage. The yield was found to be more reliable in pre-harvest (F3) stage for district Kangra. The simulated yield of 2012-13 showed 15.1 higher yield compared to actual yield of the year and more than 18.8 percent higher yield observed compared to statistical model. Hence, In order to augment the accuracy of simulated model yield more farmers’ practices are to be surveyed to include in the model inputs.
Key words: Wheat, Simulation, Validation, Yield forecast, Statistical model prediction.
Full Text : PDF; Journal doi : http://dx.doi.org/10.18782
Cite this article: Verma, K., Rana, R.S., Thakur, S. and Salaria, S., Comparison of Simulated and Statistical Model Prediction of Wheat at District Scale Yield under Sub-Temperate Climate of North Western Himalayas, Int. J. Pure App. Biosci.5(4): 1035-1050 (2017). doi: http://dx.doi.org/10.18782/2320-7051.5236
