مجله آب و خاک (Feb 2017)
Determining Prediction Model of the Canola) Brassica napus L.( Yields Based on Agrometeorological and Climatic Parameters in Mashhad Region of Iran
Abstract
Introduction: Environmental factors whichaffect crop yield areone of the most important factors in increasing yield.Accurate prediction of crop yield for economic management and farming systems is of particular importance. Materials and Methods: This research was done in order to statistically model and predict the canola growth and yield in Mashhad region based on 5 agricultural meteorology indicesand 12 climatic parameters during 1999 - 2014period. The date of planting determined with regard to the optimum temperature at planting with probability of 75% based on Weibull formula. Beginning and the end of the phenological stages of canola (germination, emergence, Single leaf, rosette, stemming, flower, poddingand ripening) were calculated on the basis of growing degree days (GDD) for each set. Calculation and statistical equations was done usingMinitab Ver. 13.0, 16.Ver SPSS and Excelsoftwares. Correlation analysis,statistical models andmultivariate models were used to determine the relationship between the annual yield of canolaand independent variables, includingclimaticparameters and agricultural meteorologyindices during the growing season between 1999- 2000 and2009-2010for each phenological stage (8stages).The bestmodel was selected with respect to the values of the coefficient of determination (R2) and root mean square error (RMSE).If the predictive power is estimated of the model RMSE values of less than 10% excellent, between 10 and 20% good, 20 to 30% average, and higher than 30% weak. The model tested by estimating the yield of canola for the 2010 to2014 years and the correction factor was calculated and the effect. Results and Discussion: Canola planting date wascalculated for 23 September in Mashhad region. The phenology of canola was calculated based on growing degree days (GDD) above 5 ° C.Germination calculatedfor25 September, emergence in 3 October, appearance single leaf in 7 October, rosette in 6 March, stemming in 4 April, floweringin 21 April, podding in 15 May and ripening in 4 Jun. The time of the phenological stages of cereals is virtually the same time. Therefore, due to the water scarcity in the studied region -canola can be used in crop rotation. Average, the highest and the lowest yield of canola were1329.5, 2159 and 835.5 kg per hectare,respectively.Canola crop yield showed a rising trend during 1999 – 2014period due toimprovingfarming techniques and mechanization. All models are significant regression coefficients were tested normal, alignment and line.Each model in the absence of proof of any of these hypotheses was removed and the 9remaining models were compared.Model 1 predicted canola crop yield in the single leaf stagewith an average yield of canola evapotranspiration ((Mpet, absolute maximum wind speed (FFabsmax) and the sum of the vapor pressure deficit (VPD).Model 5 predicted canola yield in the floweringstage based on the absolute lowest temperature (Tabsmin), average daily wind speed (FF) and total sunshine hours (SH). Model 3 predicted canola yield in the rosette stage based on the average of daily minimum temperature (Tmin), the number of days with precipitation greater than 1 mm R (day) and total pressure loss water vapor (VPD). Model 7 predicted canola yield during the whole growing season based on the average of daily maximum temperature (Tmax) and total precipitation (R).After R2 models with higher coefficient of 1, 5, 7 and 3, respectively, with coefficients of determination 0.902, 0.902, 0.868 and 0.866 respectively.Then F and RMSE were evaluated forecasting models 1 and 7 excellent, 5 good model and version 3 was average. Model 7due to lower RMSE and the number of parameters during growing season was the most appropriate model. Model validatedby means ofrecordedcrop yieldsduring 2011 and2014 years. The simulated yieldswere 1470, 1639 and 1226 with average of 1445 kg per hectare. Error percent was 45.1, 9.3 and -7.1for the following years with an average of 15.7. RMSE was 9.4, 2.6 and 2.3 with average of 7.4. The predictive value of the model was excellent for all these years. Conclusion: Model predicted the yield of canola based on the average maximum temperature (Tmax) and total precipitation (R)with error correction to reduce15.7. These variables described 86.8percent yield in the growing season and were significant at 5 percent. Canola planting date wascalculated for 23 September. Time phenology was germinated 25 September until ripening 4 Jun.
Keywords