Correlation and regression for prediction of wheat leaf rust severity in Bahawalpur and Multan using relevant meteorological data

Various meteorological variables were used as predictors to forecast wheat leaf rust disease severity in Bahawalpur and Faisalabad zones of Southern Punjab. The disease severity data was correlated with various meteorological variables viz., Tmax (maximum temperature), Tmin (minimum temperature), Tave (average temperature), Tprec (total precipitation), RHave (average relative humidity), RH0300 (relative humidity at 0300 GMT) and relative humidity at 1200 GMT to determine variables yielding highest correlation coefficients, which were included in the regression analysis with terminal disease data. Variables with best coefficient determination (R2) values were optimized to form model equations – RHave and Tprec were consistently significant with sporadic inclusions of Tmin and Tmax. The models, thus obtained were subjected to fitness and significance tests and best-fit models were obtained. RHave and Tprec tunred out to be the most detrimental factors for disease severity in each model. The models were validated by contrasting the predicted disease severity with the actual data, which yielded harmonious results indicating the practicality of the work.