SRINIVASA RAO, VSRI RAMA JOGAMBA, J2018-05-082018-05-082016http://krishikosh.egranth.ac.in/handle/1/5810044933D5300An attempt was made to study the spatial and temporal variations in area, production and productivity of tobacco in the major tobacco growing districts of Andhra Pradesh viz., Prakasam, Guntur, East Godavari, West Godavari and the Andhra Pradesh state as a whole. The study was based on 28 years of tobacco data from 1987 to 2014. The graphical analysis was used to study the variations in area, production and productivity of tobacco. An attempt was also made to measure the growth in area, production and productivity of tobacco with due consideration of discontinuity in the data. Generally, the time series data on production of crops often exhibits a discontinuity in the year to year variations. These disturbances are mainly due to the impact of technological innovations in the crop. Under this situation, the conventional time trend models fail to provide efficient forecasts, as these models are based on the assumption of uniformity in the year to year variations. To deal with this situation, the spline models were explored for forecasting the tobacco production, as discontinuity in the year to year variations is the fundamental assumption in these models. The graphical analysis indicated that the time series data on area, production and productivity of tobacco exhibited a discontinuous trend in all the districts as well as in the state as whole. The growth analysis revealed that the area of tobacco was increasing in the districts of Prakasam and West Godavari and there is a considerable decline in Guntur and East Godavari districts and in the whole state of Andhra Pradesh. Production of tobacco was increasing in all the districts, except in the district of East Godavari and there was a considerable increase in the average level of productivity over the years due to the technological innovations in the crop. The spline models were found to be relatively efficient than the conventional trend fitting models in forecasting of tobaccoen-USnullThesis