Weather data based descriptive models for prediction of coconut yield in different agro-climatic zones of India Kumar S. Naresh*, Rajagopal V., Cherian V.K., Thomas T. Siju, Sreenivasulu B.1, Nagvekar D.D.2, Hanumanthappa M.3, Bhaskaran R.4, Kumar K. Vijaya, Narayanan M.K. Ratheesh, Amarnath C.H. Central Plantation Crops Research Institute, Kasaragod, 671 124, Kerala. *Corresponding author's: E-mail: nareshkumar.soora@gmail.com
1Coconut Research Station, Ambajipeta, Andhra Pradesh. 2Regional Coconut Research Station, Batye, Ratnagiri, Maharastra. 3Agricultural Research Station, Arisikere, Karnataka. 4Coconut Research Station, Veppankulam, Tamil Nadu. Abstract Weather variables play an important role in determining the coconut palm growth, development and yield. The influence of weather on nut yield in coconut starts from inflorescence initiation and lasts till nut maturity (44 months). Historical data on weather variables and coconut nut yield from different agro-climatic zones viz., Western coastal area - hot sub-humid-per-humid (Kasaragod - Kerala; Ratnagiri - Maharastra), hot semi-arid (Arisikere - Karnataka) and Eastern coastal plains - hot sub-humid (Veppankulam - Tamil Nadu; Ambajipeta - Andhra Pradesh) of India were used for developing models for prediction of coconut yield. The prediction models with 3- and 4-year lag had high R2 values. The models differed for usage of parameters in different agro-climatic zones, indicating the relative importance of these parameters in respective conditions for realizing the nut yield in coconut. Interestingly, the parameters used in models for western coastal area - hot sub-humid-per-humid are temperature and relative humidity, as indicated even in the classification of these areas. Models were verified for 2 years and prediction of yield during 1998–99 and 1999–2000 within 10% confidence level validated these models. The study indicates that the relative humidity and temperature play important role during the ontogeny of inflorescence and nut development. The descriptive models, developed based on weather data, can be used for prediction of coconut yield two to four years in advance with in acceptable range of accuracy. The yields to be realized can be bracketed within predicted range obtained from models using one, two, three, and four year lags. These models can also be used for prediction of coconut nut yield in the situations similar to those in the present study. Top Key words Agro-climatic zones, coconut, prediction models, yield, weather. Top |