STATISTICAL VALIDATION OF ALMANAC PREDICTED RAINFALL
Loading...
Files
Date
2019-08-06
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
UNIVERSITY OF AGRICULTURAL SCIENCES GKVK, BANGALORE
Abstract
Panchang is the Hindu almanac which furnishes meteorological predictions. The
predictions are of high importance as farmers even now depend on these predictions to plan
their agricultural operations. In panchang, predictions are based on factors like ruling
planet of the year, meghadhipathy and prominent cloud type. The study was conducted to
make a comparison between the rainfall predictions made in panchang and real-time
observations. For this purpose the rainfall data of 28 years (1989-2017) for the state of
Karnataka was collected from KSNDMC. The degree of agreement between the panchang
predictions and actual rainfall occurred was estimated for annual and monsoon rainfall
data. The results indicated that for annual rainfall, 4 years had agreement greater than 75
per cent whereas for monsoon rainfall, 10 years had agreement greater than 75 per cent.
Intraclass correlation coefficient indicated that individual factors cannot decide the nature
of the rainfall for the year. The understanding of combination of these factors is likely to
help in more accurate rainfall prediction. Association rules obtained were in agreement
with panchang rules. The Nakshatra based classification of rainfall indicated that on an
average rainfall observed with less variability from Rohini to Hasta Nakshatra, which in
turn indicates that during this Nakshatra period assured quantum of rainfall is received.
MK trend test was carried out as the average rainfall observed at different ruler periods
classified into three classes in each of the Nakshatra. The results indicated significant
trends during Jeshta and Poorvabhadra Nakshatra under heavy and moderate rainfall
respectively.