author(s)
Ram Manohar Patel, Purushottam Sharma and AN Sharma

abstract
Girdle beetle happens to be a serious insect pest causing yield losses between 14 to 42 per cent in soybean. The incidence of this insect pest is related to the prevailing weather conditions. It is in this context, we intended to ascertain pertinent weather parameters that directly influence the girdle beetle incidence and the appropriate stage it infests and therefore necessary curative measures can be addressed to. Therefore, data survey was conducted on girdle beetle damage on soybean crop under Crop Pest Surveillance and Advisory Project (CROPSAP) in 20 districts of Maharashtra state during 2010-2015. The weekly data collected so were used for the analysis. The correlation analysis of girdle beetle damage with climatic variables showed the significant negative correlation with maximum temperature during current (TMax₀ =-0.44**) and previous two lag weeks (TMax_-1= -0.61**; TMax_-2=-0.54**) and significant positive correlation with relative humidity in current (RH₀=0.24*) and first lag week (RH_-1=0.29*). Separate models pertaining to two peak weeks were developed using multiple regression and assessed for its prediction accuracy. The first peak (33^rd SMW) model explained 54.2% variation in percent damage and the variables significantly influencing its infestation were TMin₀, TMax_-1, TMin_-1, and TMin_-2. The second peak model (36^th SMW) could explain 53.52% variability and TMax_-1, RF_-2, and TMin_-2 were significantly influencing the infestation. The favorable pre-disposing conditions for peak damage were found to be related to maximum temperature (27.29-31.65 ºC), minimum temperature (19.73-24.74) ºC, relative humidity (82.96-93.75%), and weekly total rainfall (15.03-141.46 mm). Weather based prediction models were also validated satisfactorily using cross-validation approach using two years (2014 and 2015) independent dataset. Small RMSE value, standardized residuals between ±3 and insignificant value of t-test signifies that there was no significant difference between observed and predicted values of girdle beetle damage. In conclusion, maximum and minimum temperature, and rainfall determine the severity of girdle beetle incidence.