2.1.19

RISK ANALYSIS FOR FORECASTING MODELS FOR ALTERNARIA BLIGHT, * ALTERNARIA BRASSICAE*, OF RAPESEED AND MUSTARD

JK DANG, MS SANGWAN, N MEHTA, CD KAUSHIK, OP SHARMA and A DHANDAPANI

CCS Haryana Agricultural University, Hisar-125 004, India

**Background and objectives**

*Alternaria brassicae* causing Alternaria blight is one of the most important disease and causes heavy losses in grain yield of rape seed and mustard [1], although this disease can be managed to some extent by the use of fungicides but a proper IPM can only be effective if quantitative relationship between disease progress and environmental factors is established [2]. The present investigations were undertaken to develop a suitable prediction model for forecasting of the disease and for better disease management.

**Materials and Methods**

Four cvs, RH-30, RH-8113 *Brassica juncea*, and YSPb-24 and BSH-1 *Brassica campestris* were sown on four different dates from middle of October to the second week of November during crop years 1993-94 to 1996-97. The crop was sown in a plot size of 3x2.10 m, with a row to row distance of 30 cm and a plant to plant distance of 10 cm, with three replications in a randomized block design. The development of disease was recorded every tenth day from the appearance of the disease until maturity (as per the standard scale). Weather parameters were recorded for the corresponding period. Advanced statistical software were used for epidemiological analysis. Different models such as Exponential, Monomolecular, Logistic and Gompertz were tried for best fit for a forewarning model which could explain realistic disease intensity using independent natural weather factors.

** Results and Conclusions **

Of the models tested, the Gompertz model was found to be the most appropriate. Weather variables such as maximum temperature, minimum temperature, rainfall, evening relative humidity (e.r.h.), wind speed and sunshine were analysed to obtain most of the variation present in weather data. Two factors were extracted. These two factors were further rotated using varimax method. These two factors explained 67% of variation. Factor loading is given by:

Factor 1=0.239*Max. temp.+0.975*Min. temp+0.69*Rainfall+0.633*e.r.h.+0.643*Wind speed-0.1 91*Sunshine

Factor 2=0.851*Max. temp.+0.231*Min. temp-0.319*Rainfall-0.757*e.r.h. -0.1 88*Wind speed+0.529*Sunshine.

The resultant factors then were used in the non-linear form of Gompertz Model. The model form, Disease=Exp[logA*log(-B*Time)]+C*Sowing-day factor 1, was found to be most appropriate. A and B are two parameters of the Gompertz+13 model and C and D are the coefficients of sowing day and factor 1. The best fitted model for each cultivar as follows:

RH 30: Y=E[log(0.620)*log(4.998*Time)]-0.890*Sowing day+0.238*Factor 1; R1 = 0.389

RH 8.113: Y=E[Iog(2.547)*iog(l.072*Time)1-0.561*Sowingday+0.086*Factor l; R2 = 0.482

BSH-1: Y =E[Iog(3.008)*log(l.335*Time)]-1.260*Sowingday+0.166*Factor l; R2 = 0.482

YSPb-24: Y=E[Iog(2.199)*log(4.827*Time)]-1.868*Sowingday+0.105*Factor l; R2 = 0.539
where Y is per cent disease intensity.

Results revealed that for a given cultivar, if the sowing day and weather parameters at a particular time is given, one can predict the likely incidence of *Alternaria* blight of rapeseed and mustard usingthe above models. This will prove to be an asset as a decision-making tool for successful disease-management strategies.

**References**

1. Dang, JK, Kaushik CD, Sangwan MS 1995. Indian Journal of Mycology and Plant Pathology 253), 184-188.

2. Kolte SJ, Awasthi RP, Vishwanath, 1987. Indian Phytopathology 49, 209-211.