Forewarning of stripe rust (Puccinia striiformis) of wheat in central zone of Punjab

Wheat is an important source of food to majority of population of many developing countries. Globally, wheat is cultivated on an area of about 219 million hectares with a production of around 763.2 million tones. India ranks second in production of wheat after China (Bhardwaj et al., 2019).Wheat is predominantly grown in the northern and north-western parts in country and is second most important staple food crop after rice. The most important wheat producing states include Uttar Pradesh, Punjab and Haryana accounting for 60 per cent area of the country. Punjab alone accounts for 13.58 per cent area and 21.77 per cent of wheat production of India (FAOSTAT 2020). During 2018-19, wheat was cultivated on 35.20 lakh ha area with a production of 182.62 lakh tones and average yield of 51.88 q ha-1 (Anon., 2020). During the last 30 years, agricultural production has been capable to maintain pace with food demand of increasing population. It has been estimated that approximately 2.5 per cent increase in cereal production will be required to meet food requirement in the next decade.

of population of many developing countries. Globally, wheat is cultivated on an area of about 219 million hectares with a production of around 763.2 million tones.
India ranks second in production of wheat after China (Bhardwaj et al., 2019).Wheat is predominantly grown in the northern and north-western parts in country and is second most important staple food crop after rice. The most important wheat producing states include Uttar Pradesh, Punjab and Haryana accounting for 60 per cent area of the country. Punjab alone accounts for 13.58 per cent area and 21.77 per cent of wheat production of India (FAOSTAT 2020). During 2018-19, wheat was cultivated on 35.20 lakh ha area with a production of 182.62 lakh tones and average yield of 51.88 q ha -1 (Anon., 2020).
During the last 30 years, agricultural production has been capable to maintain pace with food demand of increasing population. It has been estimated that approximately 2.5 per cent increase in cereal production will be required to meet food requirement in the next decade.
The susceptibility of commercially available wheat varieties to rust species is one of the serious constraint to maintain the yield and productivity of wheat under changing climatic conditions. Rusts are among the most economically significant fungal diseases in cereal crops worldwide. Stripe rust, caused by Puccinia striiformis f. sp. tritici, is an important wheat disease common in wheat growing areas experiencing cold and humid weather conditions during the crop season. In 1986, stripe rust (due to breakdown of stripe rust resistance genes YR-9) was first time diagnosed in Kenya from there its urediospores reached India in 1997-98 (Amor et al., 2008). It occurred in epidemic form in 2008-09 in Punjab and its surrounding areas and caused huge yield losses. Pannu et al., (2010) revealed about occurrence of stripe rust in epidemic form in foot hills of Punjab and neighbouring states during 2008-09 to 2010-11.
In the present situation of rising food demands with decrease in agricultural lands more efficient agricultural system will face challenge from diseases such as stripe rust. Weather plays significant role in appearance, progress and multiplication of diseases. Milus and Seyran (2004) revealed that stripe rust caused by the new isolates tends to grow faster than the previous isolates at comparatively elevated temperatures. Temperature, rainfall, humidity, sunshine duration and wind have the important effect on many plant diseases. Whenever, vulnerable host and active pathotype exist together under most favorable conditions, the probability of disease epidemic increases.
Disease forewarning is very important for the timely and effective management of rusts by using fungicides.
Many weather based regression models/equations for rust forewarning are available for various locations. But these regression models are site specific and cannot be used for different regions. The unsuitability of use of these models at different locations is due to the climate variability and some bio-physical conditions. Therefore, a need was felt to formulate weather based stripe rust forewarning model for central zone of Punjab. So keeping this in view an analysis was conducted to identify the weather variables and critical periods affecting stripe rust severity in wheat and to develop weather based prediction model for stripe rust.

MATERIALS AND METHODS
The meteorological data recorded at Agrometeorological Observatory (30°54'N latitude; 75°48'E longitude and altitude of 247 m amsl) of Punjab Agricultural University, Ludhiana for the period of (2007-08 to  was considered for the study.
The disease severity was calculated from the proportion of plant tissue infected by the disease as per Modified Mannar's scale for stripe rust of wheat (Peterson et al., 1948). Six categories of the scale were determined on the basis of the per cent area of the leaf covered by infection as follows: 5 = Upto 5 per cent leaf area infected 10 = Upto 10 per cent leaf area infected, 20 = Upto 20 per cent leaf area infected, 50 = Upto 50 per cent leaf area infected, 75 = Upto 75 per cent leaf area infected, 100 =Upto 100 per cent leaf area infected.
Stripe rust severity data was recorded at weekly interval and then percentage weekly disease severity and terminal disease severity was calculated.

Correlation study
Correlation coefficients were calculated between monthly stripe rust disease severity index and different

Model validation
To validate the regression model, disease severity and meteorological data of 2009-10 and 2019-20 (that was not used in development of model) was used by calculating the deviation of predicted value from observed value.
After that linear relationship was developed between observed and predicted disease severity.

Occurrence of stripe rust in Punjab
The disease attack on crop depends on weather conditions prevailing in the area. It is a major problem of wheat in Punjab and its surrounding areas due to prevalence of congenial weather conditions for its development.

Epidemic window of stripe rust in Punjab
Every insect-pest or disease has a specific window period during which it flourishes well. Stripe rust occurrence, development and spread is influenced by different meteorological parameters throughout disease development and spread period as shown in Fig.1 (a

Meteorological parameters and stripe rust
The pooled correlation coefficients between disease severity and different meteorological parameters were calculated on the basis of weekly disease severity and weekly weather data in different years (2007-08 to 2019-20 except 2009-10 and 2019-20) as presented in Table 1. Disease severity and weather data of 2009-10 and 2019-20 years was used for model validation so these years data was not included in pooled analysis for model development. Among correlation coefficient analysis, disease severity showed significant positive correlation with maximum and minimum temperature, sunshine hours whereas disease severity showed significant negative correlation with relative humidity. Similarly, Gupta et al., (2017) revealed that meteorological parameters viz. temperature Where, *Upward arrow indicates above normal value of parameter; **Downward arrow indicates below normal value of parameter (maximum and minimum), morning and evening vapour pressure, and micrometeorological parameters (canopy and soil temperature) showed significantly positive correlation with the stripe rust severity, whereas, morning relative humidity was negatively correlated with rust severity of wheat.

Development of a regression model
On the basis of different correlation coefficients and relationships of disease severity with different meteorological parameters, most influencing meteorological parameters were identified. These   (Fig. 4) of the model also signify good predictability of the model if a new set of meteorological data fitted into it. So, this model can be used to forewarn stripe rust incidence and thus fungicide application can be scheduled accordingly.

Validation of model
Validation of the regression model was done by calculating the predicted disease severity of two years (2009-10 and 2019-20) data. This validation indicated that developed regression model over-estimated the disease severity by 4 to 8 per cent (Fig.5) Forewarning of disease severity can be used for timely and effective management of disease so as to reduce the yield losses caused by this particular disease.