توسعه سامانه هشدار سريع براي سرمازدگي دير رس بهاره در بادامستان هاي حوضه آبريز زاينده رود

Occurrence of atmospheric hazards such as frost, late spring frost, Every year it causes a lot of losses in agriculture. Timely planning can reduce the damage caused by natural disasters. The purpose of this study is to predict the occurrence of almond orchard frost in Zayandehrood basin. This was done in three stages of study. The first step is to eva‎luate the output of the WRF model with emphasis on the minimum and maximum daily temperatures, The second step is modeling the almond flowering date And the third step is to combine the WRF model and almond flowering model. Today, using numerical weather forecasting models, we can prevent the damages caused by them. In the first phase of this research, To eva‎luate temperature prediction by WRF model at late spring frost For eleven meteorological stations in the Zayandehrud basin, They were simulated with a horizontal resolution of 1 km. Then with two point and regional approaches simulated temperatures With corresponding bony viscosity values at 24 and 48 h surface temperature forecasts (2 m), Was eva‎luated. Based on the results of the root mean square error, the modified coefficient of determination and the Mean Bias Error index. Which was better for the 24-hour simulation temperature than 48-hour was 2.8, 0.88, and 0.48, respectively. There is Statistically acceptable correlation (correlation coefficient) between independent variables, same as WRF model data And the dependent variable, which is the observed (real) data. In fact, the daily temperature output of the WRF model was performed in the study area and It was found that the model temperature prediction was statistically acceptable. Next, considering the long-term phenology of almond trees in the Najafabad region, Flowering dates were extracted and calculated based on Julius dates. Then the matrix table of thermal indexes was drawn. Among the available parameters, respectively, the number of days above average temperature, the sum of thermal units above zero (GDD> 0) and the number of days below average temperature, The highest direct correlation was significant (P-value <0.01) with flowering date That was 0.945, 0.938, and 0.921, respectively. Multivariate linear regression equations and Artificial neural network between flowering and thermal indices were also investigated. The correlation coefficients of these equations were 0.998 and 0.995, respectively. In the final phase of the study, the results showed that By combining the two phenological models and the temperature prediction by the WRF model, A 48-hour rapid al‎e‎rt to frost in the areaʹs gardens can be done with enough accuracy.

5

35705