Author
 |
SHARMA-POUDYAL, DIPAK - Washington State University |
|
Chen, Xianming |
|
Submitted to: Phytopathology
Publication Type: Peer Reviewed Journal Publication Acceptance Date: 12/16/2010 Publication Date: 5/1/2011 Citation: Sharma-Poudyal, D., Chen, X. 2011. Models for predicting potential yield loss of wheat caused by stripe rust in the US Pacific Northwest. Phytopathology. 101:544-554. Interpretive Summary: Climatic variation in the US Pacific Northwest (PNW) affects epidemic of wheat stripe rust. Previous models only estimate disease severity at the flowering stage, which may not predict the actual yield loss. To develop models for predicting potential yield loss, correlation and regression analyses were conducted using historical yield loss data and weather parameters in different seasons from 1974 to 2007. Among seasons, winter temperature variables were more highly correlated to winter wheat yield loss than from other seasons. The sum of daily temperature of the entire winter season was more highly correlated with yield loss than for one, two, and three months of the winter period. Similar results were obtained for moving average temperature and snow cover days. Within the winter season, December moving average temperatures were more or at least equally correlated to yield loss than weather variables for the one-, two-, and three-month periods. After the winter season, accumulated positive degree days and sum of daily temperatures in June had a higher degree of correlation with yield loss than the spring and fall seasons. Using the significantly correlated climatic variables, we developed models for predicting the yield loss of susceptible wheat cultivars before disease appearance in the field. The models were validated with the weather and yield loss data in 2008 and 2009. Therefore, these models can be useful in the stripe rust management for the major wheat growing areas in the US PNW. Technical Abstract: Climatic variation in the US Pacific Northwest (PNW) affects epidemic of wheat stripe rust, caused by Puccinia striiformis f. sp. tritici. Previous models only estimate disease severity at the flowering stage, which may not predict the actual yield loss. To develop models for predicting potential yield loss, correlation and regression analyses were conducted using historical yield loss data from 1974 to 2007 for winter wheat and 1979 to 2007 for spring wheat and weather parameters including moving average temperatures (MAT), sum of daily temperatures (SDT), accumulated negative degree days (ANDD), accumulated positive degree days (APDD), rainfall, snowfall, and snow cover days (SCD) in different seasons. Yield loss of winter wheat was more significantly correlated to weather conditions and had more significantly correlated weather variables than that of spring wheat. Among seasons, winter temperature variables were more highly correlated to winter wheat yield loss than from other seasons. The SDT of the entire winter season was more highly correlated with yield loss than for one, two, and three months of the winter period. Similar results were obtained for MAT and SCD. Within the winter season, December MATs were more or at least equally correlated to yield loss than weather variables for the one-, two-, and three-month periods. After the winter season, APDD or SDT in June had a higher degree of correlation with yield loss than the spring and fall seasons. These significantly correlated climatic variables during different months/seasons of a year allowed us to develop different simple and multiple linear regression equations that can be used to estimate the yield loss of susceptible wheat cultivars before disease appearance in the field. The models were validated with the weather and yield loss data in 2008 and 2009. Therefore, these models can be useful in the stripe rust management for the major wheat growing areas in the US PNW. |
