Search

In this study, an empirical assessment approach for the risk of crop loss due to water stress was developed and used to evaluate the risk of winter wheat loss in China, the United States, Germany, France and the United Kingdom. We combined statistical and remote sensing data on crop yields with climate data and cropland distribution to model the effect of water stress from 1982 to 2011. The average value of winter wheat loss due to water stress for the three European countries was about −931 kg/ha, which was higher than that in China (−570 kg/ha) and the United States (−367 kg/ha). Our study has important implications for the operational assessment of crop loss risk at a country or regional scale. Future studies should focus on using higher spatial resolution remote sensing data, combining actual evapotranspiration to estimate water stress, improving the method for downscaling of statistical crop yield data and establishing more sophisticated zoning methods. [ABSTRACT FROM AUTHOR]

• NNVAI is proposed to indicate winter wheat damage caused by a single frost event. • Growth-stage-specific relations between NNVAI and frost features were explored. • Winter wheat is sensitive to different features of frost at different growth stages. • Developed monitoring framework to do two-phase assessment of winter wheat damage. Frost events pose a threat to normal development of winter wheat and increase risk of production loss, even given the context of global warming. On the Huang-Huai-Hai Plain (HHHP), which is the largest region of winter wheat cultivation in China, frost damage has become one of the major meteorological catastrophes. In recent years, both frequency of occurrence and severity of frost events in this region have exhibited an upward trend. For timely and accurate monitoring of single-event frost damage of winter wheat, this study developed a frost-monitoring framework that considers the processes of the evolution of a frost event and the response of winter wheat, as well as the lag between them. The framework adopts a NDVI valley area index (NVAI) to represent the gradual response process of winter wheat to frost. In this framework, the normalized NVAI (NNVAI) is used to indicate the degree of damage, and a two-phased damage assessment (i.e., T1 and T2) is adopted to resolve the contradiction between the urgency of obtaining information on frost severity, the spatial distribution of frost damage, and the delayed response process of winter wheat to frost. Using prediction models based on historical statistics, assessment at the T1 stage estimates the possible damage after the end of the frost, and assessment at the T2 stage evaluates the actual damage caused by frost when the winter wheat has recovered. Application of the proposed framework to a frost event that occurred on the HHHP during March 4–9, 2018 revealed that the average absolute percentage error between actual and predicted damage across the entire region was 33 %. Moreover, the prediction of damage was made 13 days ahead of the occurrence of actual damage. Thus, the framework could provide guidance to enable action to be taken to reduce the impact of frost. The proposed monitoring framework fully considers the processes of the evolution of a frost event and the response of winter wheat to frost, and it learns potential patterns from historical frost events to provide early prediction of potential frost damage. Consequently, the framework is considered suitable for wide practical application. [ABSTRACT FROM AUTHOR]