Abstract: In recent years, temperature in the Northeast of China has been significantly increasing with global warming. However, since planting areas of relatively late-maturing varieties of maize are shifting towards north, the frequency and intensity of maize chilling disaster are increased in some areas and some years. The prevention and mitigation of chilling disaster of maize in the Northeast depend on our understanding about chilling disaster factors and the degrees of impacts. In this paper, using North East China Maize Growth Model (NEC_MaGM) and historical chilling disaster data, we analyzed temporal and spatial variations of low temperature during the grain filling period and the first frost day, and discussed the match relations between chilling years based on the two disaster factors and actual chilling years. The results indicated that NEC_MaGM had a satisfied outcome in simulating maize growth and development. The absolute deviation at simulated tasseling and maturity stages calculated by NEC_MaGM was 3.6 d and 5.9 d, respectively. The root mean square error of simulated weight of storage organ (WSO) was 839.5 kg ha^-1. The average temperature in grain filling stage of maize in the Northeast was 16.7–23.5°C. However, the rapid decrease of temperature (<16°C) that usually occurred in the late period of maize development stage had a great impact on grain filling. The average temperature on the first frost day in the Northeast was 7.4–9.2°C, which was below the threshold for maize growth. In addition, the first frost day in the Northeast had a trend of delay with time. However, maturity of new maize variety also delayed with global warming, so the first frost frequently occurred before maturity. When the average daily temperature of five successive days lower than 16°C occurred before DVS =1.9 (development stage), there was a severe impact on weight of storage organ (WSO). The years with above situation were consistent with those actually occurred in historical record. The partial early first frost occurred mainly in the period of 1.8<DVS≤1.9 in the Northeast. The numbers of sites where WSO reduced owing to the first frost were more than 11% in the Northeast for 44% of years. The years that the first frost occurred before maize maturity resulted in halt of WSO accumulation were consistent with historical record. It is concluded that low temperature in the grain filling period and the earlier first frost are the key factors of maize chilling disaster in the Northeast, which could be used as the indicators in earlier warning and prediction of chilling disaster based on the crop growth model output. Meanwhile, the crop growth model can be a powerful tool to analyze agro-meteorological disasters factors.

Key words: Maize chilling disaster, Low temperature in grain filling period, First frost, Crop growth model