Abstract: Based on greenhouse microclimate observation data in two growing seasons from December 2012 to January 2013 and from December 2013 to January 2014,the occurrence of frost damage for tomatoes in greenhouse and their influencing factors were analyzed.The diurnal temperature difference index was constructed and taken as an indicator for the happening of low-temperature and frost damage during fruit enlargement period for tomato in greenhouse.Additionally,a temperature forecasting model was constructed through a regression analysis method.The results show that the occurrence of low-temperature and frost damage for tomato in greenhouse can be attributed to the combined effects of continuous low temperature,lack of sunshine and high humidity in greenhouse.Low-temperature and frost damage happens when the number of scant-sunshine days with rain or snow outside greenhouse is more than 3,the maximum temperature in greenhouse is less than 10℃,the minimum temperature is less than 5℃,and the duration of low temperature (less than 5℃) is more than 15 h.In this study,hourly temperature data inside and outside greenhouse are used to construct a temperature forecasting model on an hourly time step for greenhouse.The forecasting accuracy for hourly temperature with an absolute error ≤3℃ from 18:00 to 08:00 reaches more than 86%.Based on the low-temperature and frost damage indicators and the hourly temperature forecasting model,an early warning for occurrence of low-temperature and frost damages can be provided.It improves the disaster prevention and mitigation capabilities in a greenhouse.

Key words: Greenhouse, Tomato, Low-temperature and frost damage indicators, Regression analysis, Forecasting model