甘藍、結球白菜及花椰菜皆為我國重要蔬菜作物，然而近幾年來，這些大宗蔬菜常因生產過盛，造成產地及市場價格大跌。一般而言，溫度被視為是影響蔬菜生長之最主要因子，且亦會影響消費者對上述蔬菜之消費行為，進一步造成蔬菜價格的變化。因此本研究發展溫度對蔬菜產量變化及價格波動影響之模式，模式中參數檢定以溫度、產量及市場價格之現地調查與統計資料進行，並進行模式模擬，模式驗證則檢視實際資料與模式預測值之百分誤差。研究結果顯示，利用動態模式，並以多項式迴歸分析進行參數檢定，可成功建立冬春季溫度與大宗蔬菜產量關係。而蔬菜價格與產量關係，亦可利用當旬之資料建立一反比模式描述之。本研究之結果可做為大宗蔬菜之超產預警及耕鋤系統啟動時機之參考依據，對於政府財政及農民權益有極大的助益。後續研究應建立大宗蔬菜產地之氣象資料監測系統，並針對其他影響蔬菜產量之因子，包括耕作面積、降雨量、降雨天數、日照時間、溫度差異等進行分析，以提升模式在應用上之可行性。

Cabbage, Chinese cabbage, and cauliflower are staple vegetables in Taiwan. In recent years, however, the market prices of these vegetables often experienced a heavy slump due to a large surplus of yield. Generally, the growth rates of these vegetables, as well as the vegetable demands of consumers, are mainly affected by the temperature of growing season. The variation between supply and demand will result in fluctuations in the market price of these vegetables. In the present study, the growth rate and the prices of staple vegetables were expressed as functions of temperature. Parameters calibration was achieved by analyzing the correlation among temperature, vegetable yields and market prices obtained from in situ investigation and/or official statistical data. Model simulation was then performed with calibrated parameters. Model validation was done by checking the percent error between the actual data sets and the predicted values. Results of this study indicated that the dynamic model, accompanied with parameter calibration by polynomial regression model, can successfully model the relationship between temperature and vegetables yields during fall and winter. The relationship between 10-day price and yield of vegetables can be expressed by an inverse model. Results of the present study can be served as an early alarm system in determining whether a surplus of vegetable yield will happen or not, which would be beneficial to the farmers and the economics of the government. Future researches are recommended to continuously and thoroughly monitor and record the weather parameters including solar radiation intensity and rainfall by in situ weather stations. The weather parameters recorded would then be incorporated into the model to refine and enhance the predictive power of the model.