2019-01-012024-05-15https://scholars.lib.ntu.edu.tw/handle/123456789/664909摘要：作物產量預測是在收穫前推估當季作物產量，透過準確地產量預測掌握各地區作物生產情形，有助於未來辦理作物保險或相關政策推動。本計畫以芒果為分析對象，擬開發利用環境因子推估芒果產量之統計模型，作為未來產量預測之依據。芒果栽培繁殖方法可分為有性生殖及無性生殖兩種，一般生產多採嫁接等無性生殖的方式，以維持優良品種特性。芒果最適生長溫度為24-27度，在適宜生長的溫度範圍內，溫度較高地區所產芒果成熟期較早、品質較好；溫度低於10度時芒果樹停止生長，開花期與幼果發育初期遭遇低溫會影響授粉率及開花，影響產量。除此之外，光照、雨量、濕度與風速對產量均有影響。本計畫預計以臺南市及屏東縣栽培之芒果園為研究對象，試行結合近年度 (1) 芒果產量資訊，與 (2) 氣象資訊，據以建立環境因子推估芒果產量之統計模型，並以交叉驗證方式將預測結果與歷史年報所列之芒果產量資訊互相比較，以驗證模型準確度。<br> Abstract: Crop yield prediction is to estimate the crop yield of the season before harvest, and to grasp the crop production situation in each region through accurate production forecast. It will help to promote crop insurance or related policies in the future. This project plans to develop a statistical model that uses environmental factors to estimate mango yield as a basis for future production forecasts. Mango cultivation and propagation methods can be divided into sexual reproduction and asexual reproduction. Generally, the method of asexual reproduction such as grafting is widely used to maintain the characteristics of excellent varieties. The optimum growth temperature of mango is 24-27 degrees. In the suitable temperature range, the ripening period of mango in the higher temperature area is earlier and the quality is better. When the temperature is lower than 10 degrees, the mango tree stops growing. The low temperature in the early stage of development affects pollination rate and flowering, which affects yield. In addition, light, rainfall, humidity and wind speed have an impact on production. This project will use the mango garden cultivated in Tainan City and Pingtung County as a research target. We will try to combine the recent years (1) mango yield information, and (2) meteorological information to establish a statistical model for estimating the mango yield by environmental factors. The cross-validation method compares the predicted results with the mango yield information listed in the historical annual report to verify the accuracy of the model.芒果產量預測氣象資訊統計模型交叉驗證Mangoyield predictionweather informationstatistical modelcross validation