玉米相對花粉濃度之轉換研究

The Study of Transformation of the Relative Pollen Concentration of Maize

10.6730/JAAT.201406_15(2).0003

余昇驊(Sheng-Hua Yu)；郭寶錚(Bo-Jein Kuo)

玉米花粉飄散 ； 相對花粉濃度 ； 高斯煙羽－指數法 ； 空間變異 ； pollen dispersal of maize ； relative pollen concentration ； Gaussian plume-exponential approach ； spatial variation

台灣農學會報

15卷2期（2014 / 06 / 01）

143 - 163

繁體中文

聶等（2013）曾以大氣高斯煙羽模式模擬玉米花粉粒的擴散趨勢，計算相對花粉濃度，並以此估計異交率，發現相對花粉濃度高估實際異交率。本研究利用2009年至2010年在台灣嘉義縣朴子所進行的玉米花粉飄散試驗資料，透過網格化田區之方式，以高斯煙羽模式，作為模擬玉米花粉粒子在大氣中傳輸的過程，並計算相對花粉濃度之累積值，再透過指數函數進行資料轉換，藉以修正過去高估玉米異交率的情形，以上修正方法稱之為高斯煙羽－指數法。經計算異交率估計值與實測值之間的相關係數達0.87以上，均方根誤差值則低於0.03，顯示本估計方法可矯正過去高估的情況，並且可同時考慮風速與風向以及實際田區的相對位置等因子。本研究亦以一般克利金法進行空間內插，用以估計接受親田區內的異交率，並呈現接受親田區異交率的空間變異。

Nieh et al (2013) employed the Gaussian plume model to simulate the dispersal tendency of the pollen grain of maize. They calculated the relative pollen concentration which was used to estimate the cross-pollination rate. This estimator overestimated the actual cross-pollination rate. In this study, the field data collected in Puzih from 2009-2010 were used. To meet the assumption of the Gaussian plume model, the total area was divided into numerous point sources to simulate the transportation process of pollen grains of maize in the atmosphere and the relative pollen concentration was aggregated. The aggregated values were transformed by an exponential function. This adjustment was termed the Gaussian plume-exponential method. It was found the correlation coefficient between the actual and predicted cross-pollination rates was 0.87 and the root mean square error was below 0.03, indicating the Gaussian plume-exponential method could adjust for the overestimation of the cross-pollination rate. In addition, the wind speed, wind direction and relative position were considered. The ordinary kriging method was also employed to interpolate the cross-pollination rate in the recipient plot. The spatial variation of the cross-pollination rate was displayed.