【作者】 林燕金；

【导师】 陈家豪；

【作者基本信息】 福建农林大学 ， 作物栽培学与耕作学， 2006， 硕士

【摘要】 香蕉是漳州的主要果树之一，而漳州处于南亚热带气候带，常有低温天气影响香蕉生产，造成严重的经济损失。低温灾害已经成为制约香蕉产业发展的主要自然灾害之一。因此，对香蕉低温害的研究和预测防御在理论研究和实地生产实践方面都具有重要的意义。本文从农业气象学和生理学角度入手，研究了8个气象低温因子对漳州市及其该市的平和县、云霄县、华安县4个区域香蕉产量的影响程度，同时，对漳州市的未来一两年冬季（12月1日-次年2月29日）的低温情况和四个区域的香蕉产量进行了预测，并对漳州产区进行了区划，提出了一些具体的防寒避寒措施。初步得出了如下研究结果： （1）利用线性回归方法分析气象低温因子及其与香蕉产量的相关性结果表明：所选择的8个气象因子间通径分析结果显示，除了x₆与x₃、x₇与x₃之间不存在线性相关关系，其余的因子两两之间成极显著相关；线性回归分析8个低温因子对香蕉产量的影响中，影响大小按位次排列为：x₃＞x₇＞x₅＞x₆＞x₈＞x₄＞x₁＞x₂，即气温日较差（x₃）与产量的相关系数最大，二者呈极显著负相关，其次是日均气温5日滑动平均值（x₇），也与产量呈极显著正相关，而日均气温（x₂）对产量的影响最小。所以，气温日较差成为香蕉遭受低温灾害时造成香蕉受害减产的最关键的气象因素。 （2）利用时间序列分析方法的季节性水平模型建立的低温预测模型分析结果表明：除个别年份的拟合值与实测值存在偏差，大多数年份的拟合值与实测值是比较接近，拟合效果比较好，其模型可以预测未来历年冬季同最低温度旬均值（（?））。历年冬季日最低温度旬均值（（?））预测的结果为：2005年12月上旬为13.8℃，中旬为9.7℃，下旬为11.2℃；2006年1月上旬为7.2℃，中旬为12.7℃，下旬为10.1℃；2006年2月上旬为9.9℃，中旬为10.9℃，下旬为9.4℃；2006年12月上旬为12.1℃，中旬为14.1℃，下旬为11.1℃；2007年1月上旬为10.3℃，中旬为8.2℃，下旬为11.4℃；2007年2月上旬为8.5℃，中旬为11.2℃，下旬为11.1℃。 （3）利用时间序列分析方法的三次指数平滑模型建立的产量预测模型分析结果表明：除个别年份由于天气及生产条件原因使得拟合值与实测值存在偏差，大多数更多还原

【Abstract】 Banana was one of main fruits of Zhangzhou where the place was of south subtropics, and low temperature affected the yield of banana frequently, which made serious economy loss. Low temperature had become the one of nature disasters restricting the development of banana industry. Therefore, it is quite significant to study and forecast low temperature disaster of banana in science theory and production practice on the spot. The text started from agricultural meteorology and physiology, the affecting degree of eight low temperature factors on the yield of banana in Zhangzhou city, Pinghe county, Yunxiao county, and Huaan county was investigated, the temperature and the yield of banana in the future in four areas mentioned were forecasted, the banana planting areas in Zhangzhou were programmed, some measures preventing damage by low temperature were put forward, and the following results were come out.（l）The result of correlation analysis among low temperature factors and between the factors and the yield of banana by linear regression showed that: accept it is not linearly correlative between x₆and x₃, so is x₇ and x₃, all the coefficients among eight factors were prominent, and the correlation between eight low temperature factors and the yield of banana gradated: x₃>x₇>x₅>x₆>x₈>x₄>x₁ >x₂. That is the correlationcoefficient between the daily difference temperature （x₃） and yield of banana was the greatest the best prominent and negative; the second was the average temperature of daily average temperature in every sliding five days （x₇ ）, but daily averagetemperature （x₂ ） was the least. Therefore, the daily difference temperature is critical oneof the factors affected the output of banana suffered from low temperature disaster.（2） The result of prediction on low temperature model established by season level model of time sequence analysis showed that: the forecasting value was close to factual value in most years except for a few years, the prediction was good, and the model couldforecast the average of daily minimum temperature in every a period ten days ( T_DX|— ) in the future. The result of prediction of the average of daily minimum temperature in every a更多还原