基于气象因子驱动的阴山地区马铃薯产量预测模型

为了及时掌握马铃薯产量,提高产中、产后气象服务能力,指导马铃薯生产,以内蒙古阴山旱作区马铃薯生产为例,利用16个旗县1980-2007年气象和产量资料及发育期等其它相关资料,采用相关和回归分析等方法,分区域分析了影响马铃薯产量的关键气象因子,并建立了气象产量预测模型.结果表明:(1)降水是影响产量的关键因子,温度次之;前山地区高温胁迫的影响大于后山,降水不足的影响则相反;干旱少雨、高温胁迫是制约该地区马铃薯产量提高的主要因素.(2)用逐步回归方法建立的幼苗期-结薯期、幼苗期-淀粉积累期、结薯期-淀粉积累期和生长季4个时间段的气象产量预测模型均达到极显著水平,拟合率75％以上,产量预测平均误差11.1％,误差变幅0.34％ ～ 27.9％,近85％的预测值准确率超出80％,对区域预测结果好于各旗县;(3)结薯期是对水热最敏感的时期;不同时间段模型中以生长季和幼苗-结薯期模型预测结果较好.所建模型可以在马铃薯产量预测业务中应用.

Study on the forecasting model of potato yield based on drive of climate factors in Yinshan area

To grasp the potato yield and to improve the meteorological service ability in the middle and late period of potato growth and to guide potato production, with the potato production of dry farming region as the research object, the key climate factors influencing potato yield were estimated and the forecasting model of potato yield was built with the cor-relation and regression method based the meteorological data of 16 weather stations and the related data of potato develop-ment in 1980-2007. The results showed that:(1)The precipitation was the key factor to restrict potato production, and the temperature secondly. The effect of high temperature stress of Qianshan area was greater than that of Houshan area, while the effect of inadequate precipitation was on the contrary. Little precipitation and high temperature were the main limiting factors. (2)The meteorological yield prediction model was established by using stepwise regression method for four periods of time, including seedling to tuber formation, seedling to stareh accumulation, tuber formation to starch ac-cumnulation and the whole growing season, reached extremely significant level, with over 75% fitting rate, 11.1% yield estimation average error and 0.34%~27.9% amplitude error, 85% estimation accuracy beyond 80%, and the estima-tion result was better than each banner county area. (3)The tuber formation was the most sensitive period to water and temperature; Among different time models, better. The estimation results of growth season and seedling to tuber formation stage were better. The model can be used on potato yield estimation business.