天津棉花气候适宜度变化特征及其产量动态预报

Climatic Suitability Variation and Yield Dynamic Prediction Model of Cotton in Tianjin

通过参数修订建立了适合天津地区的棉花各生育期温、光、水气候适宜度评价模型,计算了历年天津棉花3种气候因子的适宜度指数,在分析其变化规律的基础上,建立了可以在棉花不同生育期进行的产量动态预报模型。结果表明,1964-2009年天津地区棉花生育期的气温适宜度指数呈上升趋势,增幅为0.016/10a;日照适宜度指数呈下降趋势,降幅为0.025/10a;降水适宜度指数的线性变化趋势不显著,但具有明显的阶段性特点。利用1995-2009年资料回归分析建立的产量动态预报模型,除播种期外,其它4个生育期(苗期、现蕾期、花铃期和吐絮期)模型均通过0.05水平的显著性检验;对模型进行历史资料回代表明,2000年以后准确率较高,87.5%的年份回代准确率在80%以上,各生育期平均准确率均在88%以上;利用模型对2010、2011年棉花产量进行动态预报验证,准确率均在95.5%以上,表明模型预报效果较好,能够满足业务服务需要。

A climatic suitability evaluation model of cotton in Tianjin was established, considering temperature, sunlight and water demand. The climatic suitable index of cotton was calculated with three climatic factors in the past years, a dynamic prediction model of cotton was established. The results showed that temperature suitable index （TSI） of cotton increased at the rate of 0. 016/10y from 1964 to 2009, sunlight suitable index （SSI）of cotton decreased at the rate of 0. 025/10y from 1964 to 2009, and precipitation suitable index （PSI）of cotton changed little in the past years. The dynamic prediction model of cotton yield was established by regression analysis based on the meteorological data from 1995 to 2009. The model for seeding stage, budding stage, boll forming stage and boll opening stage passed significance testing（P 〈 0. 05 ）. The result of running the model with historical data showed that it got a higher accuracy since 2000, in which 87.5% of years accuracy rate was more than 80% , and average rate in each growth stage was more than 88%. The accuracy rate of dynamic prediction model was more than 95.5% for 2010 and 201 l, which indicated that the model could meet the demand of agro - meteorological services.