中文摘要：

通过对赤道地区两台站的扩展F实际观测值与IRI-2007模型预测值的比较，研究了该模型预测结果的地域局限性和预测的准确度．该模型建立在美洲扇区赤道附近台站资料基础上，本文选取非洲扇区大体同纬度两个台站资料进行对比研究．统计结果表明，无论布发生概率日均值的年变化趋势上，还是布扩展F发生率年均值随地方时变化以及随太阳活动的统计中，都存在着明显的差异；而对非洲两站实测值的分析和IRI-2007模型对南美巴西地区的预测又分别与当前已有的研究结果相符．分析结果表明，（1）建立在巴西台站数据基础上的IRI-2007模型的统计建模极具参考价值；（2）赤道和低纬度电离层扩展F现象有强烈的经度效应，在该模型建模方法基础上，结合不同扇区实际资料的统计，能改进该模型全球应用的普适性；（3）IRI-2007模型能正确反映巴西地区扩展F的发生概率，用该模型和其他扇区不同条件下的实测值对比，提供了一个细致研究扩展F经度效应的基础，有利于确认导致扩展F发生的基本条件和多种影响因素的具体作用．

英文摘要：

Ionospheric spread-F is a frequently-seen ionospheric phenomenon which plays an important role in the study of ionospheric physics. The latest IRI-2007 has extended a new function of spread-F occurrence prediction, which is based on the observational data of some low latitude stations around American sectors. The purpose of this article focuses on the accuracy of the model to see if there is regional restriction by comparing the actual observation data of spread-F occurrence at African sectors with prediction made from this model. The spread-F occurrence prediction model is based on the data of stations in American sector near the equator, while in this article we choose the data of the stations with almost the same latitude in African sector. The results show that there are obvious differences between the observations of these two African station and predictions made by the IRI-2007 model in the annual behavior of average daily occurrence rate, local time variations of annual average values and the trend with solar activity. On the other hand, the observational variations in the American stations of low latitudes and that of IRI-2007 model predictions are fairly in agreement. These results indicate that： （1） the statistical method of IRI-2007 spread-F prediction model based on the Brazilian stations＇ data is very good and of great reference value because it well reflects the occurrence of Spread-F in American Sector; （2） the equatorial and low latitude ionospheric spread-F phenomena show strong longitudinal effect, and the use of this model in regions other than American sector should be very careful. With different region＇s observational data, the universal usability of this model could be improved based on the same statistical method used in this model; （3） it provides a tool to study the longitudinal effects of spread-F by comparing the model predicted values and observational values under various conditions .