Data for this research come from time series of monthly average temperatures from 28 sites over the Valle del Cauca of Colombia in South America, collected over the period 1971-2002. Because of thegeographical location of the study area, monthly average temperature is affected by altitude and El Ni?o-La Ni?a (El Ni?o-Southern Oscillation, or ENSO phenomenon). Time series for some of the sites show a tendency to increase. Also, because of the two dry and wet periods in the studyarea, a seasonal pattern of behavior in monthly average temperature is seen. Linear mixed models are formulated andfitted to account for withinand between-site variations. The ENSOphenomenon is modeled by the Southern Oscillation index (SOI) and dummy variables. Spatial and temporal covariance structures in the errors are modeled individually using isotropic variogram models. The fitted models demonstrate the influence of the ENSO phenomenon on monthly average temperatures; this is seen in the maps produced from the models for ENSO and normal conditions. These maps show the predicted spatial patterns for differences in temperature throughout the study area.
展开▼
机译：这项研究的数据来自1971年至2002年期间在南美哥伦比亚的Valle del Cauca上28个站点的月平均气温的时间序列。由于研究区域的地理位置，每月的平均温度受海拔高度和El Ni？o-La Ni？a（El Ni？o-南方涛动或ENSO现象）的影响。一些站点的时间序列显示出增加的趋势。另外，由于研究区域有两个干燥和潮湿时期，因此可以看到每月平均温度的季节性行为模式。制定并拟合了线性混合模型以解决站点内部和站点之间的变化。 ENSO现象是由南方涛动指数（SOI）和虚拟变量模拟的。误差中的时空协方差结构使用各向同性变异函数模型分别建模。拟合模型证明了ENSO现象对月平均温度的影响;从ENSO和正常条件的模型生成的地图中可以看到这一点。这些地图显示了整个研究区域温度差异的预测空间格局。
展开▼
