日光温室地温预报技术研究

根据山东省莱芜市2007～2011 年冬季（12 月至翌年2 月）日光温室内、外气象观测资料，
分析了晴天、多云天、阴天不同天气条件下日光温室内地温变化特征及其与温室内外气象要素的关系；以
温室内、外气象要素为启动因子，构建了未来3 d 10 cm 逐时地温预报模型。结果表明：温室内地温与气
温关系密切，其中当日及前期室外最高气温、最低气温对10 cm 逐时地温影响较大；未来3 d 10 cm 逐时
地温预报模型中，晴天、阴天、多云天预报模型RMSE（均方根误差）分别为0.7、0.6、0.8 ℃，阴天预
报模型精度较晴天、多云天高；00：00～07：00、08：00～17：00、18：00～23：00 3 个时段预报模型
RMSE 分别为0.6、0.7、0.6 ℃，00：00～07：00、18：00～23：00 时段预报模型精度较08：00～17：00
时段高；未来1、2、3 d 预报模型RMSE 为0.5、0.6、0.8 ℃，预报精度随预报时效延长逐渐降低。

Studies on Geothermal Temperature Forecasting Technology in Solar Greenhouse

According to the meteorological data observed inside and outside the solar greenhouse
（December to next February）from 2007 to 2011，this paper analyzes the geothermal temperature changing
characteristic inside and outside the solar greenhouse，and its relationship with meteorological elements
inside and outside the solar greenhouse under different weather conditions：sunny，overcast and cloudy
days．Taking the metrological factors inside and outside of solar greenhouse as promoter，we constructed
a hourly geothermal temperature of 10 cm forecast model for 3 days in the future．The results indicated
that the geothermal temperature inside the solar greenhouse was closely related to the temperature．
The geothermal temperature of 10 cm inside the greenhouse was highly correlated with the lowest and
highest temperatures outside the solar greenhouse on that day and the day before；and in the 3 days
hourly geothermal temperature of 10 cm forecast model，the forecast model RMSE for sunny，overcast
and cloudy days were 0.7，0.6，0.8 ℃，respectively. The forecast model accuracy for overcast day is
higher than that for sunny and cloudy days．The forecast model RMSE of 00：00-07：00，08：00-
17：00 and 18：00-23：00 were 0.6，0.7，0.6 ℃，respectively．And the forecast model accuracy of
00：00-07：00，18：00-23：00 were higher than that of 08：00-17：00．The forecast model RMSE
for the next day，the next 2 days and the next 3 days were 0.5，0.6，0.8 ℃，and the forecast accuracy
was gradually decreased with the prediction time extension.