气象资料检测证明了气候模型模拟现代温度比用更复杂的湿度平衡模型好得多，而根据地质资料在山区和沿海地区重建和模型模拟的可信度要低一些，所以在比较地质资料和过去气候模拟时要考虑这个实际情况，在当前研究中我们进行末次大冰期资料一模型比较时以年均温为例，在使用所有点的资料时出现了模拟和重建之间的不一致，相反，当去掉那些模棱两可的地点后结果变得意外的好，在少数情况下结果不同时，边界条件不确定或者资料太少是模型和资料不一致的主要原因，通过使用模拟和资料的区域性平均值，我们证明了PMIP能够正确地模拟出地球陆地上区域性的年均温，在区域性研究中因经常使用未经选择的资料作为模型确认值，因而很难获得这一重要的结论。

Tests with the meteorological data demonstrate that climate models simulate modern temperature much better than more complex patterns of the moisture balance. However, the confidence level of the model simulations and reconstruction from geological records is relatively low in mountain and coastal regions. This reality should be taken into account when the comparisons between geological data and past climate simulations are performed. In the present study we performed data-model comparison for LGM with an example of mean annual temperature. The use of all available sites shows many discrepancies between simulations and reconstructions. However, results become surprisingly good when ambiguous sites are excluded from the comparison. In the few cases, when simulated and reconstructed results are different, uncertainties in the boundary conditions are considered as main factor for such discrepancies. By using regional averages of the simulated and reconstructed values we demonstrated that PMIP models correctly simulate annual temperature at the regional scale across the Earth's land. This important conclusion is difficult to get from regional studies that often use unselected data for the model validation.

Q913.84 P532