晚上收到来自英国气象局同事的一个email，转发了澳大利亚气候信息中心气候资料专家的关于气候资料均一化的事情。近日，英国卫报报道了近日对澳大利亚气象局的气候数据的质疑，气候怀疑论者怀疑，气象局对气候资料做了均一化，导致澳大利亚气候变化更接近全球气候变暖说法，因此怀疑气象局宣布的”2013年是澳大利亚最暖年份“的真实性。澳大利亚气象局公开发表了声明，认为该气候怀疑论者的指责是错误的。

全文如下：

Contrary to assertions in some parts of the media, the Bureau is not altering climate records to exaggerate estimates of global warming.

Our role is to make meteorological measurements, and to curate, analyse and communicate the data for use in decision making and to support public understanding.

To undertake these tasks, the Bureau employs highly skilled technicians and scientists and invests in high quality monitoring equipment.

The Bureau measures temperature at nearly 800 sites across Australia, chiefly for the purpose of weather forecasting. The Australian Climate Observations Reference Network – Surface Air Temperature (ACORN-SAT) is a subset of this network comprising 112 locations that are used for climate analysis. The ACORN-SAT stations have been chosen to maximise both length of record and network coverage across the continent. For several years, all of this data has been made publicly available on the Bureau’s web site.

Temperature records are influenced by a range of factors such as changes to site surrounds (eg. trees casting shade or influencing wind), measurement methods and the relocation of stations (eg. from a coastal to more inland location). Such changes introduce biases into the climate record that need to be adjusted for prior to analysis.

Adjusting for these biases, a process known as homogenisation, is carried out by meteorological authorities around the world as best practice, to ensure that climate data is consistent through time.

At the Bureau’s request, our climate data management practices were subject to a rigorous independent peer-review in 2012. A panel of international experts found the Bureau’s data and methods were amongst the best in the world.

The Bureau’s submissions to the review were published on the Bureau’s website, as were the findings of the review panel.

The Bureau’s methods have also been published in peer-reviewed scientific journals.

Both the raw and adjusted ACORN-SAT data and the larger unadjusted national data set all indicate that Australian air temperatures have warmed over the last century. This finding is consistent with observed warming in the oceans surrounding Australia. These findings are also consistent with those of other leading international meteorological authorities, such as NOAA and NASA in the United States and the UK MetOffice. The high degree of similarity is demonstrated in Figure 1 (above).

The Bureau strives to ensure that its data sets and analysis methods are as robust as possible. For this reason we place considerable emphasis on quality assurance, transparency and communication. The Bureau welcomes critical analysis of the Australian climate record by others through rigorous scientific peer review processes.

声明大致是认为（本来我全文翻译了，结果没存盘，给弄丢了，这里应该吐槽一下科学网博文没有临时存盘功能！），气象局确保资料的权威性，澳大利亚气象局的资料（ACORN-SAT）是通过严格的同行评议的，一个独立的小组评议后发现，该数据集是最好的，其论文也发表在同行评议的杂志上，经得起推敲的。

这个数据集我还真了解，其作者是澳大利亚气象学会的主席，著名资料专家Blair Treiwin博士牵头研制的，并且该数据集被同行认为是第一个第二代（逐日）的均一化气温数据集，该论文发表在2013年的Int J climatol上。但在论文发表之前，的确聘请了一个国际小组（至少我知道其中包括了来自美国，加拿大的几个老朋友，国际气候界和气候资料均一性研究的大专家）对其整个过程进行了审议，评议认为，该数据集是可信的。

正好前一段在写一个材料，对气候资料均一性的讨论也很多。算起来，对气候资料均一性研究了10多年了，虽然没有什么重大发现和骄傲的成果，但自认对这个工作还算是比较了解的。所谓气候资料的非均一性，就是因为观测站点迁址，仪器换型、环境的突然变化等等因素导致长期序列出现了不连续现象。显然，这种序列是不能作为气候分析的基础的，于是就需要做均一化。

均一化并不好做，找到好的参照是最关键的，还要有一个非常好的统计方法，一个详尽的台站沿革信息。但实际的情况是，找不到理想的参考序列，各种统计方法各有千秋，台站记录往往不全，台站环境经常不是突然变化，而是逐渐变化（如城市化）。这些都是做这个工作的障碍。目前对于突然变化的非均一性的订正还好，有几个均一化的气温数据集出来，国际上也差不多。

均一化的质量更不好评估，因为它没有实况做对照，怎么订正、订正多少虽然叫”客观化“，那是因为用的是客观统计方法，但很多时候还是靠经验，哪些点应该保留，哪些不应保留，也要靠经验。因此，之前几个硕士研究生，我都不敢让他们做资料的均一化，因为怕他们时间不够，理解不到位，反而对资料产生伤害，变成带进新的”非均一性“了。审稿中，也发现，这类文章的差别太大了，有的还可以，有的简直粗糙得要命，错的简直离谱。更要命的是，目前好像真的没有什么好的”指标“来量化评估均一化数据产品的优劣，国际上还只是依靠”同行评议“，我们的做法就是尽量公开信息，把所有的信息，订正前的、订正后的，全都给你，你自己也参与评判好坏。因此，我们2013年发表在的JGR上的关于中国第二代均一化数据的论文，我认为审稿者是审得最细致的一篇文章。

均一化，双刃剑。做好了，使气候变化更精确，做差了，是歪曲。需要慎之又慎。公众不理解，太正常了，如果做这个行当的专业人员有时也经常混淆的一塌糊涂，这就有点说不过去。