Frequently, the computer modelling of the natural and human-made environment requires localised weather files. Traditionally, the weather files are based on the observed weather at a small number of locations (14 for the UK). Unfortunately, both the climate and the weather are known to be highly variable across the landscape, so the small number of locations has the potential to cause large errors. With respect to buildings, this results in incorrect estimates of the annual energy use (sometimes by a factor of 2), or of overheating risk. Here we use a validated weather generator running on a 5 × 5 km grid to create probabilistic test reference years (pTRYs) for the UK at 11,326 locations. We then investigate the spatial variability of these pTRYs and of annual energy estimates and temperatures in buildings generated by them, both now and in 2080. Further pTRYs targeted at understanding the impact of minimum and maximum temperatures are proposed and produced at the same locations. Finally, we place these pTRYs, which represent the first set of reference weather files at this spatial resolution in the world and that include the urban heat island effect, into a publicly accessible database so researchers and industry can access them. Practical applications: Insufficiently localised weather data for building simulations have limited the accuracy of previous estimations of energy use and overheating risk in buildings. This work produces localised probabilistic test reference years (pTRYs) across the whole UK for now and future climates. In addition, a new pTRY method has been proposed in order to overcome an unexpected shortcoming of traditional pTRYs in representing typical maximum and minimum temperatures. These current and future weather data will be of interest to various disciplines including those interested in low carbon design, renewable energy and climate resilience.

中文翻译：

---

当前和将来的测试参考年（分辨率为5 km）

通常，对自然和人为环境进行计算机建模需要本地化的气象文件。传统上，天气文件基于少量地点（英国为14个）的观测天气。不幸的是，众所周知，气候和天气在整个景观中变化很大，因此，少数几个位置可能会导致较大的误差。对于建筑物，这会导致对年度能源使用量的估计错误（有时是2倍）或过热风险。在这里，我们使用在5×5 km网格上运行的经过验证的天气生成器，为英国的11,326个位置创建概率测试参考年（pTRY）。然后，我们调查了现在和2080年这些pTRY的空间变异性，年度能源估算值以及它们产生的建筑物的温度。在同一地点提出并生产了旨在了解最低和最高温度影响的其他pTRY。最后，我们将这些pTRY放置到一个公共可访问的数据库中，这些pTRY代表了世界上具有这种空间分辨率的第一组参考天气文件，其中包括城市热岛效应。实际应用：用于建筑物模拟的本地化天气数据不足，限制了先前对建筑物能源使用和过热风险的估算的准确性。这项工作为整个英国提供了针对当前和未来气候的本地化概率测试参考年（pTRY）。此外，为了克服传统pTRY在代表典型的最高和最低温度方面的意料之外的缺点，已经提出了一种新的pTRY方法。这些当前和未来的天气数据将吸引各个学科的兴趣，包括那些对低碳设计，可再生能源和气候适应力感兴趣的学科。