Chan Duo, Gebbie Geoffrey, Huybers Peter, Kent Elizabeth C
School of Ocean and Earth Science, University of Southampton, Southampton, UK.
Department of Physical Oceanography, Woods Hole Oceanographic Institution, Woods Hole, 02543, USA.
Sci Data. 2024 Aug 30;11(1):953. doi: 10.1038/s41597-024-03742-x.
Accurate historical records of Earth's surface temperatures are central to climate research and policy development. Widely-used estimates based on instrumental measurements from land and sea are, however, not fully consistent at either global or regional scales. To address these challenges, we develop the Dynamically Consistent ENsemble of Temperature (DCENT), a 200-member ensemble of monthly surface temperature anomalies relative to the 1982-2014 climatology. Each DCENT member starts from 1850 and has a 5° × 5° resolution. DCENT leverages several updated or recently-developed approaches of data homogenization and bias adjustments: an optimized pairwise homogenization algorithm for identifying breakpoints in land surface air temperature records, a physics-informed inter-comparison method to adjust systematic offsets in sea-surface temperatures recorded by ships, and a coupled energy balance model to homogenize continental and marine records. Each approach was published individually, and this paper describes a combined approach and its application in developing a gridded analysis. A notable difference of DCENT relative to existing temperature estimates is a cooler baseline for 1850-1900 that implies greater historical warming.
准确的地球表面温度历史记录是气候研究和政策制定的核心。然而,基于陆地和海洋仪器测量的广泛使用的估计值在全球或区域尺度上并不完全一致。为应对这些挑战,我们开发了动态一致温度集合(DCENT),这是一个由200个成员组成的相对于1982 - 2014年气候学的每月表面温度异常集合。每个DCENT成员从1850年开始,分辨率为5°×5°。DCENT利用了几种更新或最近开发的数据均一化和偏差调整方法:一种用于识别陆地表面气温记录断点的优化成对均一化算法、一种基于物理知识的相互比较方法来调整船舶记录的海表面温度中的系统偏差,以及一个耦合能量平衡模型来均一化陆地和海洋记录。每种方法都已单独发表,本文描述了一种组合方法及其在开发网格化分析中的应用。DCENT相对于现有温度估计的一个显著差异是1850 - 1900年的基线温度较低,这意味着历史变暖幅度更大。