Datasets for RHALS with EW Paper

This data contains measurements of the mass spectra of highly oxidized multifunctional molecules (HOMs) from the paper “Ambient Measurements of Highly Oxidized Gas-PhaseMolecules during the Southern Oxidant and Aerosol Study (SOAS) 2013” by Massoli et. al (2018). These measurements were taken at a forest site in Alabama in the summer of 2013. Positive Matrix Factorization (PMF) was run on the data, finding 6 underlying factors to the data, which contained anthropogenic components. The data consists of a large 27,336 by 1059 matrix, where the rows correspond to the time stamps, and the columns correspond to the mass spectra. There is an adjacent uncertainty matrix of the same size, where each entry corresponds with the estimated uncertainty (standard deviation) of each measurement. The specific timestamps and mass spectra compounds are also given. Additionally, the six factor solution described in the paper is presented as both its time series and mass spectra components. Taking a linear combination of the outer products between corresponding estimates will give a low rank approximation of the original data matrix.

本数据集收录了Massoli等人2018年发表的《2013年南部氧化剂与气溶胶研究（SOAS）期间大气高氧化态气相分子原位测量》论文中的高氧化多官能团分子（highly oxidized multifunctional molecules, HOMs）质谱测量数据。该测量工作于2013年夏季在阿拉巴马州的一处森林站点完成。研究人员对该数据集应用正矩阵分解（Positive Matrix Factorization, PMF）方法，从数据中解析出6个包含人为源组分的潜在因子。数据集主体为规模达27336×1059的矩阵，其中行对应时间戳，列对应质谱。数据集附带一个相同维度的配套不确定度矩阵，矩阵内每个元素对应单次测量的估计不确定度（标准差，standard deviation）。本数据集同时提供了具体的时间戳与质谱化合物信息。此外，论文中所述的6因子解析结果以时间序列与质谱组分两种形式予以呈现。对对应估计值的外积进行线性组合，即可得到原始数据矩阵的低秩近似。