Arctic Atmospheric River Labels and Climatology Based on 3-hourly ERA5 and MERRA-2 From 1980 to 2019

<div class="pub-content">
<div class="pub-content">
<p><strong>A. Background</strong></p>

<p>In "Climatology and decadal changes of Arctic atmospheric rivers based on ERA5 and MERRA-2" by Zhang, Tung, and Cleveland (Environmental Research: Climate, 2023, <a href="https://doi.org/10.1088/2752-5295/acdf0f" rel="nofollow noopener noreferrer noopener noreferrer" target="_blank">https://doi.org/10.1088/2752-5295/acdf0f</a>), we analyzed Arctic atmospheric river (AR) patterns from 1980-2019 using twelve sets of AR labels derived from the ERA5 and MERRA-2 reanalyses. Employing a multifactorial approach, ARs were identified by analyzing vertically integrated water vapor and its transport in the atmosphere using three climate thresholds (75th, 85th, and 95th percentiles). Our findings indicate an upward trend in AR events since the mid-1990s. Labels based on the 75th percentile were more sensitive to Arctic surface warming and revealed an expansion of AR-affected areas. Furthermore, the maximum frequency of ARs shifted towards the Arctic Ocean from 2000-2019. There was a notable increase in AR occurrence, surface temperature, and moisture across the Atlantic, Arctic, and Pacific Oceans. ARs also contributed to increased moisture transport into the Arctic. While the rise in AR occurrence was mainly linked to long-term Arctic surface warming and moistening, local atmospheric circulation changes were also significant, such as over the Chukchi Sea. Teleconnection patterns influenced AR activity. The extreme events identified by the 95th-percentile labels displayed the most significant changes and were most influenced by the teleconnection. The AR labels and graphics presented in this study can aid in understanding and assessing Arctic ARs and their impacts in ongoing and future research.</p>

<p><strong>B. Contents of Repository</strong></p>

<p><strong>1. Readme File</strong></p>

<p>The file <em>Readme.txt</em> is the document about each of the twelve datasets for Arctic Atmospheric River Labels and the Supplementary Figures.</p>

<p><strong>2. Twelve Sets of Arctic Atmospheric River Labels  </strong></p>

<p>These are binary (0 or 1) labels indicating the presence of atmospheric river at a latitude-longitude location from 35◦ N to the North Pole. The spatial resolution of the ERA5-based labels is 0.25<sup>◦ </sup>longitude × 0.25<sup>◦ </sup>latitude. The spatial resolution of the MERRA-2-based labels is 0.625<sup>◦</sup> ×0.5<sup>◦</sup>. The temporal range is from 1980 to 2019, with a 3-hourly temporal resolution for all labels.</p>

<p>The AR detection and tracking algorithm followed a similar multifactorial design as <a href="http://doi.org/10.1029/2020JD033667" rel="nofollow noopener noreferrer noopener noreferrer" target="_blank">Zhang, Tung, and Cleveland (2021)</a>, in which we found that the choices of moisture fields and climate thresholds strongly dictated the summary statistics of the identified ARs and AR-related surface hydrometeorological effects. In this work, we created six sets of Arctic AR labels from ERA5 reanalysis and six sets from MERRA-2 reanalysis, a total of twelve sets of labels from January 1980 to December 2019. The Figure 2 in <a href="https://doi.org/10.1088/2752-5295/acdf0f" rel="nofollow noopener noreferrer noopener noreferrer" target="_blank">Zhang, Tung, and Cleveland (2023)</a>, as shown above, summarizes the Arctic AR detection and tracking algorithm, which is further detailed in the appendix A of the paper.</p>

<p>The AR detection and tracking, and the ensuing statistical analysis were performed via distributed-parallel computing, especially the divide-and-recombine approach using the open-source R version 3.4 (<a href="https://www.R-project.org/" rel="nofollow noopener noreferrer noopener noreferrer" target="_blank">R Core Team 2019</a>) and the R-Hadoop Integrated Programming Environment (Rhipe, <a href="http://www.datadr.org" rel="nofollow noopener noreferrer noopener noreferrer" target="_blank">Guha 2018</a>), back-ended by a Hadoop system version 2.7.4 (<a href="https://doi.org/10.1002/sam.11242" rel="nofollow noopener noreferrer noopener noreferrer" target="_blank">Cleveland and Hafen 2014</a>, <a href="https://doi.org/10.1007/s42081-018-0008-4" rel="nofollow noopener noreferrer noopener noreferrer" target="_blank">Tung et al 2018</a>).</p>

<p>The names of these files follow a format convention of: <em><strong>AAAAA</strong>_<strong>BB</strong>th<strong>CCC</strong>.tar</em>, where:<br />
<em><strong>AAAAA</strong></em> is the name of the reanalysis data source, either ERA5 or MERRA2.<br />
<em><strong>BB</strong></em> is the value of a climate threshold, a number being 75, 85, or 95.<br />
<em><strong>CCC</strong></em> is the moisture field based on which the climate threshold was applied, either ivt or iwv.</p>

<p><strong>3. Supplementary Figures</strong></p>

<p>An extensive set of figures (Figures S1 to S85) that visualizes and compares the AR labels is presented in the Supplementary Figures file available <a href="https://doi.org/10.1088/2752-5295/acdf0f" rel="nofollow noopener noreferrer noopener noreferrer" target="_blank">online</a>. Due to the size limitation at the journal site, only a coarse resolution document is provided.</p>

<p>Here, we provide the individual figure files in high resolution, archived in a <em>Supplementary_Figures.tar</em> file. These figures support the analysis and discussions in the section 3 of <a href="https://doi.org/10.1088/2752-5295/acdf0f" rel="nofollow noopener noreferrer noopener noreferrer" target="_blank">Zhang, Tung, and Cleveland (2023)</a> on AR occurrence frequency, moisture, moisture transport, trends, and modulation by the teleconnections with internal climate variability.</p>

<p><strong>C. References</strong></p>

<p>Cleveland W S and Hafen R 2014, Divide and recombine (D&R): Data science for large complex data. Stat. Anal. Data Min. 7 425–33 https://doi.org/10.1002/sam.11242</p>

<p>Guha S 2018, Rhipe: R and Hadoop Integrated Programming Environment R package (version 0.75.2) (available at: http://www.datadr.org)</p>

<p>R Core Team 2019, R: a language and environment for statistical computing (Vienna, Austria: R Foundation for Statistical Computing) (available at: www.R-project.org/)</p>

<p>Tung W-W, Barthur A, Bowers M C, Song Y, Gerth J and Cleveland W S 2018, Divide and recombine (D&R) data science projects for deep analysis of big data and high computational complexity. Japan. J. Stat. Data Sci. 1 139–56 https://doi.org/10.1007/s42081-018-0008-4</p>

<p>Zhang C, Tung W-wen and Cleveland W S 2021, In search of the optimal atmospheric river index for US precipitation: a multifactorial analysis. J. Geophys. Res. 126, 10, e2020JD033667 https://doi.org/10.1029/2020JD033667</p>

<p>Zhang C, Tung W-wen and Cleveland W S 2023, Climatology and decadal changes of Arctic atmospheric rivers based on ERA5 and MERRA-2. Environ. Res.: Climate https://doi.org/10.1088/2752-5295/acdf0f</p>
</div>
</div>