KNMI-LENTIS large ensemble time slice dataset description
<p><strong>1. Contents </strong></p>
<ul>
<li><strong>Available variables in KNMI-LENTIS</strong>
<ul>
<li>request-overview-CMIP-historical-including-EC-EARTH-AOGCM-preferences.txt</li>
</ul>
</li>
<li><strong>Where is the data deposited on the ECWMF's tape storage (section 4)</strong>
<ul>
<li>LENTIS_on_ECFS.zip </li>
</ul>
</li>
<li><strong>Data of all variables for 1 year for 1 ensemble member (section 5)</strong>
<ul>
<li>tree_of_files_one_member_all_data.txt</li>
<li>{AERmon,Amon,Emon,LImon,Lmon,Ofx,Omon,SImon,fx,Eday,Oday,day,CFday,3hr,6hrPlev,6hrPlevPt}.zip</li>
</ul>
</li>
</ul>
<p><strong>2. Description of this Zenodo dataset</strong></p>
<p>This Zenodo dataset pertains to the full KNMI-LENTIS dataset: a large ensemble of simulations with the Global Climate Model EC-Earth3. The periods are for the present-day period (2000-2009) and a future +2K period (2075-2084 following SSP2-4.5). KNMI-LENTIS has 1600 simulated years for both the two climates. This level of sampled climate variability allows for robust and in-depth research into extreme events. The available variables are listed in the file <strong>request-overview-CMIP-historical-including-EC-EARTH-AOGCM-preferences.txt. </strong>All variables are cmorised following CMIP6 data format convention. Further details on the variables and their output dimensions is available via the<a href="https://clipc-services.ceda.ac.uk/dreq/mipVars.html"> following search tool</a>. The total size of KNMI-LENTIS is 128 TB. KNMI-LENTIS is stored at the <a href="https://www.ecmwf.int/en/computing/our-facilities/data-handling-system">high performance storage system of the ECMWF (ECFS)</a>. </p>
<p>The Global Climate Model that is used for generating this Large Ensemble is EC-Earth3 - VAREX project branch <a href="https://svn.ec-earth.org/ecearth3/branches/projects/varex">https://svn.ec-earth.org/ecearth3/branches/projects/varex</a> (access restricted to ECMWF members).</p>
<p>The goal of this Zenodo dataset is :</p>
<ol>
<li>to provide an accurate description and example of how the KNMI-LENTIS dataset is organised. </li>
<li>to describe in which servers the data are deposited and how to gain access to the data for future users</li>
<li>to provide links to related git repositories and other content relating to the KNMI-LENTIS production</li>
</ol>
<p><strong>3. How KNMI-LENTIS is </strong><strong>organised</strong></p>
<p>KNMI-LENTIS consists of 2 times 160 runs of 10 years. All simulations have a unique ensemble member label that reflects the forcing, and how the initial conditions are generated. The initial conditions have two aspects: the parent simulation from which the run is branched (macro perturbation, there are 16), and the seed relating to a particular micro-perturbation in the initial three-dimensional atmosphere temperature field (there are 10). The ensemble member label thus is a combination of: </p>
<ul>
<li>forcing (<em>h</em> for present-day/historical and <em>s</em> for +2K/SSP2-4.5)</li>
<li>parent ID (number between 1 and 16)</li>
<li>micro perturbation ID (number between 0 and 9)</li>
</ul>
<p>In this Zenodo dataset we publish 1 year from 1 member to give insight into the type of data and metadata that is representative of the full KNMI-LENTIS dataset. The published data is year 2000 from member <em>h010</em>. See Section 4 </p>
<p>Further, all KNMI-LENTIS simulations are labeled per the CMIP6 convention of variant labelling. A variant label is made from four components: the realization index <em>r</em>, the initialization index <em>i</em>, the physics index <em>p</em> and the forcing index <em>f</em>. Further details on CMIP6 variant labelling be found in <a href="https://pcmdi.llnl.gov/CMIP6/Guide/modelers.html">The CMIP6 Participation Guidance for Modelers</a>. In the KNMI-LENTIS data set, the forcing is reflected in the first digit of the realization index <em>r </em>of the variant label. For the historical simulations, the one thousands (r1000-r1999) have been reserved. For the SSP2-4.5 the five thousands (r5000-r5999) have been reserved. The parent is reflected in the second and third digit of the realization index <em>r</em> of the variant label (r?01?-r?16?). The seed is reflected in the fourth digit of the realization index <em>r</em>: (r???0-r???9). The seed is also reflected in the initialization index <em>i</em> of the variant label (i0-i9), so this is double information. The physics index <em>p5</em> has been reserved for the ECE3p5 version: all KNMI-LENTIS simulations have the <em>p5</em> label. The forcing index <em>f </em>of the variant label is kept at 1 for all KNMI-LENTIS simulations. As an example, variant label r5119i9p5f1 refers to: the 2K time slice with parent 11 and randomizing seed number 9. The physics index is 5, meaning the run is done with the ECE3p5 version of EC-Earth3. </p>
<p><strong>4. Where is the data deposited on the ECWMF's tape storage</strong></p>
<p>In this Zenodo folder, there are several text files and several netcdf files. The text files provide</p>
<p>Data from KNMI-LENTIS is deposited in the <a href="https://www.ecmwf.int/en/computing/our-facilities/data-handling-system">ECMWF ECFS tape storage system</a>. Data can be freely downloaded by to those who have access to the ECMWF ECFS. Else, the data can be made available by the authors upon request. </p>
<p>The way the dataset is organised is detailed in <strong>LENTIS_on_ECFS.zip. </strong>This contains details on all available KNMI-LENTIS files, in particular details for how these are filed in ECFS. The files on ECFS are tar zipped per ensemble member & variable: these contain 10 years of ensemble member data (10 separate netcdf files). The location on ECFS of the tar-zipped files that are listed in the various text files in this Zenodo dataset is</p>
<p>ec:/nklm/LENTIS/ec-earth/cmorised_by_var/ </p>
<pre><code class="language-bash">#!/bin/bash
#-------------------
# script to write out LENTIS details on ECFS
#-------------------
for freq in AERmon Amon Emon LImon Lmon Ofx Omon SImon fx Eday Oday day CFday 3hr 6hrPlev 6hrPlevPt; do
for scen in hxxx sxxx; do
els -l ec:/nklm/LENTIS/ec-earth/cmorised_by_var/${scen}/${freq}/* >> LENTIS_on_ECFS_${scen}_${freq}.txt
done
done</code></pre>
<p>Further, part of the data will be made publicly available from the <a href="https://esgf-node.llnl.gov/projects/cmip6/">Earth System Grid Federation (ESGF) data portal.</a> We aim to upload most of the monthly variables for the full ensemble. As search terms use <strong>EC-Earth</strong> for model and <strong>p5</strong> for physical index to locate the KNMI-LENTIS data. </p>
<p><strong>5. Data of all variables for 1 year for 1 ensemble member</strong></p>
<p>The netcdf files of the data of 1 year from 1 member <em>h010 </em>are published here to give insight into the type of data and metadata that is representative of the full KNMI-LENTIS dataset. The data are in zipped folders per output frequencies: AERmon, Amon, Emon, LImon, Lmon, Ofx, Omon, SImon, fx, Eday, Oday, day, CFday, 3hr, 6hrPlev, 6hrPlevPt. The text file <strong>request-overview-CMIP-historical-including-EC-EARTH-AOGCM-preferences.txt </strong> gives an overview of variables available per output frequency. the text files <strong>tree_of_files_one_member_all_data.txt </strong>gives an overview of the files in the zipped folders. </p>
<p><strong>6. Related links</strong></p>
<p>The production of the KNMI-LENTIS ensemble was funded by the KNMI (Royal Dutch Meteorological Institute) multi-year strategic research fund <a href="https://www.knmi.nl/research/weather-climate-models/projects/mso-climate-variability-and-extremes-varex">KNMI MSO Climate Variability And Extremes (VAREX)</a></p>
<p>GitHub repository corresponding to this Zenodo dataset: <a href="https://github.com/lmuntjewerf/KNMI-LENTIS_dataset_description.git">https://github.com/lmuntjewerf/KNMI-LENTIS_dataset_description.git </a></p>
<p>Github repository for KNMI-LENTIS production code: <a href="https://github.com/lmuntjewerf/KNMI-LENTIS_production_script_train.git">https://github.com/lmuntjewerf/KNMI-LENTIS_production_script_train.git</a></p>
opencc-by-4.0Jan 2023View details →