<div>Hi Dennis,</div><div><br></div><div>Thank you! The differences do relate to weights.</div><div>When both calculating without (or with) weights, the two outputs are exactly the same.</div><div><div><br></div><div><br></div><div style="font-size: 12px;font-family: Arial Narrow;padding:2px 0 2px 0;">------------------&nbsp;Original&nbsp;------------------</div><div style="font-size: 12px;background:#efefef;padding:8px;"><div><b>From: </b>&nbsp;"Dennis Shea";&lt;shea@ucar.edu&gt;;</div><div><b>Date: </b>&nbsp;Fri, Nov 20, 2015 12:51 PM</div><div><b>To: </b>&nbsp;"LI Qi"&lt;liqi123sh@qq.com&gt;; <wbr></div><div><b>Cc: </b>&nbsp;"ncl-talk"&lt;ncl-talk@ucar.edu&gt;; <wbr></div><div><b>Subject: </b>&nbsp;Re: [ncl-talk] different RMSE values calculated by NCL and CDO</div></div><div><br></div>I don't have time to look at this right now. So, some speculation ....<br><br>[1] NCL's&nbsp;&nbsp;&nbsp;&nbsp; dim_rmsd_n&nbsp; does no&nbsp; weighting (no areal weighting).<br>&nbsp;&nbsp;&nbsp;&nbsp; It is just arithmetic.<br><br>[2] I speculate that the CDO does some areal weighing (say, cos(lat)<br>of the differences)<br><br><br><br>==<br>NCL has the following functions that may be better for comparing.<br><br>http://www.ncl.ucar.edu/Document/Functions/Built-in/wgt_arearmse.shtml<br>http://www.ncl.ucar.edu/Document/Functions/Built-in/wgt_arearmse2.shtml<br><br>Good luck<br><br>On Thu, Nov 19, 2015 at 7:04 AM, LI Qi &lt;liqi123sh@qq.com&gt; wrote:<br>&gt; Dear NCL community,<br>&gt;<br>&gt; I would like to calculate spatial RMSE of simulations and observations.<br>&gt;<br>&gt; the NCL function I use is:<br>&gt; rmse = dim_rmsd_n(obs, sim, (/1,2/))<br>&gt; where (/1,2/) represents the spatial dimensions lon,lat<br>&gt;<br>&gt;<br>&gt; while the CDO command is:<br>&gt; cdo sqrt -fldmean -sqr -sub sim.nc obs.nc rmse.nc<br>&gt;<br>&gt; However, the CDO output (64.44424) is different from the NCL one (63.95646),<br>&gt; and I couldn't figure out why.<br>&gt;<br>&gt; Attached are the two .nc files (sim var=rain_tot(time, lat, lon); obs<br>&gt; var=crain(time, lat, lon))<br>&gt;<br>&gt; Any hint is much appreciated!<br>&gt;<br>&gt; Best,<br>&gt; Qi<br>&gt;<br>&gt; _______________________________________________<br>&gt; ncl-talk mailing list<br>&gt; ncl-talk@ucar.edu<br>&gt; List instructions, subscriber options, unsubscribe:<br>&gt; http://mailman.ucar.edu/mailman/listinfo/ncl-talk<br>&gt;<br></div>