<div dir="ltr"><div>Yes. No significance would be my assessment.</div><div><br></div><div>I speculate there are refereed papers on this topic.</div><div><br></div><div> <br></div></div><br><div class="gmail_quote"><div dir="ltr">On Mon, Nov 26, 2018 at 12:48 AM Sujata Mandke <<a href="mailto:amin@tropmet.res.in">amin@tropmet.res.in</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div style="font-family:arial,helvetica,sans-serif;font-size:12pt;color:#000000"><div><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">Respected Dennis Shea sir,</span></span></span></span></p><p class="m_603853978113569708western" style="margin:0px;font-weight:normal;line-height:100%;text-decoration:none" align="justify"><br></p><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">I have included all your suggestions in the ncl script (modified ncl script is attached). </span></span></span></span></p><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal"><br></span></span></span></span></p><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">The correlation plot using this ncl script is also attached.</span></span></span></span></p><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal"><br></span></span></span></span></p><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">For statistical significance - </span></span></span></span><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">T</span></span></span></span><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">he estimate of number of independent values</span></span></span></span></p><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal"> using "<a href="http://www.ncl.ucar.edu/Document/Functions/Built-in/equiv_sample_size.shtml" target="_blank">equiv_sample_size</a>" for snow over Eurasia </span></span></span></span><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">=64 </span></span></span></span><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">and for SST over </span></span></span></span></p><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">East Pacific </span></span></span></span><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">=7 </span></span></span></span><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">. Minimum of the two, that is 7 is used for statistical significance </span></span></span></span></p><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">testing using "rtest". Probability values representing statistical significance of </span></span></span></span></p><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">correlation (for maximum lag=12) are high (ascii file for the same is </span></span></span></span></p><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">attached for total lag=12*2+1) and therefore none is </span></span></span></span></p><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">significant (not even near significance). </span></span></span></span></p><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal"><br></span></span></span></span></p><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">Does this indicate that none of the correlation is statistically</span></span></span></span></p><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal"> significant or there is a problem in the way it is estimated?</span></span></span></span></p><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span style="text-decoration:none"> <span lang="zxx"><span style="font-weight:normal">With best regards</span></span></span></span></p><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">Dr. Sujata Mandke</span></span></span></span></p><p class="m_603853978113569708western" style="margin:0px;line-height:100%" align="justify"><span style="color:#000080" color="#000080"><span lang="zxx"><span style="text-decoration:none"><span style="font-weight:normal">scientist, IITM,PUNE, INDIA</span></span></span></span></p><br></div><div><br></div><hr id="m_603853978113569708zwchr"><div><b>From: </b>"Dennis Shea" <<a href="mailto:shea@ucar.edu" target="_blank">shea@ucar.edu</a>><br><b>To: </b>"S.Amin" <<a href="mailto:amin@tropmet.res.in" target="_blank">amin@tropmet.res.in</a>><br><b>Cc: </b>"ncl-talk" <<a href="mailto:ncl-talk@ucar.edu" target="_blank">ncl-talk@ucar.edu</a>><br><b>Sent: </b>Wednesday, November 21, 2018 3:34:30 AM<br><b>Subject: </b>Re: [ncl-talk] plot lead/lag correlation with statistical significance<br></div><br><div><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div>My opinion:</div><br><div>[1] You should remove the climatological annual cycle from each dataset.</div><br><div><a href="http://www.ncl.ucar.edu/Document/Functions/Contributed/clmMonTLL.shtml" target="_blank"><b>http://www.ncl.ucar.edu/Document/Functions/Contributed/clmMonTLL.shtml</b></a><br></div><div><a href="http://www.ncl.ucar.edu/Document/Functions/Contributed/calcMonAnomTLL.shtml" target="_blank"><b>http://www.ncl.ucar.edu/Document/Functions/Contributed/calcMonAnomTLL.shtml</b></a><br></div><br><br><div>[2] Typically, the 'swe' file contains a _FillValue (-999.0) when there is no snow cover. I suggest setting this  to 0.0.</div><br><div>   sn= in1->swe(iStrt:iLast,{50:70},{20:140})  ;only specific region averaged over total Eurasia<br>   sneu r=<b> wgt_areaave_Wrap</b>(sn,1.0,1.0,0);;;create area averaged snow over Eurasia<br>   sneur = <b>where</b>(<b>ismissing</b>(sneur), 0.0, sneur)</div><div>   <b>printVarSummary</b>(sneur)<br>   p<b>rintMinMax(</b>sneur,0)<br>   print("---")<br>   ymd    := <b>cd_calendar</b>(sneur&time, -2)  ; yyyymmdd<br>   print(ymd+"  "+sneur)<br>   print("---")<br></div><br><div><div>[3] You *know* the climatology. You want to know the feedback of sst/sneur anomalies on each other.</div><div>Use the anomalies in the correlation calculations.</div><br></div><div>[4] There are statistical issues. EG: the number of independent values to be used in the testing.</div><br><div>Typically, successive monthly values of SST or SNE are not independent. If (say) a January SNEUR is very large, then the February SNE will likely be large also. So estimating the number of independent values is an issue. <br></div><br><div> <b> <a href="http://www.ncl.ucar.edu/Document/Functions/Built-in/equiv_sample_size.shtml" target="_blank">http://www.ncl.ucar.edu/Document/Functions/Built-in/equiv_sample_size.shtml</a></b><br></div><br><div>Estimate for the SNEUR/SST anomalies separately. Use the smaller value.</div><br><div>[5] I am sure the IITM has staff members who know more about statistical testing/inference than ncl-talk. I suggest talking with them.</div><br><div>Good luck<br></div></div></div></div></div></div></div><br><div class="gmail_quote"><div dir="ltr">On Tue, Nov 20, 2018 at 6:13 AM Sujata Mandke <<a href="mailto:amin@tropmet.res.in" target="_blank">amin@tropmet.res.in</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Dear NCL community,<br>
Greetings!<br>
 I have ploted lag/lead correlation (cc) between two area-averaged <br>
(Eurasian snow and East Pacific SST) monthly time series as xy plot (line).  <br>
Further, I want to indicate which part of cc is statistically <br>
significant on this line using line markers. <br>
<br>
I had tested statistical significance of cc using “rtest”. <br>
 In the ncl script(attached),  I have plotted  line for<br>
 lead/lag correlation. How to mark part of this cc (line), <br>
which is statistically significant by using line markers. <br>
<br>
My questions are:<br>
(i) Is my statistical significance testing correct?<br>
<br>
(ii) how to indicate those points (by marker or any other way),<br>
 that are statistically significant on the   lead/lag cc line plot.<br>
<br>
I had extensively searched NCL-talk archives but <br>
did not find solution to my problem. My guess is that,<br>
 i had to use “where” function and then gsn_add_polymareker,<br>
 but do not know how to implement this in the ncl script.<br>
I am using NCL version 6.4.0 on linux machine. <br>
<br>
Any suggestion would be of great help.<br>
Many thanks in advance.<br>
With best regards<br>
Dr. Sujata Mandke<br>
Scientist, Indian Institute of Tropical Meteorology<br>
PUNE, INDIA_______________________________________________<br>
ncl-talk mailing list<br>
<a href="mailto:ncl-talk@ucar.edu" target="_blank">ncl-talk@ucar.edu</a><br>
List instructions, subscriber options, unsubscribe:<br>
<a href="http://mailman.ucar.edu/mailman/listinfo/ncl-talk" rel="noreferrer" target="_blank">http://mailman.ucar.edu/mailman/listinfo/ncl-talk</a><br>
</blockquote></div><br></div></div></div></blockquote></div>