<p><span style="font-family:Arial;font-size:12.0pt;">Hello NCL experts,</span></p><p> </p><p><span style="font-family:Arial;font-size:12.0pt;">I'm using the NCL function pdfx (https://www.ncl.ucar.edu/Document/Functions/Contributed/pdfx.shtml) to compute probability density functions for a number of datasets. I think there is a problem with how the function deals with missing values.</span></p><p> </p><p><span style="font-family:Arial;font-size:12.0pt;">I'm attaching two plots: Volcano_tempmovingtrend_timeseries_MMMpdf.png is obtained from ensemble-mean datasets (no missing values), while Volcano_tempmovingtrend_timeseries_MEMpdf.png is obtained from individual ensemble members of the same datasets (loads of missing values). The red lines in Volcano_tempmovingtrend_timeseries_MEMpdf.png are really messed up, probably because of the great amount of missing values in that dataset.</span></p><p> </p><p><span style="font-family:Arial;font-size:12.0pt;">Is pdfx skipping missing values in the computation? What workaround can I implement to avoid the issue?</span></p><p> </p><p><span style="font-family:Arial;font-size:12.0pt;">Many thanks,</span></p><p><span style="font-family:Arial;font-size:12.0pt;">Giorgio</span></p>