Vishali,<div><br></div><div>I'm also not certain what you have in mind, but I'm guessing it's something like<a href="https://code.zmaw.de/embedded/cdo/current/cdo.html#x1-3650002.8.22"> CDO's seasmean operator</a>, which computes a timeseries of three-month seasonal means (DJF,MAM,JJA,SON) from a timeseries (which could be 3-hourly, daily, monthly, whatever). For example, if you have a 10-year record of daily winds, the operator would reduce it to a record of 40 timesteps, each of which contains a three month mean. (BTW, I would call what Dennis is describing the "annual cycle of daily winds" but this may just be a function of the local dialect among the folks I work with.)</div>
<div><br></div><div>I'm not aware that NCL has a comparable operator, but <a href="http://www.ncl.ucar.edu/Document/Functions/Contributed/month_to_seasonN.shtml">month_to_seasonN()</a> should work if you're starting with monthly data. For daily or higher-time-resolution data, you'll probably need to iterate through your timeseries, computing each three-month mean separately. How you do this is going to depend on the particular structure of of your data, the time format, etc.</div>
<div><br></div><div>~Rob</div><div><br></div><div><br>-- <br>Robert Nicholas<br>Department of Atmospheric Sciences<br>University of Washington, Box 351640<br>Seattle, WA 98195-1640<br><a href="mailto:rnicholas@atmos.washington.edu">rnicholas@atmos.washington.edu</a><br>
tel 206-438-3509, fax 206-543-0308<br>
</div>