Intro:

• velocity changes greater than 30Mph/sec. These are glitches.
• All velocities less than .1 mph. The wind meter is often offline/disconnected while the computer continues to log the data.

Last 7 days

• The top plot has the wind velocity binned to 1 minute steps. The black lines in the average over each minute, while the red line is the peak hold over each 1 minute.
• The bottom plot is the median direction for each 1 minute of the day. The direction is where the wind is coming from.

---

Monthly plots: (top)

• The average and peak velocities versus date. The data is averaged and peak held over  15 minute steps.
• The average and median velocities by hour of day for the month. The resolution is 15 minutes.
• The average and median velocities by direction. The data is averaged over 1 degree steps. Data with velocities less than 1 mph are not used (since direction needs the wind to turn the vain).
• A histogram of the wind velocity for the month. 1 Mph bins are used. It is normalized to give the fraction of time each velocity occurs.
• Cumulative velocity distribution. It is the fraction of time wind velocities exceed the velocity of that bin.

Notes: (top)

• I think there is a digitization error in the direction. You can see evenly spaced dropouts. Histograms of the sampled data show jumps at evenly spaced levels.
• There are occasional large glitches in the data (2 to 4 seconds per month). I remove them by not allowing the wind acceleration to exceed 30 mph/sec (i don't think the wind meter can respond that fast anyway).
• There are few boxes mounted near the wind meter that may be biasing the direction.
• The data files contain an ascii string data and a mjd date. The mjd date comes from the single board computer (sbc) that reads the data. The ascii string time is provided by the computer the writes the data to disc (obsdisplay). The time on the sbc drifts by many minutes while obsdisplay is running the ntpd. The times used in the plots are from obsdisplay (ntp based).

---

Yearly plots:  (top)

• The Average daily velocity. Days with less than 50% of the normal samples are excluded and marked with a red +.
• The maximum daily velocity.
• The monthly average wind velocity versus hour of day. The different months are color coded. Nov and Dec use a dashed line (since there are only 10 colors in use).
• The monthly histogram of wind velocity. The y axis is fraction of time a velocity within the 1 mph bin occurred during the month. Months with less than 25% of the normal samples are excluded.
• The monthly average wind velocity versus wind direction.
• Monthly histograms of the wind direction. This is the fraction of time the wind direction fell within the 10 degree bins.

 

Yearly wind plots

2002 (.ps) (.pdf)	2003 (.ps) (.pdf)	2004 (.ps) (.pdf)	2005 (.ps) (.pdf)	2006 (.ps) (.pdf)
2007 (.ps) (.pdf)	2008 (.ps) (.pdf)	2009 (.ps) (.pdf)	2010 (.ps) (.pdf)	2011 (.ps) (.pdf)
2012 (.ps) (.pdf)

processing: x101/wind/plots_year_all

Wind during storms

• 22-24aug12 - tropical storm isaac
• Wind speed,direction at AO during tropical storm isaac (.ps)  (.pdf)
• 21-22aug11 - tropical storm/hurricane irene
• Wind speed,direction at AO during hurricane irene (.ps)  (.pdf)
• on page 22 (22aug11) you can see the wind change direction as it passed ao.
  
• 30-31aug10 - hurricane earl
• Wind speed,direction at AO during hurricane earl (.ps)  (.pdf)
• The velocity plots show the 1 second samples.
• Note that the wind direction may be off by some..
• 17-18aug07   - hurricane dean
• Wind speed,direction at AO during hurricane dean(.ps)   (.pdf)
• The velocity plots show the 1 second samples.

History:  (top)

• 07feb11: obsdisplay rebooted 31jan11 pm. windmeter did not restart. Restarted it on 07feb11
  
• 02sep10: for 7 day plots switched from median to mean for direction data.
  
• 31aug10: wind meter direction reset.
  
• 29may09: wind meter disconnected to test distomat.
  
• jun08: wind meter reinstalled. jorge installs new version of program. The direction output is filtered. The old program delayed between the samples before averaging. The new program samples in a burst and then averages so it gets the same value. This is why the direction is now discrete.
• 19mar08: wind meter cpu fails.
  
• aug07: windmeter reinstalled on platform after painting.
  
• 23may07: windmeter brought back down to the lab to check it out....
  
• may07: windmeter moved to new position because of painting.
  
• 02mar06: 10am wind meter raised 3 feet to be above local obstructions. The windmeter direction was recalibrated so north was north. When the current wind meter was installed in jun05 they aligned it by pointing it at Tower 12 rather than parallel to the T12 direction.
• 15feb06: noticed that the wind direction if off by about 90 degrees. Don't know how long it's been that way.
• 01feb06: started the wind monitor pages.
• jun05-aug05: the wind meter broke. A new one was purchased and installed on 17aug05.
• jun04-sep04: wind meter broke. We fixed it and it went back into operation on 16sep05.

 home_~phil