I am analysing wind speeds and directions from the Newcastle Nobby’s Signal Station AWS. This station is situated at 33 m above mean sea level and 10m above ground, at the Hunter River mouth (NSW; AUstralia). It is a coastal weather station, being located on a headland facing the ocean, south of Stockton Bight, a 34km long embayment.
As you can see from my image "WindSpeedLongTerm", I calculated the annual means for every year I had. From 2000 on the annual mean wind speeds (WS) dropped significantly to very low values.
This in turn, greatly affects the sand Drift Potential (see DP_Year image) that I calculated using Freyberger's method, essentially taking into account the wind speed threshold to mobilise sand of the granulometry, type and roughness occurring in this area, and frequency of wind above the threshold occurring from each direction.
I need your help to figur out why wind speed and sand drift potentials dropped so dramatically during 2000-2015 period.
I checked the equipment history report and except some equipment replacements and maintainence, nothin relevant seems to be happened.
The accuracy is still the same (.00) and always 3h readings (thus 8 wind speeds per day).
From 2000 on more Null Values were found (red Nas in the first image), but in my opinion not so many to justify such a drop.
I checked the Southern Oscillation Index (SOI) to see if something strange happened but nothing unusual to my knowledge (I didn't check correlations, only visual plots comparison).
According to the SOI and the Australian Bureau of Meteorology data, a prolonged moderate La Niña period occur prior 2001, then near to neutral conditions established, except some weak El Niño conditions in 02-03 and 06-07. After 06-07, La Niña and neutral conditions dominated, with the exception of 09-10 weak to moderate El Niño event. La Niña then dominates especially from 2010-12, where SOI indicated a strong El Niña period.
2012, 2013, 2014 were ENSO neautral all the time, except end of 2014 when very near El Nino thresholds were reached, Then early May 2015, El Nino conditions were up again.
Was the period 2010-15 a particular La Niña/neautral period in some way I don't understand?
Thx for any suggestion.
Nic
Dear Kenneth,
Thx for pointing me to the NOAA ONI indicator, to compare with the SOI I was using.
And also for providing me the charts with the right period I am considering! Because I wrote a wrong period in the question (2010-15), while actually the "ambiguous" period starts from 1999-2000, intensifies on 2010-11 then "recovers" in 2015.
Thus, it is more complicated than before!
By looking at your charts (ONI) and the SOI and comparing them with the long-term series of wind speed or drift potential seems that:
In fact, seems that La Niña dominates both the wind speed (WS) in this area, and the sand drift potential (DP).
But it seems that a negative correlation between La Niña strength and DP exists. The stronger La Niña occurs, the lowest DP it is.
When La Niña is weaker, DP and WS are the highest...
Considering that, according to the Australian Bureau of Meteorology (BoM), La Niña in Australia is linked to:
and that "The 2010–11 and 2011–12 La Niña events were two of the most significant in Australia’s recorded meteorological history", I am now more confused than before.
How is possible that when La Niña is strongest, DP and WS are lowest in this region?
Is the strenght of La Niña (as calculated by ONI and SOI) inversely related to the strength of the trade winds in this region?
I hope, because only like this it would make sense!
Nic
I've been looking at so many wind observations and this here is a classical case. Your wind data looks like there is a break change in 2001 leading to a drop of around 1.5 m/s in annual mean. This is supported by the systematic change in N observations. I'm quite sure that the data is not homogeneous. You should try to get comparable data from coastal stations nearby (maybe there is even one from the open sea?).
On this website linked below you can get neighboring stations. I would just repeat the same procedure you did with your station and compare them. From the German North Sea coast we know that changes in N observations and/or small movements of the station have very large changes, see e.g. Lindenberg et al. (2012).
http://weather.gladstonefamily.net/site/94774
https://www.schweizerbart.de/papers/metz/detail/21/76604/Representativity_of_near_surface_wind_measurements_from_coastal_stations_at_the_German_Bight
Article Representativity of near surface wind measurements from coas...
Dear Frederik,
thx a lot for the suggestions and the resources attached.
Unfortunatly, I cannot acces nearby weather station wind data for comparisons right now, as the request (to the Australian Bureau of Meteorology, BoM) would take some time and a payment that I cannot afford.This is only a small fraction of my Master project, I would do so many things, but I have no time or resources.
I would have loved to compare my data with similar coastal stations along NSW, but on BoM website they provide only monthly means of wind speeds at either 9am or 3pm. That is not the raw data I need to replicate calculations and compare wind speed time series.
Just a precisation: What do you mean by "not homogeneous"?
There is a sharp change in number of wind speed and direction readings per day in October 1988, when from 2 readings (9am and 3pm) passed to 7 readings (3h step). And yes, NAs values appear synchronously to the drop of 1,5 m/s (2001).
So, NAs have such a great influence on yearly means, even if they represent a small portion of the total observations per year?
Thank for your help Frederik,
best regards
Nic
There is usually a significant diurnal cycle in wind speeds. At coastal stations, the diurnal cycle is often even much larger than on land due to the sea-breeze system. So if you have less observations, you may end up using the time of the day where wind speeds are on average always higher or lower than using another time of the day etc. Changing the number of observations will always change the mean value systematically.
As an example, see the figure for the daily cycle of different wind farms and imagine you change the time or sampling rate... It will significantly impact the daily/annual mean. I just took one of the first examples I found on that - it may be very different for your location.
https://carboncounter.files.wordpress.com/2015/07/average_hour_rescaled2.png
https://carboncounter.wordpress.com/2015/07/10/are-wind-farms-more-productive-at-night/
Thank you for your expertise Frederik.
Extremely important points and interesting. I will give a better look at sampling times and averaging methods adopted by the weather station system, particularly at changes from 2001 on.
I think I observed on UAV imagery smaller parabolic ripples superimposing on massive transverse coastal dunes in this area. I need to check. This could indicate that daily sea breezes are particularly important, thus, changes in number of observations or time of the day might considerably affect the homogeneity (now I got what you meant) of the data.
Thank you again! Extremely useful.
Nic