Skip to content

The effect of wind direction on annual temperatures at Jurmalciems, Latvia

March 6, 2018

By Paul Homewood




It only takes a glance at the CET record to see that there was a sharp step up in temperatures in the late 1980s, since when they have remained essentially flat.

Two other things stand out though:

1) Apart from two years, we have seen annual temperatures at similar levels occasionally in the past.

2) There have been similar leaps in the trend in the past.


Both these factors suggest that what we have seen in the last thirty years is “weather”, rather than an underlying warming trend.

I have also seen similar trends in other parts of NW Europe.


I am hoping to do some more analysis on this and post up in a day or two. But reader Paul Weldon has written a rather interesting paper about the relationship of temperature trends and wind patterns in Latvia:


The effect of wind direction on annual temperatures at Jurmalciems, Latvia.



Most people will be aware that weather on any day is affected by wind direction. For those living in Europe, if the wind were blowing from the north, for example, it would always be colder than if it were blowing from the south. South-westerlies are associated with mild, wet and windy weather. Weather systems and their movement through a region determine the wind direction and bring with them typical meteorological characteristics and, over the longer term determine the climate of a region. There is annual variability in temperature which can be influenced by criteria such as the North Atlantic Oscillation (NAO) and Atlantic multi-decadal oscillation (AMO). These affect the annual temperatures in the same way that El Nino Southern Oscillation (ENSO) affects the tropical Pacific.

This paper sets out to determine how closely the relationship between wind direction and temperature affects variation in annual temperatures, with a view to eventually looking further back into past records to see how decadal variations in wind patterns have affected past weather.

For this paper, I have my used my own records of temperature and wind direction taken from the Baltic coast in Latvia and cover the 11-year period 2007-2017. Their accuracy may be brought into question, but if a significant relationship between annual temperatures and wind direction can be established, then the next step will be to use official records from Latvia to see if the relationship stands up to better quality data and previous years.



The average daily temperatures were taken for each individual month (January – December) and allocated to wind direction. The average temperature for wind direction was then used to produce a theoretical average temperature for the years of the study.

Where there was more than one wind direction recorded for a specific day, then the average wind direction was taken if the directions were close, for example a recording of SE/SW was given a reading of S (south). For a larger swing in daily wind direction, for example SW/NE then the reading was deleted from the data. Likewise, a recording of ‘variable’ was also deleted. The incomplete year of 2005 was used to find the average temperature for month but was not included in the final annual results as the year was incomplete. The results were then plotted on a graph to show the correlation (or lack of it) between actual and calculated temperature



In winter, average temperatures for easterly winds (NE-SE) were much colder than those from westerly directions (SW-NW). In summer, the reverse is true. However, the temperature difference in wind direction is far greater in winter than in summer; 12.3 degrees C in January as opposed to 3.8 degrees C in July (see figure 1).

The annual average temperature for wind direction showed little difference (2.8 degrees C) with the warmest winds being from the south, and the coldest from the north. It is also worth noting that in the spring and autumn winds from the sea are slower to warm/cool showing the longer time needed for the water temperatures to adjust to difference in insolation.



Winds from the SW are seen to be the most dominant (prevailing wind) direction, making up on average 28% of wind direction/days. This percentage varies between 22% and 33%. Although there is an apparent peak in Jul/Aug, there is no obvious trend. Winds from the north have the lowest occurrence, with an unexplained peak in Aril/May. Winds from an easterly (NE/SE) direction appear to peak in the first 3 months of the year and are at their less frequent in the summer months (see figure 2).


Comparison of actual annual temperatures with the theoretical temperature calculated from wind direction only shows a good correlation (see figure 3):



Figure 3 Graph showing actual and calculated average annual temperature


The close correlation substantiates the claim of average temperatures being strongly influenced by wind direction. Annual temperature values rise and fall together from 2007 – 2013 and from 2014-2017. However, this is not the case for 2013-2014. The values are nevertheless very similar especially 2008, 2012 and 2017 where the values are virtually the same.



The temperature at a given location is not specific to its location but influenced by weather systems bringing influences from other areas. Differences in insolation with latitude create a transfer of heat from the equator to the poles with a general rule that, in the northern hemisphere, winds from the south bring warmer temperatures than winds from the north. There is also the difference between winds from the sea and winds from the land. Wind direction reflects both these variables and their frequency has a marked effect on average annual temperatures. It is therefore no surprise to find that the relationship is strong between these two variables.

Wind direction is not the only variable, insolation and humidity also play a marked role. Nevertheless, wind direction contains some of the effects of these variables: winds from the sea will generally have a higher humidity than those from the land and be associated with increased cloud cover (less insolation). Water vapor is the most important greenhouse gas and night-time temperatures will reflect their abundance in the atmosphere.

The results show some limitations when using wind direction only as a proxy for temperature and were apparent in the study. Sea breezes caused by the locations proximity to the coast made it difficult to attribute one daily wind direction. These changes (often from west to east with light winds) were ignored as being part of the conditions typical of the conditions for the general wind direction. Another limitation is that wind from a specific direction can be the result of two different weather systems and sources of air.

Figure 3 shows that most of the year to year changes in actual temperature are matched by a similar change using calculated temperature. The exception is the year 2014, and there appears to be no obvious reason for this cause. It must be said that there have been 3 or 4 changes of thermometer during the recorded period, although when these changes occurred was not recorded. Bearing in mind the relatively small changes in average annual temperature then it is quite possible that readings for this period have become biased.

Temperature readings were for the previous 24 hours, but wind direction was taken as the average for the date given. The relationship between the two would therefore be a little out of alignment, especially if the weather were unsettled.



Comparison of wind direction and temperature show a strong correlation. Analysis of these 2 parameters from weather station data from the past can therefore be used to show how weather patterns in the past have affected annual temperature variation. Records show that there has been a general warming of the Earth over the last 150 years, and this has been attributed to global warming. It is therefore worth extending the time scale of this paper back to show how much of that warming is due to changes in wind direction rather than a general warming of air temperatures per se.

Winter temperatures dominate the average annual temperature. So if the winter months are dominated by winds coming from the sea (S – NW) and comparatively mild (0.6 – 3.1 C) rather than those from the land (NE – SE) and much colder (-5.3 – -9.3), then the result will tend to be a low average annual temperature. This fits well with the idea that variations in annual temperature will be governed mostly by weather patterns as opposed to a general warming of the global temperatures. The NAO is such a system that will drive the variation in temperatures.

This paper has highlighted several problems with using personally collected data, and it is therefore necessary to replicate the same analysis with officially recorded data before embarking on a much larger project back into the 19th century.


A lot of this is really intuitive when you think about it. And certainly HH Lamb was a big believer in the role that weather patterns and winds had on climate trends in the past, as opposed to just daily weather.

He researched and wrote extensively about how these long term changes affected climate around the world.

I wonder what he would make of today’s climate?

  1. Chris, Leeds permalink
    March 6, 2018 7:39 pm

    On your point about ‘similar leaps in trend in the past’ it has always struck me that if you look at the full CET series (back to 1659) there were really extraordinary ‘convulsions’ of temperature. If you look at the ten year running means the ‘decade’ ending 1700 had a mean temperature of 8.07C, whereas the ‘decade’ ending 1738 had a mean of 9.87C – that’s an increase in CET of 1.8C in 38 years. That’s a bigger increase than we have seen in the 20th/21st Century. Similarly – and in contrast – there were steep drops in CET in the 1670s and in the 1740s

    And within shorter periods there was a 1C rise in a decade between the 10 year mean of 8.07C for the decade (1691-1700) and 9.18C for the decade 1701-1710…. That is entirely compatible with the ‘step change’ we have seen in the 1990s. Moreover. the ‘warm years’ of the 1730s would not have been so very different from the ‘warm years’ that have become common in the last two decades. What was happening then? And why cannot a similar dramatic change now been an entirely natural occurrence?

  2. jim permalink
    March 6, 2018 7:59 pm

    Please include summer maximums and winter minimums in your analysis.
    Please also include measurememt accuracy which must be a minimum of +-0.5C given accuracy of most years in the data series.
    I think you will find that summer maximums lie within the error band of measurement accuracy. And any recent increases in means is due to increases in winter minimums.
    Also I am sure you will confirm that exceptional months ( especially Octobers) account for much of any rapid change, which may well lead back to wind direction.

    • paul weldon permalink
      March 7, 2018 8:51 am

      thanks for your comments, Jim. The objective of the analysis was to find out how strong the correlation is between wind direction and annual temperatures. I therefore tried not to digress from the main issue and therefore did not consider max and min temps, although that would have been an interesting exercise. I have no way of knowing how accurate the measurements were, and I think I have given ample mention in the text as to the possible irregularities in the data. What is important to me is the correlation, and would appreciate comments on my findings. I have not carried out any statistical analysis on it at this stage, I think that it is obvious just by looking at the graph.

  3. MrGrimNasty permalink
    March 6, 2018 8:39 pm

    The 1980’s step change coincides with instrument changes and massive urban development. It always looks more like a data discontinuity than a genuine change to me.

  4. The Old Bloke permalink
    March 6, 2018 10:26 pm

    Well said MrGrim. We all went from eyeballing the mercury every 3 hours to having automatic recording every second. Stations were also moved from rural locations to ones of urban. Stations once had their thermometers inside Stevenson Screens, now they are stuck on poles, walls, well anything really in direct sunlight.

    • John F. Hultquist permalink
      March 7, 2018 3:52 am

      “… now they are stuck on poles, walls, well anything really in direct sunlight.

      I’m trying to decide whether your are a true curmudgeon or just grumpy today.

      • The Old Bloke permalink
        March 7, 2018 9:01 am


  5. March 6, 2018 10:34 pm

    The sudden step up in the late eighties is very evident in spring months, it looks like an inhomogeneity, but can’t be because every record shows the same thing:

  6. Geoff Sherrington permalink
    March 7, 2018 12:23 am

    Does the global wind direction sum to nil at any given time?
    If so, how can it affect global temperatures in the long term? Geoff.

    • John F. Hultquist permalink
      March 7, 2018 3:56 am

      Perhaps not “direction” but otherwise:
      Objects executing motion around a point possess a quantity called ANGULAR MOMENTUM. This is an important physical quantity because all experimental evidence indicates that angular momentum is rigorously conserved in our Universe. It can be transferred, but it cannot be created or destroyed.

    • paul weldon permalink
      March 7, 2018 10:00 am

      AN interesting question, Geoff! Just a couple of thoughts – An El Nino makes a substantial effect on the Earths climate system, but overall the winds will balance. The recent cold periods in the Eastern part of the US have been balanced by warmer temps in Alaska, Are the temps in the northern hemisphere colder or warmer when the Polar front meanders? And are our records influenced by the distribution of weather stations? It would seem to me that although the wind direction balances overall, there is still the possibility that the regional temperatures added together may not.

  7. quaesoveritas permalink
    March 7, 2018 10:23 am

    I have long considered that it would be interesting to correlate wind direction (or rather the ultimate source of the wind) with CET, but unfortunately wind data are not available in sufficient detail to do that.
    I think it is remiss of the UKMO that it does not produce wind speed and direction data in the same detail as temperature and rainfall data.

  8. March 7, 2018 1:22 pm

    In order to have meaningful data, the scientific method requires as much consistency as possible and a standard or “control” position with which your data is compared. Sensing for temperature at the point of wind sensing should be in an open area without obstructions unless you are testing local conditions with buildings, etc. For example, Wall St. in NYC would have stronger wind than other nearby areas in the city due to the clustered presence of tall buildings which provide a tunnel effect.

  9. Douglas Brodie permalink
    March 7, 2018 6:17 pm

    “Both these factors suggest that what we have seen in the last thirty years is “weather”, rather than an underlying warming trend.”

    The trustworthy UAH satellite temperature series suggests the same thing, see

    Looking back over the last 20 years, a line drawn from half way up the steep El Nino induced warming peak in 1998 to the prevent time is flat. The warming from about 2015 to 2017 was due to a large and prolonged natural El Nino which had nothing to do with man-made CO2 and temperatures have since then fallen back to the 2014 levels when even the alarmist Met Office conceded there had been a standstill in global warming from the turn of the century.

    Simple eyeballing of the graph from 1979 shows no discernible slow but steady alleged man-made global warming at the model-predicted rate of about 0.2ºC per decade, just warming and cooling fluctuations at much more rapid rates of change which are clearly natural rather than due to man-made CO2, in fact they map onto the Multivariate ENSO index, see

    These fluctuations culminate in the huge natural jump of over 0.7ºC in less than a year in the 1998 El Nino followed by a cooling La Nina then a couple of years later by another sudden El Nino which lifted global temperatures to a new, slightly higher thermodynamic equilibrium which has endured to the present time.

    It seems unscientifically obvious that the global warming since 1979 has been overwhelmingly if not wholly due to just a few strong, natural El Nino warming events which have not been counteracted by equally strong La Nina cooling events.

Comments are closed.

%d bloggers like this: