Some longer term trends
I have been asked if I have fitted any trend lines to the data series from 1993. At that time the answer was "No", but that was then and this is now.
There are many possible data items that could be used and - even using EXCEL - many trendlines that could be fitted. To kick things off I have shown below:
The values of r2 for those series look reasonable (I don't have measures of significance available).
The value for r2 is quite low, suggesting no trend in these data.
The data comes from two sites in Carwoola which give pretty consistent readings, so I don't think the changes of site in 2014 (for temperature) or 2007 (for rainfall) affected matters greatly.
Following on from comments on the initial publication of this post a question was asked about the cool values for 1993. One way of addressing this is to look at BoM observations for the (now closed) site Canberra Airport Comparison NSW (Station Number 070014) which has data from 1949 to 2010. I have charted temperature data for that site and the Carwoola records for the period 1993 to 2010. Note that the data for Carwoola for the period is all from a single site (not on my property so I can't comment on detail).
In both sets of data the correlation coefficient between the series is ~0.70 suggesting the overall pattern is similar.
As Carwoola is about 200m higher than the Canberra Airport I'm not surprised that the minimum temperatures are lower. As the Carwoola data is not on my property I can't comment on the reason for the rise in minimum temperatures between 1997 and 1999 but the observer at that time provided data from a weather station. He has also commented in the past that minima differ across the property with low areas collecting cold air: I suspect growth of vegetation in some way changed airflow over the period.
I only have measured annual average temperatures for Carwoola for 3 years so have approximated average temperatures as the mean of average maximum and average minimum for the period 1993 to 2016.
As would be expected that is influenced by the low minima recorded prior to 1997. Disregarding those years eyeballing the chart suggests variation around a quite flat line for the period 1999 to 2016
There are many possible data items that could be used and - even using EXCEL - many trendlines that could be fitted. To kick things off I have shown below:
- Data series for annual average minimum and maximum temperatures
- A data series for total annual rainfall;
- With a 4th order polynomial trend fitted. I find that order tends to give a fairly high value of r2, without butchering the degrees of freedom too badly.
The data comes from two sites in Carwoola which give pretty consistent readings, so I don't think the changes of site in 2014 (for temperature) or 2007 (for rainfall) affected matters greatly.
Following on from comments on the initial publication of this post a question was asked about the cool values for 1993. One way of addressing this is to look at BoM observations for the (now closed) site Canberra Airport Comparison NSW (Station Number 070014) which has data from 1949 to 2010. I have charted temperature data for that site and the Carwoola records for the period 1993 to 2010. Note that the data for Carwoola for the period is all from a single site (not on my property so I can't comment on detail).
As Carwoola is about 200m higher than the Canberra Airport I'm not surprised that the minimum temperatures are lower. As the Carwoola data is not on my property I can't comment on the reason for the rise in minimum temperatures between 1997 and 1999 but the observer at that time provided data from a weather station. He has also commented in the past that minima differ across the property with low areas collecting cold air: I suspect growth of vegetation in some way changed airflow over the period.
I only have measured annual average temperatures for Carwoola for 3 years so have approximated average temperatures as the mean of average maximum and average minimum for the period 1993 to 2016.
Comments