Annual Weather Report 2017.
The aim of this report is to try to place 2017 into some sort of historical context. The monthly reports tend to place each month into a context within a year,
I will comment beforehand that IMHO (and LHOSM - the Less Humble Opinion of Serious Meterologists) there is no doubt about anthropogenic climate change. However I am far from convinced that the time series I have available are sufficient to demonstrate this.
The most notable feature (IMHO ) is the very low mid year peak in 2017. This is not surprising as we only received 16.2 mm of rain between 21 May and 30 July as against an average of close to 100mm for that period.
This leads me to think about the relationship between 1500 Hr Humidity and rainfall. Looking at Annual averages for rH and mm of rainfall the relationship isn't strong: r2 =0.12. I have marked the point for 2017 with a red ring.
I will comment beforehand that IMHO (and LHOSM - the Less Humble Opinion of Serious Meterologists) there is no doubt about anthropogenic climate change. However I am far from convinced that the time series I have available are sufficient to demonstrate this.
Rainfall
At one point it seemed that we were going to have the lowest annual rainfall since our records began in 1993. However good falls in the last three months of 2017 averted that. We ended with 496 mm, the 4th lowest fall.
There is a faint cycle visible in the data but the value of r2 shows it to be very weak.
Temperature
I'll begin with a caveat about maximum temperatures. As explained in this post there seems to have been something abnormal with recording of maximum temperatures prior to 2010. So for maximums I am only going to examine the years since that date. I compiled three summary series of the average of daily maxima and the highest and lowest maximum for each year.
I checked a trend line for the average value and it has a low value of r2 indicating that there is no trend in these data.
I did much the same for minimum temperatures, but extended over the whole period since 1993 as I don't have so much doubt about the data. The average minimum for 2017 of 6.58oC was the second highest recorded (after 6.66oC in 2010).
The values of r2 for the trends in average minimum and overall minimum for the year are both quite high, suggesting that there is a trend evident. It suggests we are warming by ~0.15oC per year which doesn't seem like much but could be expressed as increasing by 1oC in 7 years which is a worry (but see the second paragraph in this post). It is interesting that between 1993 and 2016 the population of Canberra grew from 300,000 to 400,000: is there now a heat sink just over escarpment?
As a further comment on minimum temperatures, all 2017 monthly minimum temperatures were above average (in the case of March by 3.94oC and December by 3.21oC).
In terms of extreme temperatures I use Heat Wave Days as an indicator of heat and two values of frost for cold.
I use a variant of the BoM measure of a Heat Wave with my measure being 3 consecutive days or more with a maximum of >30oC and a minimum >10oC. I then count the number of days falling within a Heat Wave. For 2017 we recorded 39 HWD which is exactly the average since 2010.
Please excuse the lurid appearance: I have just discovered the glow function in EXCEL! Prof. Tufte would be appalled!
For frosts a distinction is made between the temperature at the standard height of the measuring device (known as a screen) and the derived temperature at ground level. When the screen temperature is +2oC the ground temperature will be 0oC. in the UK my Dad used to call this a ground frost, while a screen temperature of 0oC was an air frost. I think BoM use the terms Light Frost and Hard Frost for the same types of events.
Rather than simply show the number of days this chart shows the number of frosts of each type as a 5 of the average number of frosts of that type. Clearly there was well under the average number of ground frosts (96 in 2017 vs 133 avg) while the 71 air frosts were very close to average (due to a low number of these events in May (6 vs 10) and October (0 vs 30) balancing the higher numbers in other cooler months).
The ground frost values are very consistent with my common theme of increasing minimum temperatures.
I did much the same for minimum temperatures, but extended over the whole period since 1993 as I don't have so much doubt about the data. The average minimum for 2017 of 6.58oC was the second highest recorded (after 6.66oC in 2010).
As a further comment on minimum temperatures, all 2017 monthly minimum temperatures were above average (in the case of March by 3.94oC and December by 3.21oC).
In terms of extreme temperatures I use Heat Wave Days as an indicator of heat and two values of frost for cold.
I use a variant of the BoM measure of a Heat Wave with my measure being 3 consecutive days or more with a maximum of >30oC and a minimum >10oC. I then count the number of days falling within a Heat Wave. For 2017 we recorded 39 HWD which is exactly the average since 2010.
For frosts a distinction is made between the temperature at the standard height of the measuring device (known as a screen) and the derived temperature at ground level. When the screen temperature is +2oC the ground temperature will be 0oC. in the UK my Dad used to call this a ground frost, while a screen temperature of 0oC was an air frost. I think BoM use the terms Light Frost and Hard Frost for the same types of events.
Rather than simply show the number of days this chart shows the number of frosts of each type as a 5 of the average number of frosts of that type. Clearly there was well under the average number of ground frosts (96 in 2017 vs 133 avg) while the 71 air frosts were very close to average (due to a low number of these events in May (6 vs 10) and October (0 vs 30) balancing the higher numbers in other cooler months).
Humidity
BoM usually cite the relative humidity as the value for 1500 Hrs (ie 3PM). This is very reasonable as the humidity varies greatly through the day, but there is little variation between months in the early morning.
I only have 5 years of data for humidity and a few months of that are missing as actual values. I estimated the missing stuff based on the monthly averages and nearby observations. They are shown in the following chart.
This leads me to think about the relationship between 1500 Hr Humidity and rainfall. Looking at Annual averages for rH and mm of rainfall the relationship isn't strong: r2 =0.12. I have marked the point for 2017 with a red ring.
The value for 2013 (48% rH and 586mm of rain) looks interesting. Applying Trewin's Hypothesis ("Any interesting statistic is probably a processing error.") I have looked at the data for that year and concluded that the rainfall is biased by very heavy falls (>60mm) on one day each in January and September which months were otherwise quite dry. However even reducing the rainfall by 100mm doesn't give a very strong relationship! So I'll put this down as needing further study (and additional data).
There are two aspects to wind: speed and direction. I suspect my site is too protected, so the speed is understated (but probably relatively OK - it may read 20kph when the gust is actually 40kph but it shows a windy period vs a calm period which is what I need).
Here are the average gusts for the last 4 years.
These data suggest that the last two years have been windier that the previous 2. As BoM only have 13 months of data on wind speed (at least that is all available at no cost) I can't validate this.
WRT to wind direction every time I check my weather station console against what actually seems to be going on outside the window it looks roughly OK. However, I am of the view that the prevailing wind here is from the NW (with an Easterly change in Summer evenings). This rosette shows what my weather station thinks is happening!
If anyone else has other ideas about what is going on I'd be pleased to hear from them!
Wind
A very vexed topic as the anemometer on my weather station is on top of the station, at about 1.5m above the ground rather than 10m. It is probably also subject to interference by vegetation. In other words the site is appalling. However it is in the same position so the data should be comparable.There are two aspects to wind: speed and direction. I suspect my site is too protected, so the speed is understated (but probably relatively OK - it may read 20kph when the gust is actually 40kph but it shows a windy period vs a calm period which is what I need).
Here are the average gusts for the last 4 years.
WRT to wind direction every time I check my weather station console against what actually seems to be going on outside the window it looks roughly OK. However, I am of the view that the prevailing wind here is from the NW (with an Easterly change in Summer evenings). This rosette shows what my weather station thinks is happening!
Comments