Moments of Manilla’s Yearly Rainfall History

Manilla Annual rainfall history: Four moments

Comparing all four moments of the frequency-distributions

Annual rainfall for Manilla, NSW, has varied widely from decade to decade, but it is not only the mean amounts that have varied. Three others measures have varied, all in different ways.

I based the graph on 21-year sub-populations of the 134-year record, centred on consecutive years. I analysed each sub-population as a frequency-distribution, to give values of the four moments: mean (drawn in black), variance (drawn in red), skewness (drawn in blue) and kurtosis (drawn in magenta).

[For more about the moments of frequency-distributions, see the recent post: “Kurtosis, Fat Tails, and Extremes”. See also the Note below: “Instability in the third and fourth moments.”]

Each trace of a moment measure seems to have a pattern: they are not like random “noise”. Yet each trace is quite unlike the others.

The latest values are on the right. They show that the annual rainfall is now remarkable in all four respects. First, the mean rainfall (black) closely matches the long-term mean, which has seldom happened before. By contrast, the other three moments are now near historical extremes: variance (red) is very low, skewness (blue) very negative, and kurtosis (magenta) very positive.

To my knowledge, such a result has not been observed or predicted, or even suspected, anywhere.

Manilla Annual rainfall history: Mean

The mean yearly rainfall (the first moment)

As I have shown before, the mean annual rainfall was low in the first half of the 20th century, and high in the 1890’s, 1960’s and 1970’s. Rainfall crashed in 1900 and again in 1980.

Manilla Annual rainfall history: Variance

Yearly rainfall variance (the second moment)

Continue reading


Manilla’s automated rain gauge down again

Photo of rain gauges

The rain gauge we had

[In the post “No rain, and no rain gauge” of 18 March 2016, I reported that the Manilla Post Office rain gauge, after 132 years, was no longer being read. Now the automated gauge that replaced it is not being read either.]

Manilla’s official rainfall record since 2015

For the last eighteen months, Manilla’s official rainfall record has depended on an automatic rain gauge. The Bureau of Meteorology moved this gauge from the Post Office to a nearby yard of the Manilla Historical Society Museum. The gauge, which had provided flood warnings only, became also the general-purpose rain gauge for Manilla.

From the date when manual readings ceased, 26 March 2015, there was no Manilla rain gauge for 424 days. From 23 May 2016, the re-purposed, and re-located automatic rain gauge then operated as Station 055031, Manilla Post Office, for 137 days to 7 October 2016.
Due to a fault, there were no readings for 161 days to 17 March 2017. After repair, the gauge then operated as Station 055312, Manilla (Museum), for 191 days to 24 September 2017, when it failed again. At the present date (5 November 2017) it has been out of service for 42 days.


Since Manilla rainfall readings became automated eighteen months ago, 38% of the readings have been missed, missing months at a time. This is appalling. When the Manilla rain gauge was read by the Postmaster, from 1883 to 2015, far less than 1% of readings were missed, never for more than two days at a time.

Without records from a rain gauge that is recognised officially, Manilla residents will have no evidence to prove the severity of the next drought. If people think this is important, it seems they would be far better-served by a local volunteer than by a Bureau of Meteorology that cannot afford to keep the automatic rain gauge running.

October 2017: no drought

Grevillea robusta flowers

Flowers of Silky Oak

No temperatures were extreme in this month. In the second week, the mean weekly temperature was four degrees above normal, rather like the last week of September.
Dry air on the 6th and on the 31st made the dew point eight degrees low, but humid air on the 11th made it seven degrees high.
My rain gauge registered six rain days, with high readings of 38.5 mm on the 9th, 22.0 mm on the 12th, and 16.8 mm on the 21st. (The automatic gauge at the Museum remained down.)

Weather log for October 2017

Comparing October months

As shown by the arrow on the second graph, October months became warmer and more moist with each year from 2012 to 2015. October 2016 was very cool, then this month was again warm. The trend to more moisture continued through all six October months from 2012 to 2017. It was shown not only by rainfall, but also by cloudiness, dew point, and narrowing daily temperature range. No other calendar months had this trend.
The high total rainfall of 84.1 mm (80th percentile) wiped out the serious and severe rainfall shortages seen in September. Now, the lowest percentile value is that for the 4-month total (117 mm). Being at the 15th percentile, it does not rate as serious.

Climate for October

Data. A Bureau of Meteorology automatic rain gauge operates in the museum yard. From 17 March 2017, 9 am daily readings are published as Manilla Museum, Station 55312.  These reports use that rainfall data when it is available.  The gauge last reported on 24 September 2017.

All other data, including subsoil at 750 mm, are from 3 Monash Street, Manilla.

3-year trends to October 2017

Avoiding drought

3-year trends to October 2017

October raw anomaly data (orange)

October 2017 was moist: all moisture indicators had dropped sharply down the graphs, retreating from the aridity of August and September. Daily maximum temperature anomaly (x-axis in all graphs) fell towards normal, while that of the subsoil (lower right graph) remained low. Daily minimum temperature anomaly (lower left graph) jumped from extremely low to extremely high.

 Fully smoothed data (red)

The latest fully-smoothed data point is that for April 2017.
At that time, the climate was warm and almost static. There was a pause in a drift towards aridity.


Fully smoothed data – Gaussian smoothing with half-width 6 months – are plotted in red, partly smoothed data uncoloured, and raw data for the last data point in orange. January data points are marked by squares.
Blue diamonds and the dashed blue rectangle show the extreme values in the fully smoothed data record since September 1999.

Normal values are based on averages for the decade from March 1999.* They appear on these graphs as a turquoise (turquoise) circle at the origin (0,0). A range of anomalies called “normal” is shown by a dashed rectangle in aqua (aqua). For values in degrees, the assigned normal range is +/-0.7°; for cloudiness, +/-7%; for monthly rainfall, +/-14 mm.

 * Normal values for rainfall are based on averages for the 125 years beginning 1883.

House June warmth profiles: IV

Part IV: Solar gain in the clear-story

In a solar-passive house, do clear-story windows trap much heat?
How about overcast days?

Graph of clear-story temps, 2 days

[This post repeats some data of an earlier post, headed  “Part III: Daily temperature cycles, east wing”. Please refer to that post for more details.]

The graph above shows records of temperature for two days in mid-winter. Records of cloud cover (plotted in purple) show that the first day was overcast and the second mainly sunny.
Through the sunny second day, the temperature readings taken just inside the clear-story windows (black) rose and fell just like the outdoor temperature (red), but they were much higher. I have drawn a dotted red line at a temperature 13.5° higher than outdoors. It fits well to the clear-story temperature (black) on that day. During the previous day, which was overcast, the dotted red line does not fit. It is about 6° higher than the actual clear-story temperature.
By experiment, I found that I could make a model (plotted in green) that would match the actual clear-story temperature as the cloud cover changed. As well as adding 13.5° to the outdoor temperature, I subtracted two thirds of the cloud cover measured in octas. As plotted (green), this model matches the clear-story temperature through both days. At two data points there was a mis-match: those points have not been plotted.

 Graph of clear-story temps, 5 days

The second graph shows all five days of the experiment. My model of temperature in the clear-story space (plotted green), fits the actual readings (black) on all days.
The model includes one other feature: the maximum temperature that I allow is 26°. That also matches. As mentioned in Part III, a thermostat turns on fans at 26°. That prevented the temperature from rising higher.


A solar passive house is likely to gain more winter heat if it has north-facing windows in a clear-story above room level. It may also lose more heat. If so, the cost of the clear-story design may not be justified.
This experiment shows that, in this particular house during one harsh winter, the clear-story performed very well.

People may be as surprised as I was at the closely-matching pattern of outdoor and clear-story temperatures in mid-winter, and at how very much warmer the clearstory was: more than thirteen degrees warmer in fine weather.
It may also provoke some thought that the match persisted in overcast weather, but with the clearstory being only eight degrees warmer than outdoors in that case.

Back to Part I: Average temperature values.
Back to Part II: The two-storied west wing’s daily temperature cycles
Back to Part III: the single-storied east wing’s daily temperature cycles

September 2017 even more arid

Trellised vine photo

Blooming Wonga-Wonga Vines

Despite more cloud, September was even more arid than August. Weekly temperatures remained normal until the last week, which was remarkably warm. There were some notable events. The 13th, the first 30°-day of the season, was followed by a day more than 16° cooler. Extremely low humidity early on the 20th made the dew point (-8.8°) almost as low as the record set last month. Among minimum overnight temperatures that were near zero, the one on the 24th (22.8°) set a record by being 14.7° higher than normal.
The number of frosts (below +2.2° in the screen) was 13 (a September record), almost as many as in August. Perhaps the frost on the 20th was the last of the season. That is the normal date for it.
There was 5.2 mm of rain on the 14th, and an estimated 0.3 mm on the 29th. (The automatic gauge at the Museum was down by then.)

Weather log

Comparing September months

The mean daily maximum of 25.0° was rather high, fully 5° higher than last year. With a mean daily minimum (5.7°) that was rather low, the daily temperature range reached the record wide value of 19.3°.
Extremely dry air was shown by a mean early-morning dew point of 2.7°, the lowest September value, 8.1° below normal.
The total rainfall of 5.5 mm (estimated) was very far below the September average (41 mm), at the 8th percentile. That is a serious rainfall shortage. The current rainfall totals for four months (95 mm) and for six months (175 mm) are also serious rainfall shortages. Even worse are the totals for two months (19 mm) and for three months (33 mm): they are severe rainfall shortages.
Similar severe shortages occurred in October 2013, May 2008, and May 2005. Extreme shortages last occurred in the six-month drought of 2002.

Climate graph for September

Data. A Bureau of Meteorology automatic rain gauge operates in the museum yard. From 17 March 2017, 9 am daily readings are published as Manilla Museum, Station 55312.  These reports use that rainfall data when it is available.  The gauge last reported on 24 September 2017.

All other data, including subsoil at 750 mm, are from 3 Monash Street, Manilla.

A drought has begun

A year ago, I showed that Manilla was far from being in a drought. That is not true now. There are severe shortages of rain.

Rainfall status at Manilla, September 2016 and September 2017.

The first graph has rainfall totals up the left margin. They are not expressed in millimetres but as percentile values, Along the bottom margin is the number of months included in calculating each rainfall total.

On the graph, I have compared the rainfall situation today, September 2017, plotted in red with that of September 2016, plotted in grey. Much has changed.

Take, for example, the 12-month (one-year) rainfall total. Rainfall totals for 12 month periods are directly above the value “12” at the bottom of the graph, near the label “Number of Months included”. In data for the month of September 2016 (grey), the 12-month total (actually 802 mm) had been at the 80th percentile, which was very high. In up-to-date data for the month of September 2017 (red), the 12-month total (actually 484 mm) is at the 17th percentile, which is very low.
Although rainfall totals for  periods longer than 12 months have not fallen so much, nearly all of them have fallen. Three that have not are those for 30 months, 36 months and 42 months. They were already low, due to including in them some months of low rainfall several years ago, in 2013 and 2014.

So far, real shortages have occurred mainly within the last 12 months. Beyond that, the two-year rainfall total of 1285 mm, for example, is still near normal, plotting at the 48th percentile.

The second graph shows in detail how shortages that are serious or severe have developed during the last six months. These were the monthly rainfall amounts, with the normal amounts in brackets:

April: 24.0 mm (39.3);
May: 55.6 mm (40.3);
June: 62.8 mm (44.3);
July: 13.2 mm (41.4);
August: 13.8 mm (39.5);
September: 5.5 mm (41.2).

As a result, the current situation is as shown below. There are already severe rainfall shortages, at the 2nd or 3rd percentile, in the two-month and three-month totals to date. There are also serious shortages, at the 8th and 9th percentiles, in the four-month and six-month totals to date.

Drought status at Manilla in September 2017

I will update these graphs each month to show how the situation changes.