Low Dew Points and cold days persist
Manilla Smoothed Monthly Anomalies of Climate Variables
Parametric Plots
Update for January 2012


Except for humid November 2011, Dew Points at Manilla (Centre left graph) have now shown remarkable aridity for nine months. The July 2011 value for Dew Point anomaly, now fully smoothed, requires the blue line to be moved yet again. It sets a new lower limit for smoothed Dew Point anomaly: minus 1.70 degrees.
Daily maximum temperature (X-axis, all graphs), from a remarkably low value (-4.6) in December, rose only to -3.7 in January.
On the top left graph, the rainfall anomaly returned from a record positive value in November to stongly negative within two months.
Cloudiness went from very high to extreme.
Daily temperature range (centre right graph) and daily minimum temperature (bottom left graph) moved back from the extreme values of December.

Manilla January weather precis here.

Note:
New data for January 2012 allow updating with more smoothing applied to all months back to July 2011, which is now fully smoothed. Fully smoothed data - gaussian smoothing with half-width 6 months - are plotted in red, partly smoothed data uncoloured, and unsmoothed 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.

I discussed these graphs here.

Cloud further increased

This post updates that of 31/6/10 on "Cloudiness".
The percentage of cloudy mornings observed at Manilla, NSW has increased again. The average for the 21 months from April 2010 is now nearly twice what it was in the eight years to August 2007.
This seems to me to be the largest sustained climate change reported anywhere during the last twelve years. Am I wrong?

As before, cloud amount in octas is observed each morning, generally 8am to 9am EST. Each month, the percentage of "cloudy mornings" with cloud exceeding 4 octas is tabulated. Mean values are calculated for the calendar months in a standard period of a decade beginning June 1999. These mean values are subtracted from actual values to give a "cloudy mornings % anomaly".
Using the CUSUM method, the cumulative sum of these anomalies is graphed and trend lines are fitted.
Further to data presented earlier, a third, even steeper trend line begins in April 2010. Trend line slopes are:
Jun 1999 - Aug 2007: -2.46 percentage units per month;
Aug 2007 - Apr 2010: +9.08 percentage units per month;
Apr 2010 - Jan 2012: +19.70 percentage units per month.

These trend line slopes represent, for the three time intervals, the deviation from the mean of the standard period. On the second graph, they are the horizontal lines. Their values, as listed above, can be read on the vertical axis.

Smoothed anomaly data, shown by purple curves, oscillated pseudo-biennially in the first time interval, merely fluttered in the second, then returned to a pseudo-biennial oscillation in the third. In late 2010, smoothed data showed their highest peak, 25 percentage points above the mean for the standard period.

The horizontal red lines relate closely to simpler data: the recorded values of percent cloudy mornings. This is so, because the percent cloudy mornings varies little with the month of the year. I have been able, with little error, to label each horizontal line with the actual mean percent of cloudy mornings in the time interval:
Jun 1999 - Aug 2007: 27.4%;
Aug 2007 - Apr 2010: 37.7%;
Apr 2010 - Jan 2012: 48.8%.

When the average number of cloudy mornings almost doubles, I would expect drastic changes affecting our lives: growth of crops and insect and fungus attack on them, and reduced solar power generation.

Just a personal observation SB but i think you might be onto something with your cloud cover although you will have to take the La Nina effect into account.
We have had very little rain down here in Horsham in west Vic over the entire summer.
No rain for February so far, 8.6 mms in Jan, 31.2 mms in December with 29 mms of that in one storm and in November dribs and drabs that came to nearly 50 mms.

BUT unlike the last couple of decades of clear skies for days on end, there has been cloud of some sort nearly every day down here for months. It is not continuous but often considerable patches of high level cirrus plus some stratus is drifting through nearly every day.
But the interesting thing which as an old glider pilot you will probably appreciate is that this year for the first time for at least 20 years or maybe closer to 30 years we are getting CU nearly every day and some of that is with an 8 -9000 foot base. and some pretty good soaring with decent thermals instead of the miserable broken, rough as guts, clear sky stuff we have been flying in for the last 20 years at least.
This year was starting to show signs of the cycle returning to the 1970's boomer soaring years with those high cu's of the times at 10 to 12000 feet, booming lift and rocket ship like speeds or at least they were for Boomerangs and things like Libelles and etc.

And it has definitely reinforced my view that we are now entering into a cooling and much wetter phase here in Australia which will run for at least for most of a decade at least with a lot more cloud, much better soaring and maybe stinking wet cold winters again the norm, similar to the 1950's and 1970's.
The location of the critical southern summer trough lines has moved considerably further west into central Victoria from the long time eastern Victorian position and unlike past years there has been a proliferation of almost daily troughs forming somewhere over SE Australia and over WA.

To do all of this the local atmosphere must be getting an injection of summer tropical moisture in a big way compared to the couple of decades past so the future will be interesting to say the least.

I believe we have turned the corner into a new weather and climate cycle not that dissimilar to the cooling 1970's when the belief was that there was a new Ice age just around the corner. And that new cycle might well be ongoing for many years into the future here in western Vic and SE Australia

Thanks, ROM.
How do you suggest I should "take the La Nina effect into account"?
I don't see that the two sudden leaps in cloudiness relate to La Nina at all.

Without siting down and figuring the mechanism, I figured that as the La Nina was pumping more water vapour down into eastern Australia and hence much greater rainfall that might be one reason at least for greater cloudiness.

Incidentally this evening when i had look at the WZ Australian radar I boggled at the extent of the lightning across Australia. Things are getting damn unsettled and unstable across an enormous area of the continent.
Results/ who knows but it could really get wet in the near future.

Sorry Queenslanders and NS Welshmen , the big city papers and media don't even mention you and your ongoing soul destroying battle with floods any more.
We country folk are the forgotten people the third class residents who are only of use to wring cheap food and exports out of without bothering to give anything in return, exports which are used to pay for all the imports and luxuries which the cities don't bother to produce any more.

Maybe Keating was right after all with his Australians becoming the poor white trash of Asia.

Manilla Smoothed Monthly Anomalies of Climate Variables
Parametric Plots
Update for February 2012


Winter 2011 a time of trend reversal
Fully smoothed data points are now available for Winter 2011. Most variables show stable near-zero anomaly values at a time of trend reversal:
Max temp anomaly (x-axes, all graphs) had been rising since cold October 2010, but peaked in June 2011 while still negative.
Rainfall anomaly, which had reached a minimum just below zero in May, rose quite rapidly during the winter.
Cloudiness varied little around a minimum in July.
Uniquely, Dew Points were very low: the anomaly reached a new record low smoothed value of -1.70 degrees in July. This beat the drought year smoothed value of -1.48 degrees of September and October 2002.
Daily temperature range was slightly low and falling; so was daily minimum temperature. Subsoil temperature was slightly high and falling.

February 2012 very cold and wet
Recent partly-smoothed data points generally diverge rapidly from the near-neutral values of winter. February raw anomalies include extremely high rainfall and extremely low daily maximum and minimum temperatures.
Low humidity persists, marked by negative Dew Point anomalies, despite extreme rainfall and cloud cover.


Note:
New data for February 2012 allow updating with more smoothing applied to all months back to August 2011, which is now fully smoothed. Fully smoothed data - gaussian smoothing with half-width 6 months - are plotted in red, partly smoothed data uncoloured, and unsmoothed 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.

A Decade of Summer Climate Change

I have used mean data from 13 summers (December, January and February) at Manilla, NSW to show changes in summer climate that have happened in the first decade of the 21st century.
Each climate variable is expressed as as an anomaly (Note 1.) I have divided the variables into two groups: Heat Indicators and Moisture Indicators. For the moisture indicators I have devised a "Moisture Index" as a summary variable (Note 2.)
Graphs show year to year variation in the anomaly values. Trends do not seem to be linear, so I have fitted second degree (parabolic) trend lines.

Heat Indicators
Short-term variation in summer heat made the 1st, 3rd, 6th, 9th and 13th summers on the graph cooler than their neighbours by a degree or more. Warmer summers formed two groups of two, then the 10th, 11th and 12th as a group of three.
Trend lines were all convex upwards. They all began half a degree low at the start of the decade ('00-'01). By the middle of the decade ('05-'06) they were about half a degree high: summer climate had warmed by one degree. Differences had appeared:
* daily min temp peaked in this summer ('05-'06) at +0.6 degrees;
* daily max temp had peaked earlier, in '04-'05 at only +0.4 degrees;
* daily mean temp was between the two;
* subsoil temp peaked with daily min temp, but lower.
At the finish of the decade ('10-'11), the trends of daily min temp and subsoil temp had fallen back to the values they had at the start (-0.5). However, the others ended much colder: the daily max anomaly was down to -2.0 degrees, and the daily mean anomaly down to -1.3 degrees.
Judging by the daily max trend-line, at the end of the decade summer days were 2.4 degrees colder than they had been six years earlier, and 1.6 degrees colder than at the start of the decade.

Moisture Indicators
Short-term variation in summer moisture tended to alternate wet and dry summers. The variables did not always move together, particularly in later years.
All trend lines for moisture indicators rose through this decade. Cloudy days and rainfall rose the most (2.2 "degrees") and Dew Point the least (0.2 degrees). The two trend lines for rainfall and Dew Point were convex upwards, like those of the heat indicators. This made the middle of the decade moister than the mean of the ends. Trend lines for cloudy days and (minus) temp range (Note 3.) were concave upwards, making the middle of the decade less moist than the mean of the ends.
By chance, the composite Moisture Index trend is a straight line, rising by 1.6 "degrees" in the decade. At the finish of the decade, two moisture variables stood away from the steadily-rising Moisture Index: cloudy days were very high and rising very rapidly; Dew Points were low and falling.

Note 1. Anomalies are values with the mean value of a reference period subtracted. For rainfall, the reference period is the current Australian standard period 1961-1990. For other variables that I have observed myself, the reference period is the decade March 1999 to February 2009.
Note 2. The Moisture Index is a weighted mean of the anomaly values of variables that are moisture indicators. The weighting re-scales the spread in data values so as to give similar spreads on the graph. The Moisture Index is expressed in degrees Celsius, like anomaly values for Dew Point and daily temperature range.
Moisture Index = ((Dew Point Anomaly) + (Minus Daily Temp Range Anomaly) + (Percent Cloudy Mornings Anomaly/10) + (Season Rainfall Anomaly in mm/50))/4.
Note 3. Daily temperature range is a negative moisture indicator: a narrower range indicates greater moisture. The sign of the anomaly is changed here to agree with the other moisture indicators.

A Decade of Winter Climate Change

My trend analysis of climate variables at Manilla NSW in 13 summers (above)
is matched to some extent by my earlier analysis of trends in 13 July months.
However,it is clearly not comparing "apples to apples" as the phrase goes.

The following graphs and notes for 13 winters allow 1:1 comparison of summers and winters. The method and "Notes" given in the previous post apply.

Heat Indicators
Short-term variation in winter heat was chaotic. Often the anomaly of daily maxima was in the opposite sense to that of daily minima.
Contrary to summer results, trend lines were all concave upwards. They were all near zero just before the middle of the decade. Max temps began the decade at +0.4 degrees and ended it at -0.4 degrees for a total fall of 0.8 degrees. Min temps rose a total of 1.3 degrees, starting at -0.1 and finishing at +1.2. As a result, mean temps rose about 0.2 degrees from start to finish, but the rise from the minimum at mid-decade to the end was +0.4 degrees. Subsoil temps started and finished with mean temps, but were 0.1 degrees lower at mid-decade.


Moisture Indicators
As in summer, short-term variation in moisture tended to alternate wet and dry winters.
Trend lines for moisture indicators in winter varied in every way, but moisture generally increased. The only trend that was convex upwards, showing relatively higher moisture at mid-decade, was rainfall. Rainfall anomalies were near -0.5 "degrees" at the decade start and finish, and -0.1 at the middle.
The trend of (minus) daily temperature range increased uniformly, and more steeply than in summer, starting at -0.5 and finishing at +1.5. (That is, the winter daily temperature range fell from 16 degrees to 14 degrees through the decade.)
The remaining two moisture variable anomalies showed concave-upward trends. Cloudy days increased almost as much as in the summer trend, rising rapidly at the end. Dew Point anomalies fell from +0.8 at first to -0.5 in the 7th year, then rose slightly to -0.4.
The composite "moisture index" trend was slightly concave upwards, steady at -0.2 "degrees" at first and ending at +0.5.

Manilla Smoothed Monthly Anomalies of Climate Variables
Parametric Plots
Update for March 2012

March 2012 retreated from cold and wet extreme.
New raw data for March 2012 anomalies show a retreat from low extremes towards normal for max temps, min temps, and temperature range. Rainfall (perhaps temporarily) moved from extremely positive to negative, and subsoil temperature from negative to positive. Cloudiness remained very high, and Dew Point very low.

Fully-smoothed data points for September 2011 include a new 13-year record high monthly rainfall anomaly of +28.6 mm (first graph). This beats the record of +20.0 mm set in November 2005. The new record is sure to be beaten by October and November 2011 when their values can be smoothed.
Although the smoothed value for Max Temp in September 2011 was not very low, the smoothed value for December 2011 may reach a record low.


Note:
New data for March 2012 allow updating with more smoothing applied to all months back to September 2011, which is now fully smoothed. Fully smoothed data - gaussian smoothing with half-width 6 months - are plotted in red, partly smoothed data uncoloured, and unsmoothed 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.

Recent Variation in Humidity

I have noted recent very low values of Dew Point anomaly at Manilla in parametric plots posted above. The centre left graphs present early morning Dew Point anomaly vs. daily maximum temperature anomaly. July 2011 had a 13-year record low value for smoothed monthly Dew Point anomaly.
Two things make this data hard to interpret.
1. From month to month, Dew Point tends to vary directly with air temperature, not inversely as implied by my model (blue line) of (QBO) oscillation between "droughts" with high temperature and low Dew Point and "flooding rains" with the reverse.
2. Early morning Dew Point does not relate nearly as well to daily maximum temperature as it does to daily minimum temperature, which occurs at the same time of day.

I have made a new graph for the latest data set, plotting early morning Dew Point anomaly against daily minimum temperature anomaly. Recognising the direct relationship between them, I have now arranged the axes in the normal sense.
Because the observations of Dew Point and temperature were made at the same time of day, points on the parametric plot show the Dew Point Depression by the relation of y-values to x-values.
The graph includes lines of equal Relative Dew Point Depression plotted in green. Plotted values depend on the fact that the blue marker showing zero minimum temperature anomaly and zero Dew Point anomaly (for the standard decade beginning March 1999) does not mark equal values of the two variables. Decade mean Dew Points are 2.5 degrees lower than mean minimum temperatures. A line of constant Dew Point Depression passing through the graph origin has the value 2.5 degrees.

While Dew Point is an index of absolute humidity, Dew Point Depression is an index of relative humidity. Approximately, Dew Point Depressions of 0, 2, 4, and 6 degrees mark Relative Humidities of 100%, 90%, 80%, and 70%.

This graph shows that the record low Dew Point anomaly of July 2011 was not as extremely arid in terms of Relative Humidity as September 2009, which led into the heat-wave of Novenber 2009. Between those two dates, Relative Humidity rose rapidly to a high peak in November 2010, then fell again just as rapidly.
The persistently low Dew Point anomalies of the summer just past did not cause low Relative Humidity, because the air temperature anomalies were also very low.

Humidity variation since 1999

Variations in Relative Humidity at Manilla, expressed as Dew Point Depression, were traced for the last 36 months in Post #1089152.
The graphs here extend back to September 1999, the limit of my gaussian-smoothed data.

The first observations have the highest Relative Humidity on the record, showing early morning Relative Dew Point Depression (RDPD) of less than half a degree, or a Relative Humidity of more than 97%.
From November 2000 to October 2001, while both (minimum) temperature and Dew Point fell, RDPD stayed constant at 1.7 degrees. RDPD then rose rapidly to 4.4 degrees, showing extreme aridity in the drought of October 2002, and rapidly fell back to normal by March 2003. Values then remained normal to the finish point of the first graph.
On the second graph, Only the final three data points for winter 2009 (also shown in Post #1089152) have high RDPD values showing relative aridity. All the other points have moderate values of RDPD between 1.7 degrees and 3.3 degrees.
This long period of stable RDPD values began in data plotted in the first graph: in total, it lasted for more than six years from February 2003 to May 2009. While RDPD (and Relative Humidity) remained normal, both temperature and Dew Point cycled through a large range as a Quasi-Biennial Oscillation.

The trace, as plotted, cycles anti-clockwise in most cases. The typical cycle has a maximum value of (minimum) temperature anomaly followed after a month or so by a maximum value of Dew Point anomaly. Similarly, a minimum value of (minimum) temperature anomaly is followed after a month or so by a minimum value of Dew Point anomaly.
In this record, there were two large excursions into aridity in terms of Dew Point Depression, peaking in October 2002 (4.3 degrees) and in September 2009 (4.5 degrees). These cycles also were anti-clockwise but the minima of Dew Point anomaly came nearly a year after minima of (minimum) temperature, and were promptly followed by temperature maxima. A later very brief excursion into aridity came in July 2011.

Climate in one summer like climate in the winter before?
In the first post in this thread (#3744) I suggested discussing: "whether a warm summer follows a warm winter, and such questions." With thirteen years of data now, I'll have a go at it. (Answer: Perhaps they do!)

Graphs here repeat parts of graphs in Post #1095485 and Post #1094508.

Temperature anomalies
The first graph brings together the seasonal anomalies of mean daily temperature for thirteen winters (bottom axis) and thirteen summers (top axis). Winters are in the same columns as the next summers.
To the eye, the sequence of temperature anomalies of winter matches that of the next summer well. Some extremes also match: the two warmest winters are followed by the two warmest summers, and the coolest winter is followed by the second-coolest summer.
Lines linking successive values either rise or fall (except in two cases).
Links move in the same sense in 8 cases (of 12);
Links move in the opposite sense in only 2 cases.

The second graph moves the winter data one year to the left, so that each column compares a summer with the next winter.
To the eye, the sequences do not match, nor are they clearly opposed to each other.
Looking at the links (11 cases):
Links move in the same sense in 4 cases;
Links move in the opposite sense in 5 cases.

Moisture anomalies
Rather than rainfall anomaly, I have chosen to use my composite "Moisture Index" anomaly (See Note below.).
The third graph is like the first, but shows Moisture Index rather than Mean Temperature anomaly.
To the eye, the sequence of moisture anomalies of winter matches that of the next summer well. Some extremes also match: the two most moist winters are followed by the two most moist summers, and the winter of 2006 and the following summer are both the second "driest".
Lines linking successive values either rise or fall.
Links move in the same sense in 9 cases (of 12);
Links move in the opposite sense in only 3 cases.

The fourth graph, like the second, compares a summer with the next winter.
To the eye, the sequences do not match; to some extent, they seem to be opposed to each other.
Looking at the links (11 cases):
Links move in the same sense in 4 cases;
Links move in the opposite sense in 7 cases.

Conclusion
For this site, on the North-west Slopes of NSW, there is a tendency for winter temperature and moisture anomalies to predict temperature and moisture anomalies in the same sense in the following summer. Summer anomalies do not predict those of the following winter.

Note. "Moisture Index" is defined in Note 2 of Post #1094508, . Rainfall amount does not always reflect other measures of moisture. In this case, rainfall in winter 2003 was less than expected from other measures, and rainfall in summer 2011-12 was more than expected from other measures, particularly the very low Dew Point.
In this data set, Moisture Index gives a much better winter-to-summer match than does rainfall anomaly.

Re: Climate in one summer like climate in the winter before?

For those who like to see predicted variables regressed on predictor variables:

Mean Temperature anomalies in degrees:
Summer on preceding winter:
y = 0.60x - 0.34 (R^2 = 0.05)

Winter on preceding summer:
y = -0.06x + 0.14 (R^2 = 0.02)

Moisture Index in "degrees":
Summer on preceding winter:
y = 0.67x + 0.20 (R^2 = 0.38)

Winter on preceding summer:
y = 0.06x + 0.05 (R^2 = 0.00)

Season rainfall anomaly in mm:
Summer on preceding winter:
y = 0.34x + 10.00 (R^2 = 0.03)

Winter on preceding summer:
y = 0.02x - 18.84 (R^2 = 0.00)

Since the factor next to the "x" should be around 1.0, and the "R-squared" value should be not much less than 1.0, there are only two equations here that are perhaps not trivial:
For mean temperature anomalies, summer on preceding winter, the regression explains 5% of the variation
For Moisture Index anomalies, summer on preceding winter, the regression explains 38% of the variation.
Both these relations are direct, with slopes of 60% and 67%.

Why do winter climate anomalies predict summer climate anomalies?

At my site, temperature and moisture anomalies of one winter predict those of the following summer in some years. Why is it so?
Do these anomalies stay steady from winter, through spring to summer? No, they do not. Often, anomalies go to peaks or troughs in spring. I noted this in the Opening Post, but I can try to make a better case.

These two graphs are logs of my record of smoothed monthly anomalies of mean temperature and "Moisture Index". Similar, earlier graphs are in Posts #1003011 and #1003132 .
On these graphs, I have marked July values, representing winter, with blue squares and January values, representing summer, with orange squares. Red arrows from blue squares to orange squares show the change in anomaly from winter to the following summer. Many arrows are nearly level, showing that the anomalies are almost the same.

Where the arrows are nearly level, the anomaly curve from winter through autumn to summer is far from level: usually it loops through a maximum or minimum value in autumn that is quite different to the winter and summer values.
This is true for both the temperature and the moisture graphs. Years that show it best for temperature are 2000, 2001, 2002, 2004 and 2009. Those for moisture are 2005, 2008, 2009 and 2010. There are a few years when steeply-sloping arrows show that summer anomalies are completely different from those in the preceding winter. However, if you try inserting arrows in the same way between one summer and the following winter, you will find that nearly all arrows slope steeply: there is no relation to be found.
Since similar anomaly values in winter and a following summer do not occur with similar values also in spring, their similarity must be due to being symmetrically placed on rising and falling slopes to a peak or trough in spring.

Which are the preferred months for peaks of climate anomalies?

Graphs in the previous post identified months of peaks and troughs of climate anomalies at Manilla, NSW in the past 12 years. The first two graphs here summarise the data by calendar months, with a smoothed ((1:2:1)/4) red line showing which times of the year are preferred. Peaks and troughs of temperature occur most often near mid-October; those of Moisture Index near early October. Posts earlier in this thread anticipated both the mid-spring timing and the time for moisture being ahead of that of temperature.
Peaks and troughs in these climate anomalies are about three times more common in spring than at any other time of year.

For comparison, I have made a similar graph for ENSO for the same period. Data are taken from the attached graph that logs the variation of ENSO, with the peaks and troughs of the oscillation marked.
Again, ENSO peaks and troughs occur mainly in only one or two months: November and December in this case. The annual pattern is similar to that of peaks and troughs of temperature and moisture anomalies observed at Manilla. As noted before , ENSO peaks and troughs lag far behind those of climate variables observed at Manilla: 5 weeks behind temperature, and 7 weeks behind moisture.

hi surly bond
l was searching the www for dew point anomalies for queensland and l found a link to your study in manilla nsw
noticed a very high peak in the latest data after a strong dip. it looks very dramatic? your commentary?
could you give any explanation for this
is it a la nina phenomena?
dew points have been around 19-20 deg c everyday in coolangatta for nearly 2 weeks now . looking for some anomaly data for these dew points in april regards

Hi crikey
I can't work out what Dew Point figures of mine you are looking at. The graph you copied is not current, it does not show Dew Points, and the last points are extreme because they are not smoothed.
Plots including Manilla Dew Points up to March this year are in this thread:
http://forum.weatherzone.com.au/ubbthrea...var#Post1097154
and:
http://forum.weatherzone.com.au/ubbthrea...var#Post1098152
Notice that the Dew Point scale is inverted in the first one (centre left graph).
As you see, we have been in a phase of very low Dew Points at Manilla for a whole year, despite high rainfall and very cloudy skies (and despite La Nina).
This month, April, began with two weeks of low early morning DP's, about 5 degrees, but in the last week they have been 12 degrees. Neither is far from the April normal of 9 degrees.

A fabulous study you have going on here surly bond..
does your latitude and longitude contribute to some of those lower dew points? are you inland? flat landscape?
i am in qld atm and from Vic . I am just fascinated in the change of dp's as you head coastal north,
I don't know much about sub tropical weather but l suppose this long run of high dp's on south coast qld are related to persistent easterlies coming in from the tasman sea ? which may have warm anomalies. I am just speculating?
Is it the cool southerlies that bring your dp's down in Manilla?
Factors that effect your dp's?

Anomaly values, especially Dew Points

It may not be quite clear that nearly all the monthly variables that I post are not actual values but anomaly values. The anomaly value "-2.0 degrees", for example, is two degrees lower than the normal value for that month of the year.
My normal values for monthly rainfall are means for the 125-year record to 2008.
Normal values for variables that I have observed myself are based on means for the decade beginning March 1999.

Since "crikey" is interested in Dew Points, I will list the Manilla normal values that I use. These have a degree of fudging, since I used an inadequate psychrometer from 10/7/05, replaced with a better one only from 1/12/09. I have adjusted the earlier values. Before 10/7/05, when I had no observations, I have adopted Tamworth 6am values with 0.8 degrees added.

Manilla, NSW normal values for Dew Point at the time of early morning temperature minimum:
Jan 14.2
Feb 14.9
Mar 13.0
Apr... 9.0
May.. 5.2
Jun... 3.7
Jul.... 2.4
Aug.. 2.5
Sep.. 5.4
Oct... 8.3
Nov 11.2
Dec 13.2

Note that Manilla has an extremely wide range of monthly mean Dew Points: 14.9 - 2.4 = 12.5. This is in keeping with the wide range of monthly temperatures. Coastal places, and many tropical places would have much narrower ranges in both.