Major Error in EC Arctic Temp Data?

See this article here:

http://wattsupwiththat.com/2010/04/22/dial-m-for-mangled-wikipedia-and-environment-canada-caught-with-temperature-data-errors/

What Can We Expect for Summer 2010?

Of course predictions of future climate is always wrong, to some degree. But it is often said by meteorlogists you see on TV things like "we had a bad summer last year, so it will be good this year". Or "we had a hot spring so expect a cold summer" and so on.



But what does the data actually tell us? Does what happen in the spring affect what happens that summer? That is, is the deviation from spring's average affect the deviation from the summer average?



Spring average is tough since the temperature is increasing during the months of March, April and May. So to get an "average" for spring and see the deviation of each year (spring anomaly) I did an average for each week, subtracted each week's average from that and averaged the anomalies for those weeks into one number for the spring. This way I could easily break spring up into smaller segments if need be.



Summer deviation from the average was simply the subtraction of each day June, July and Aug from the average temp for those months.



In all cases above the max temp was used, not the mean. Station 4333 (Ottawa) was used.



This is the scatter plot of the spring anomaly vs the summer anomaly.








Notice that there are equal number of points in each quadrant. Thus the spring has no effect on what the summer will be. Hotter springs can have hot or cooler subsequent summers.



But does a hot summer mean the following year's summer will not be as hot? To do that you again have to plot each summers max temps from the average of the summer max temps to see deviations (anomalies). Does a negative deviation in one year mean a positive deviation in the following year? Yes it generally does.




From just 1930 onward, shows a flatter trend.





Checking against the mean and lowest of the max temps we see the same back and forth swinging around the baseline for the anomaly.










Is this surprising? Not really. Any time there is an average, with a variation in each year, it will have to swing back and forth around that average.



So what will happen this summer? Is it possible to calculate or at best estimate if it will be hot or not?



Plotting only those years above the baseline against the next years anomaly shows this scatter.








It shows that a hot summer has just as much of a chance to produce another hot summer in the following year as a cool one.



However, things are a little different if you look at the mean temps in this manner. This plot is only those summer years where the deviation is positive (above normal) plotted against the deviation of the following year. When the deviation has a small positive, there is equal chance the next year will be positive or negative. But those years with a high anomaly, the following year tended to be cooler. 2C+ anomaly for a summer mean and the following year will be below the baseline. This needs to be checked with other stations, but it would be nice if we had more than a few points of data to confirm this.




2009 summer highest max temps was below the baseline, 2008 was even lower. If you look at the means, 2005 was the last above the baseline, with each subsequent year below. Looking only at that graph one could claim that this summer should be at or above the baseline giving us an average max temps between 25 and 27C, and a max temp of around 35.



This could be a good tomato season (last year was too cool).



Come Sept we will know for sure.

2000 paper on Canadian Surface Temps

In 2000 this paper appeared:


Temperature and Precipitation Trends in Canada During the 20th Century
Xuebin Zhang, Lucie A. Vincent, W. D. Hogg, Ain Niitsoo
AtmosphereOcean (2000)Volume: 38, Issue: 3, Pages: 395–429



http://www.cmos.ca/Ao/articles/v380301.pdf


What is interesting in this paper is the conclusion (best to show that first):

Like other parts of the world, Canada has not become hotter (no increase in higher quantiles of maximum temperature), but has become less cold.

Well, isn't that interesting. Exactly what I'm seeing. Let's continue as there is more interesting stuff here.

Based upon the gridded datasets, the time series of annual mean temperature anomalies relative to the 1961–1990 mean, are computed and shown for southern Canada (1900–1998) in Fig. 3. There is a statistically significant positive trend, which accounts for an increase of 0.98C, for the region during the period. The linear trend is not exactly monotonic. The rises of temperature prior to the 1940s and after the 1970s account for the significant trend. There is a modest decrease during 1940–1970. Trends differ for different regions, and for different seasons, as well as for daily maximum and minimum temperatures.

Yep, that's what I see in the all the stations. That 1940-1970's pause and/or drop in the average. But there is more:

The greatest warming, which is in the Prairies, is about 1.58C over the 99-yr period. The spatial patterns of the trends differ from season to season. The mean daily maximum temperature has increased over all of southern Canada in both
winter and spring. However, it has increased in some areas but decreased in other
areas during summer and fall. Among the four seasons, spring shows the greatest
warming. The spatial pattern in this season is similar to the annual one, except that
spring warming is stronger and the area with significant upward trend has expanded
from the Prairies to include northern B.C. and Manitoba. The greatest warming during spring is well over 28C for the 1900–1998 period in the Prairies. Warming during winter is more than 1.58C during 1900–1998 in western Canada, however, the trends are significant only in southwestern B.C. Summer maximum temperature
shows significant positive trends in Quebec and the Prairies, and significant negative trends in southwestern B.C. Fall shows no significant warming or cooling trends. It is apparent that warming in spring maximum temperature contributed the
most to the positive trend in the annual mean of daily maximum temperature.
Strong warming is the sole characteristic of the minimum temperature. This is
clearly shown in Fig. 5, where no negative trends can be found. Annual mean minimum temperature has increased from about 1 to 2.58C during the last 99 years, with strongest warming in the Prairies and southern Quebec. The trends are statistically significant over all of southern Canada. Spring minimum temperatures have warmed the most, with a rate over 38C during 1900–1998 in the northern Prairies. Winter has the second highest warming rate with some areas in the Prairies and B.C. reaching as high as 38C, however, some of the warming in Manitoba, Ontario and along the east coast is not significant. In summer, the spatial pattern is quite uniform with significant trends from 1.3 to 2.08C. During the fall, the minimum temperatures have warmed with significant increase in eastern Canada and along the west coast. Overall, these results clearly show that daily minimum temperatures, indicators of nighttime temperatures, have significantly increased throughout southern Canada over the past century.

...

Significant positive trends were also found in the annual and seasonal daily mean
temperatures. The spatial patterns (not shown) are similar to those of minimum temperatures but the trends are of a lesser magnitude. This suggests that it is the strong and positive trends of the minimum temperatures, especially during spring and summer, that contributed the most to the trend in the mean temperatures.

...

Greater warming in the minimum than in the maximum temperatures results in
significant decrease in the daily temperature range since the beginning of this century (Fig. 6). Significant decrease in the DTR is observed from coast to coast and for all seasons, with a trend of –0.5 to –2.08C during 1900–98. We can conclude that night-time temperature has increased more than daytime temperature in all seasons in the southern part of Canada during the last century. Most of the decrease in the DTR occurred prior to the 1950s, especially late in the first half of the 20th century, coinciding with an increase in total cloud amount in Canadian mid-latitudes during the first half of the 20th century (Henderson-Sellers, 1989; McGuffie and Henderson-Sellers, 1988). Henderson-Sellers (1989) did not propose any specific reasons for the cloud increase. Significant decrease in the DTR did not occur in the second half of the century when the greatest increase in greenhouse gases took place. This suggests that trends in the DTR are closely related to changes in total cloud amount. The trends in both DTR and total cloud cover differ from one season to the other. Future investigation into the relationships between the changes in DTR and cloud cover is needed.

It needs to be understood that they gridded the country as such:

To reduce the amount of calculation, trend analysis was performed on the time series of a coarse 200 × 200 km grid obtained by averaging the values of 16 grid points from the 50 × 50 km grid. This procedure further smoothes the fields. Since the climate observation network was not established in northern Canada until the late 1940s, there were large portions of the country with no data during the first half of the century. Based loosely on error analyses reported in Milewska and Hogg (unpublished manuscript) and those reported here, interpolation limits were established. No more than five stations with the shortest distance (less than 750 km for precipitation and 1000 km for temperature) to a grid point are used to interpolate the value for the grid. The distance limits are seldom reached. For example, more than 80% of grid values for precipitation have stations within 200 km. No grid values were generated for northern Canada (north of about 608N) prior to 1950s due to insufficient data. Trend analysis was performed on datasets for 1900–1998 for the southern part of the country (south of about 608N) where grid values were complete for both temperature and precipitation, and for 1950–1998 for the nation as a whole.

Ok, let's take that on face value. Though from what I've seen with differences in close by stations, I'm not happy with their approach. Seems this differences between close by stations to them is "noise" that needed smoothing out.

Regardless of this methodology to smooth out the data, their conclusion is clear:

Annual mean temperature has warmed an average of 0.98C in southern Canada
over the last century. Associated with this increase in mean temperature is a relatively smaller increase in daily maximum temperature and a larger increase in daily minimum temperature. In this century, the increases have resulted in a decrease in diurnal temperature range by 0.5 to 2.08C. The bulk of decline in DTR occurred during the first half of the century, coinciding with an increase in cloud cover during that period. Both of these results are broadly consistent with greenhouse gas induced climate change, but the timing of the changes, coming prior to the most significant increase in greenhouse gases, suggests that other mechanisms may be responsible. Examining the areas affected by abnormal and extreme temperature confirmed the above analysis. It also suggested that the probability distribution of minimum temperature has shifted with a higher mean but only the left-hand side of maximum temperature distribution has been shifted upward. This indicates that southern Canada has not become hotter but less cold.

At least they are honest.

More studies are needed before we can conclude that such changes are the manifestations of anthropogenic climate change.

Canadian Summer Trends

This paper Detection and attribution of climate change: a regional perspective (http://www3.interscience.wiley.com/cgi-bin/fulltext/123310513/PDFSTART) claims to evaluate the tends in temperatures as being caused by human factors (specifically AGW).

This is common verbiage:

This study found a detectable change over the 20th century in decadal mean temperatures over each of the six populated continental areas (Europe, North America, South America, Asia, Australia, and Africa), and furthermore found that these changes could only be reproduced with the inclusion of anthropogenic
greenhouse gas emissions.

They did this comparing of the temps to computer models of forcings. Well, that sums up their findings right there. Computer models, nothing more than very expensive "what if" computer games (and I'm a software developer.)

They go on to say:

Jones et al.33 examined summer (June–August) mean temperatures over the past
century over a set of standard subcontinental regions of the Northern Hemisphere. These subcontinental regions divide each of the six continental region into a small
number (between two and six) subregions chosen to represent different climate regimes.34 When signals were regressed individually against the observations,
an anthropogenic signal was detected in each of 14 regions except for 1, central North America, although the results were more uncertain when anthropogenic
and natural signals were considered together.

Ok, we have this data for Canada, let's have a look at July and August temps for all regions. (top line is the highest of the max temp, the upper orange line is the upper standard deviation of the max temps, the black line is the average of the mean, the lower orange line is the lower standard deviation of the min temps and the bottom line is the lowest of the min temps) Thus 65% of the temperature data falls within the upper and lower standard deviations.

From west to east:

BC: BELLA COOLA (Stn 380) Converge in year 2644 at 22C


BC: VICTORIA GONZALES HTS (Stn 113) Converge in year 2850 at 23C


ALB: CALGARY INT'L A (Stn 2205) Converge in year 3100 at 23C


SASK: CHAPLIN (Stn 3080) Converge in year 2681 at 12C


MAN: BIRTLE (Stn 3469) Converge in year 3195 at 14C


ONT: OTTAWA (Stn 4333) Converge in year 2682 at 19C


ONT: LONDON (Stn 4798) Converge in year 3660 at -7C


The rest of the data is far too short, missing records, so how they can possibly justify their claim, with so much missing data, is a mystery.

This study also claims summer temps are increasing, from anthropic forcings, and thus is causing more forest fires: http://www.cccma.ec.gc.ca/papers/ngillett/PDFS/2004GL020876.pdf

The result is clear from the actual station data. The highest of the max temps in the summers (day time highs) in all regions of Canada are dropping. The lowest of the summers (night time lows) are increasing. Thus the range of extreme temps is converging with a meet some 800 years in the future when the nightime and daytime temps would be the same. Since this is NOT physically possible, thus these trends MUST change direction at some time before then and start to diverge. Thus this is only part of a cycle.

Thus this change in summer temps has NOTHING TO DO WITH ANTHROPIC FORCINGS.

Southern Ontario April Heat Wave

Record-breaking temperatures this weekend: Environment Canada

Said the Toronto Star: http://www.thestar.com/news/canada/article/789040--record-breaking-temperatures-this-weekend-environment-canada?bn=1

"In Ontario and Quebec, forecasters are expecting temperatures to exceed seasonal norms by as much as 15 to 20 degrees, he added.

“Smashing the records, clobbering it, pulverizing, whatever you want to say ... These are things that you should see two, two-and-a-half months from now,” Phillips said in a telephone interview."

Really? Well, let's see.

Ottawa hit 28C today. A record breaker -- for that day. But is this beyond normal for highest temps reached in the first week of April? How many times has the temp been above 20 in the first week of April in the past 110 years? 7 with 8 days above 20C.




Making this year above "normal". Normal would be within 1 standard deviation of the temps around the average since 1900.




The average is 11.5 +/- 5.02. Two standard deviations puts the highest at 21.5C. So 28 in the first week of April for Ottawa is truly abnormal statistically speaking.

However, things change dramatically if the last week of March and the first week of April are looked at (the week before and after today).








We now get 11 years, with 20 days. Note the interesting pattern.





1945-1948 cluster of days 20C or more is at the top of the warm trend from 1990 to 1945, then a 29 year gap of no days above 20C, right during the downturn of temps between 1945 and 1975, then out of the blue we go to every 5 to 8 years these nice warm springs.

None of them in the 28C range. Again making this week unusual.

We need to next check to see of the first 2 weeks in April shows anything unusual with this day. That is, is spring shifting earlier as we have noted before.




Normal is 15.1+/-4.75, upper second standard deviation is 24.6C.




An interesting way to see these unusual high temps for the first 2 weeks of April looks like this:




Notice the interesting gaps every few decades when 10 or 15 years have no first two weeks 20C or more.




Looks like normal variation, with other cycles added in. No evidence of global warming trends here. This is just a fluke year in the normal fluctuations caused by climate chaos.

Just to make sure, let's see the entire range of temps for the first week of April:




Interesting indeed. Temperatures for the first week of April swing wildly between day time highs and night time lows. Note specifically the nighttime lows (must be as there are no highest days in that first week below zero) have been warming over the years. More evidence that it isn't the high temps that are changing, but the low extreme temps are rising, pushing the average to rise.

It will be interesting to see how this year fairs with the arrival of spring (last frost night in the spring).

This is the trend for spring arrival for Ottawa.



Clearly spring has been arriving earlier since 1920's but the same as it was at the beginning of the 20th century. The last frost night since the 20's and 30's was from the middle of May then to the end of April now. This extends the spring growing season in the Ottawa area by some 15 days.

This is just a normal cycle. No correlation with CO2 levels.

Next we need to compare the rest of Southern Ontario stations. This plot shows over the years all stations that had a temp 20C or more in the first week of April. Each dot is a station. Temps on the Y axis, years on the X axis:



Gee, we had temps in Ontario in the high 20's and low 30'sC back in the 1920's and 1930's in the first week of April.

This plot shows all stations maximum temps in the first week of April. Notice the wide range and the low R2.



Yes, the trend is to warmer first weeks in April. But we already knew that as spring is generally coming earlier, so yes, the temps during this transition phase of the year would be increasing.

What I find humerus about this is the interviews on TV about how people are liking this warm weather. Every one! Oh, sorry people. You are not allowed to be enjoying this. This is prima facia evidence that humans are heating up the planet and that is bad. So you should be feeling guilty about this warm weather, it's bad for the environment.

Yeah, right...