Anthony Watts and his critical thinking cherry blossom fail

Anthony Watts runs a climate conspiracy blog called WUWT. He dismisses most climate science as "claims" and thinks that global warming is caused by Russian steampipes. Anthony often suffers confirmation bias, meaning he doesn't understand what he reads because his mind, shaped by his personal view of the world, twists it into something else altogether.

Credit: Uberlemur 

Yesterday he decided to take on the National Parks Service and plant researchers, and disputed the main causes of trends in blossoming time of cherry trees in Washington DC. Incidentally, Anthony is still fighting World War II, though it took place before he was born. He wrote:

We can assume that because the Washington cherry trees are curated and budding times tracked by the National Park Service that they get plenty of water and nutrients, after all we can’t have dead and dying cherry trees on the mall. That would insult Japan, who made them a gift in 1912 before they decided to surprise bomb the crap out of us at Pearl Harbor.

Night light affects the timing of budburst in four deciduous species in the UK

But Anthony's harboured resentment of Japan is not the main point of this article. This is about his ignorance of plant science. Anthony found a paper by Richard H. ffrench-Constant et al, which showed that night lights in urban centres can significantly affect the timing of budburst. Interesting. Many flowering plants demonstrate a sensitivity to day length and this has been known for a very long time. Photoperiodism in soybean and tobacco was reported by Garner and Allard in 1920. The unique point about this recent paper was that it studied the effect of artificial light at night time.

What the researchers did was look at the difference of the effect of night time illumination on the timing of budburst, over a 13 year period. They looked across the UK, and matched budburst from four species of deciduous trees with satellite imagery of night-time lighting and average spring temperature. They found that night time lighting and temperature affected the timing of budburst. In the abstract, the authors state:

We find that budburst occurs up to 7.5 days earlier in brighter areas, with the relationship being more pronounced for later-budding species. Excluding large urban areas from the analysis showed an even more pronounced advance of budburst, confirming that the urban ‘heat-island’ effect is not the sole cause of earlier urban budburst. Similarly, the advance in budburst across all sites is too large to be explained by increases in temperature alone.

Now if you read the paper or the press release, you'll notice a number of things:

• The researchers looked at 13 years of data - from 1999 to 2011.
• The species were European sycamore, European beech, Pedunculate oak and European ash, none of which are close relatives of Washington's cherry trees.
• The researchers were comparing areas where there is night lighting with areas without.
• The researchers assumed that higher temperatures (from urban heat island effect) would also affect the timing of budburst, and allowed for that in their analysis.

The main thing that you'll notice is that this is a bit like the urban heat island effect. It will only make a difference to the timing of budburst if there's a change. That is, if night time lighting increases, then it's expected that budburst would occur earlier, all else being equal. That's the same with the urban heat island effect. City temperatures don't just keep increasing relative to rural areas, unless the area is being built up. (Even then, in some cities there's an "urban cool island" effect, where built up areas are cooler than the surrounding rural regions.)

Anthony Watts visits Dunning and Kruger

Anthony, little Dunning-Krugerite that he is, doesn't understand any of that. He barges in and disputes the expert opinion from the National Parks Service. Anthony decides that the new paper means that it's not warmer spring weather that has caused the cherry trees in Washington to bloom earlier over time, it's night time lighting. Anthony jumps in feet first and says that Jason Samenow is wrong in a Washington Post article he wrote four years ago. Now Jason quoted the National Park Service’s chief horticulturist Rob DeFeo as saying that March temperatures are the key predictor of peak bloom dates. Jason also quoted a peer-reviewed paper, which referred to the factors affecting the timing of cherry blossoms in Washington. So Anthony is rejecting all the experts in favour of his own misinterpretation of another paper, where the authors also know what factors affect budburst, while adding another factor.

NPS prediction of cherry blossom time in Washington

This year the NPS predicted that the peak blossoming would take place between March 18 and March 23, amending the date forward because of unseasonal warm weather. On 8 March this year, the NPS released a statement that:

Driven by this week's sustained, unseasonably warm temperatures, the projected peak bloom dates for the cherry blossoms has been revised to March 18-23. Although the National Park Service factored above average March temperatures into the original prediction date, potentially record-setting temperatures, averaging nearly 20 degrees above normal for the next week, have greatly accelerated the bloom watch. In response to the earlier peak bloom, the Tidal Basin Welcome Area, located near the paddle boats at Maine Avenue SW and Raoul Wallenberg Place SW, will now be open from March 18 –April 3 (it was previously scheduled from April 2-17).

The actual date of peak bloom this year was reported by NPS as occurring on March 25, which is pretty close, and closer than April 2-17. I'd be surprised if the street lighting was turned up extra high, as part of a "climate hoax" conspiracy.

Early spring temperatures affect the timing of peak blossom in cherries

Like many temperate flowering trees, the timing and proliferation of bud burst in Washington's cherry trees (Prunus×yedoensis ‘Yoshino’ and Prunus serrulata ‘Kwanzan’) are affected by three main factors. Provided there is sufficient water and nutrients, the main factors affecting the onset of bud burst are:

• the duration and extent of chilling in winter (cold temperatures for winter dormancy), 
• the temperature in spring (how much it warms), and 
• the length of daylight.

With global warming, when there is sufficient chilling over winter, bud burst will occur earlier in warmer springs. (Daylight length over the year isn't affected by global warming.)

Back in 2001, Abu-Asab et al looked at changes over the period 1970 to 1999 in the onset of flowering of 100 plant species (44 families) in Washington DC. Taking all 100 species, the first flowering advanced by 2.4 days over the 30 year period. A subset of 89 species had an advance of 4.5 days in first flowering.

In 2007, Miller-Rushing et al reported findings on the impact of rising temperatures on cherry trees in Japan. The abstract describes how the rise in February-March temperatures led to earlier flowering:

Climate change is affecting plant phenology worldwide. Phenological responses vary among species, but it is not clear how responses differ among closely related species. We examined a 25-yr record (1981-2005) of flowering times for 97 trees, representing 17 species and hybrids of cherry (Cerasus sp. or Prunus sp.) grown at Mt. Takao, in Tokyo, Japan. The cherry trees flowered earlier over time, by an average of 5.5 d over the 25-yr study. Earlier flowering was explained largely by a 1.8°C increase in February-March mean monthly temperatures. Most species and hybrids flowered 3-5 d earlier for each 1°C increase in temperature, but early-flowering taxa flowered as much as 9 d earlier for each 1°C increase in temperature. Flowering durations and differences in flowering times among species were greater in warm years than in cold years. Species and individual trees also flowered longer in warm years. These results show that the flowering times of closely related species may change similarly in response to climate change, but that early-flowering species may diverge from the overall trend in a predictable way. Such changes in flowering may affect gene flow and pollination as the length of the flowering season increases.

Then in 2011, Chung et al wrote a paper about predicting of the blooming time of Washington's cherry trees. In the abstract they pointed out:

For its sensitivity to winter and early spring temperatures, the timing of cherry blossoms is an ideal indicator of the impacts of climate change on tree phenology. 

They also described the trees in Washington DC:

Despite over 60 years of peak bloom data, no study to our knowledge has attempted to forecast the future bloom dates of the cherry trees at the Tidal Basin of Washington, DC. These cherry trees were propagated in 1911 from scions from 12 selections from the Ekita-mura area of Japan and planted in the spring of 1912[14]. The Yoshino (Prunus×yedoensis ‘Yoshino’) cherry is the most abundant cultivar in the Tidal Basin. It is a hybrid of unknown origin from Japan with a significant historical, cultural, and economic importance for the region [20]. Significant genetic similarities have been reported between P.×yedoensis accessions in the Tidal Basin and P. serrulata var. spontanea from Korea [21]. The Kwanzan (Prunus serrulata‘Kwanzan’), a double flowering pink cherry, is another abundant cultivar with 44 trees at the Tidal Basin. These two varieties are also widely cultivated throughout the U.S.

Anthony Watts should take his own advice

Anthony Watts is a prime example of the Dunning-Kruger effect. He thinks he knows more than all the experts. Here is some of what he wrote:

Oh this is funny. I pointed out the other day on Twitter that “Global Warming is the universal boogeyman of the left”. Almost anything that seems slightly out of the ordinary in the natural world now seems to have global warming or climate change immediately slapped on it as the cause. I blame the education system and the lack of teaching critical thinking. 

That's when he launched into an attack on Jason Samenow, who correctly wrote about how cherry trees will bloom earlier when March is warmer. Anthony quoted Jason writing:

In other words, real-world data support the overall idea that the D.C.’s March climate is warming and the blossoms’ bloom dates are shifting earlier in response.

Remember, the NPS  chief horticulturist Rob DeFeo said the same thing back in 2012, as did the NPS press release this year. That doesn't matter two hoots to a pseudo-scientist like Anthony Watts who couldn't think critically if he tried. (The word "pseudo-scientist" is flattery. Anthony doesn't even rank as a pseudo - no-one would mistake him for a scientist.)

Anthony bored ahead, quoting Richard Feynman (one telltale sign of deniers is their penchant for quoting Richard Feynman at the drop of a hat):

Problem is, that’s not science, it’s activism. Dr. Richard Feynman once wrote: "The first principle is that you must not fool yourself and you are the easiest person to fool."

More fool Anthony.

After that he launched into Liebeg's Law of the Minimum, as if scientists have never heard of it before. Pity he doesn't apply it to his "CO2 is plant food" meme.

Next he gets himself all tied up in knots. What Anthony wrote was how temperature and available sunlight are the key determinants in growth. Anthony doesn't bother to distinguish between budburst and overall growth, by the way. After claiming that Jason Samenow was talking through his hat when he reported scientists and the NPS Chief Horticulturist saying that March temperatures are the key determinant of peak blossom timing, Anthony then admits that temperature is a factor.

Thing is, he blames the rise in temperature on urban heat island effect, without checking anything. Now that probably would have played a part in Washington. Thing is, the trend since 1946 isn't all that evident. Some recent March's have been the hottest in the record (since 1946), but the trend is weak (r^2 is very low). Here's the the month of March anomaly from the 20th century mean, in Fahrenheit (from NOAA):

Figure 1 | Washington DC temperature for March only. Data source: NOAA

The recent years that stand out as having a warmer March are in 2000, 2010, 2012 and 2016. Below is a plot of the the peak bloom date each year, expressed as the difference in days from the average for the first ten years (1992 to 2001). Looking at when the timing was earliest, you can see that 2000 is the year when it blossomed earliest, followed by 2012 and 2016. This is consistent with March temperature having an impact on time of blossoming. The years that buck the notion are 1998 (not an especially warm March) and 2010 (when blossoming wasn't particularly early).

Figure 2 | Peak bloom date expressed as the difference in days from the average peak bloom date 1992-2010. Data source: National Cherry Blossom Festival

Who turned the lights on and off in Washington DC?

The question on everyone's mind at WUWT is surely this. Who turned on the lights in 2000, 2012 and 2016? Who turned off the lights in 2005 and from 2013 to 2015? Is there a gremlin in the street lighting in Washington DC?

Dumbo Watts quoted the new paper:

The comparison revealed that in areas with the highest levels of night-time light exposure, citizen scientists reported spotting emerging leaves on trees up to 7.5 days earlier.

Then he commented:

Up to 7.5 days earlier? Gosh, that’s even worse than Samenow’s story claiming 5 days. Better yet, this data was collected by citizen scientists, according to the press release. I’ll bet some of them might even be climate skeptics.

Can you see Anthony's mistake? The paper on lighting is between areas with night-time light and areas without. The article by Jason Samenow was about the trend over time between 1921 and 2011. Anthony has mixed up trends over time with an either/or in the present. In effect he's assumed that Washington DC has gone from no lights to bright night lights and this is the main factor affecting the timing of blossoms. (Refer this paper that goes back to the 9th century - which doesn't support Anthony's claim.)

Anthony blithely put up charts showing not the difference in lighting over Washington, but the difference in lighting over the USA as a whole. It's hard to tell if the night time light over Washington has changed much in recent decades. He did put up a pretty picture of cherry trees at night, writing:

Gosh, global night brightening  seems to be growing faster than global warming. And in Washington, those cherry trees sure seem to get a lot of light at night as this photo demonstrates:

What he didn't demonstrate was that those lights were turned off between 2013 and 2015, or that the lights were turned up very high in the years when March, by pure coincidence I'm sure, happened to be rather hot.

What's interesting...

Note that I'm not disputing the paper about night lighting. I think it's really interesting. What this article was about was how Anthony Watts lacks any skill in the critical thinking that he demands in other people. He also shows a distinct bias in the scientific papers that he likes. Anything he thinks he can twist toward supporting his "climate hoax" conspiracy theory, he'll use, or mis-use, even though it doesn't support his "hoax" theory. Otherwise he dismisses all mainstream science as merely being a "claim".

Citizen science - plant watching

If you want to get involved in phenology (seasonal changes), check out Project BudBurst if you're in the USA. There are probably similar projects elsewhere in the world.

From the WUWT comments

MarkW has a question about budburst:

June 29, 2016 at 9:28 am
Would I be correct in assuming that the “budburst” being talked about here would be flower buds? Not leaf buds.

DHR didn't read Jason Samenow's article very carefully. He was not writing about the average temperature of the contiguous USA. He was writing about the trend in the average March temperature in Washington DC. Not only that, but the average temperature for the USA has gone up by a lot more than 0.5 F in the last 90 years. Last year the temperature was 2.39°F above the average for the twentieth century.

June 29, 2016 at 9:32 am
NOAA data says the contiguous US has warmed about a half degree F (or perhaps one degree F, its iffy) during the past 90 years. I do not know where Mr. Stabenow got his DC temperature data of a rise of 2.3F during that time, but it clearly could not be global warming.

Funny, Mark from the Midwest doesn't know that the cherries bloomed quite a bit earlier this year. Funny, he doesn't realise that Jason Samenow's article is four years old. He wrote:

June 29, 2016 at 9:32 am
Funny, I live in “The Cherry Capital,” (more or less), and everything has been pretty much on schedule for some time now, even a bit late in a few of the past 10 years. Apparently Samenow didn’t read the memo.

Steve Reddish doesn't keep up with research.

June 29, 2016 at 9:38 am
I find it funny that earlier blossom bloom time today is claimed to be proof of global warming caused by humans, but not proof of a MWP in Japan. I don’t think light pollution was an issue back then.
SR

As well as the paper about Japan I mentioned up top, there are lots more in the literature. Here's another one for Steve, from the abstract of an article looking at very long term records of cherry tree phenology in Japan (from the 9th century). Note the opening sentence, which supports Jason Samenow and the experts he cited, and doesn't say much for Anthony Watts' research skills:

The full-flowering date of cherry trees fluctuates in accordance with temperature conditions during February and March. Full-flowering dates were closely related to the March mean temperature by means of a temperature accumulation index, in which plant growth is considered to be an exponential function of temperature

References and further reading

Garner, W. W. And H. A. Allard. "Effect of the relative length of day and night and other factors of the environment on growth and reproduction in plants." Journal of Agricultural Research 18:553-606. (1920) (pdf here)

ffrench-Constant, R. H., R. Somers-Yeates, J. Bennie, T. Economou, D. Hodgson, A. Spalding, and P. K. McGregor. "Light pollution is associated with earlier tree budburst across the United Kingdom." Proceedings of the Royal Society B (2016) DOI: 10.1098/rspb.2016.0813 (open access)

• Night-time light pollution causes spring to come early - press release at ScienceDaily.com

Abu-Asab, Mones S., Paul M. Peterson, Stanwyn G. Shetler, and Sylvia S. Orli. "Earlier plant flowering in spring as a response to global warming in the Washington, DC, area." Biodiversity & Conservation 10, no. 4 (2001): 597-612. (pdf here)

Chung, Uran, Liz Mack, Jin I. Yun, and Soo-Hyung Kim. "Predicting the timing of cherry blossoms in Washington, DC and mid-Atlantic states in response to climate change." PLoS One 6, no. 11 (2011): e27439. doi: 10.1371/journal.pone.0027439 (open access)

D.C.’s cherry blossoms have shifted 5 days earlier: what about global warming and the future? - article by Jason Samenow in Washington Post, March 2012

Aono, Yasuyuki, and Keiko Kazui. "Phenological data series of cherry tree flowering in Kyoto, Japan, and its application to reconstruction of springtime temperatures since the 9th century." International Journal of Climatology 28, no. 7 (2008): 905-914. (open access)