Urban Night Lighting Observations Demonstrate The Land Surface Temperature Dataset is ‘not fit for purpose’


Crossposted from https://wattsupwiththat.com/2022/12/18/urban-night-lighting-observations-challenge-interpretation-of-land-surface-temperature-observations/

Urban Night Lighting Observations Demonstrate The Land Surface Temperature Dataset is ‘not fit for purpose’

Foreword by Anthony:

This excellent study demonstrates what I have been saying for years – the land surface temperature dataset has been compromised by a variety of localized biases, such as the heat sink effect I describe in my July 2022 report: Corrupted Climate Stations where I demonstrate that 96% of stations used to measure climate have been producing corrupted data. Climate science has the wrongheaded opinion that they can “adjust” for all of these problems. Alan Longhurst is correct when he says: “…the instrumental record is not fit for purpose.”

One wonders how long climate scientists can go on deluding themselves about this useless and highly warm-biased data. – Anthony


Guest essay by Alan Longhurst – From Dr. Judith Curry’s Climate Etc.

The pattern of warming of surface air temperature recorded by the instrumental data is accepted almost without question by the science community as being the consequence of the progressive and global contamination of the atmosphere by CO2.   But if they were properly inquisitive, it would not take them long see what was wrong with that over-simplification: the evidence is perfectly clear, and simple enough for any person of good will to understand.

In 2006 NASA Goddard published two plots showing that the USA data[1] did not follow the same warming trend as the rest of the world. Rural data numerically dominate the USA archive, while urban data massively dominate almost everywhere else.   Observations began very early in the USA – being introduced by Jefferson in 1776 – and that emphasis had already then been placed on providing assistance to farmers.

They are consistent with the ‘global warming‘ that so worries us today being an urban affair, caused not by global CO2 pollution of the global atmosphere but by the heat of combustion of petroleum we burn in our vehicles, our homes and where we work – all of which is additive to the radiative consequences of our buildings and impermeable cement and asphalt surfaces. However, towns and cities in fact occupy only a very small fraction of the land surface of our planet, about 0.53% (or 1.25%, if their densely populated suburbs are included) according to a recent computation done with rule-based mapping. But it is in this very small fraction of land surfaces that most of the data in the CRUTEM or GISTEMP archives have been recorded.

Consequently, very few surface air temperature observations have been made in the small villages which, with their farms and grazing lands, are scattered in the otherwise uninhabited grassland. forest, mountain, desert and tundra.  Nor is it widely understood that our presence there has been associated with progressive change since the introduction of steel and steam to plough the grasslands and to cut forests for timber.[2] 

A measure of the brightness or intensity of night lighting, the BI index, was derived by NASA from the work of Mark Imhoff, who calibrated and ranked night lights in seven stable classes – one rural, two peri-urban and four urban.[3]   The BI indes for airport of Toulouse is at 59 and the central district of Cairo is at 167.  Care must be take with apparent anomalies similar to that of Millau which is an active little town of 20,000 people but it has a BI = 0, as does Gourdon which has only 4000.  This is because the MeteoFrance instruments at Millau have been placed on a bare hilltop on the far side of a deep, unbuilt valley adjacent to the town and so they record only the  conditions of the surrounding countryside.

It is not only in major cities that the effects of urbanisation can be detected; this effect can also be detected in data from some very small places that would otherwise be considered rural as at Lerwick, a port in the Orkney Islands with a population of <7000.  Here, the GHCN-M data from KNMI show a warming of about 0.9oC over the period 1978-2018, while during the same period the day/night temperature difference increased by 0.3oC.  Retention of heat at night is characteristic of urban warming.

But Gourdon, a compact little rural village not far from my home in western France has a BI of only 7 for a population of only 3900.  It is situated in farmland that was abandoned 150 years ago when the vines died, and it is now given over to sheep, goats and scrub vegetation.   Little hamlets in this region are now often dark at night and their road signs may warn you that you are entering a ´Starlit village´.

Despite its deep isolation, there is a manned Meteofrance data station in Gourdon which over a 60-year period has recorded a very gradual and small summer warming since mid-20th century, associated with perfectly stable winter conditions.

Since buildings and human activity have undoubtedly changed at Gourdon in this long period, perhaps especially by the growth of rural tourism, this effect was probably predictable.  The same is seen in data from other small places such as Lerwick, a port in the Orkney Islands with a population about twice that of Gourdon.  Here, GHCN-M data from KNMI show a warming of about 0.9oC over the period 1978-2018 while during the same period the day/night temperature difference increased by 0.3oC.

The BI values for night lighting are in no way influenced by fact that the thermometric data with which each is associated have later been merged with data from another station to achieve regional homogeneity.   Consequently, it is appropriate to associate them with night-light data in the hope of isolating the effects of local combustion of hydrocarbons in towns and cities, from what we must attribute to solar variation.    The consequences of homogenisation on the surface air temperature data is avoided here by the use of GHCN-M data from the KNMI site – which are as close to the original observations, adjusted only for on-site problems, as is now possible to get.

The urban warming phenomenon has been observed and understood for almost two hundred years.  Meteorologist Luke Howard (quoted by H.H. Lamb) wrote in 1833 concerning his studies of temperature at the Royal Society building in central London and also at Tottenham and Plaistow, then some distance beyond the town:

But the temperature of the city is not to be considered as that of the climate; it partakes too much of an artificial warmth, induced by its structure, by a crowded population, and the consumption of great quantities of fuel in fires: as will appear by what follows….we find London always warmer than the country, the average excess of its temperature being 1.579°F….a considerable portion of heated air is continually poured into the common mass from the chimnies; to which we have to add the heat diffused in all directions, from founderies, breweries, steam engines, and other manufacturing and culinary fire..’ [4]  

To Luke Howard’s list must now be added the consequences of the combustion of hydrocarbon fuels in vehicles, mass transport systems, power plants and industrial enterprises located within the urban perimeter, cement/asphalt surfaces and their relative contributions day and night.[5]

The energy budget of the agglomeration of Toulouse in southern France is probably typical of such places: anthropogenic heat release is of order 100 Wm2 in winter and 25 W m-2 in summer in the city core, and somewhat less in the residential suburbs.  Observations of resulting evolution of surface air temperatures in central Toulouse are compatible with the anticipated effect of the inventory of all heat sources seasonally.  Below the urban canopy layer, a budget for heat production and loss through advection into surrounding rural areas has been computed and it is found that this loss is important under some wind conditions.  In this and many other urbanisations, there is also an important seasonality of heat release by passing road traffic that forms a major component of the heating budget, since national highway systems commonly pass close to major centres of population.[6] 

Larger cities, larger effects: in the core of the city of Tokyo during the 1990s the seasonal heat flux range was 400-1600 W.m-2 and the entire Tokyo coastal plain appears to be contaminated by urban heat generated within the city, especially in summer when warming may extend to 1 km altitude, much higher than the simple nocturnal heat island over large cities.[7]   The long-term evolution of urban climates is well illustrated in Europe where, in the second half of the 20th century when their natural association with regional climate was abruptly replaced by a simple warming trend that took them almost 2oC above the base-line of the previous 250 years.

Although, globally, the energy from urban heat is equivalent to only a very small fraction of heat transported in the atmosphere, models suggest that it may be capable of disrupting natural circulation patterns sufficiently to induce distant as well as local effects on the global surface air temperature pattern.  Significant release of this heat into the lower atmosphere is concentrated in three relatively small mid-latitude regions – eastern North America, western Europe and eastern Asia – but the inclusion of this regional injection of heat (as a steady input at 86 model points where it exceeds 0.4W m2) has been tested in the NCAR Community Atmospheric model CAM3. 

Comparison of the control and perturbation runs showed significant regional effects from the release of heat from these three regions at 86 grid points where observations of fossil fuel use suggest that it exceeds 0.4 Wm-2.  In winter at high northern latitudes, very significant temperature changes are induced: according to the authors, ‘there is strong warming up to 1oK in Russia and northern Asia…. the north-eastern US and southern Canada have significant warming, up to 0.8 K in the Canadian Prairies’.

The suggestion that the global surface air temperature data – on which the hypothesis of anthropogenic climate warming hangs – are heavily contaminated by other heat sources is not novel.  The map below shows the locations of 173 stations used by MacKittrick and Michaels for a statistical analysis of the contamination of the global temperature archives by urban heat., using which they rejected the null hypothesis that the spatial pattern of temperature trends is independent of socio-economic effects which was, and still is, the position taken by the IPCC – for which MacKittrick was then a reviewer.[8]

In the present context, this study seemed worth repeating, so a file of 31 clusters of BI indices was gathered from the ‘Get Neighbours’ lists that are shown when accessing GISTEMP data.  These clusters comprise 1200 data files representing 776 towns or cities and 424 rural places – of which 355 are totally dark at night.  They therefore represent a wide range of individual station histories – many longer than 100 years – and are sufficient for the task.   Just 53 of the 540 rural sites listed are in Western Europe, the remainder being located in the vast, night-dark expanses of Asia – where the data based on the arctic island of Novaya Zemyla includes only three with significant night lights,  of which one is the city of Murmansk.

 The cluster centred southeast of Lake Baikal includes two cities (329,000 and 212,000 inhabitants having BIs of only 28 and 13) together with 39 small places – of which 28 are totally dark at night  – while that immediately to the west of Baikal includes 19 such places.   But not all bright locations have large populations, because intensive industrial farms – solar panel energised – can dominate regional night lighting as it does at in some Gulf States: an experimental farm alone here generates a BI of 122, while the 3012 people who live at Shiwaik generate a BI of 181.

The map below indicates the central locations  of 30 clusters in relation to the distribution of native vegetation type. [9]

                                  Central stations of each cluster 

        Place name                              Radius km   BI=0 BI>25  Npop<1K    N       E      

1   Gourdon, France                                  288       5           1               6         44.7  01.4

  2   Valentia Observatory, S. Ireland    400       14         2              14        51.9  10.2

  3   Santiago Compostella, Spain           406      7          23              2          42.9   06.4

  4   Muenster, Germany                         109      1          7               0          52.4   07.7

  5   Innsbruck, Austria                           107      9           2               4          42.3   11.4

  6   Bursa, Turkey                                  224     12         1                2          40.4   25.1 

  7   El Suez, Egypt                                 532       7       21                0      25.4   32.5

  8   Abadan                                             628        6       17                0         30.4   48.5

  9   Gdov, Russia                                    224      14         5             10          58.7   27.5

10  Saransk. W Russia                            434        9         9                1          54.1   45.2

11  Tobolsk, Russia                                 482         8        7                5          58.1   68.2

12  Lviv, Ukraine                                    293      10        5                2           49.8   23.9     

13  Simferopol, Crimea                           397       14        4                2           44.7   34.4            

14  Tulun , Russia                                    485      19        4               9            54.0   98.0

15  Tatarsk, Russia                                    308      14         1               6           55.2   75.9

16  Krasnojarsk, Russia                            391     13         2               7            56.0   92.7

17  Ostrov Gollomjanny, Russia i             277      38         2            24           79.5   90.6

18  Malye Kamakuki, Russia                     82      30         1            23            72.4   52.7

19  Kokshetay, Kazakstan                          460      15         3              2          53.3   69.4

20  Cardara, Russia                                   212       12         0              1           41.3   68.0

21  Nagov, Russia                                     696      30         0              4          31.4   92,1

22  Selagunly, Russia                                846      26        0              5            66.2  114.0

23  Loksak, Russia                                    493       31         0            11          54.7  130.0

24  Gyzylarbat, Russia                              636       20         5             5          38.9    56.3

25  Ust Tzilma, Pechora Basin                 451        16         1           7           65.4    52.3

26  Cape Kigilyak, Kamchatka               1055        37        0            9          73.3  139.9

27  Dashbalbar, Mongolia                       435       29         1            6           49.5  114.4    

28  Guanghua, China                               465       17         2             ?           32.3  111.7   

29  Youyang, S. Korea                            417       26          0            ?          28.3  108.7

30  Poona, N. India                                 681         4          7            0          18.5   73,8

31  C. India                                             601         1        17            0          23.2   71.3     

32  Mai Sariang, Burma                          57         10          4            1          68.2    97.9

33  Central Japan                                   203          5       13            1          34.4  132.6

These data may be used to investigate the supposed warming of Europe and Asia that so worries the public.  In far eastern Russia and neighbouring territories 8 clusters are listed  which include 296 place-names lacking any night-lighting at all, together with just five small towns having night-light indices of only 1.  In such places, it is the natural cycle of climate conditions – modified locally by progressive anthropogenic change in ground cover – that dominates the global pattern of air temperature, and in rural regions there is a rather simple relationship between population size and BI.

Towns and villages occupy only a very small fraction of the continental land surface of our planet, currently about 0.53% – or 1.25% if their densely-populated suburbs are included – according to a recent study using rule-based mapping.  Although it is peripheral to the present discussion, it must be emphasised that conditions in the sparsely-inhabited rural or natural regions are not static at secular scale – everywhere, including in Asia, grasslands and prairies have been grazed or ploughed, and forests clear-cut and replaced with secondary growth.

Consequently, the distribution of population is highly aggregated and associated – as it must be – with regional economic development.  This is illustrated in the images below which show that in western Europe access to the sea is critical, as it is in Japan, while in night-dark Ukraine and Russia it is the zones of temperate broadleaf forest and temperate steppe in which settlement and urban development has been most active.[10]  The arctic tundra belt is very sparsely populated but does includes a few industrialised cities, of which Archangelsk is the largest.

Although, globally, the energy from heat of combustion is equivalent to only a very small fraction of the energy transported in the atmosphere, models suggest that it may be capable of disrupting natural circulation patterns sufficiently to induce distant as well as local effects on the global SAT pattern derived from observations.  Significant release of this heat into the lower atmosphere is concentrated in three relatively small mid-latitude regions – eastern North America, western Europe and eastern Asia – but the inclusion of this regional injection of heat (as a steady input at 86 model points where it exceeds 0.4W m2) in the NCAR Community Atmospheric model CAM3 has important but distant regional effects, especially in winter. 

Comparisons of control and perturbation runs show significant regional effects from the release of heat from these three regions at 86 grid points at which observations of fossil fuel use suggest that it exceeds 0.4 Wm-2: specifically, in winter at high northern latitudes, very significant temperature changes are induced: according to the authors, ‘there is strong warming up to 1oK in Russia and northern Asia…. the north-eastern US and southern Canada have significant warming, up to 0.8 K in the Canadian Prairies’.  Especially in northern North America, where the instrumental record is excellent, this effect is readily observed night lighting is highly aggregated and associated – as it must be – with regional economic development.  This is illustrated in the image above which shows that in western Europe access to the sea is critical, as it is in Japan, while in night-dark Ukraine and Russia it is the zones of temperate broadleaf forest and temperate steppe in which settlement and urban development has been most active.[11]

In eastern Asia, 8 clusters include 268 places that are dark at night, together with just 47 having some night-lighting, mostly of intensity <20.  They include only one city (BI = 153).   In such regions, it is the multi-decadal cycle of solar brilliance that dominates the evolution of air temperature, modified by local effects of change in vegetation and ground cover.

But it is really a misuse of the term ‘rural’ to apply it to the small inhabited places scattered across northern Asia, for this implies some similarity with landscapes such as surrounds Gourdon, devoted now or in the past to farming and herding.  But small villages in asiatic Russia have nothing to do with rurality: their houses and streets have simply been set down in natural terrain – in the wildlands, if you will – that is subsequently ignored; there are no crops, gardens or greenhouses, and the activities of the population are not clear.  The wide unpaved streets bear very few motor vehicles – and there is no street lighting.  Many are described as administrative centres and some have a small dirt runway for light aircraft, while a few seem not to be connected to the rest of the world by dirt roads even seasonally,

Here are two small places in northern Siberia with very different seasonal temperature regimes, of which one is clearly well on its way to urbanisation.  Each lies between 65-70oN on the banks of the river Lena.

 Zhigansk is a long-settled little town founded in 1632 by Cossacks sent to pacify and tax the region; it is now an administrative centre housing 3500 people., laid out beside the river on a rectangular grid.  Until the Lena freezes, it has no road access to the outside in winter.

Kjusjur, just south of the mouth of the Lena in a subarctic environment, was founded in 1924 as the administrative centre for this region, and has a population of 1345; routine meteorological data began to be collected in 1924 and continues today.  About 100 small houses and one larger building are set on unpaved streets beside the stony bank of te river; it has neither runway nor river landing place, but rough tracks leave the settlement to north and south which must be impassable much of the year.[12]

Two motor vehicles can be seen in Kjusjur and a few small boats are pulled up on the beach, while there are about ten motor vehicles in Zhigansk and neither place has any street lighting. Zhigansk has a dirt airstrip with a radar installation that perhaps also houses the meteorological station.   Each has a temperature regime appropriate to its situation, and although it was what I was looking for, I am surprised by the strength of the response to urbanisation at Zhigansk.   I was also expecting that each would respond – at least in very general terms – to solar forcing, and so it does: the cooling of the 1940s and 50s which caused us so much concern in those years about a coming glaciation is clear.

  A compilation of arctic data and proxies took 64oN as the limit of the Arctic region, within which 59 stations were used to analyse the pattern of regional co-variability for SAT anomalies based on PCA techniques.[13]   This demonstrated quasi-periodicity of 50-80 years in ice cover in the Svalbard region: at least eight previous periods of relatively low ice cover can be identified back to about 1200.

Hindcasting climate states is not easy: a recent synthesis of tree-ring data from the Yamal peninsula rashly states that in Siberia the ‘industrial era warming is unprecedented…. elevated summer temperatures above those…for the past seven millennia‘.  However, documents and observations show that this is one generalisation too far.  In summer 1846, as recorded by H.H. Lamb, warming across the arctic extended from Archangel to eastern Siberia, where the captain of a Russian survey ship noted that the River Lena was hard to locate in a vast, flooded landscape and could be followed only by the ‘rushing of the stream’ which ‘rolled trees, moss and large masses of peat’ against his ship, that secured from the flood ‘an elephant’s head’.

The temperature reconstruction below is from annual growth of larches on the Yamal peninisula at the mouth of the Ob.[14]  It testifies that the early decades of the 19th century did indeed include a period of very cold conditions on the arctic coast, while supporting the reality of periods of warmth likely to caused melting of the permafrost of tundra regions.

 In any case, irruptions of warm Atlantic water into the eastern Arctic – including the present one – are well recorded in the archives of whaling, sealing and the cod fisheries.  The present period of a warm Arctic climate is not novel and there is an abundant record from the cod fisheries in the Barents Sea and beyond, not to speak of the documentation concerning the intermittence of open seas from the sealers and whalers in northern waters.

The surface air temperature data are dominated by observations made in towns and cities so that the secular evolution of the climate is determined not by the gaseous composition of the atmosphere, nor by solar radiation: instead, it is dominated by the consequences of our ever-increasing combustion of fossil hydrocarbons in motor cars, public transit and home heating systems, as well as in the industrial plants and factories  where most of us must work.  To this must be added the daily accumulation of solar heat in the stonework or cement of our buildings facing each other along narrow passages.

One conclusion is unavoidable from this simple exploration of the surface air temperature archive: as used today by the IPCC and the climate change science community the instrumental record is not fit for purpose: it is contaminated by data obtained from that tiny fraction of Earth’s surface where most of us spend our brief span of years indoors.  

Footnotes

[1] Hansen, NASA press release and J. Geophys. Res. 106, D20, 23947-23963.

[2] Ellis, E.C. et al. (2010) Glob. Ecol. Biogeog. 19, 589-606

[3] R.A. Ruedy (pers. comm)- see GISS notice dated Aug 28, 1998, at the Sources website

[4] from H.H. Lamb

[5] see for example, Li, X et al. (2020) Sci. Data 7, 168-177.

[6] Pigeon, G. et al. (2007) Int. J. Climat. 27, 1969-1981

[7] Ichinose, T.K et al. (1999) Atmosph. Envir. 33, 3897-3909, Fujibe, F. (2009) 7th Int. Conf. Urban Clim., Yokohama

[8] McKittrick, R.R. and P.J. Michaels (2004 & 2007) Clim. Res. 26 (2) 159-273 & J.G.R. (27) 265-268

[9] Map is from Gao and O’Neil (2020) NATURE COMMUNICATIONS |11:2302https://doi.org/10.1038/s41467-020-15788, image is from eomages.gf.nasa.go

[10] Map from Gao and O’Neil (2020) NATURE COMMUNICATIONS |11:2302https://doi.org/10.1038/s41467-020-15788, image is from eomages.gf.nasa.gov

[11] Ellis, E.C. et al. (date) Global Ecol. Geogr. 19, 589-60, and “Anthropogenic biomes: 10,000 BCE-2025 CE (doi.3390/land9050129v

[12] Images from Google Maps software

[13] Overland, J.A.. et al. (2003) J. Clim. pp-pp

19 Polyakov, I.V. et al. J. Clim. 16, 2067-77

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.