A ‘mosaic’ perspective of climate due to natural and societal influences

By E. Craig Jowett

Earlier this year, UN General Assembly President María Fernanda Espinosa Garcés stated that only 11 years remain to avert global climate change catastrophe. (www.un.org/press/en/2019/ga12131.doc.htm) This statement got me interested in learning about the effects of human population on climate and the ability to sustain ourselves. As such, I undertook an extensive analysis of the available government information in these areas. In doing so, I now believe that some of the ongoing finger pointing is misdirected, and that we will be just fine, even after 11 years.

Setting the global scene

World population correlates very tightly with the National Oceanic and Atmospheric Administration (NOAA) World CO₂ levels (Figure 1), and a case for causal connection is inferred because of the close association, “time order” with population increasing before CO₂, and a valid mechanism for humans producing CO₂ in their daily lives.

But with poor correlation, or correlation with no mechanism or time ordering present, a causal connection cannot be inferred between any two variables, even CO₂ and temperature.

Should we worry about this increase of CO₂ (and of population), and what other variables are connected, such as air temperature? What emergencies are there? Do we really have only 11 years left?

Agricultural food crops

For sustainability, one essential consideration would certainly be world food supply. United Nations Food & Agriculture publishes annual data for food products around the world. For 20 of the top agricultural food products, including corn, wheat, rice, potatoes, beans, cassava, sorghum and yams, plus sugar cane, fresh vegetables and fruits, production has increased an average of 54% from 2000 – 2017. Only sweet potato production has decreased.

The top 10 staple crops listed by National Geographic have increased by 45% in the same time period, and animal production for meat has increased by 29%, (statistics in www.fao.org). From 2000 – 2017, World population has increased by 23%, indicating that production of food crops and meat has more than kept pace with population. This all means that UN data indicates no agricultural emergency from the climate or from population growth.

Forest fires

Many things are attributed to atmospheric warming, including forest fires. Ontario has a great website for residences near forests. “Each year, in Ontario, homes are damaged or destroyed by wildfire, yet nearly all of these homes could have been saved if owners had followed a few simple rules.” (www.ontario.ca/page/firesmart-landscaping)

Natural Resources Canada (NRC) points out that, “lightning causes about 50% of all fires but accounts for about 85% of the annual area burned”. The total number of wild fires in Canada has dropped 40 – 50% since 1990 (www.nfdp.ccfm.org/en/data/fires.php). The area burned varies widely but is “stable” over that time period, with larger fires being in remote areas further from firefighters (www.nrcan.gc.ca). From 1980 – 2017 there is no correlation between area burned and NASA’s North America Land-Ocean temperature index (R2 = 0.01). For interest, CO₂ doesn’t correlate with fires either (R2 = 0.02).

Regional geotectonic flooding

When sea level rise and flooding are considered in the climate discussion, isostasy effects on the Earth’s crust are understated. As a field geologist in the Arctic Islands, traversing old beach strands with whalebones high above the sea ingrained in me how mobile the Earth is.

Expanded seawater volumes will raise levels, but land subsidence adds to the problem of coastline flooding and it has happened already. Subsidence can be human related and permanent, as in New Jersey where groundwater withdrawal is lowering land level, or natural and temporary as in Texas where massive rainfall depressed the land by a couple of centimetres (www.nasa.gov).

Or, it can be natural, permanent, and continent-wide, due to melting of the Ice Age glaciers. The Hudson Bay area continues to rise 10 – 12 mm/year even though the 2 – 3 km thick glacier that had once depressed the land is long gone.

However, the surrounding envelope of land is subsiding due to the same isostasy, in effect tilting the land away from Hudson Bay. Southwestern Georgian Bay is affected with relative water levels rising as it is being tilted slowly southward. (www.ontariobeneathourfeet.com)

Rainstorms and urban flooding

In the Yukon’s Mackenzie Mountains, I once witnessed intense flash flooding that almost took out our camp. Rains hadn’t caused flooding until the mountain scree slope opposite us was sun-dried and hardened into an impervious layer, allowing sheet runoff and flooding with very powerful force.

Certainly, with urbanization and population increase, impervious surface areas from roofs, sidewalks, driveways, roads, and parking lots have also increased very much since the 1950s. Stormwater cannot penetrate the soil as it used to. Instead it courses quickly along the impervious surfaces, causing flooding, erosion, and contamination from animal fecal matter and general garbage.

The US Geological Survey publishes research on this topic, and their Fact Sheet 076-03 is a good start for the causal connection between urbanization and flooding. Floods need impervious surfaces, or they don’t happen as much. However, we need impervious surfaces, or our cars would get stuck in the mud and our roofs would leak.

The frequency of floods may or may not be affected by changes in climate, but more inhabitants and urban paving do certainly facilitate them.

Urban climate islands

Gardeners know that “micro-climates” can be made with walls, wind breaks, etc., to trap heat and enable warmer-climate plants to thrive. By their design, cities similarly trap heat, but they also produce heat to make their own “heat islands”.

An extensive study in China (Zhou et al., 2015, Sci. Rep. 11160) showed measurable urban heat effects continued outwards 3 – 5 km from city perimeters. This heat effect can be visualized as a permanent dome of heat energy over the city, pushing outward and upward to dissipate heat and warm the countryside and atmosphere. City air temperatures can be more than 5 – 10°C higher than the surrounding countryside (www.epa.gov), and paved surfaces up to 50°C hotter than shaded surfaces.

One thing that surprised me is the tremendous heat picked up by rain as it tracks over hot asphalt, heating up by 10 – 15°C or so. Heated stormwater negatively impacts surrounding streams (and oceans presumably) by warming them.

Looking at Southern Ontario

Toronto’s downtown Bloor and St. George weather station operated by Environment & Natural Resources (ENR) from 1840 – 2006, and rural Owen Sound’s ENR and MOE stations from 1878 – 2007.

Omitting the 1883 – 1886 eruption of the Krakatau volcano in the Dutch East Indies, Toronto’s average temperature was 6.7°C for 1840 – 1877, 7.7°C for 1878 – 1916, and 8.9°C for 1917 – 2002. Owen Sound averages were 6.9°C for 1878 – 1916 and 6.9°C for 1917 – 2002, more than a 1.9°C difference since WWI. (The boundary of 1916 – 1917 was chosen because of the sudden and sustained jump in overall temperature difference of 0.9°C in 1916 – 1.9°C in 1917.) Figure 2 demonstrates the divergence in temperatures between the two cities after WWI and includes the cold years caused by the Krakatau volcano eruption.

Compared to the 40 years prior to WWI, Toronto has warmed by 1.19°C since WWI, and Owen Sound by 0.05°C. The change in Toronto is more from warmer winters than hotter summers, and in Owen Sound, winters got warmer but summers actually cooled. Note that temperature fluctuations are coincident in Owen Sound and Toronto records (also for other Georgian Bay weather stations) and reflect regional and global incidents.

Toronto has been 1.9°C warmer than Owen Sound for a century, and since 1940 it averages 3.0°C warmer than Wiarton and Chatsworth. It is a sustained, year-round, 2 – 3°C heat anomaly, as Toronto continuously produces its own heat.

Toronto as A CO₂ dome

While cities appear to be their own furnaces, they also appear to form their own “CO₂ domes” (e.g., Phoenix, AZ; Idso et al. 2001, Atmos. Env. 35: 995), attributed mainly to vehicular traffic. Environment and Climate Change Canada (ECCC) monitors CO₂ across the country, and although each site has its own specific source and sink characteristics, it is intriguing to contrast Toronto’s CO₂ with surrounding areas. (Since CO₂ behaviour is complex, a proper quantitative comparison requires meteorological modelling, so this is simply a first look at isolated sites for qualitative comparison.)

Using the linear trend of total hourly unfiltered data, the ECCC Downsview site in central Metropolitan Toronto reached the 400 ppm CO₂ mark back in 2005, and rises steadily at 20 ppm/decade (Table 1). Egbert, southwest of Barrie, lags 5 years behind and is rising at 18 ppm/decade, with Fraserdale, north of Timmins, lagging 8 years and rising at 20 ppm.

Hanlan’s Point on Toronto Islands is anomalous as it lagged 6 – 7 years behind Downsview, but is now rising very rapidly at 41 ppm/decade. Turkey Point on Lake Erie passed 400 in late 2012, a lag of 7 years, and is rising at 24 ppm/decade, a little higher than the norm.

Compared to Downsview and Hanlan’s Point, surrounding stations have lower overall values of CO₂ and are less “noisy”. Toronto’s CO₂ fluctuations have transitory peaks of >100 ppm above average, similar to Hanlan’s Point. Alert, Nunavut, and Sable Island, Nova Scotia, especially are very quiet with regular fluctuations of 15 – 20 ppm. This difference indicates that Toronto has substantial nearby influences, and that it is also an urban CO₂ dome.

Toronto has been a “heat dome” since WWI and also appears to be a “CO₂ dome”, consistently 2 – 3°C and ~10 ppm CO₂ above rural Georgian Bay which is only 150 km northwest. When both heat and CO₂ (and water) are generated in situ by combustion of various carbon-based compounds (natural gas, gasoline, propane, wood, coal, oil), the greenhouse effect of CO₂ cannot, in itself, explain these higher city temperatures. Toronto’s environment is akin to a potbelly stove effect rather than a greenhouse effect, a key distinction.

But has atmospheric CO₂ influenced air temperatures outside of the urban heat zones; i.e., is there evidence for a substantial greenhouse effect?

Temperature history of southwest Georgian Bay

Owen Sound: Annual temperatures increase in a seesaw pattern from 1898 to 1997, followed by a decrease from 1998 – 2018 (Figure 2), and show the same warming and cooling trends with respect to atmospheric CO₂ concentration (in Figure 3 combined with Chatsworth and Wiarton data).

It is always a challenge to select start and end points for trends to avoid artificial trends. Both upward and downward trends can be artificially induced, or correctly uncovered, depending on how end points are chosen. A separation in the data is made at 1997 – 1998, when temperatures jumped due to El Nino (www.ec.gc.ca/meteo-weather), and which is also evident in the Toronto, Chatsworth and Wiarton data. The anomalously hot year 1921 is recorded in Toronto, Owen Sound and Collingwood. The volcanic winter years of 1883 – 1885, due to the Krakatau eruption, are visible for Toronto and Owen Sound, but would be an artificial starting point, for instance.

All data used herein are sourced from websites of government agencies, or have been requested and obtained from the same. Years with one or more months missing are omitted, and any omission or separation of data is explained.

Chatsworth: Historical Chatsworth temperatures rise in a fairly distinct seesaw manner from 1958 – 2006, when the station closed. A single linear trend to the whole population of data is not always the best representation of the data. When separated at the El Nino jump of 1997 – 1998, temperatures increase slightly with increasing World CO₂ until 1997 – 1998, and then decrease, similar to the Owen Sound data. If the single anomalously low data point of 1962 is removed, the 1958 – 1997 trends are flat, with no increase or decrease with CO₂ at all. Temperatures are seen to rise and to fall with increase in CO₂, in an independent manner.

Collingwood: The ENR station in Collingwood operated from 1891 – 1974, but there are three large gaps in the record. The gap from 1920 – 1935 seems to be a reasonable separation in the record, and is used here. Warming occurred for 1891 – 1921 and very slight cooling to no change for 1935 – 1974. As in Owen Sound, 1921 stands out as a warm year in Collingwood, and 1904 a cool year.

As mentioned, trend lines are sensitive to subjective start and end points in data. As an example, when the single 1921 data point is omitted from the 1891 – 1921 period, the relation of temperature to time changes from a positive +0.03 slope to a negative –0.001 slope, and the same for temperature versus world CO₂. This demonstrates how trends can easily be changed and even misrepresented, meaning results are not always as they may appear. In either case, however, there is no clear relationship between warming temperatures and increasing CO₂ in the air.

Wiarton: This rural weather station has provided an uninterrupted record since 1947. Temperatures fall noticeably from 1947 – 1980, followed by overall rise from 1981 – 2018. Wiarton is embedded in the Georgian Bay trends in Figure 3.

Looking at only the last four decades, one would conclude that rising temperature corresponds to increasing world population, rising world CO₂, and rising Ontario CO₂ values. For the whole time period of 1947-2018, a linear trend line would steadily increase with respect to world CO₂ at a +0.01 slope.

However, for the three decades before 1980, it is quite the opposite with a – 0.05 slope followed by a +0.01 slope after 1980, demonstrating that CO₂ content corresponds with both warming and cooling trends, and not just warming.

For 1947 – 1980, annual temperatures decrease 0.50°C for every 10 ppm increase in CO₂ over the first 34 years, and decrease 0.63°C for every billion more people in the world. During the period of 1981 – 2018, temperatures increase 0.13°C for every 10 ppm rise in CO₂, and increase 0.28°C per billion rise in population. At the High Arctic military base of Alert, Nunavut, this same pattern of cooling and warming is seen with a –0.04 slope from 1951 – 1979 and +0.07 from 1980 to 2018.

Figure 4 depicts the importance of carefully choosing the start and end points for correlation, as the results can be fundamentally different, and all from the same database. In this case, there is no positive correlation between temperature and CO₂ if the 1980 hinge and 1998 El Nino years are taken as data separators.

There is no systematic, straightforward relationship between atmospheric warming and CO₂ change in the SW Georgian Bay area. A good correlation can imply a causal relationship, as is very reasonable with world population and world CO₂ in Figure 1. However, where there is no correlation, a causal relationship cannot exist, with the variables being independent of each other. For cause, there must be a time order and a mechanism, as well as association. But without association in the first place, there is no case for CO₂ causing temperature variation.

The very tight relationship between world population and world atmospheric CO₂ values can be inferred to be causal, i.e., more people beget more CO₂ in their day-to-day lives. However, in the SW Georgian Bay area (which also records world climate events) and in general, the effects of this increasing CO₂ appear to be less than worrisome.

There is no issue with world food crop and meat production not meeting population growth, and forest fires are not increasing in Canada. Forest area burned varies considerably year to year, has an unclear relationship to El Nino years, and no relationship to NASA’s North America Land-Ocean temperature index. Land subsidence, natural or human-induced, adds to regional sea level rise. Impervious surfaces in the urban built-environment and paving over increase flooding and warm and degrade surface waters.

Urban centres like Toronto are by design permanent heat traps, permanent heat sources, and permanent CO₂ sources that impact surrounding rural areas. This permanent city-rural difference suggests that a more “mosaic approach” of cities and rural entities may be more appropriate than a global homogeneity approach.

Owen Sound and Toronto both recorded the Krakatau volcano eruption; Owen Sound, Collingwood and Toronto recorded the very hot year of 1921; Owen Sound, Toronto, Chatsworth and Wiarton all recorded the very hot year of 1998. This connectivity suggests that local records are also regional records, and likely global records.

Random Walk Behaviour?

Historically, temperatures in SW Georgian Bay have been rising and falling substantially every year, and show no “hockey stick” shape. Population, time, and CO₂ are variables that increase steadily without decreasing, whereas temperatures rise and fall widely on an annual or multiple-decade basis, more akin to a stock market trend, and sometimes jump up as in 1921 and 1998, also like stocks.

This behaviour has been referred to as a “Markov chain” or “random walk” process (e.g., Gordon 1991, Am. Meteorological Soc. 4; 589), which comprises moving from a first value to a second value, with the direction determined by chance. The next move is taken from position two and not from the first position so that the immediate and ultimate values do not depend on past history.

This continuation from the latest position is important to temperature trends, as both the decades-period trends and the year-to-year changes, as seen for Wiarton, for example, can develop. Gordon also discusses climate being “subjected to periodic shocks of random warming and cooling impulses”. Owen Sound impulses are seen at 1904, 1921, 1940, and El Nino 1998, among others. It does look like the record is periodically “knocked off balance” but then carries on as usual from the new position. World CO₂ and Ontario CO₂ behave differently than that, independent of temperature.

A new perspective?

Georgian Bay temperatures do not correspond with CO₂ strongly enough to infer a causal connection. There will be other factors that impact climate, as is reasonable with such a complex open system like the Earth within its solar system. Local weather stations are the only historical basis for prediction we have, as there are no “global measurements”.

Global events like the Krakatau volcano eruption and the hot years of 1921 and 1998 are reflected at the local stations, and so to be accurate, any global predictive model needs to return the favour, and reflect local historical records during calibration. An over-averaged model would miss the interplay of urban and rural entities, and that would be unhelpful.

One representation identified here would be depicting large cities as stationary source nodes of permanent heat and CO₂, broadcasting outwards into their surrounds. This perspective is not just looking at local weather, since it includes both permanent effects of urban anomalies on the regional background climate through the seasons, but is akin to a series of local climate mosaics instead of a global melting pot. It is also potentially uncovering the important causes and not just overall symptoms.

It appears from my investigation that the impervious and heat-producing urban environment has more impact than solely a CO₂ induced greenhouse effect. There appears to be less than a catastrophe looming. Thus, media and political alarms can be less fervent and refocus on appropriate responses for future warming or cooling.

E. Craig Jowett, PhD, P.Eng. did his Masters and PhD at Toronto between working stints as a field geologist from coast to coast in Canada and the Arctic Islands, Eastern & Western Europe, and SW USA, followed by NATO Science Fellowships at Michigan and Cornell. After a time with the University of Waterloo as a research professor, he founded his own Ontario-based wastewater manufacturing and research company. This article appears in ES&E Magazine’s August 2019 issue.