Urban fires

Detection of light produced directly from urban lighting has been proposed (Loeb and Turner, Reference Loeb and Turner2012). In pre-industrial societies, widespread street lighting would likely be combustion-based (although chemoluminescence is possible), and would therefore be of sufficiently low intensity that interplanetary detection would be implausible. However, the Great Fire of London shows that large urban fires are possible. The light output from such fires would be many orders of magnitude brighter than that from combustion-based street lighting, of similar area Fig. 1.

Fig. 1. Illustration of a series of terrestrial events, potentially detectable at interstellar distances with future technologies.

Large fires are necessarily easier to detect than small fires. However, pre-industrial cities were generally small in size, limited by their agricultural hinterlands and resultant transport networks. Furthermore, urban fires may be instantaneously limited in size, even as they spread through the urban environment. Any very large pre-industrial city would likely require a waterborne transport network (Chant & Goodman, 2005), which would itself reduce the likelihood that fires would become extensive (both by providing firebreaks and offering a ready supply of water for extinguishing fires).

It is postulated that (particularly in the event of military action, terrorism or similar) a widespread or multi-seat urban fire could exist on such a scale as to be directly observed at interplanetary distances, given a very large city (or network of cities).

Proving a fire to be urban, as opposed to a wildfire, is problematic. Indeed, it is likely that wildfires are ordinarily far larger than urban fires (Williams, Reference Williams1982). Nevertheless, it is possible that an urban fire may occur on a planet that is either rainy or has fire-resistant vegetation – thus allowing the isolation of any urban fires as atypical for that planet. However, this leaves the difficulty of distinguishing urban fires from infrequent wildfires. This type of detection would require long observation periods, or the simultaneous observation of large numbers of target planets; colossal urban fires are presumed to be generally rare (decadal to centurial, and longer on sparsely-populated or primitive planets with few cities).

Conceivably, the combustion products of large urban fires may differ from those of natural fires:

1. (1) Combustion gas spectroscopic characteristics, or aerosol spectropolarimetric properties may differ from those of natural fires – potentially due to drier combustion material, or the different temperature and oxygen levels in urban fires (Fields et al., Reference Fields, Cole, Summers, Yalcintas and Vaughan1989).

2. (2) Some specific artificial (but pre-industrial) substances may have combustion products that may be detectable through transit spectroscopy (e.g. from burned paints, leathers etc.) (DeHaven, Reference DeHaven1958).

However, these differences would naturally be minimised in pre-industrial societies – in which the predominant sources of flammable materials would presumably be plant- or animal-based. These would consequently mimic the signatures of natural burning – save for their dry state, which may result in detectibly different combustion products (Brode and Small, Reference Brode, Small, Solomon and Marston1986) – albeit likely in small quantities. However, these combustion signatures are not necessarily distinguishable from dry materials in desiccated natural environments. The accurate distinction of urban fires from natural ones would therefore ideally require a number of large urban fires to occur on a single planet. This would require long observation periods or the chance observation of a period of severe conflict.

Land use change

The dawn of farming is associated with land surface changes resulting from the rise of agriculture and pastoralism (Marshall, Reference Marshall2020) – i.e. substituting grasslands and forests with croplands and pastures (Lambin et al., Reference Lambin, Turner, Geist, Agbola, Angelsen, Bruce, Coomes, Dirzo, Fischer, Folke, George, Homewood, Imbernon, Leemans, Li, Moran, Mortimore, Ramakrishnan, Richards, Skånes, Steffen, Stone, Svedin, Veldkamp, Vogel and Xu2001), species introduction (Reidsma et al., Reference Reidsma, Tekelenburg, van den Berg and Alkemade2006), or fire-clearance (although this pre-dates formal agriculture). These are all processes that can be observed with the use of remote sensing methods, through the changes in surface albedo they produce (Boisier et al., Reference Boisier, de Noblet-Ducoudré and Ciais2013; Wang et al., Reference Wang, Schaaf, Kim, Erb, Román, Yang, Taylor, Masek, Morisette, Zhang and Shirley2017). Fires may be detected by other methods, as previously described.

The spread of human societies into new areas has often been accompanied by land-use change, predominantly for agricultural use, or to optimise hunting lands. This involves a variety of processes: forest clearance (Clark, Reference Clark1947; Clement and Horn, Reference Clement and Horn2001); irrigation (Clark, Reference Clark1947; Doolittle, Reference Doolittle2014); scrub burning (Ryan et al., Reference Ryan, Knapp and Varner2013), marsh drainage (Menotti, Reference Menotti2012) etc., whilst such changes are significant at a local scale, they would have to be continental in scale to have a realistic chance of being detected at interplanetary distances. Fortunately, the historical and prehistorical record on earth offers examples – e.g. the European colonization of the New World, and the much earlier arrival of humans in Australia. The migration of settlers from Western Europe triggered rapid changes in the land-use of the Americas. Firstly, introduced diseases contributed to the collapse of existing civilizations (King, Reference King2015), which resulted in the rewilding of farmlands. Then, as more settlers arrived and spread, the land was manipulated by controlled grazing and cultivation – resulting in a very different pattern of land-use than was previously the case. Similarly, in Australia, the landscape was shaped by fire (Bowman, et al., Reference Bowman, Walsh and Prior2004) and by hunting of megafauna (Johnson, Reference Johnson2005). These changes may have occurred around 10× as long ago as the rise of agriculture – and, conceivably, could have been practised to a detectable extent by archaic hominins.

Both expansion and contraction of societies may affect land-use. Pandemics (Little, Reference Little2007), migration (Conn et al., Reference Conn, Segura, Wilkerson, Schlichting, Póvoa, Wirtz and de Souza2002), war (Van et al., Reference Van, Wilson, Thanh-Tung, Quisthoudt, Quang-Minh, Xuan-Tuan and Koedam2015), genocide (Robinson et al., Reference Robinson, Boardman, Evans, Heppell, Packman, Leeks and Acreman2000), continental discovery (Sandor et al., Reference Sandor, Bettelheim and Swingland2002) etc., have all driven land-use change in human history and pre-history. In modern times, there have been more subtle changes, due to increasing concentrations of CO₂ in the atmosphere (Zhu et al., Reference Zhu, Piao and Myneni2016). Use of artificial fertilizers has resulted in a widespread greening (Johansson et al., Reference Johansson, Hall, Prentice, Ihse, Reitalu, Sykes and Kindström2008). There has also been an accompanying increase in eutrophication (Jørgensen and Richardson, Reference Jørgensen and Richardson1996), and an extension of ocean dead zones (Gupta, Reference Gupta2015), which provide alternative proxies. Some of these changes, such as the mining of near-surface coal seams (Dodson et al., Reference Dodson, Li, Sun, Atahan, Zhou, Liu and Yang2014) and the utilisation of phosphate fertilizers (Rutherford et al., Reference Rutherford, Dudas and Samek1994), can conceivably be conducted at scale by pre-industrial societies.

Resulting continental-scale changes in reflected or absorbed spectra, and observed aerosol proxies, would likely emerge on centurial or millennial timescales – requiring extended or parallel observation. Such signals would be difficult to distinguish from natural ecosystem changes. Aggregate analysis of a number of planets may be required to extract an amount of data sufficient to justify any robust conclusions. Notably, changes in levels of greenhouse gases may result from pre-industrial land-use change (e.g. deforestation) (Pielke et al., Reference Pielke, Marland, Betts, Chase, Eastman, Niles and Running2002) – and these are potentially capable of detection at interstellar distances (see section ‘Climate change and fossil fuel use’).

More speculatively, the deliberate creation of huge geoglyph monuments cannot be ruled out. Much as the Nazca lines are visible from the air, greatly amplified versions of this phenomenon could conceivably be detectable at interstellar distances. Such monuments could be created by modifying the landscape surface, or the distribution of dyes. In deserts, such modifications may be persistent. This concept stretches both the capacity for creation and detection to its limit. Notably, such monuments may be created incidentally, or with the deliberate intention of communication; speculative consideration of extra-terrestrial life pre-dates industrialisation by centuries (Scharf, Reference Scharf2019); this discussion continued as astronomy and science developed at around the time of the industrial revolution (Crowe, Reference Crowe1986). Similarly, a reverse process is possible – where civilization may take care to reduce or disguise detectable signs of their society's existence or activities.

Introduced species and extinctions

The effect on natural biota as a result of the introduction of invasive species is comparable to anthropogenic land-use change. Modifying species, such as beavers (Anderson et al., Reference Anderson, Griffith, Rosemond, Rozzi and Dollenz2006); predators, such as rats (Abdelkrim et al., Reference Abdelkrim, Pascal, Calmet and Samadi2005); or the pathogens and vectors responsible for plant diseases like Dutch Elm Disease (Castello et al., Reference Castello, Leopold and Smallidge1995), have often resulted in rapid ecosystem changes. Conceivably, a change may have a very broad impact on a biome scale – e.g. instigating a transition from deciduous to coniferous woodland. In that example, winter loss of leaves may be detectable, by looking for seasonal changes in leaf cover. This can be achieved by observing the particular optical signature of chlorophyll, or other light-harvesting molecules; or alternatively by looking for the winter signature of bare ground, or snow.

Whilst these changes are conducted by natural processes, they are triggered by an anthropogenic event. Conclusions in the section on ‘Land use change’ therefore also apply to the effects of introduced species. Such signals would be particularly useful in conjunction with other indicators of expanding civilization.

Such effects may also apply to bodies of open water, as well as the land surface. An extra-terrestrial world could be imagined, in which oceans were separated and radically different. For example, a planet may have two large oceans, only one of which has surface photosynthetic life (e.g. azolla). Any anthropogenic activity that transferred photosynthetic species from one ocean to the other would trigger potentially-detectable changes over a number of timescales, as the introduced species spread: annual blooms (Platt et al., Reference Platt, White, Zhai, Sathyendranath and Roy2009), a decadal increase of extent, and longer-term planetary biogeochemical change (Asner and Vitousek, Reference Asner and Vitousek2005). Notably, spectropolarimetry may be particularly useful for monitoring the presence or characteristics of oceans of exoplanets – due to the polarisation of light reflected from the ocean surface. Additionally, ocean processes (aerosols and gas release) influence cloud formation – giving another route to detect ocean changes.

The opposite of species introduction is extinction. Whilst the introduction of pathogens (as briefly discussed above) is one possible cause, there are many other reasons for extinction. The loss of megafauna species to hunting has been strongly associated with human expansion – particularly in prehistoric times (Duncan et al., Reference Duncan, Blackburn and Worthy2002). More modern analogues exist in the near-extinction of the American bison (Taylor, Reference Taylor2011) – which is a component of the New World changes discussed. In all such cases, it is at least conceivable that ecosystem changes from comparable events could be detectable at interplanetary distances – provided that these were both pronounced, and continental in scale. To give a specific example: the loss of large grazing animals may result in the conversion of grassland to the forest (Gibson and Brown, Reference Gibson and Brown1991). As described above, such a transition may be detectable by leaf colour, seasonal fall, or the presence of monoterpene hazes (Kerr and Ostrovsky, Reference Kerr and Ostrovsky2003).

Seasonal agricultural changes

Large-scale agricultural monocultures have specific optical signatures. Ploughing, growing, flowering, ripening and harvesting are all potentially detectable – and highly temporally reliable. The more extensive a monoculture, the larger the signal would be. A planet with a wheat monoculture would transition from a near-white ripened crop to near-black bare earth (Lobell et al., Reference Lobell, Asner, Ortiz-Monasterio and Benning2003), accompanying the harvesting and ploughing season. As discussed above, a sudden loss of chlorophyll-containing surface vegetation would potentially be observable. By contrast, natural ecosystem processes rarely expose extensive bare soil in a sudden change, nor would they result in a sudden loss of surface chlorophyll (fires and snow excepted). However, careful parameterisation of leaf fall changes would be required, to demonstrate clear and detectable differences from ploughing events. Comparably, a canola crop would have vivid yellow flowers in season (Rabe, Reference Rabe2003). On a planet with extensive continents and pervasive monocultures, the agricultural cycle would offer a potentially strong signal, which is robustly periodic over annual timescales. However, other planets may lack strong annual seasonality, so such periodicity is not guaranteed.

Albedo changes due to harvesting would only be a function of the fraction of the planet covered in the land (L_FRAC), the fraction of land covered by the crop (L_CROP) and the changes in albedo before and after harvest (dA). Assuming a change in albedo as high as 0.2 (Davin et al., Reference Davin, Seneviratne, Ciais, Olioso and Wang2014; Seneviratne et al., Reference Seneviratne, Phipps, Pitman, Hirsch, Davin, Donat, Hirschi, Lenton, Wilhelm and Kravitz2018) and the conditions of a single supercontinent covering 50% of the planet completely covered in the same crop, this yields a potentially measurable change in the albedo of around 10%. This figure depends on the precise position of the continent, and the geometry of the planet's position relative to its sun and the human observer.

With the European Space Agency-Switzerland optical photometric space mission CHaracterizing ExOPlanet Satellite (CHEOPS) (Cessa et al., Reference Cessa, Beck, Benz, Broeg, Ehrenreich, Fortier, Peter, Magrin, Pagano, Plesseria, Steller, Szoke, Thomas, Ragazzoni and Wildi2017), Serrano et al. (Reference Serrano, Barros, Oshagh, Santos, Faria, Demangeon, Sousa and Lendl2018) reported that it would be possible to detect the albedo of a planet depending on the ratio of its radius compared to that of its star. In particular, they report that an albedo similar to that of the Earth is detectable when this ratio exceeds 0.04, but the noise level remains too high to determine a 10% change in albedo for a radius up to 0.07. This ratio for the Earth is <0.0092. Currently, the largest candidate to have been found with a density similar to the Earth has been Kepler 10c, with a radius 2.32 times bigger than the Earth (Rajpaul et al., Reference Rajpaul, Buchhave and Aigrain2017); current satellite capabilities would not be capable of resolving such a change, given a star like our Sun. Smaller stars, however, such as M-Dwarf stars (Red Dwarf), could present a chance at detection. This kind of star, which accounts for 85% of all stars in the Milky Way, would tend to have tidally locked orbiting planets (Kasting et al., Reference Kasting, Whitmire and Reynolds1993) – meaning that any seasonality would not arise from comparable mechanisms to that found on earth. While this could make the emergence of plant-like life harder than around Sun-like stars, recent results have shown that atmospheric collapse in the presence of tidal locking may not happen (Merlis and Schneider, Reference Merlis and Schneider2010); more complex life might indeed be possible (Tarter et al., Reference Tarter, Backus, Mancinelli, Aurnou, Backman, Basri and Young2007), and detectable seasons would still occur if the orbit of the planet is not circular. In this case, considering that the size of M-dwarf stars lies between 0.1 and 0.6 Sun radii, detection might be possible with current capabilities for Earth-sized planets if the star lies at the lower end of the spectrum. Dragomir et al. (Reference Dragomir, Teske, Günther, Ségransan, Burt, Huang and Stassun2019) detected an Earth-sized planet orbiting such a star, but its closeness to the star makes it unsuitable for life. Future planned missions (such as the PLAnetary Transits and Oscillations of stars PLATO – Catala, Reference Catala2008) might lower the detectability threshold further.

While white dwarfs would also be good candidates for this kind of detection – given their small size and dimness – the red giant phase of their evolution tends to destroy close-orbiting planets (Villaver and Livio, Reference Villaver and Livio2009). It is therefore unlikely that anything but cold and distant planets will remain in their orbit (Vanderburg et al.,Reference Vanderburg, Rappaport, Xu, Ian, Crossfield, Becker, Gary, Murgas, Blouin, Kaye, Palle, Melis, Morris, Kreidberg, Gorjian, Morley, Mann, Parviainen, Pearce, Newton, Carrillo, Zuckerman, Nelson, Zeimann, Brown, Tronsgaard, Klein, Ricker, Vanderspek, Latham, Seager, Winn, Jenkins, Adams, Benneke, Berardo, Buchhave, Caldwell, Christiansen, Collins, Colón, Daylan, Doty, Doyle, Dragomir, Dressing, Dufour, Fukui, Glidden, Guerrero, Guo, Heng, Henriksen, Huang, Kaltenegger, Kane, Lewis and Lissauer2000), although the possibility of smaller planets remaining cannot be completely excluded (Barnes and Heller, Reference Barnes and Heller2013).

Beyond the direct optical detection of agricultural land manipulation, there is a possibility of detecting secondary outputs of agriculture. Stubble burning was extensive in the UK until outlawed (Clapp, Reference Clapp2014). If such burning was planetary in scale, atmospheric smoke loading may be observable (Smith et al., Reference Smith, Adams, Maier, Craig, Kristina and Maling2007), e.g. via spectropolarimetry. Again, combustion products could potentially be distinguished from those of natural fires – meaning that the type of smoke, and not merely its extent, could be used as a detection strategy. Methane produced by ruminants may also constitute a detectable agricultural proxy (see section ‘Climate change and fossil fuel use’).

Climate change and fossil fuel use

Climate change may be caused by the exploitation of carbon-rich fossil fuels, which were extracted before the industrial era (Low Tech, 2011; Pirani, Reference Pirani2018).

Climate change can also be caused by the expansion of the number of ruminant livestock (Ripple et al., Reference Ripple, Smith, Haberl, Montzka, McAlpine and Boucher2013), and land-use change (Flannery, Reference Flannery2005). Climate change can potentially be detected at interstellar distances, e.g. by using the secondary eclipse method (Baskin et al., Reference Baskin, Knutson, Burrows, Fortney, Lewis, Agol and Showman2013), as well as the measurement of greenhouse gas concentrations, as previously discussed.

Whilst we associate climate change with the industrial era on earth, it is potentially possible that different geologies may facilitate sudden, large-scale exploitation of fossil fuel resources. For example, if settlement of the Americas was accompanied by discovery and exploitation of large and easily-accessible fossil fuels reserves, a scenario is conceivable where detectable climate change was a result – without assuming industrialisation.

Separately from climate change, direct signs of fossil fuels combustion may be detected. For example, the use of such fuels on earth is associated with mercury and sulphur pollution, as well as smoke. Pre-industrial fossil fuel use, particularly in colder climates, was predominantly for heating. This offers a potentially detectable seasonal cycle.