The best studies agree---sampling biases ruin their research.

Yet another problem with the assumption that catastrophic human-caused environment change simply causes civil war, as Salehyan and Hendrix noted, is that violence at the scale of a civil war requires significant resources. In their view, civil wars are more likely to occur in times of relative abundance. While “riots and protests, may emerge from conditions of scarcity,” they argue, “sustaining a militant organization requires considerable planning and resources” (Salehyan and Hendrix 2014, p. 240). Reasons to fight might exist. For this to turn into civil war, however, people “also need the capability to do so, and environmental scarcity may limit such capability, thus undermining the resource base necessary for mobilizing armed violence” (Salehyan and Hendrix 2014, p. 240).

A related debate concerns what Adams and colleagues have claimed to be a sampling bias in studies of the connection between environment change and armed conflict (Adams et al. 2018). Levy accepts the existence of some sampling bias but rejects the view that this bias results in an overstatement of the connection between environment change and conflict. “Knowing that case selection is biased is useful, but not a reason to lower our estimate of the climate’s impact on conflict” (Levy 2018, p. 441).

In responding to Levy’s criticism, authors claiming bias wrote that they did not “deny a link between climate change and conflict in principal. Indeed, some of our own recent work indicates that such a link exists, but it is highly conditional.” Their problem with the research being done in this field was that “sampling biases… increase the risk that such links are overstated, that crucial world regions do not receive sufficient attention and that little knowledge is produced on peaceful adaptation” (Ide et al. 2018, p. 442– 3).

After reviewing the literature on the relationship between climate change and violent conflict, Sakaguchi, Varughese and Auld concluded that the “current literature offers mixed evidence. This makes it difficult to render a definitive statement about the climate-conflict relationship” (2017, p. 640). While they pointed out that just over 60% of the studies they reviewed found “that climate change variables are positively correlated with higher levels of violent conflict,” Sakaguchi, Varughese and Auld also argued that “many subtleties and countertrends underlie this overall pattern” (2017, p. 640). Thus, even though “a majority of reviewed studies envision climate variables influencing conflict through a causal pathway, … these pathways are often theoretically underspecified and have only weak empirical support” (Sakaguchi et al. 2017, p. 641).

As Koubi put it, the research that has been done on this question “provides some evidence that climatic changes could act as a ‘threat multiplier’ in several of the world’s regions. In particular, the extant literature shows that climatic conditions can lead to conflict in agricultural-dependent regions and in combination and interaction with other socioeconomic and political factors” (Koubi 2018, p. 200). After having claimed that, to their knowledge, “no one in the field of climate research has suggested that climate change could be the ‘sole cause’ of war, violence, unrest or migration”, Butler and Kefford recommended “viewing climate change instead as a risk multiplier, influencer or co-factor … In this way of thinking, environmental and ecological factors interact with social determinants, including those that are economic, demographic and political, to produce phenomena such as migration, conflict and famine” (2018, p. 587).

There can be no doubt that conflict will increase during the final hundred years of humanity. But it will result from a complex interaction of socio-political factors and a catastrophically changed environment. It may not go beyond conflict between different groups or between the government and opposition groups and become civil war. This depends on the capacity of those opposition groups. In many cases, they will lack the resources to conduct a civil war. The Syrian war is itself a good illustration of the problem, as the groups opposed to the Syrian government have only been able to conduct the extended civil war in which they have been engaged with the support of outside groups. (Mazzetti and Apuzzo 2016).

The question of whether civil war will break out is something that can only be answered “region by region” and the answer must be based on “knowledge of pre-conflict geographies, such as drivers of resilience and vulnerability” (Farbotko 2018). Sometimes governments may abandon territory and opposition groups can seize control of that land. But it is likely to be land that is suffering worst from the effects of catastrophic human-caused environment change and will not be habitable. To replace an existing government or take control of a region within a country through civil war is no simple thing. It may happen. But it will not happen on the scale that some people have predicted. And it will not happen just because of the weather.