A polluting war: Risk,experts,and the politics of monitoring wartime environmental harm in Eastern Ukraine

Abstract

War always harms the environment. As the fog of war produces unreliable data, it also obstructs our capacity to monitor those harms. While some call for more data collection to advance a clear narrative of the origins of environmental harm, sociologists of risk and professional risk assessors find that the urgency of environmental hazards depends not on data alone but on who has the authority to define those risks. Without a clear understanding of how environmental experts engage in the socio-political struggles over the interpretations of risk during armed conflict, we may undervalue, first, how assessments and adequate action remain practically and politically difficult, and second, how data may be misused. Drawing on interviews with environmental experts and reports, I examine the politics of environmental expert knowledge on conflict pollution in the war in Donbas (2014–2022), a region hosting over 4500 hazardous industrial enterprises. By going beyond technological evidence collection, the article broadens our understanding of the obstacles of knowledge production and the attribution of environmental harm in highly politicized and violent contexts. Based on insights from environmental politics, the article finds that experts manage three issues undermining the reliability of environmental risks in an active warzone: pre-existing industrial pollution, environmental damage spread across government and non-government-controlled territories, and disinformation.

Introduction

War always pollutes the environment, especially when it damages hazardous industries and toxic waste storage facilities. While we have a growing understanding of the impact of conflict pollution or toxic remnants of war—“any toxic or radiological substance resulting from military activities that forms a hazard to humans and ecosystems” (Kellay and Weir, 2014: 13)—on the environment, ecosystems, and public health (Austin and Bruch, 2000; Briggs and Weissbecker, 2012; HLSIHRC and CEOBS, 2020; Zwijnenburg et al., 2020), we know less of, first, how experts behave in practice to assess the risks of conflict pollution in active armed conflicts, and second, how they overcome obstacles of knowledge production in highly politicized contexts like Donbas; a region with preexisting heavy industrial pollution and environmental hazards spread across areas controlled by Ukraine and occupied by Russia. ¹

Studies on conflict pollution acknowledge the difficulties in data collection in conflict zones due to ongoing violence, the collapse of environmental monitoring infrastructure, and lack of access to sites. They often call for more technical data collection and the use of new methods to push for effective policy solutions and fill the “evidence gap” (Zwijnenburg et al., 2020). In contrast, sociologists of risk and environmental politics scholars argue that how environmental data can be instrumentalized depends on who has the authority to define risks and losses and less on available technological data (Beck, 2009; Fesenfeld and Rinscheid, 2021; Wynne, 2002). Such struggles over authority often emerge in the wake of environmental disaster, when states and corporations downplay the risks of toxic pollution despite available evidence (Auyero and Swistun, 2009; Cartwright, 2019; Kuchinskaya, 2014; Nading, 2020).

While studies on environmental monitoring of disaster often assume that scientific expertise renders pollution visible, leading to clear policy solutions (Hammond, 2021; Lahsen and Turnhout, 2021), they rarely take on an actor-oriented approach to explain how experts produce knowledge while navigating particular contexts (Lidskog and Sundqvist, 2018: 318). ² These socio-political struggles over the interpretations and definition of environmental risks do not change the real risks themselves but do affect how they are perceived and acted upon. How environmental harm can be exposed depends on authority figures—such as experts and their identified role in defining risks and losses, but also politicians or industrial managers—who may attempt to influence the interpretation of risk in a process detached from technical risk assessments.

Drawing on reports by Intergovernmental Organization (IGO) and nongovernmental organization (NGO) reports, and semi-structured expert interviews, this article contributes to studies on the environmental consequences of armed conflict by examining how experts make sense of wartime environmental harm in the Donbas region in Eastern Ukraine (2014–2022), covering the temporarily occupied areas of Donetsk and Luhansk Oblasts. ³ Home to around 4500 hazardous mining, metallurgical, and chemical enterprises, the Organization for Security and Co-operation in Europe (OSCE) labelled Donbas “Europe’s most significant man-made environmental burden” (OSCE, 2017). Up to 80% of these industrial sites have hazardous production facilities (Chernenko et al., 2019: 6), posing significant risks to the environment and public health. Russia’s invasion in Ukraine on 24 February 2022 impact on the environmental risk lies beyond the scope of this article, but environmental monitors continue to map out the grave environmental risks the war causes. ⁴ International NGOs such as PAX, CEOBS (Conflict and Environment Observatory), and the Zoï Environment Network continue to report that Russia has shelled power plants, fuel pipelines, and other hazardous industrial sites throughout Ukraine (CEOBS and Zoï Environment Network, 2022a, 2022b, 2023; PAX, 2022a). The flood following the detonation of the Kakhovka dam in 2023 will have environmental consequences beyond Ukraine's borders.

By examining how experts practice environmental monitoring in an active conflict, this article contributes to studies on wartime environmental harm by characterizing the difficulties of collecting data around environmental harm in a highly politicized context. The article finds that experts engage in three socio-political struggles around evidence collection: how to collect reliable data on conflict pollution in a region with heavy industrial pollution dating before the war, second, how to get access to data when environmental harms are split between government and non-government-controlled areas, and third, how to cope with the use of environmental data as a tool to blame or spread disinformation? In this environment, rather than the actual environmental risks and losses rising or falling, the perception of urgency rises and falls according to the behavior and interpretations of actors—environmental experts and other actors who attempt to determine or undermine the urgency of environmental harm— and how they attempt to define the public perception of risk.

Conflict pollution, experts, and the sociology of risk

Armed conflict pollutes the environment when armed groups use toxins to poison foliage or when damaged industries, war rubble, or undetonated ordnance (UXO) leak toxins and harmful minerals into the environment (Austin and Bruch, 2000; Kellay and Weir, 2014; White and Désoulières, 2017). War can lead to deforestation, the introduction of new pollutants in the air and water, terminal illnesses, birth-defects, and wildlife decline (Briggs and Weissbecker, 2012; Brito et al., 2018; Nixon, 2011; Zwijnenburg et al., 2020). In recent years, international organizations nurtured a blooming optimism for legal accountability and remedy for environmental destruction during armed conflict—for instance, the recent development and codification of the principles for the protection of the environment in relation to armed conflict by the UN International Law Commission (the PERAC principles) and the new legal definition of ecocide (Stop Ecocide Foundation, 2021).

Because some effects of environmental harm are slow to emerge (Nixon, 2011), the impact of war on public health or the environment may still be missed in post-conflict assessments (Briggs and Weissbecker, 2012: 176). Zwijnenburg et al. (2020) add that due to small sampling sizes, lack of access, and multiple potential causes of public health risks, establishing the causes of pollution and its effects on human health remains challenging (also see HLSIHRC and CEOBS, 2020). Conflicts often destroy environmental monitoring infrastructure and experts are forced to flee their country and lose their jobs, which further deteriorates the capacity to collect evidence especially in countries coping with corruption and weak governance (Williams and Dupuy, 2017). Assessments of the impact of conflict on the environment or public health are therefore often prone to under or over-estimation (Austin and Bruch, 2000).

Sociologists of risk and professional risk analysts are therefore pessimistic of how environmental risks and hazards can be contained (Beck, 2009; Kammen and Hassenzahl, 2018). The long-term health effects of pollution or radiation are often slow to emerge and imperceptible through the human senses or “out of sight” (Nixon, 2011). The scope of human-made environmental risks often cannot be fully known or avoided with the tools available to us, blocking our ability to attribute responsibility to any single actor (Beck, 2009). While slow to surface and often disputed, it does not mean that the effect of conflict pollution and toxic remnants cannot be determined. Rather, expertise may only be recognized, and assessments may only be carried out, once human health is being affected and becomes visible to a wider audience. For example, the Kakhovka dam break and major flood in Ukraine in June 2023 exposed the interlinked public health and environmental harms to a wide audience, even though those risks were present much earlier, since 2014.

Because we still need practical solutions to solve looming environmental catastrophes, an increasing amount of technological and expert knowledge is necessary to predict and contain environmental hazards (Beck, 2009). Responses to environmental disaster often remain framed as science-based: science can both identify and lead to effective policy solutions for environmental problems (Lidskog and Sundqvist, 2018: 310). In contrast, environmental politics scholars (Hammond, 2021; Lahsen and Turnhout, 2021) argue that assessing risks involves not only overcoming collecting new conclusive data which could lead to societal transformation, but also involves overcoming disputes over authority and evidence (Lidskog and Sundqvist, 2018). Carrying out risk assessments and estimating the impact of conflict pollution on the environment depends on political willingness, resulting in delays and a lack of acknowledgment for victims of toxic exposure.

While we have a growing understanding of the impact of conflict on the environment, ecosystems, and public health, the study of the impact of conflict pollution remains understudied from an actor-oriented approach to explain how experts produce knowledge while navigating high-risk contexts (Lidskog and Sundqvist, 2018: 318). ⁵ Without a thorough understanding of the how environmental and public health harm of toxic remnants during armed conflict are assessed, we risk undervaluing how experts overcome obstacles in practice when evaluating the environmental and related public health harms.

Much of the expertise around the knowledge of toxic exposure arises slowly through social interactions: it often only arises in the aftermath of a disaster and requires sustained advocacy (Fortun, 2009). Expertise is not only acquired knowledge and skills, expertise is a specialist competence which “is achieved through socialization and appropriation within particular groups and contexts, and needs to be recognized and affirmed by others” (Eyal, 2013; Lidskog and Sundqvist, 2018: 327). For instance, expert strategies to expose environmental harm to the public may be strained by the material conditions they work under and pressures on their careers (Tousignant, 2018); conditions that are even more straining during wartime. Moreover, experts often face industrial management and policy makers who deny toxic exposure or actively obstruct experts—even when the harmful properties of chemicals are well-known (Auyero and Swistun, 2009; Cartwright, 2019; Nading, 2020).

By adopting an actor-centered and environmental politics approach to expert knowledge production, this article further finds that wars amplify these disputes around authority and the collection of evidence collection. Data is scarce or difficult to collect, evaluations of risk assessments are disputed, and data can be used in disinformation campaigns by the polluter. By examining how environmental experts manage these contestations, we can uncover how expertise on conflict pollution is produced—and how environmental and public health harms can be contained—in violent conflicts.

Floods, disruption, collapse: Monitoring environmental harm in Donbas 2014–2022

In 2013 and 2014, after the Ukrainian government decided to forego on signing an Association Agreement with the European Union, protesters organized demonstrations in favor of closer ties with the EU. In response, and with Russia’s covert support, protests, and occupations of government buildings in Donbas escalated into a war between the Armed Forces of Ukraine and militia backed by Russian troops and paramilitaries. In 2014, the Russian military occupied the Crimean Peninsula. In the same year, and even though denying its involvement, Russia de-facto invaded Donbas, establishing effective control over parts of the region by installing proxies, providing weaponry, and having a military presence. Since then, the self-proclaimed Donetsk People’s Republic (DPR) and Luhansk People’s Republic (LPR) in Donbas had effective control over the region and carried out numerous grave human rights violations, such as torture, forced displacement, and enforced disappearances (Amnesty International and Human Rights Watch, 2016; UHHRU, 2017). On top of shelling, which never ceased, people living in Donbas also had to cope with cuts in power supplies, heating, and drinking water. Targeting such supply lines often undermines humanitarian assistance and cumulates into widespread human suffering (Sowers and Weinthal, 2021).

The international community attempted to reach a truce between the de facto authorities in Donbas and the Ukrainian military forces in the Minsk I and II agreements, but both agreements were doomed to secure peace. On 5 September 2014, the “Trilateral Contact Group” (consisting of representatives from Ukraine, Russia, and the OSCE to facilitate diplomatic negotiations) signed the Minsk Protocol to bring a halt to the armed conflict. The Trilateral Contact Group had several working groups on economic issues and humanitarian issues. The Russian Federation, the LPR, the DPR, and the OSCE signed the document. The ceasefire did not last, and the new Minsk II agreement called for another ceasefire in the Donbas region. The Minsk agreements were doomed to fail due to their unclarity over Ukraine’s sovereignty and territorial integrity: “Ukraine [saw] Minsk-2 as a means to restore its sovereignty, but Russia [viewed] it as a tool with which to cripple Ukraine’s sovereignty” by forcing Ukraine to accept the undefined “special status” of the DPR and LPR through a constitutional amendment (Allan and Wolczuk, 2022).

These unclarities over territorial integrity did reverberate to environmental monitoring as large parts of the existing industrial monitoring in the heavily industrialized Donbas region ceased to be accessible by Ukrainian authorities. The de facto authorities - the Russian installed proxies in Donbas - had effective control over the region, the conflict separated the environmental monitoring system between territories under control of the Ukrainian government and those that were not. Ukraine had a solid environmental monitoring system in place to monitor environmental risks before the conflict, but these protections have fallen since 2014 as environmental monitoring systems and governance structures in Donbas collapsed when Ukraine lost effective control over these territories and effective control over environmental monitoring (Hook and Marcantonio, 2018, 2022: 3).

Already before the conflict started, the region coped with heavy pollution from its metallurgical and coal sector (Environment People Law, 2015; Kirky et al., 2011; State Environmental Protection Administration in Donetsk Oblast, 2010). Donbas is Europe’s and Ukraine’s most heavily industrialized region. Donetsk and Luhansk oblasts are home to 4500–5300 polluting industrial sites and toxic waste disposal facilities (Denisov et al., 2015; OSCE, 2017), many of which were in the vicinity of the front line of the conflict—or, what used to be the Contact Line. Some are abandoned, in disrepair, and are at risk of flooding and leaking harmful toxins to the ground water (OSCE, 2017; Zoï Environment Network, 2019). Many of these are high-risk facilities (Law 2245-III63, ‘On High-Rish Facilities’, 18 January 2001, cited in Environment People Law, 2015), with around 80% of those facilities having installations harmful to the environment (Chernenko et al., 2019).

During the war, both the shelling of these industrial sites and the disruption of their industrial production caused most environmental harm (Denisov et al., 2015; also see OSCE, 2017). Between 2014 and 2017, enterprises in Donbas reported 500 operational disruptions (OSCE, 2017), for instance, disruption of the pumping of mines or breaks in supply chains of high-grade coal, which forced industries to shift to lower grades of coal (Denisov et al., 2015). Mines that are not pumped can flood, which could potentially pollute water supplies (OSCE, 2017). Toxic spills from Tailing Storage Facilities (TSFs)—ponds storing the liquid and sludge waste of industries and mines—could lead to a domino effect of transboundary pollution that could spread across the Black Sea Region (Nikolaieva et al., 2019; OSCE, 2017). In 2019, Ukraine counted 465 TSFs, storing over six billion tons of toxic waste; 200 of which are located in the Donetsk and Luhansk regions and 75 were in the territory controlled by Ukraine (Nikolaieva et al., 2019).

In the period before February 2022, international organizations have been paralyzed to provide technological and financial aid to curb environmental risks in Donbas (Hook and Marcantonio, 2022). During a period of 8 years, monitoring happened in an already uncontrolled environmental situation. Most environmental monitoring has been carried out by a small group of NGOs, national environmental experts, and the OSCE which outsourced studies to independent environmental safety experts (Denisov et al., 2017; Weir, 2017). One of their earliest warnings of environmental harm came from Environment People Law (EPL) and the Environmental Law Alliance Worldwide (ELAW), which asked for a stop of the military aggression and bring a halt to the environmental destruction in the conflict-zone (Environment People Law, 2015: 58). In 2015, the Toxic Remnants of War Network found that still relatively little was known about the chemical impact of the conflict and that a sampling study of EPL confirmed some “war chemicals” from the use of conventional weapons alongside larger quantities of damaged equipment and war rubble (Denisov et al., 2015). In 2015, the World Bank-EU-UN did an early analysis of the environmental impact of the conflict to formulate environmental recovery measures after the conflict is over (Weir, 2017). Other studies from UNICEF, UN Office for the Coordination of Humanitarian Affairs (UN OCHA), the Trilateral Contact Group and the Henri Dunant Centre for Humanitarian Dialogue are restricted and not public (Weir, 2017). The OSCE found in its 2017 report “Environmental Assessment and Recovery Priorities for Eastern Ukraine” (2017: 34) that it is likely “to expect a release of pollutants into the environment over a long period of time.”

Driven by a nationalist imperialist desire to bring Ukraine into a “pan-Russian nation” that has been years in the making (Kuzio, 2023), Russia launched a full-scale invasion of Ukraine in February 2022. Russia’s dehumanizing discourse about Ukrainians led to widespread war crimes, including torture, abductions, rape, enforced disappearances, and extrajudicial executions (Busol, 2023; Kuzio, 2023). Russia’s indiscriminate attacks on towns, hospitals, civilian infrastructure, water treatment facilities, and industrial enterprises with hazardous facilities have caused a humanitarian crisis and an environmental catastrophe (Busol, 2023; CEOBS and Zoï Environment Network, 2022b; Denisov et al., 2022; PAX, 2022a, 2022b; Racioppi et al., 2022; Weir, 2022). While the environmental risks of the war were present since 2014, the Nova Kakhovka dam disaster in June 2023 cast environmental harm caused by the invasion in full view and has consequences not just for Ukraine, but for larger parts of Europe and the Black Sea Region.

Methodology

The article is part of an interview study on environmental monitoring during armed conflict. Drawing on 12 semi-structured interviews from a set of 21 interviews with environmental safety experts, environmental activists, and other experts working on environmental pollution during armed conflict on the international level and in Eastern Ukraine. The interviews were conducted between 2018 and 2022 and therefore do not cover the period after Russia’s invasion of Ukraine in February 2022. The article also draws on NGO reports, IGO reports, as well as media reports on environmental hazards in Donbas. Due to the Covid-19 pandemic and travel restrictions, most interviews were conducted online. One interview was conducted in person in Georgia. One interview was conducted in Russian, others in English. Informed consent was provided by interviewees after they were explained the study’s aims.

Conducting interviews with people from conflict or post-conflict settings, consists of risks: interviewees may be hesitant to speak about politically sensitive issues, which could introduce bias to data (Naylor and O’loughlin, 2021). If their security is not considered, speaking openly in a highly politicized environment carries real risks for the interviewees. Online interviewing has drawbacks for a method that relies on rapport. Furthermore, online communications could also pose security concerns, especially when interviewing people who have worked in the “hybrid warfare” context in Ukraine. However, security issues are not exclusive to online interviewing. In the present study, most interviewees have been in the public and international eye because of their writing and reporting. All informants’ names have nevertheless been anonymized for security reasons. This study does not cover interviews with informants who live in the conflict zone or in the occupied territories in Donbas.

Experts were interviewed not just to uncover their specific knowledge, but to assess the decision power they possess (Krause, 2014). The interviewees were selected on four factors: first, their experience (10 to 30 years of experience in environmental monitoring or environmental safety), second, whether they have had influence in shaping the narrative of conflict related environmental harm inside and outside of Ukraine through reporting or other media presence, third, their personal involvement in environmental monitoring in the region, and fourth, their affiliation with either IGOs, NGOs, or national institutes. This qualitative interview study did not intend to sample a large population but offers insights on how a small group of experts reflect on their own practices of environmental monitoring in the war in Donbas since 2014. The interviews were semi-structured, and the interviewees were asked to reflect on how monitoring has developed over time since 2014, their experiences of working in the armed conflict, the obstacles in gathering data, and their views on what is the main source of conflict pollution. Data from the interviews was further triangulated with data from NGO or intergovernmental reports and other interviewees.

A dearth of data: Legacy pollution and leveraging access to industries

Experts are involved in two strategies of the politics of conflict pollution and risk measurement: bridging legacy pollution and leveraging access. First, they balance the urgency of data on conflict pollution in an already polluted region. Second, as environmental harms are divided over territories under control and not under control of the Ukrainian government, and the politicized access to industrial sites, environmental experts who work on the ground must leverage their networks to gain access, or sometimes access sites without permission. Experts gage whether industrial managers—who often depend on their relationships with local politicians—are open to external assessment.

In 2017, the OSCE reported a dearth of reliable information as “currently, entire sections of the territory lack any environmental monitoring activities, and many industries are operated under a regime of non-transparency, offering little in the way of reliable information on the nature of damages to their facilities and related infrastructure” (OSCE, 2017: 19). The conflict destroyed monitoring stations and environmental observation data sets and experts moved away from the conflict area (Hook and Marcantonio, 2022; OSCE, 2017: 65). With the exception of high-risk industries, the Ukrainian government suspended regular inspections (OSCE, 2017) and relocated expert staff (Environment People Law, 2015). An NGO representative explains that part of the lack of data originates, first, from the lack of access to the monitoring stations on the territory not under control of Ukraine, and second, from the high background levels of preexisting pollution—or, legacy pollution—which could render the urgency of conflict pollution invisible:

One of the difficulties of dealing with the Donbas conflict is that Donbas was quite polluted before the conflict. So, the background levels of pollution were pretty high. Therefore, “normal” hostilities do not add that much. So, if it were a pristine area, it would have been different, but it was not. Having said that, Donbas did have a fairly developed system of environmental monitoring before the conflict […] we even worked on putting this information on the public domain. The system almost worked, but then the whole thing started. So, in a way, there is enough data. Ok, there is never enough data, but I’ve seen much worse. There is enough data to look at the situation before the conflict. ⁶

A lack of access to assess data in the temporarily occupied territories in Ukraine further hampered articulating imminent dangers: “A lot of that was of course lost there […] It’s hard to say whether these posts for example still exist […].” ⁷ Zoï Environment Network reports that, after the start of the conflict, the Ukrainian government lost control over most of these stations; some have stopped transferring data altogether. For instance, data on the reduction of emissions following the reduction of industrial production are unavailable (OSCE, 2017). The OSCE found that, since 2014, the number of businesses in Donbas that report to the State Statistics Service of Ukraine “dropped from 40,000 to 14,000, with the number of major industrial facilities submitting environmental pollution reports dropping from 131 to 37” (OSCE, 2017: 37). Since 2017, available environmental data in government controlled areas has increased (Zoï Environment Network, 2019).

Without access to the non-government-controlled area a comprehensive analysis of environmental hazards and risks in the mining sector seemed impossible. As the Donbas area is porous—many of the mines shafts and tunnels are connected—flooding in one mine shaft can spill over to other shafts, which means that assessments of individual industrial sites falls short to capture the environmental risks and domino effects (Truth Hounds, 2021). A focus on single industrial enterprises in the parts of Donbas not under control of Ukraine ignores that mineshafts are connected underground:

We do not have access to the non-government-controlled area. But we communicate, there are organizations and of course special monitoring missions which are working on the non-government-controlled area. We present all the studies to them, and hope that we will be able to complete the study once we have access to the area. But, this is a huge problem, and it could take years to assess and analyze the state of the mines and flooding. They are connected and they have hydrological connections with each other, and they need to be analyzed in complex […] There should be a political decision for dialogue and communication. Unfortunately, at the moment we only can use available materials with no physical access and physical assessments. ⁸

Environmental hazards across the government-controlled and non-government-controlled areas lead to further risks. For instance, the flooding in mineshafts across the contact line could not be coordinated from the area under control of the government. The negotiation necessary before gaining access to data or industrial sites, some of which are abandoned or are in disrepair, further hampered the monitors’ ability to gain a complete picture. ⁹ A lack of conclusive data of data pushed some environmental experts to mobilize old contacts in the territory not under control of Ukraine to get informal access to data:

The problem is that even if any environmental monitoring is taking place there, we have no data. There is no data on water or air, because all the state services financed by Ukraine have now left, they are in Ukrainian-controlled territories […] We have almost no information on mine flooding or groundwater quality from that side […] We get it either from open sources or “through old connections”, let’s say it is a private source of information, and of course it is difficult to analyse everything with only half of the picture. ¹⁰

The NGOs that carried out early risk assessments had similar issues with gaining access to industrial sites, sometimes relying on personal contacts to get access:

We have engaged three local actors connected to the protection of the environment. [names contacts]. They have connections, they know people, they know how to approach them. They know the ecologists working on the different enterprises. There have always been ecological units in these plants, and you can approach them to ask them questions […]. ¹¹

Experts faced similar issues with gaining access to TSFs, as industries are, in principle, free to block access to their property. ¹² One interviewee criticized the lack of clear regulations that would grant researchers prompt access to sites, especially those which are close to Ukrainian military sites. Because of the military presence, experts had to go through lengthy procedures of the Ukrainian authorities before they got approval to get access to sites near the contact line. ¹³ They had to follow procedures of several governmental entities before final approval. This resulted in a tight timeframe in which they could gain access to sites when emergencies would arise. ¹⁴ These procedures could take weeks and delays at one entity would mean that they would have to restart the procedure. ¹⁵

Some NGO representatives declare that enterprises were open to inspections because the conflict created a space to solve problems related to industrial production, while others found that they were at the mercy of their willingness to cooperate with environmental safety assessments: “We have to cooperate with the military personnel, and it’s a vulnerable zone. But when we speak about enterprises, this is…maybe due to the conflict, they are open. No problems with access to the enterprises and to analyse the situation on government-controlled area close to the contact line.” ¹⁶ In contrast, other interviewees argued that industrial sites are not always be open to outside assessments. ¹⁷ A report on TSFs close to the contact line found that some enterprises did not give the experts full access to internal reports (Nikolaieva et al., 2019).

Environmental safety experts gain access to sites in several ways, either through negotiation with both local political and business elites or using connections with IGOs as leverage:

[Access is] an issue of communication and experience in Ukraine because a private company can [deny access] under our national legislation. Some companies are interested in such research because they would like to raise their issues at international organisations such as the OSCE. If you can communicate that to the top management […] it’s one way [to get access]. The second way is through governmental officials. But, if you like governmental officials to get involved, you need to go to the local level […] Operators and local authorities depend on each other. In Ukraine, they're friends, usually. But we have […] cases when we didn’t get absolute permission to conduct our research in the enterprises’ facilities because the quality of the management is so low that they are afraid to show. ¹⁸

With or without official access, some experts risk accessing sites anyway; risk assessments include an amount of political risk-taking of trespassing or not working under regulations: “for me the biggest problem was to not see [the sites]. I will not do research without site visits.” ¹⁹

Managing uncertainty around attributing environmental harm

The politics of environmental damage measurement includes managing uncertainty over how environmental harm can be attributed, interpreted, and presented to the public. The OSCE finds that, for instance, the Donetsk Airport and coal-mining infrastructure near Kyryllov had been intentionally destroyed in the fights between Russian backed proxies and government forces, but underlines that, while the area is dense with hazardous industry, “a sustained, targeted assault on industrial facilities” has been absent from the history of the conflict (OSCE, 2017: 40). Quoting a 2014 Ukrainian government report based on localized monitoring done in Shchastia, the OSCE found no increases in pollutants during the shelling (OSCE, 2017: 40). An earlier 2015 report by the EPL group does underline a direct connection between fighting and the release of harmful substances (Environment People Law, 2015), recommending people to stay inside and “breathe through wet gauze bandage, which binds acid gases.”

These disputes deepen when, according to some interviewees, organizations take a more alarmist stance to attribute harm to the fighting. Especially in the early years of the war, some argued that alarmist simplifications in the prevailing media narrative over the origin of environmental hazards linked to fighting might overlook the indirect—and far more damaging and long-term—environmental damage resulting from disrupted industrial enterprises (Denisov et al., 2015). An environmental expert alludes that those who engage in early alarmism might discredit the environmental data itself, indicating that there is a political element to the timing of measurement and the dissemination of findings on the origin of harm:

Like, when the OSCE did field research, in 2017, one thing we found out was that the impact of hostilities on chemical pollution is actually very low, apart from a few cases […] Whereas some of the NGOs were really crying out that now the area is badly polluted because of the combat. It was not because of the combat. One of the reasons is that Donbas was polluted before […] So, these kinds of things, are just too dangerous, because they [b]asically disrupt the trust of people in environmental data, environmental monitoring. They should be careful not to [raise] false alarms. ²⁰

The level of risk and the choice of sites was also prone to debate. In 2017, a Swiss mission at the United Nations Office for the Co-ordination of Humanitarian Affairs (OCHA) conducted a fact finding mission that was later quoted by the OSCE, which had studied enterprises in a 20 mile zone along the contact line and found that, “while the number of environmentally hazardous industrial sources is certainly significant, yet the general history of the conflict testifies to the absence of a sustained, targeted assault on industrial facilities (OSCE, 2017: 40). Relying on the Minsk Agreement, the report deduced that “the category of heavy weaponry capable of inflicting major damage to industrial infrastructure […] is banned on the territory of the conflict [which] severely curtailed its use (OSCE, 2017: 40). Alongside other considerations, the report acknowledged that when these circumstances change the risks could be severely high, but that the risk of industrial damage leading to environmental harm is “low to medium” (OSCE, 2017). ²¹

Some of the interviewees working in Ukraine for domestic environmental monitoring agencies, recounted being surprised by the findings of OCHA and were sceptical of the effects of the report on the urgency of environmental harm in the region, especially by diminishing the international interest in environmental harm in the conflict. ²²

To me [and anonymous] it was really a surprise, because by this time we had already been exploring the risks for a long time, we knew this region. I was born in this region. I know the Donetsk region; I know its industry. The risks were great even before the fighting. During the fighting, they increased significantly. And we’ve seen accidents that have polluted the environment. I don’t know why such conclusions were made by the experts, but indeed they were a surprise to us. And we strongly disagreed with that conclusion. But, I don’t know what the consequences are [laughter]. I think that at least these conclusions - they postponed the environmental projects of the international organization in Donbas for several years. ²³

For these experts, the definition of risk depends partly on whether those with more local knowledge have greater insights into how to assess it. Not only the risk itself, but also risk assessments have a political life of their own (Elliott, 2018). In the opinion of the expert above, the risk assessments, allegedly, might have influenced future assessments.

Beck finds that “the notion of an objective yardstick against which degrees of risk can be measured overlooks the fact that risks count as urgent, threatening and real or as negligible and unreal only as a result of particular cultural perceptions and evaluations” (Beck, 2009: 13). Reflecting on the OCHA assessment, an expert working for an IGO responded that the risk assessment reached that conclusion because the study only investigated a handful of well-prepared industrial enterprises in government-controlled area. ²⁴ The expert both confirms that risk assessments are informed by evaluations and material limitations, simultaneously acknowledging the limitations of risk assessments and protecting their political weight:

Ultimately it is up to the people involved in the risk assessment to define what is an acceptable risk […]. I agree with their assessment that, if you look at those plants, in that situation, the risk of something happening to them is low to medium […] So, I don’t think you can take it as the OSCE says that the environmental risks in Ukraine is low […] It was, you know, a very initial assessment that kickstarted a risk assessment process more broadly. A risk assessment process is extremely complicated and politicised because it involves both state actors and non-state actors, and everyone has their own opinion as well. ²⁵

Humanitarian programs and IGOs that conduct humanitarian and environmental risk assessments possess the authority to assign and select which health or environmental risk, and which losses are prioritized or obscured, further affecting which programs are funded (Krause, 2014; Kuchinskaya, 2014). Which of those losses takes priority in policy making or environmental policies is up for debate. Whether these evaluations hampered international monitoring is unclear, however, until Russia’s full-scale invasion in 2022 there has not been a comprehensive and long-term approach to environmental monitoring and protection in the region in the conflict zone. The Conflict and Environment Observatory found that due to a lack of donor interest and persistent insecurity in the region, a UN-led comprehensive assessment of the environmental risks in the conflict has not happened (Denisov et al., 2017; Hook and Marcantonio, 2022).

Even without an UN assessment, according to CEOBS, which analysed all available reports on the environmental impact of the war up to 2017, concluded that in those reports “the methods, whether analytical or fieldwork-based, provide a ‘rough yet consistent’ picture of the environmental impact of the conflict, with plenty of room for improvement” (Weir, 2017). The various NGO and intergovernmental reports on the conflict in Eastern Ukraine generally overlap in their findings (Weir, 2017). Looking at the various reports published between 2015 and 2021, they divide the causes of environmental damage, broadly speaking, in three categories: (a) damage resulting from shelled industries and forest fires, (b) damage resulting from a disruption of industries and mines, which could cause toxic waters to flood and spill into the groundwater or rivers, and (c) a rise in poaching and illegal wood felling.

In those assessments, combat damage is never ignored and often merges with other risks, such as the disruption of production. Later reports on the risks of combat damage to TSF (Nikolaieva et al., 2019) shows that there are significant risks to military activity around TSFs. Damage to the industrial sector and toxic storage facilities can lead to significant environmental damage and domino effects which would result in transboundary pollution and the pollution of the Sivirskyi Donets river (Nikolaieva et al., 2019; OSCE and Ministry of Ecology and, 2019).

In April 2014, Russian backed fighters captured Avdiivka, a town in the Donetsk oblast. The fights between the UAF and the Russian backed fighers which ensued over the following months damaged the nearby Avdiivskyi Koksohimichnyi zavod (Avdiivka Coke and Chemical Plant)—Europe’s largest Coke plant operational since the 1960s. The plant stopped working in August 2014, right after fighting started. A report by Truth Hounds, a Ukrainian NGO working on the conflict, found that 329 shells had hit the chemical plant in the beginning of 2015, without damaging the high-risk production units (Duprey and Bondarenko, 2019: 11), but killing 12 workers (Nikolaieva et al., 2019). A report on TSFs of the Avdiivka plant found that the plant had restarted 15 times and was de-energized over 200 times (Nikolaieva et al., 2019: 18). The damage to the plant lead to cuts in the gas supply to households for several days to months (Duprey and Bondarenko, 2019). According to Truth Hounds, the shelling also resulted into the release of benzol and coke gas, which has high concentrations of benzene, toluene, and ammonia, among other potentially harmful substances (Duprey and Bondarenko, 2019).

Disputes around the origin of environmental damage also arise when enterprises try and attribute responsibility of incidents to the conflict instead of preexisting maintenance issues. For instance, the Zasyadko mine, one of Europe’s most dangerous mines, experienced several lethal incidents due to underground explosions. During the conflict, in March 2015, a methane explosion killed 33 out of the 200 miners working in the mine. While the mine’s management attributed the 2015 methane explosion to shelling at Donetsk Airport (Denisov et al., 2015), previous incidents—also at other coal mines in the region— have been attributed to the mining sector’s lack of financing and safety precautions (RFE/RL, 2011). In peacetime, enterprises might often use diversion tactics to diffuse responsibility, and deny information around pollution (Auyero and Swistun, 2009; Shriver et al., 2020). In conflict zones where conflict pollution and industrial pollution overlap, enterprises hover between transparency to get public and government support and obfuscation to hide preexisting mismanagement, which might further obscure accountability for, and the origins of, harm.

Environmental risks and disinformation

In international politics, disasters and “shock events” are moments where opposing parties blame each other for political gain (Toal and O’Loughlin, 2018). Environmental threats and disasters are similar stages on which conflicting parties can sow distrust and disinformation to advance their interest. Furthermore, as the fog of war hampers scientific evidence collection on environmental damages, this disinformation further distorts the reliability of environmental data even when the risks of environmental disaster are well-known. Evidence collection in Eastern Ukraine happens in a context of “hybrid warfare” or, conventional warfare combined with other strategies, including cyberwarfare and disinformation campaigns (Hoffman, 2007). Environmental disinformation undermines environmental protection in armed conflict (Darbyshire, 2021; Lizotte, 2021; Weir and Denisov, Nickolai, 2019). Disinformation—“misleading information that has the function of misleading” (Fallis, 2015)—includes the spread of government propaganda or other false advertising, often on social media platforms. Orchestrated disinformation campaigns are often linked with “internet trolls” in authoritarian regimes (Guriev and Treisman, 2015), but over 70 countries use bots or employees of government agencies, military, or private companies who spread false information and manipulate public opinion over social media (Bradshaw and Howard, 2017: 3, 2019). ²⁶

A growing awareness of environmental harm of armed conflict, the breakdown of environmental monitoring in conflict zones, and struggles over the validity of data are breeding grounds for false information. The urgency of environmental risks backed up by technological data are not immediately apparent (Lahsen and Turnhout, 2021), which could lead to politicians, the military, and other groups to use environmental harm as narrative to discredit or twist the risks of environmental harm in war. In Ukraine, environmental risks and information have been used in several ways: disinformation by armed groups to attribute environmental harm to Ukrainian NGOs, misappropriation of environmental data to shift responsibility for environmental harm, and spreading false information about chemical incidents used as a provocation.

In the conflict in Eastern Ukraine, both Russia and Ukraine have accused each other of harming the environment or using environmental risks as a tool to provoke. In 2021, Minister of Defense of the Russian Federation, Sergei Shoigu, accused Ukraine and “Private military companies [from the United States] located in Donetsk, to prepare a provocation with chemicals in Eastern Ukraine” (RIA Novosti, 2021). A month before Russia’s invasion of Ukraine, Ukrainian military intelligence alleged that leaking containers with ammonia were delivered to “Concern Stirol” in occupied Gorlovka: “the man-made disaster caused by the actions of the Russian occupiers can be sued to accuse Ukraine of using toxic chemicals and as a pretext for expanding armed aggression against our state.” (Defence Intelligence of the Ministry of Defence of Ukraine, 2022).

Conflict and Environment Observatory reports that, in 2018, a dispute occurred in which false documents claimed that the Siverskyi Donets River was contaminated with radioactive material. In 2018, the CyberBerkut hacking group claimed that, with the help of the US secret services, Ukrainian authorities had poisoned water supplies in areas not controlled by the Ukrainian authorities with nuclear waste from the Vakelenchuk storage facility (Gumenyuk and Nazarov, 2018; Weir and Denisov, Nickolai, 2019). Conflict and Environment Observatory reports that StopFake news monitor disputed the story, arguing that the storage facility had all nuclear waste removed in 2017 (Weir and Denisov, Nickolai, 2019). The false documents also spoke of how “following the deliberate contamination, a broad information campaign spearheaded by international NGOs would be unleashed against the de-facto authorities in Donetsk and Luhansk, as well as against Russia, to blame them of incompetence, the inability to protect their people, and a breach of international law” (Weir and Denisov, Nickolai, 2019).

Conflict and Environment Observatory also alleges that the disinformation around this deliberate contamination could have been a pre-emptive measure. Just before the leak of the documents, the Donetsk de-facto authorities had announced that they would stop pumping water from the Yunkom coal mine, which contained radioactive material from a 1979 Soviet nuclear explosion to free trapped gas (Gumenyuk and Nazarov, 2018; Hook and Marcantonio, 2018; Weir and Denisov, Nickolai, 2019). The mine was abandoned and allegedly contains normal level of radiation, but stopping the pumping led to the flooding of the mine and a potential spill into the environment and Siverskyi Donets river (Weir and Denisov, Nickolai, 2019). Floods in connected mineshafts might also be a cause for flooding in the mine, and the potential spread of radioactive water (Darbyshire, 2020).

Disinformation also includes the appropriation and misrepresentation of environmental research. Environmental research conducted by the Ukrainian NGO EPL was appropriated by the environmental rights expert of the Russian Federation’s Presidential Council for the Development of Civil Society and Human Rights, Elena Esina, who presented the results of the studies as her own. ²⁷ Environment People Law’s study covered the soil contamination and environmental harm caused by shelling and explosions in the vicinity of Savur-Mohyla. Results of that study were published in an article on the environmental impact of shelling for “Ukrainian Truth” (Norenko, 2015). Environment People Law alleged that Russian officials used the material to blame Ukraine for environmental harm (Environment People Law, 2016).

These incidents affect both the actual and perceived reliability of data collection and environmental monitoring. While environmental monitors were forced to negotiate access or mobilize older contacts to assess environmental hazards spread across the government-controlled and non-government-controlled areas, the accusations and misrepresentation of environmental data sowed distrust. An NGO representative explains that the de facto authorities both expressed a desire—or, feigned interest—to cooperate on environmental issues, and that Ukraine and the self-proclaimed authorities blame each other for being responsible for environmental harm:

The other sides, the de facto authorities, continuously claim that Ukraine is damaging Donbas environment. They say that Ukrainian enterprises pollute the air too much, the water too much, produces too much waste. […] The de facto authorities wanted to set up a mixed group, with internationally accepted organizations, NGOS and you name it, and then go across the line of contact and inspect how well the environment is protected on the Ukrainian side. Of course, clear propaganda, everybody understands this would never happen. ²⁸

This back-and-forth of allegations also breeds distrust within the interaction between environmental monitors and experts. While interviewees were interested in gaining access to data from beyond the contact line, they were simultaneously sceptical about the intentions of organizations who wished to share data. The interest by the de-facto authorities to cooperate with international organisations is partly because of seeking some recognition. One NGO representative reflects:

I personally got approached several times by NGOs from the other side, even kind of authorities from the other side, asking whether we have access to the international community, whether we would be interested in data they have on the environmental situation, so that we can share with the international community. The point being, again, trying to get to the international circles with the message that Ukraine is damaging the Donbas environment. […] So, of course, we’ll look at your data, but we’ll also look at other data, and if you’re happy with this, then let’s cooperate. And they never came back. ²⁹

Another environmental expert was also not confident that Ukraine could rely on the cooperation of the Russian Federation in case of transboundary spills. ³⁰

The environment often features in environmental peacebuilding research as a strong opportunity for peacebuilding policies and cooperation: “shared environmental challenges are potential entry points for cooperation between groups, even if their relations are hostile or characterized by mistrust” (Ide et al., 2021). Nevertheless, shared environmental harm may equally be politicized and distorted in any direction: as an opportunity for intimidation, sabotage, and disinformation. Besides the examples given above, in the beginning of the conflict, engineers often could not reach the industrial sites that require maintenance located in the areas in between armed forces, because both sides in the conflict suspect that a cease fire could be used as an opportunity to regain territory (Duprey and Bondarenko, 2019: 13).

Conflict does not render the environment an apolitical place where opposing parties can cooperate; it can be a site for sabotage, crime, and disinformation, in part because of the resources it provides, but also because of the political leverage environmental harm has in international politics (Harris, 2015). Experts navigate both the difficulties of data collection as well as these political struggles to shape a clear narrative of conflict pollution and its detrimental effects on the environment and the livelihoods and health of people living in conflict zones.

Conclusion

This article examined the obstacles to knowledge production that environmental experts face when monitoring conflict pollution in Donbas, Ukraine (2014–2022). The article finds that experts manage three obstacles undermining the reliability of environmental risks in an active warzone: pre-existing industrial pollution, environmental damage spread across government and non-government-controlled territories, and disinformation.

The article broadened our understanding of these obstacles of knowledge production and the attribution of harm in highly politicized contexts in two ways. First, while some assume a linear path from technological data to policy making on environmental harm, what constitutes the practice of environmental expertise and data collection is often ignored (Lidskog and Sundqvist, 2018). Moving past technological skills, this article found that environmental expertise and risk assessment in high-risk settings involves the capacity to manage political disagreements over who has the power to assign threats or allow access to affected sites, and with that, access to environmental data. Environmental experts hover between calls of alarmism and scientific rigor: whether to call for environmental catastrophe on premature assessments or carry out more sampling and collect more information knowing that a comprehensive assessment can only be carried out after the conflict. Second, environmental experts must overcome disagreements on the origin of environmental pollution and whether it has been connected to the fighting or connected to the heavily polluted industrial zone fraught with safety issues, lack of access to sites, and gaps in environmental monitoring. Both dilemmas fuel disinformation campaigns: the environment can be a field of cooperation between conflicting parties, but environmental hazards may equally be politicized and provoke further distrust when used as a tool of intimidation, in sabotage, and in disinformation.

The war in Ukraine happens at a geopolitical juncture. With the development of new legal principles, international organizations nurture a blooming optimism for legal accountability and remedy for environmental destruction during armed conflict. Owing to an increased technological capacity to monitor the environment, domestic and international audiences are more aware of the environmental consequences of war and its related public health impact (Weir, 2022). In contrast to other wars where significant environmental damage occurred, the Ukrainian authorities are keen to foster international cooperation to monitor the environment, promote green reconstruction, and launch ecocide cases. Besides that, environmental harm has equally become a playing card in geopolitics and war; conflicting parties use risk assessments as tools to discredit one another, especially during war, as with the flooding of the Yunkom mine. Russia’s targeting of energy and industrial infrastructure is another strategy of war to pressure Ukraine. Even though Russia was occupying the Kakhovka dam at the time, it has blamed Ukraine for exploding the dam in June 2023 leading to widespread human suffering and environmental deterioration across the Black Sea Region.

Experts have to navigate these emerging power domains. As the impact of war on the environment is increasingly reckoned with by policy makers, NGOs, and international organizations, what we know about environmental damage in high-risk settings and how environmental experts navigate obstacles will likely influence how data can be used in compensation programmes or remedies for people living with conflict pollution. Not unlike the chemical disaster in Bhopal (Anderson, 1993; Fortun, 2009), the industrial disaster and toxic pollution in Ukraine already is the grounds for new expertise. Ukraine will also be a key case where new ways to remedy environmental destruction during and after war are developed (as with the establishment of the Register of Damage for Ukraine in The Hague), and a test whether new international legal principles could work in practice around the globe.