The Russian Filtration System and the Mixed Messaging

According to a U.S. Department of State report, “Russian officials…in Russia-controlled areas of Ukraine are undertaking a monumental effort to ‘filter’ the population as a means of suppressing Ukrainian resistance and enforcing loyalty among the remaining population.”

Within the Russian filtration system, Ukrainian civilians are first screened and documented at a filtration site, and in some instances undergo interrogation, collection of biometric data, and invasive searches. After processing, civilians face one of three outcomes depending on the threat they pose to the Russian war effort: 1) detainment and imprisonment in Eastern Ukraine or Russia, 2) forced deportation to Russia, or 3) documentation and free movement. For those who are forcibly deported to Russia, the displaced persons are held at Russian-operated temporary accommodation centers (TACs).

Despite Western exposure of the human rights abuses at the filtration sites, on September 7, 2022, Vasily Nebenzya, Russian representative to the United Nations, publicly denied their existence at a U.N. Security Council Meeting and accused the West of an attempt to spread disinformation. Moscow spun positive narratives on filtration sites, referring to them instead as TACs, likely to portray the sites as having a humanitarian purpose.

Monitored Activity at Alleged Filtration Sites

We analyzed ten alleged filtration sites (see Figure 1 below) using commercial imagery and publicly available reporting. This analysis included four sites in Donetsk Oblast, Ukraine that were initially identified by the Conflict Observatory in an online report dated August 25, 2022. To contribute to the body of analysis, GDIL conducted further monitoring of the activity at the four sites using updated imagery, provided additional illustrative data originally excluded from the Conflict Observatory report, and assessed the activity trends to evaluate the regional shifts in filtration activity.

The Conflict Observatory report’s filtration identification methodology aided GDIL in setting key indicators for filtration activity (e.g., location and proximity to major areas of conflict, use of tents and/or expansion of infrastructure, vehicle queues, high volumes of parked passenger vehicles and high-capacity vehicles, and human queues). In addition, GDIL gleaned a baseline understanding of the levels of vehicle and pedestrian activity expected at a filtration site. To identify new sites, GDIL explored areas outside Donetsk that were not previously reported, searched for the key indicators in imagery, and sought to corroborate possible filtration operations with open-source reporting. As a result, GDIL identified six new potential filtration sites at or near checkpoints into Russia or Russian-occupied Ukraine, monitored the site activity over time, and evaluated how regional filtration activity trends related to anticipated, active, or settled conflict areas. The ten sites include the following locations:

Donetsk Oblast, Ukraine

• Manhush (two sites)
• Nikolske
• Bezimenne

Rostov Oblast, Russia

• Veselo-Voznesenka

Belgorod Oblast, Russia

• M-2 Highway (near the Nekhoteevka checkpoint)
• Logachevka

Crimea, Russian-occupied Ukraine

• Armyansk
• Perekop
• Peredmistne

We analyzed imagery of four sites in the southern Donetsk Oblast region (see Figure 1), which were identified and reported on by the Conflict Observatory as locations with filtration operations. GDIL’s goal was not to duplicate Conflict Observatory’s findings. Rather, we aimed to add to the analysis by monitoring activity at the four sites using updated imagery, providing additional data not originally included in the report, and identifying and incorporating site activity trends into GDIL’s overall analysis of how the trends relate to areas of anticipated, active, or settled conflict.

All four sites began demonstrating signs of declining vehicle activity or the removal of tent structures between May and July 2022, according to imagery reviewed by GDIL. The Conflict Observatory reported the dismantling of deployed tents at the site in Bezimenne, Donetsk Oblast as of June 2022; however, GDIL’s analysis of imagery identified a reduction in vehicle activity at the other three sites in Manhush and Nikolske. Since these three sites were housed in permanent buildings, we analyzed vehicle activity over time to evaluate the trends in filtration site activity. Moscow likely responded to the decreased civilian traffic in the southern Donetsk region by dismantling filtration tents where they were deployed and leveling down operations in permanent buildings, reflecting their reduced need for filtration in the region. Russian state media reports corroborated this decline in civilian intake. On April 2, 2022, TASS, a Russian state-owned news agency, reported that 14,168 people and 3,300 vehicles had crossed into Russia from Ukraine in the past 24 hours. On October 6, 2022, TASS reported that refugee intake was “on the decline” and claimed over 3,500 people had crossed in the same period.

Manhush, Donetsk Oblast, Ukraine 

Using testimonies and video media, the Conflict Observatory identified two filtration sites in the town of Manhush, located in the southern Donetsk Oblast region: 1) A police checkpoint and 2) the Manhush District Department of Internal Affairs (2 km east of the police checkpoint). GDIL is contributing imagery analysis of both sites to evaluate the changes in site filtration activity over time. Since both sites used permanent structures for filtration operations, we used levels of vehicle activity as a proxy for the level of filtration site activity. GDIL observed an overall decline in vehicle activity at the two locations, which started between April and May 2022 and continued into October 2022.

At the police checkpoint, imagery analysis from August 30, 2021 (pre-invasion) to April 3, 2022, suggested a significant increase in vehicle activity, which likely reflected the civilian migration spurred by the fighting in the nearby city of Mariupol in the spring of 2022. In August 2021, there was little traffic passing through the checkpoint (see Figure 2). In April 2022, we identified approximately 20 white vans parked inside the checkpoint and 76 vehicles queued at the checkpoint headed west (see Figure 3). Between April 3 and May 8, 2022, we first noticed a decline in vehicles, with approximately 13 parked vans inside the checkpoint and 27 vehicles headed in both directions on May 8, 2022 (see Figure 4). In addition, the extensive vehicle queue, observed in April 2022 heading toward Ukraine, was not present. Between May 8 and October 12, 2022, we observed a further reduction in vehicle traffic, with the absence of the vans and the vehicle queue and only some commercial vehicles heading in both directions on October 12, 2022 (see Figure 5).

Similarly, in our review of the imagery capturing the Manhush District Department of Internal Affairs filtration site, we observed an increase in vehicle activity between August 30, 2021 (pre-invasion) and April 3, 2022. In August 2021, there was little traffic in the areas around the site (see Figure 2). In April 2022, we identified approximately 150 vehicles near the site (see Figure 3). Between April 3 and May 8, 2022, the vehicle activity at the site began to decline with approximately 37 vehicles in the same site vicinity on May 8, 2022 (see Figure 4). Between May 8 and October 12, 2022, we saw a relative increase in the number of vehicles parked in the lot on the east side of the filtration area on October 12, 2022 (see Figure 5). However, there was an overall decline in vehicle activity between April 3 and October 12, 2022, with an estimated 72 vehicles present in the same areas on October 12, 2022, a 50% reduction in vehicle activity since April 2022.

The reduced vehicular traffic at the two Manhush sites is part of an assessed downward trend in civilian movement through the filtration process in the southern Donetsk Oblast region. This downward trend appeared to start when the battle in nearby Mariupol was drawing to a close in May 2022 and the Kremlin secured control of the region (see Graphic A and Figure 1).

Nikolske, Donetsk Oblast, Ukraine

The Conflict Observatory used open sources and imagery analysis to identify the Nikolske Police Department as a filtration site. In their report, they provided imagery from March 19 and April 3, 2022, to demonstrate the increasing volume of vehicle activity at the site between the two months. GDIL estimated vehicle counts from the April 3, 2022 imagery, which were not included in the earlier Conflict Observatory reporting, analyzed new imagery from July 31 and November 15, 2022, and assessed an overall declining trend in vehicle traffic between April and July 2022. We estimated approximately 54 passenger vehicles and seven white buses in the areas around the Police Department on April 3, 2022. According to imagery acquired by GDIL, this was followed by decreases in traffic on July 31, 2022 (see Figure 6) and November 15, 2022, with approximately 25 vehicles and two buses in the same areas (see Figure 7).

Similar to that experienced by the neighboring town of Manhush, the end to the fighting in Mariupol in May 2022 likely slowed the migration of civilians fleeing battle zones through filtration sites in southern Donetsk such as the one in Nikolske (see Graphic A).

Bezimenne, Donetsk Oblast, Ukraine

The Conflict Observatory report used civilian testimonies and imagery to identify the filtration operation at a location in Bezimenne. They also revealed the decline in vehicle activity at the Bezimenne filtration site between March and June 2022 and the dismantling of the 30 site tents as of June 2022. According to GDIL’s analysis of imagery dated October 3, 2022 (see Figure 8), we observed the continued absence of tents and vehicle traffic at the site.

Of the four sites in the Donetsk Oblast included in this assessment, Bezimenne is the only site that relied on temporary infrastructure to conduct filtration operations. At the sites in Manhush and Nikolske, which were housed under permanent structures, we monitored changes in filtration activity using vehicle activity as a proxy indicator. The significance of including the Conflict Observatory’s findings on the vehicle activity levels and use and subsequent dismantlement of the Bezimenne tents are two-fold: 1) it was a strong visual indication of Moscow’s lack of need for the filtration site which added support to our assessment of the decreasing civilian migration and filtration activity in the southern Donetsk region and 2) it led GDIL to use tent deployment and dismantlement and levels of vehicle activity as signatures of filtration operations in our search for new unreported sites along the Ukraine-Russia border.

Activity Variations at Possible Filtration Sites along the Ukraine-Russia Border

We extended our analysis into areas outside of the Donetsk Oblast region. Using our baseline of key signatures for filtration operations, we identified three new, unreported possible sites at or near Russian checkpoints along the Ukraine-Russia Border in the Rostov and Belgorod Oblasts (see Figure 1). GDIL observed the deployment of tents and/or vehicle traffic in imagery from April to May 2022 at the three locations, which are characteristic of filtration operations.

At two of the alleged filtration sites (Veselo-Voznesenka checkpoint in Rostov Oblast and the M-2 Highway site in Belgorod Oblast), the removal of tents and decreasing vehicular traffic from May 2022 to August 2022 may have illustrated a decline in displaced civilian movement through these areas by August 2022 and Moscow’s decision to reallocate filtration resources to other areas. In contrast, imagery of the Logachevka site in Belgorod Oblast from April to September 2022 suggested that the Russians did not reduce their filtration operations at the checkpoint in the same time period.

Veselo-Voznesenka, Rostov Oblast, Russia 

Our analysis of imagery located a possible filtration site at the Russian Veselo-Voznesenka checkpoint, located 26 km east of Bezimenne off a major highway. We identified filtration site indicators such as tent encampments and vehicle queues. Between March 14, 2021 (pre-invasion) and April 6, 2022, imagery analysis suggested the installation of approximately 20 tents near the right of the checkpoint by April 6, 2022 (see Figure 10) and a large queue of approximately 100 vehicles transiting from Ukraine into Russia. We are uncertain of the type of filtration activity at the checkpoint. Western news media referred to the Veselo-Voznesenka checkpoint as a “transit point”; however, given that the checkpoint is the closest crossing into Russia for civilians leaving the town of Mariupol in the southern Donetsk Oblast region, we assess with moderate confidence that Russian authorities registered and screened Ukrainians prior to entry into Russia.

Analysis of imagery between April 6, 2022, and July 18, 2022, the site downsized from 20 to 15 tents, and the vehicle queue decreased from approximately 100 to 44 vehicles entering Russia. In Sentinel-2 satellite imagery from August 2022, the tent footprint was non-existent, suggesting that authorities dismantled the tents between August 6 and 16, 2022 (see Figure 11). Due to the poor resolution of the August 2022 images, we were not able to conduct a comparative analysis of the vehicular traffic at the site from July to August 2022. Sentinel 2-A satellite imagery dated November 14, 2022, confirmed the continued absence of the tents at the site; however, due to the low resolution, we were not able to review the site vehicle activity. The removal of the tent cluster as seen in Figure 11 and the downward trend in vehicle activity at the site were consistent with the trends seen at the Manhush, Nikolske, and Bezimenne filtration sites, which are all located in the southern Donetsk Oblast region and within 92 km west of the Veselo-Voznesenka checkpoint. We assess with moderate confidence that the end to the fighting in Mariupol in May 2022 led to the decrease in civilian movement through the checkpoint, and Russian authorities no longer saw the need for the filtration tents.

M-2 Highway, Belgorod Oblast, Russia

Through imagery and open media analysis, we located a possible filtration site along the M-2 Highway in the Belgorod Oblast, between the Ukrainian town of Kharkiv and the Russian town of Belgorod. Imagery analysis included tent encampments and vehicle activity similar to what was observed at the identified filtration camps in the Donetsk Oblast. In July 2022, a Russian official reported the relocation of a mobile “TAC”  to a spot 8 km from the Nekhoteevka checkpoint on the Ukraine-Russia border in Belgorod Oblast, Russia. The alleged M-2 Highway filtration site is located generally in that vicinity, approximately 12 km from Nekhoteevka. Between April 26, 2019 (pre-invasion) and May 20, 2022, imagery analysis suggested a deployment of tents by May 2022 (see Figures 12 and 13). Although the Russian official referred to the site as a TAC, filtration was probably performed at this location because it would be the first tent encampment encountered on a major highway leading out of the city of Kharkiv, which saw intense fighting up until May 2022. The situation is comparable to the Manhush, Bezimenne, and the Veselo-Voznesenka filtration sites, which are also located on major highways leading out of intense fighting in Mariupol. Furthermore, open-source reporting mentioned the existence of a filtration site between Kharkiv and Belgorod where civilians endured “fingerprinting… and an interrogation with a ‘psychologist’.” Due to the identification of several key filtration site indicators and corroborating open-source reporting, we assess with high confidence that the M-2 Highway location was a filtration site.

According to GDIL imagery analysis, the number of tents and parked vehicles more than doubled from May 20, 2022, to July 6, 2022. In that time frame, the number of tents increased from four to nine, and the number of parked vehicles from approximately 13 to 36 (see Figures 13 and 14). In July 2022, we also observed a white bus at the driveway into the tent encampment (see Figure 14). Although we observed increases in vehicle activity, the overall numbers are relatively low compared to the vehicle traffic observed at filtration sites in southern Donetsk and at the Veselo-Voznesenka checkpoint. These sites experienced numbers ranging from approximately 60 to 150 queued and parked vehicles and high-capacity vehicles at peak activity, compared to the M-2 Highway site’s peak volume of approximately 37 vehicles on July 6, 2022. Between July 30 and August 9, 2022, authorities dismantled the northern collection of tents according to imagery from the Sentinel-2A satellite (see Figure 15). In that same period, a Russian regional official described a “slight decrease [of those staying from Ukraine]” at the site and stated that they would remove the mobile site once it was “no longer needed.”

The Kremlin likely anticipated a flux of displaced civilians from the fighting in Kharkiv between February and May 2022 and deployed the filtration site to process Ukrainians (see Graphic A), but dismantled the site due to a reduction in civilian traffic or an underwhelming turnout at the site in late Summer 2022. Despite the continued Ukrainian counteroffensives between September and October 2022 in the Kharkiv Oblast region, commercial imagery analysis between August 9 and October 12, 2022, suggested that the Russians did not reconstruct the alleged filtration tents or “mobile TACs” at the location, which would have been well positioned to process Ukrainian civilians fleeing the resurging conflict. As of October 12, 2022, only two temporary structures adjacent to the gas station were still present (see Figure 16). We assess that the Ukrainian advancement north of the city of Kharkiv in September 2022 likely slowed an already low civilian migration to Belgorod, Russia. When the area north of Kharkiv was under Russian control in early September, the Russian authorities likely urged civilians to flee to Russia. As Ukraine retook territory in Kharkiv, this likely eliminated Russian authorities’ influence on civilian migration to Belgorod. Thus, the Kremlin likely did not expect significant civilian flows to Belgorod through the M-2 Highway site; and therefore, did not reconstitute the filtration infrastructure at the location.

Logachevka, Belgorod Oblast, Russia

The Logachevka border crossing checkpoint, located between Kharkiv Oblast, Ukraine, and Belgorod Oblast, Russia, is a possible site where Moscow is using tent deployments combined with existing infrastructure to conduct filtration operations. Between August 6, 2020 (pre-invasion) and September 9, 2022, imagery analysis of the Logachevka site suggested the presence of three key indicators for filtration operations: 1) the deployment of tents at the checkpoint, 2) extensive vehicle queues, and 3) human queuing at tent structures. On April 9, 2022, we first observed the tent deployment at the checkpoint and a 3.1 km long passenger vehicle queue attempting to enter Russia (see Figure 18). On July 8, 2022, although the extensive vehicle queue primarily consisted of commercial trucks, we identified a potential area of human queuing at the blue tent (see Figure 19). On September 9, 2022, we measured a vehicle queue approximately 2.5 km in length and continued to observe the presence of the tent cluster (see Figure 20). We cropped the images in Figures 18 through 20 to fit them in the report; therefore, the entire vehicle queues were not included. We lacked open-source reporting to directly support our assessment of filtration activities at the Logachevka checkpoint, but there had been general statements by Ukrainian media of Russian soldiers interrogating Ukrainians in Belgorod. Based on the three identified key indicators for filtration activity at the site, paired with limited open-source reporting, we assess with low to medium confidence that the Logachevka checkpoint was used for filtration.

Imagery analysis from April to September 2022 of the Logachevka site suggested that the Russians have sustained filtration operations at the checkpoint, with the continued presence of tents at the site until the most recently available and discernable imagery on September 9, 2022 (see Figure 20). We reviewed site imagery from November 15, 2022; however, the poor quality of the image precluded further evaluations of the site. Long passenger vehicle queues observed in April 2022 did not reappear until September 2022 with long vehicle queues primarily comprised of commercial trucks. Furthermore, between May 1, 2022, and June 30, 2022, there was an absence of vehicle queues crossing through the checkpoint. Based on these findings, GDIL assesses that the Logachevka checkpoint experienced waves of civilian migration triggered by the fighting in Kharkiv Oblast between February and May 2022 and then in early September, when the Ukrainians launched their counteroffensive on September 6, 2022 (see Graphic A). In between those times, the commercial transport of goods between Russian-controlled territories and Russia appeared to dominate travel through the Logachevka checkpoint. In addition, due to the checkpoint’s location and proximity to the estimated boundary of claimed control by Ukraine and Russia (see Figure 1), we assess with moderate confidence that Moscow anticipates a continued flow of displaced civilians into Russia through the checkpoint with the possibility that Ukraine may seek additional territorial gains in the area.

We assess that the difference in activity between the two sites located near or at Russian checkpoints is related to the Ukrainian counteroffensives between September and October 2022 and the assessed territorial control in the region. In contrast to the M-2 Highway site, we observed consistent long vehicle queues (passenger and/or commercial) and the continued use of deployed tents between April and September 2022 at Logachevka. In this same time period, the territory that would feed displaced populations through the Logachevka checkpoint had remained in Russian control. In these Russian-controlled regions, authorities likely transited willing and unwilling civilians into Russia, not into Ukraine, and sustained the use of the site filtration tents due to an expected reliable flow of civilians through the site.

Expansion of Possible Filtration Sites at Crimean Border Checkpoints

We continued our pursuit to identify additional unreported filtration sites in other Russian-controlled territories. Through our analysis of available imagery and corroborating open-source reports, we identified three new possible filtration sites at the Armyansk, Perekop, and Peredmistne border crossing checkpoints between Ukraine and Crimea.

Through imagery analysis from April to September 2022, we observed key visual indicators of filtration sites at the checkpoints, including expanding infrastructure at all three crossings. A lack of tent deployments at the locations as well as the absence of key filtration site indicators within 50 km of these three checkpoints, suggested that the filtration operations occurred under the permanent structures as civilians passed through in their vehicles or by foot. As illustrated in open-source interviews with Ukrainians fleeing into Crimea, one man stated that the Russian authorities confiscated his cell phone at a Crimean checkpoint to review his contacts and social media accounts and purportedly downloaded his data. Separately, in an interview with a Ukrainian human rights organization that collected multiple civilian testimonies, the interviewee pointed to the Crimean border crossing at Armyansk as a filtration site. Based on the imagery analysis, publicly available reporting, and proximity to the Ukrainian southern counteroffensive, we assess with moderate confidence that there were filtration operations at the three checkpoints into Crimea.

The Russians may have expanded their filtration operations with the construction of new permanent structures in September 2022 due to the anticipated fighting in Kherson Oblast, where the Ukrainians launched a southern counteroffensive in August 2022 and continued to advance as of early November 2022 towards the Russian-controlled city of Kherson (see Graphic A). At the time, Russian authorities likely anticipated an influx of displaced civilians and an increasing need for filtration operations at the Crimean checkpoints due to the fighting.

Armyansk, Russian-Occupied Crimea

As previously discussed, we identified the Armyansk border checkpoint, located on the Crimean Peninsula between Kherson Oblast and Crimea, as a site with possible filtration activity. Between May 5, 2021 (pre-invasion) and May 29, 2022, imagery analysis revealed an increase in activity from little to no visible vehicle or human queues at the Armyansk checkpoint before the war to a queue of approximately 52 passenger vehicles and possible human crowds waiting to enter Crimea a year later (see Figures 21 and 22). Analyzing imagery between May 29 and September 26, 2022, we identified the reconstruction and relocation of a new service road bypassing the checkpoint on both sides of the border, which made way for a new permanent structure, and a high volume of pedestrians queuing at a small, green-roofed station before the crossing into Crimea (see Figures 22 and 23). Due to the identification of several key filtration site indicators, including proximity to the Ukrainian southern counteroffensive, vehicle and pedestrian queues, and the use and expansion of infrastructure, as well as corroborating open-source reports directly naming Armyansk as a filtration site, we assess with moderate to high confidence that there were filtration operations at this location.

Imagery analysis between April 29, 2022 and September 26, 2022 showed queues that ranged from 23 to 120 passenger vehicles, averaging approximately 62 passenger vehicles awaiting entry into Crimea. With the Ukrainian southern counteroffensive approaching the town of Kherson in late August 2022, we assess with moderate confidence that Russian authorities constructed the new permanent structure, as opposed to deploying tents, to accommodate the existing and expected high volumes of queued vehicles and the long wait times due to the filtration operations. We also assess that the expanded service road was likely constructed for civilians or military personnel with the appropriate credentials and/or documentation to bypass the filtration activities occurring under the permanent structures. As anticipated, between September 26 and October 20, 2022, we observed an increase in the number of vehicles in queue from approximately 23 to 61 passenger vehicles awaiting entry through the checkpoint. This increasing activity likely aligned with the urgency to move civilians out of Kherson due to the tense military situation and the advancement of Ukrainian forces in late October 2022. Between October 20 and 29, 2022, we noticed a substantial shift in the vehicle queues, with vehicles lining up to re-enter Ukraine and a few vehicles, primarily commercial trucks, awaiting to enter Crimea.

Perekop, Russian-Occupied Crimea

As previously discussed and similar to the Armyansk location, we identified the Perekop border checkpoint, located on the Crimean Peninsula between Kherson Oblast and Crimea, as a site with possible filtration activity. Between May 5, 2021 (pre-invasion) and May 25, 2022, imagery analysis revealed an increase in activity from little to no visible vehicle or human queues at the Perekop checkpoint before the war to a queue of approximately 40 passenger vehicles (see Figure 24). Compared to the Armyansk checkpoint, we observed similar construction of new permanent infrastructure and vehicular queues at the Perekop border crossing in the same time frame. Between August 29 and September 6, 2022, analysis of Sentinel 2-A imagery showed the construction of a new processing station similar to the one observed at Armyansk (see Figure 25). Due to the identification of several key filtration site indicators, including proximity to the Ukrainian southern counteroffensive in Kherson Oblast, the presence of extensive vehicle queues, and the expansion of infrastructure, as well as a corroborating open-source report describing filtration activities at Crimean checkpoints, we assess with moderate confidence that there were filtration operations at this location.

Post-invasion analysis of imagery from April 29, 2022, to July 21, 2022, showed a range of 20 to 40 queued vehicles, averaging approximately 33 passenger vehicles, waiting to cross the border. With the Ukrainian southern counteroffensive approaching the town of Kherson in late August 2022, we assess with moderate confidence that Russian authorities constructed the new permanent structure, as opposed to using tents, to accommodate the existing and anticipated high volumes of queued vehicles and the long wait times due to the filtration operations. As anticipated, analysis of imagery from September 25, October 22, and November 5, 2022, showed a steady increase from approximately 18 to 23 to 66 queued vehicles, respectively, waiting to enter through the Perekop checkpoint (see Figures 26 and 27). Similar to Armyansk, the increase in vehicle activity at the Perekop border checkpoint was likely due to the Russian authorities urging civilians to flee Kherson in late October 2022.

Peredmistne, Russian-Occupied Crimea

As previously discussed and similar to the situation at the Armyansk and Perekop locations, we identified the Peredmistne border checkpoint, located on the Crimean Peninsula between Kherson Oblast and Crimea, as a site with possible filtration activity. We had limited access to clear, post-invasion imagery of the border crossing with only one clear image from October 2022 and a few low-resolution images from the Sentinel 2-A satellite between June 5 and August 29, 2022. In our analysis of the low-resolution Sentinel 2-A imagery from between June 5 and August 29, 2022, we identified the expansion of the service roads to the west of the checkpoint which allowed space for the construction of a new processing station between August 9 and 29, 2022 (see Figure 28). We evaluated the vehicle activity at the site to determine if any vehicle queues were associated with the migration of displaced persons. Although we observed approximately two dozen vehicles queued to enter Crimea on October 8, 2022, pre-invasion imagery from 2019 and 2021 showed similar, if not longer vehicle queues. For example, pre-invasion site imagery from June 26, 2019, revealed a significant vehicle queue waiting to enter Crimea (see Figure 29); therefore, we could not assess whether the vehicles queuing at the checkpoint on October 8, 2022, were attributed to fleeing civilians or reflected typical vehicle movement through the checkpoint. At Peredmistne, we identified only one key filtration site indicator – the use and expansion of infrastructure (see Figure 30 for a clear image of the new permanent structure and the service roads as of October 8, 2022). The location of the checkpoint is located further from the Ukrainian counteroffensives in Kherson, approximately 264 km, as compared to its counterparts at Armyansk and Perekop, which measure about 130 km from Kherson. We used an open-source report that described filtration activities at Crimean checkpoints, but it did not specify at which location. With this limited body of evidence, in combination with the consideration that the Peredmistne border checkpoint experienced infrastructure expansions and timelines that paralleled those at the other two Crimean checkpoints, we assess with low to medium confidence that there were filtration operations at this location.

We also assess with low to moderate confidence that Russian authorities constructed the new permanent structure due to an anticipated increase in civilian migration from the conflict in Kherson Oblast, Ukraine, similar to the situations at Armyansk and Perekop. Due to the limited amount of imagery and the lack of a pre-war baseline to evaluate conflict-induced civilian migration, we were not able to confirm if the site experienced increased civilian movement after the expansion, and thus, if the Russians properly judged the need for the checkpoint expansion.

Imagery Appendix

See and click through the carousels to compare imagery.

Contribution Note

This report was made possible by the research contributions of GDIL researchers: Danny Moriarty, Nayla Borrell, Sofia Sandrea, Kevin Barrett, Varij Shah, Kameesh Karim, Carter Hull, and Wyatt Krieg.