After analyzing over 170 high-resolution commercial satellite images of North Korea’s Yongbyon Nuclear Scientific Research Center captured between August 2012 and November 2019, Stanford’s Center for International Security and Cooperation (CISAC) has observed activity at the Experimental Light Water Reactor (ELWR) that may provide new insights into its operational status.
Construction of the ELWR began in late 2010 and, according to conversations between a visiting Stanford University delegation and Yongbyon officials in November 2010, the North Koreans expected the reactor to commence operation by the end of 2012. While it has appeared externally complete since early 2013, analysts in the open-source community have not observed definitive visual signatures that the ELWR has reached full operational status. Moreover, the North Korean government has not spoken publicly about the reactor or its status since November 2011.
Commercial satellite imagery from earlier this year, however, reveals a narrow but steady liquid effluent likely trailing from a pipeline stemming from the Turbine-Generator Building of the ELWR for at least three months. This new activity follows and builds upon notable developments at the ELWR since 2017, ostensibly aimed at preparing the reactor for start-up operation. They included the frequent movement of vehicles, cranes and equipment around the reactor’s entrance, the emplacement of a transmission tower and electrical transmission lines in 2017 and the construction of a dam and spillway by early 2018 to control water flow for the cooling system intake located upstream.
This is not the first time effluent has been observed being discharged in this area. Past instances, however, were likely associated with the Supply Facility’s wastewater pipeline, which shares an endpoint with the ELWR’s cooling system pipelines. Due to the increased volume and duration of effluent observed earlier this year, it is reasonable to conclude that does not emanate from the wastewater pipeline but is associated with the ELWR’s cooling system instead. If this is the case, at a minimum, this signifies ongoing testing or preparation of the ELWR’s cooling and/or service water systems for operation of the reactor.
A start of ELWR operations could have significant implications for North Korea’s nuclear weapons program and would complicate any denuclearization process. Although the stated purpose for the ELWR is electricity generation, the reactor could be operated to produce weapons-grade plutonium or tritium for boosted fission or hydrogen bombs. The potential start-up of the reactor also raises questions about whether or not a civil nuclear energy program should be included in a denuclearization process. If so, the disposal of spent fuel and the decommissioning of the reactor would become complicating factors that do not exist today.
ELWR and Its Cooling System
North Korea began construction on its ELWR in late 2010 and allowed a visiting Stanford University delegation to observe the progress. The delegation was told that the reactor was designed to be a 100 MW-thermal pressurized light water reactor (LWR), which translates to roughly 25-30 megawatt-electric (MWe), and that it was intended to be a prototype for a larger electricity-producing LWR. While the North Koreans expected the ELWR to be operational within two years of that Stanford visit, and satellite imagery showed the exterior of the reactor, including its containment structure, was completed by the end of 2012, no definitive signs of full reactor operations had been observed by the end of 2018.
One possibility for a delayed start at the ELWR involves complications and delays in the finalization of its cooling system, in addition to potential delays in the production of low-enriched ceramic fuel and technical issues involving the new LWR technology. The ELWR’s cooling system, like the intermittently-operational 5 MWe reactor just north of the ELWR, uses water from the adjacent Kuryong River to cool the reactor and discharges warm water resulting from reactor operations back into the river. Satellite imagery taken during the construction of this cooling system reveals a complex system of pipes, trench work, pump houses and cisterns built to carry out reactor cooling operations. Analysis of water discharge has yielded unparalleled insight into the activities of both the 5 MWe reactor and the ELWR.
At present, imagery analysis conducted in concert within 38 North, CISAC and subject matter experts concludes there are two known discharge locations for the tertiary cooling loop of the ELWR’s cooling system. The main discharge point is a white riverside building—what the authors believe to be a pump house—located 65 meters southeast of the ELWR’s Turbine-Generator Building and connected by piping (ELWR Main Discharge Line) built in 2013. A second discharge point is located 155 meters south-by-southeast of the ELWR. In early 2013, pipes (ELWR Secondary Discharge Line) were laid from the turbine-generator complex to the banks of the Kuryong River downstream from pump house complex, to provide a secondary discharge pathway.
Figure 1. Overview of piping emplacement for main discharge line.
Figure 2. Overview of piping emplacement for secondary discharge line.
Past Effluent at the Main and Secondary Discharge Line
Visible liquid outfall from the main discharge location near the ELWR pump house complex would provide a strong indication of testing or operation of the ELWR’s cooling system. To date, there has been no observed significant liquid outfall from the ELWR Main Discharge Line that would suggest such activity, with the exception of a short period between the end of 2017 and early 2018 when a stream of liquid discharge is seen emanating from the pump house complex., 
Visible liquid outfall has previously been observed at the identified secondary discharge point for the ELWR. Shortly after the ELWR Secondary Discharge Line was emplaced in mid-2013, significant discharge was seen at the endpoint of the pipeline, flowing across the riverbank into the Kuryong River. This liquid effluent is most notable in an image from June 10, 2013, in which white foam is visible at the discharge point. This may have been an instance of preliminary testing or flushing of the ELWR’s water service or cooling system immediately following the emplacement of the pipeline.
Figure 4. Prominent discharge with foam likely from secondary discharge line visible.
It is important to note, however, that a wastewater pipeline from the nearby ELWR Support Facility connects to the endpoint of the ELWR Secondary Discharge Line, forming a “dual discharge point.”
Figure 5. Wastewater line visible from ELWR support facility.
Some water discharge was observed emanating from the ELWR Support Facility wastewater line prior to the emplacement of the ELWR Secondary Discharge Line, but in significantly smaller quantities and for shorter periods of time than were observed after the ELWR Secondary Discharge Line was installed.
Figure 6. Faint liquid outfall from ELWR support facility wastewater line visible.
After the ELWR Secondary Discharge Line was emplaced and tested, fluid effluent at the dual discharge point was observed only periodically (15 out of 115 images) between mid-2013 and early 2019 and not at the level of prominence observed during mid-2013 (described above) and in mid-2019 (described below). In short, outside these two short periods of time, discharge observed during this time frame was more likely coming from the Support Facility than the Secondary Discharge Line.
Recent Effluent at the Dual Discharge Point
In March through June of 2019, a significant increase of effluent at the dual discharge point was observed. On 21 high-resolution satellite images obtained by CISAC taken between March 22 and June 30, 2019, liquid discharge has been apparent along the bank of the Kuryong River at the dual discharge point. The consistency of this effluent and the length of time that it is visible distinguish it from previous cases of effluent seen intermittently at this location.
Figure 7. Overview of effluent near ELWR support facility.
During the increased, consistent liquid discharge between March and June 2019, the effluent from the dual discharge point was significant enough to destroy a part of the riverbank road. By June 30, North Korea began reconstruction of this riverbank and road section, an effort that continued until July 14, 2019. The effort may have involved the emplacement of a culvert under the road prior to the road restoration. After July and through early November, liquid effluent has remained visible at this point but its presence is muted. The images show liquid pooling on the inland side of the Kuryong River and offer no clear indication of ongoing discharge or movement of the liquid across/under the road into the Kuryong River. This could signify no new or ongoing discharge, less discharge, or less visible discharge after the road repairs and the possible emplacement of an underground culvert obstructed our view of the discharge pathway.
Figure 8. Overview of road reconstruction by riverbank.
Figure 9. Overview of liquid pooling inland of road by riverbank.
Interpreting Observed Effluent Activity at the Dual Discharge Point
We cannot completely rule out that the effluent from the dual discharge point is emanating from the Support Facility wastewater line and not the ELWR Secondary Discharge Line, given that both lines share the same endpoint and we know that the wastewater line is capable of liquid discharge. However, our judgment is that the most recent liquid discharge is not coming from the wastewater line. The effluent observed between March 22 and June 30, 2019, is greater than the intermittent light trickle of liquid discharge that we have historically observed to be solely associated with the ELWR Support Facility. Furthermore, we have not observed a noticeable change or increase in activity at the ELWR Support Facility coinciding with, and plausibly generating, the increased liquid discharge observed between March and June 2019.
We believe instead, with a measured degree of confidence, that the effluent observed between March 22 and June 30, 2019 is associated with the ELWR and its secondary line. We base this judgment on the strength of the effluent flow and the length of time, at least three months, that the effluent is present. If true, at a minimum, the effluent indicates ongoing engagement of, testing of or flushing of water through the cooling system or other service water system to prepare for eventual operation of the ELWR. While the pipeline in question is likely not the main discharge pathway for hot water from the reactor’s cooling system, it does probably serve at least some discharge function for service water systems of the ELWR, if not some function related to the tertiary cooling loop of the reactor. Activity at this location, while indeterminate, is still notable insofar as it relates to the overall testing or functioning of the ELWR’s support systems.
We also consider whether the liquid discharge indicates previous testing or current full operation of the ELWR. That is, could the effluent indicate active cooling of the reactor because of an earlier effort to test full operation—necessitating ongoing decay heat removal—or because the reactor has been brought into continuous operation? The likelihood of previous or current full reactor operation is low. As stated above, the secondary discharge line is likely not the main cooling/discharge system for the ELWR. The main discharge of the tertiary cooling loop is likely near the pump house complex, and we have not seen recent evidence of discharge at that location, as we would expect in the case of full operation.
In summary, it is our judgment that the liquid discharge observed from March to at least June was likely emanating from the ELWR secondary discharge line. And depending on how one interprets the visual signatures in images taken after June, it could be reasonable to conclude that the liquid discharge has continued through early November. Effluent of this volume and duration has only been observed once before, from March to June 2013, when the ELWR hot water cooling/discharge lines were built and the secondary line was likely subjected to preliminary testing involving the flushing of river water through the line. If effluent is indeed emanating from one of the ELWR’s cooling lines, at a minimum, this signifies ongoing testing or preparation of the ELWR’s cooling and/or service water systems for operation of the reactor. Given the significance of the ELWR’s operational status for North Korea’s nuclear program and the complications a burgeoning nuclear energy program would add to nuclear negotiations, further monitoring of this location for changes to, testing of and operation of the ELWR’s cooling system is needed.
Contributors to this analysis include Chaim Braun, Nick Hansen, Frank Pabian, and Siegfried Hecker.
For more detail on some of these developments, see Allison Puccioni and Elliot Serbin, “North Korean ELWR makes progress toward operations,” Jane’s Intelligence Review, January 3, 2018, https://www.janes.com/images/assets/853/76853/North_Korean_ELWR_makes_progress_towards_operations.pdf.
Other organizations have reported on instances of possible discharge from the ELWR’s cooling system indicating testing of the reactor in 2017 and 2018. The UN Panel of Experts reported on September 5, 2017, that “on at least one occasion in 2017, hot water was discharged through the drainage channel of the light water reactor, suggesting a possible test of the cooling system of the reactor,” though it is not clear whether this report is referencing the main discharge location we have identified. Separately, the Institute for Science and International Security issued a report on April 5, 2018, suggesting that recent activity at the pump-house complex possibly related to the secondary cooling system could indicate that the reactor “is already operating or the cooling system is currently being tested.”
This analysis dovetails with an earlier assessment of the ELWR cooling system conducted by 38 North, which found that testing of the main ELWR cooling line may have started in March 2018 and continued for at least part of 2018. Our analysis shows that this suspected discharge of liquid from the main cooling line might have started prior to March 2018, as early as November 2017.
This could be a case of obfuscation of the discharge terminus, mirroring North Korea’s previous obfuscation of hot water discharge line stemming from the nearby 5 MWe reactor.