Following the Fukushima nuclear disaster of 2011, it was rapidly discovered that owing to the unfortunate location of the plant and its construction, its buildings’ basements had become flooded by groundwater ingress, which subsequently became highly contaminated. In order to avoid reverse diffusion of the contaminated water into the environment, those managing the site were compelled to continually pump out and treat the contaminated water, at a rate commensurate with its inflow. It was anticipated or perhaps it would be better stated as ‘earnestly hoped’, that by keeping the water level in the flooded building basement below ground water levels that contamination would not defuse out of the flooded basement. Naturally as a consequence TEPCO are accumulating and endeavouring to store and decontaminate the net amount of water ingress each day.
To facilitate containment necessary for the safe decommissioning of the immediately contaminated reactor buildings in September 2013 TEPCO commissioned the construction of their controversial ‘ice-wall’. Installation of the facilities to create the ice-wall commenced in June 2014 and was completed on February 9, 2016 at an estimated to cost some ¥34.5 billion (circa $339 million). Activation was on March 31, 2016, with commencement of the freezing of the seaward side wall. Freezing of the land-side wall commenced on June 6, 2016, with the secondary phase of sealing the last openings in the land side wall commencing on December 2, 2016. At this point we should note that the ice-wall in not penetrating to the depth of the aquifer, has no base to its containment, thus the wall is little more than a skirt, with water free to percolate in and out from below the contaminated site.
We now find ourselves in the spring of 2017, with the ice-wall’s chillier plant having run flat out for a year with seemingly little net impact on water ingress. Frustrated by this apparent lack of progress, on December 26, 2016 the Japanese Nuclear Regulatory Authority (NRA) citing “limited, if any effects,” advised TEPCO that the “frozen soil wall” should be relegated to a secondary role in reducing contaminated groundwater at the Fukushima No. 1 nuclear plant. Yet TEPCO still persisted in asserting that the ice-wall was effective stating “We are seeing certain results.” Which begs the questions: What results were they seeing and as TEPCO’s response would suggest, have the NRA been too presumptive in dismissing the ice-wall’s impact and groundwater ingress? Or perhaps TEPCO’s engineers, being so bought into their radical ice-wall concept they don’t want to ‘lose face’ or perhaps they have simply lost the plot?
In a bid to head of criticising of their activities for being less than transparent and tardy in properly advising the public, TEPCO have conveniently put certain of their findings into the public domain, in the form of press releases. From this data, it’s possible to get a rudimentary grasp of what’s going on beneath TEPCO’s ice-wall. Regular updates on volumes of contaminated waters pumped from drainage wells and the reactor buildings’ basement, along with local rainfall have been regularly published. These indicated the seasonal cycle of rainfall in the Fukushima area and further show a relationship between local rainfall and the volumes of water, (Figure 1).
Working on the basis of the limited available data and an anticipated lag between rain falling and its impact on groundwater, and assuming a direct relationship between water ingress and the total amount of water transferred or pumped out of the system, we can drive a relationship between the averaged daily water transfer (a measure of approximate water ingress) and the rainfall total for the prior month, (Figure 2). These criteria show very plausible cause effect linear correlation (i.e. of the type, y = mx + c), (Figure 3). Thus, we can envisage the contributions to groundwater flow within the aquifer beneath Fukushima being comprised of two portions (a) a large steady flow arising from rainfall which may have fallen years to decades ago on the mountains to the west of the site and equating to the linear equation’s constant and (b) a highly variable amount of flow arising from recent rainfall, predominantly within the last month.
Whilst the linear relationship between the phenomena is simplistic, on the available data application of 2nd or 3rd order polynomial curve fitting does not give any significant improved correlation coefficient (R). Given we have identified the correlation and observe seasonality, we can factor out the seasonality and project rolling annualised rainfall and water transfer (Figure 4).
Within the scope of natural variance, the annualised rainfall at Fukushima shows no significant long term trend, being flat and circa 1.5 metres per year. The water transfer level does show some improvement and notwithstanding the slightly higher than average autumnal rains in 2016, water transfer levels are on the decline. Alas given the magnitude of that decline in relation to that hoped for by the ice-wall’s advocates to 50 tonnes per day, it was understandable that the NRA were rather less than impressed.
We also have to consider that our original correlation between rainfall and implied water ingress was conducted on all available data. The reality is several operational events were being executed over the period, such as the commencement of 24 hour pumping from inland relief wells with the aim of reducing groundwater around the stricken buildings, as well as the phased installation of the ice-wall itself. Thus our initial correlation is a composite of parallel events. If we reapply our linear relationship model on a rolling 12 monthly period, to exclude any rainfall seasonality, we see some interesting features, (Figure 5).
Had the ice-wall achieve a positive effect we should observed both a reduction in total amount of water transferred (y) being made up by a reduction in the overall basal flow (c) and of course a reduction in the recent rainfall component as reflected in a reduction of its independent variable (m). We see a reduction in apparent basal flow. As this reduction has occurred in isolation with the independent variable increasing over time, we can attribute reduction in ‘c’ in good measure to the impact relief wells. However, the overall amount of water being pumped out of the stricken buildings has remained high and it has done so because the aquifer has become more susceptible to the impact of recent rainfall. This suggests that the aquifer adjacent the site has become more porous and not less porous over the last few years. Had the ice-wall had a positive effect, a decline in the independent variable ‘m’ over time should be observed.
I would conjecture that if such is the case what could have caused this effect. It is possible that the installation of the coolant pipe-work has caused significant sub-soil disturbance, coupled with the degradation of the substrate rock texture by ground heave. The above should effectively have been self repaired when the ice-barrier froze. However, in this circumstance, owing to the size of the ice-wall and it lack of capacity to freeze the entire depth of the aquifer, it is likely that the aquifer disruption at its margins has resulted in increased porosity in the aquifer directly beneath the wall. Furthermore, given that the wall is incomplete and operating at the extent of its capacity, and that the site is subject to seasonal warming, and has had operational outages it is highly likely that the freeze thaw cycling peripheral to the ice-wall has cause deterioration to the aquifers subsoil texture and cohesion, thereby giving rise to localised increase porosity of the aquifer. As such I am not of the opinion that the installation of the ice-wall has had a ‘limited impact’. I believe it has had a ‘significant and negative impact’ on the porosity of the aquifer local to the site of contamination, and I believe it has added circa 20% to the volume of contaminated water generated since its installation.
But there again, that’s just one persons musings and opinion, and I dare say other will disagree and think I’m writing bollocks. Either way, I would be fascinated to see what “certain results” the TEPCO engineers saw. And if what they saw was good, I’d like a double of whatever they’d been drinking…
 11 July 2016, ‘Fukushima’s Ice-Wall a Fridge Too Far’ Peter J. Hurley, Linkedin.com https://www.linkedin.com/pulse/fukushimas-ice-wall-fridge-too-far-peter-j-hurley
 December 27, 2016 Kohei T., The Asahi Shimbun ‘NRA: Ice wall effects ‘limited’ at Fukushima nuclear plant’: http://www.asahi.com/ajw/articles/AJ201612270056.html