Increasing benthic vent formation: a threat to Japan’s ancient lake

An autonomous underwater vehicle (AUV) was deployed in Lake Biwa from 2000 to 2012. In December 2009, ebullition of turbid water was first found in the deepest area (> 90 m) of the North Basin. Follow-up investigations in April and December 2010 and January 2012 confirmed the existence of benthic vents similar to the vents observed in other deep lakes. Importantly, vent numbers per unit travel distance in Lake Biwa dramatically increased from only two vents (0.37 vents km−1) in December 2009 to 54 vents (5.28 vents km−1) in January 2012, which could be related to recent tectonic activity in Japan, e.g., the M9.1 Tohoku earthquake in March 2011 and slow earthquakes along the Nankai Trough from 2006 to 2018. Continuous back-up investigations from 2014 to 2019 revealed additional benthic vents in the same area. The sudden increase in benthic vent activity (liquid and gaseous ebullitions) have significant potential to alter lake biogeochemistry and, ultimately, degrade Japan’s major drinking water source and may be a harbinger of major crustal change in the near future.

www.nature.com/scientificreports/ nutrients, from the bottom sediment to overlying water (Fig. 2c,d), and may enhance water column microbial production. Indeed, average phytoplankton production in Lake Biwa increased from 574 mgC m −2 day −1 in 2006/2007 to 1104 mgC m −2 day −1 in 2017/18 16 . ROV monitoring of benthic vents was near the deepest point of Lake Biwa on 15 May 2014. Several vents were found (Fig. 2b) where gas bubbles emerged intermittently at nearly 10 s intervals. Many large sized Daphnia species were observed around these vents, possibly Daphnia pulicaria, which is a zooplankton species that suddenly appeared after 1999 in the offshore zone of Lake Biwa 17 . We also found some microbial mats around the holes, which are typical features of hydrothermal vents in lakes Baikal 10 , Tanganyika 11 and Taupo 12 . We collected two gas samples with the ROV and analyzed the gas bubbles emitted from the Lake Biwa holes. Methane was the major constituent (> 99%), which is known to support bacterial growth and could explain the large Daphnia spp. population that was observed 18,19 . Some dissolved methane concentrations in Lake Biwa were measured before 20,21 , but it is difficult to take gas samples without any special devices such as an ROV attached sampler or human occupied vehicle 22 .
Gas plumes (Fig. 2c) have also been observed at several sites using a quantitative echo sounder from 2012 to 2019 as further evidence of vent existence. As these gas plumes were identified with target strengths between − 60 dB and − 50 dB, the bubble size was estimated as ~ 40 mm in diameter 23 . The bubbles spread horizontally along the bottom of the thermocline in the stratified season (Fig. 2c) and reached the water surface during the mixing season (Fig. 2d) 24 . Such gas movement is known to contribute to vertical material transport, which can be another important mechanism of internal nutrient loading 25 , and can additionally contribute to atmospheric greenhouse gas production. Gas ebullition from the sediments of wet ecosystems is a GHG pathway that has long been underestimated but generally dominates emissions 26 .

Alternative origins and locations of vents
To examine the possibility that the marked high turbidity near the bottom of Lake Biwa may be due in part to sediment re-suspension arising from shear instability in the benthic boundary layer, we deployed an acoustic Doppler current profiler (ADCP) to measure water velocity profiles every half an hour for 551 days from 2010 to 2012. Bottom shear stress was calculated from 31,898 field data profiles 27 . The results showed that only 29 profiles (21 in 2010, 1 in 2011 and 7 in 2012) were > 10 Nm −2 , the critical bottom stress needed to resuspend bottom sediments 28 , while 96.3% (30,727 profiles) were < 1.0 Nm −2 . The maximum value greater than critical shear stress was 19.95 Nm −2 measured in 2010. The probability of generating high bottom stress due to shear instability is < 0.1%, which clearly does not explain the increased and continuing turbidity from 2010 to 2012.
Sediment ebullition could also be the result of deep groundwater inflows. As Lake Biwa is surrounded by mountains, groundwater seepage can be found in many places, but these are mainly concentrated in areas < 40 m www.nature.com/scientificreports/ deep 29 near the shore. The recent benthic vents, however, are unlikely to be caused by ordinary groundwater inflows because the recent rapid increase of lake turbidity and the ebullition locations in areas > 80 m deep do not coincide with the area of groundwater seepage previously observed 30,31 . Thus, we conclude that neither groundwater inflows or bottom shear stresses were responsible for the sudden recent turbidity observed in the Tantan surveys. Rather, we believe that the evidence suggests that benthic vents are the cause, and when seeking possible mechanisms of their formation in Lake Biwa, the lake's geologic history and formation 2,32 becomes crucial.

Geologic history of Lake Biwa
The present area of Lake Biwa is believed to have been mountainous terrain a few million years ago, but then subsided at about 0.74 mm year −1 . This occurred because it lies on the northwest section of a crustal block that was affected by a change in direction of compression, first from south to north then from east to west, owing to complex plate tectonic interactions in the central island 2 .
Comparing the thickness of accumulated sediments from about 1.5 Ma to the present day 2,33 with the locations of the benthic vents detected between 2009 and 2012 shows that the vents are located along the area of shallow sedimentation, corresponding roughly to the location of the former mountain range (Fig. 3). A large negative gravity anomaly of -60 mGal was measured around Lake Biwa, indicating a thin crustal layer under the lake or the presence of low-density materials due to faults 34 . This area of the lake bottom not only has many faults but has also been subjected to large earthquakes (> M7.0) every 400-500 years 35 . These background data suggest that the benthic vents found in Lake Biwa did not originate from ground water, but are related to active tectonic movements.   Uemura and Taishi, 1990). The green circles indicate large ebullitions of gas and water observed using a quantitative echo sounder from 2012 to 2019. Noteworthy is that these vents are located along the shallow sedimentation region (< 500 m) indicated by light yellow, which may correspond to an ancient mountain range of several million years ago that gradually sank to form the present lake basin due to tectonic movement. The lake sediments accumulated during the last several hundred thousand years. (b) An enlarged map of way points (grey rectangle), water temperature inversion points measured by Tantan (red crosses) and vents from 2009 to 2012 (measurement dates are given in legend). A sediment temperature profile in Fig. 4 Fig. 1b); i.e., the water temperature near the bottom was slightly but significantly higher than the temperature of the overlying water column. The water temperature inversion points are plotted in Fig. 3. Temperature inversion is usually unstable in freshwater lakes, as mixing of the water column would usually be expected to remove such an inversion. However, if highly turbid water occurs in the benthic boundary layer, the density profile near the bottom can be stable (Fig. 5). Moreover, if the lake bottom is continuously heated, the unstable condition can persist. In order to verify this situation, we deployed 6 RBR thermistors to measure sediment temperature gradients vertically in the bottom sediment near the deepest point in Lake Biwa from 1 July 2010 to 4 January 2012. The results showed a large mean temperature gradient of 0.2 K m −1 , where the upward gradient was positive (N = 69,455 and SD = 0.014 K m −1 ), indicating strong upward heat flux in the bottom sediments at the deepest point in Lake Biwa (Fig. 3) over the two-year deployment period from 2010 to 2012 (Fig. 6), which was almost 5 times greater than the heat flux outside the benthic vent area (about 0.04 K m −1 ). Irrespective of the cause for these observed physical changes in the lake bottom, the rapid increase of vents and the associated increased turbidity and GHG releases in Lake Biwa requires systematic ongoing surveillance. The increases in turbidity in the deep benthic layer were not due to benthic shear stress or groundwater inflows, but mainly due to ebullition created by the recently increasing number of vents. These dramatic changes have the potential to not only indicate possible major crustal change but also significantly impact the health of Lake Biwa by increasing influxes to the water column of nutrients, particles and greenhouse gases, and leading to biogeochemical and ecological changes within the lake and ultimately deterioration in lake environmental quality. Indeed, satellite observation of the lake from 2002 to 2018 has indicated a steady increase in chlorophyll concentration, concomitant with algal production (Goto et al. personal communication, 2019). This could significantly impact millions of people that depend upon Lake Biwa for a range of ecosystem services and drinking water, and thus is an environmental problem and priority demanding close monitoring with advanced technology such as an AUV 40 .

Methods
The world's first autonomous underwater vehicle (AUV) " Tantan  www.nature.com/scientificreports/ A key feature of Tantan is the ability to take two high-definition digital video movies and one digital still picture at the same time 4 . These video cameras faced forward and covered almost the same view. Tantan is also equipped with sensors to observe benthic environmental water temperature, pH, turbidity, dissolved oxygen, conductivity and chlorophyll-a, which were used to determine the environmental conditions in different water bodies during the investigations. Using the two video cameras (SONY HDR-SR12), we obtained video images of vents and were able to identify their position with SSBL (SGK System Giken) and DVL (RDI Doppler Velocity Log). The transect line densities of vents after 2010 in Table 1 were evaluated from linear regression of the accumulated numbers of vents against travel distance.
We used Tantan to measure the vertical water temperature gradient at 1 m above the lake bottom along a transect line with two high resolution temperature probes (RBR with accuracy of 0.002 K). One thermistor  www.nature.com/scientificreports/ was attached to the top foredeck of Tantan, while another was fixed to its bottom surface, and the gradient was calculated over the height difference of 0.58 m between top and bottom probes. The thermistor positions were carefully fixed in order to avoid wrong readings due to AUV battery heating. We also used a multi-beam echo sounder (MB1, Teledyne, 170 kHz-220 kHz), a quantitative echo sounder (Kaijo, KFC-3000, 70 kHz), and a ROV (NHK original assembly) to identify the positions of vents and gas ebullitions. To identify bottom sediment resuspension, we deployed a 1200 kHz acoustic RDI Doppler current profiler (ADCP) fixed facing downward from 10 m above the bottom at 90 m from September 2010 to January 2012. Gases were collected on two occasions using an original sampling device attached to the ROV, developed by Ritsumeikan University's robotics team, and analyzed by gas chromatography.
Temperature profiles in bottom sediments were measured with RBR probes located at 0.0, 0.60, 0.80, 1.00, 1.20 and 1.40 m below the water-sediment interface near the deepest point where vents were found. The sampling interval was set to 10 min. The east-west direction shrinking distance rate [SDR (mm year −1 )] of Lake Biwa was estimated with field data from two stations, Kutsuki (west side) and Hikone (east side), within the dense Global Positioning System (GPS) network in Japan 41 . In Fig. 5, turbidity was measured with a fine-scale profiler (F-probe) at 1 m height above the lake bottom. FTU (Formazin Turbidity Unit) is the most widely used measurement unit for turbidity.

Data availability
All data used here are available from the first author.