Natural attenuation and remobilization of arsenic in a small river contaminated by the volcanic eruption of Mount Iou in southern Kyushu Island, Japan.

The active volcano Mount Iou, in the southern part of Japan, erupted in 2018 for the first time in approximately 250 years. Geothermal water discharged from Mount Iou had high concentrations of toxic elements, such as arsenic (As), which could seriously contaminate the adjacent river. In this study, we aimed to clarify the natural attenuation of As in the river through daily water sampling for approximately eight months. The risk of As in the sediment was also evaluated using the sequential extraction procedures. The highest As concentration (2000 µg/L) was observed upstream but typically remained below 10 µg/L downstream. Dissolved As was the main form in the river water on non-rainy days. Arsenic concentration in the river naturally decreased through dilution and sorption/coprecipitation with Fe, Mn, and Al (hydr)oxides during flow. However, peaks in As concentration were frequently observed during rainfall events, possibly due to sediment resuspension. Furthermore, the range of pseudo-total As in the sediment was 4.62-14.3 mg/kg. Total As content was highest upstream before decreasing further along the flow. When using the modified Keon method, 44-70% of the total As existed as more reactive fractions associated with (hydr)oxides.

Review on stabilization/solidification methods and mechanism of heavy metals based on OPC-based binders.

Stabilization/solidification (S/S) with ordinary portland cement (OPC)-based binders is a suitable method to remediate heavy metal (HM)-contaminated soil and reuse resources of industrial wastes. In industrial wastes, alkaline wastes such as red mud (RM), soda residue (SR), pulverized fly ash (PFA), and alkalinity granulated blast furnace slag (GGBS) can immobilize HM ions (Pb2+, Zn2+, Cd2+, Cr3+, and Cu2+) by precipitation. However, some HM ions (such as AsO43-) would redissolve within the strong alkali environment. In this case, PFA, GGBS, metakaolin (MK), and incinerated sewage sludge ash (ISSA) which have low pH, can be used to immobilize HM ions or added to the OPC-based binders to adjust the pH in the soil products. Moreover, the calcium silicate hydrate (CSH), calcium aluminum silicate hydrate (CASH), ettringite (AFt), and calcium monosulfoalumiante hydrates (AFm) generated during the pozzolanic reaction can also immobilize HM ions by adsorption on the surface, ion exchange, and encapsulation. SR and GGBS can be used to immobilize the HMs (such as CrO42- and AsO43-), which are mainly affected by AFt and AFm. For those not affected by AFt and AFm but related to immobilization by precipitating (such as Mn2+), other wastes except SR and GGBS are suitable for treating contaminated soil. Nevertheless, the formation of AFt is also instrumental for soil product strength. There are several factors affecting soil product strength. In the future, the influence of different hydration products on the S/S effects, competitive adsorption of HM ions, effects on long-term HM stabilization, and novel materials are worth being explored by researchers.

Forecasting the seasonal pollen index by using a hidden Markov model combining meteorological and biological factors.

The seasonal pollen index (SPI) is a continuing concern within the fields of aerobiology, ecology, botany, and epidemiology. The SPI of anemophilous trees, which varies substantially from year to year, reflects the flowering intensity. This intensity is regulated by two factors: weather conditions during flower formation and the inner resource for assimilation. A deterministic approach has to date been employed for predicting SPI, in which the forecast is made entirely by parameters. However, given the complexity of the masting mechanism (which has intrinsic stochastic properties), few attempts have been made to apply a stochastic model that considers the inter-annual SPI variation as a stochastic process. We propose a hidden Markov model that can integrate the stochastic process of mast flowering and the meteorological conditions influencing flower formation to predict the annual birch pollen concentration. In experiments conducted, the model was trained and validated by using data in Hokkaido, Japan covering 22 years. In the model, the hidden Markov sequence was assigned to represent the recurrence of mast years via a transition matrix, and the observation sequences were designated as meteorological conditions in the previous summer, which are governed by hidden states with emission distribution. The proposed model achieved accuracies of 83.3% in the training period and 75.0% in the test period. Thus, the proposed model can provide an alternative perspective toward the SPI forecast and probabilistic information of pollen levels as a useful reference for allergy stakeholders.