With II treatments, flooding and drying events induced redox fluctuations. During drying events, diffusion of atmospheric oxygen into the anoxic soil results in a rapid increase in redox potential , with increased drying time leading to more oxic conditions. These redox fluctuations directly impact the speciation of many elements in the soil, including As and Fe, which are critically important for this study. As expected, II treatments reduced As accumulation in plant tissue and grain. The high severity dry down treatments were the most effective at reducing As concentrations in plant biomass. Several studies have found that II is an effective method for reducing As accumulation in rice , but only a few have measured soil redox potential and evaluated field scale temporal data that predicts the changes in redox active species. In 2018, the pH increased over the growing season , with a more alkaline pH observed for the CF treatment in comparison to II treatments. Higher pH increases AsV solubility and can lead to As release into the soil solution and increase its bio availability for plant uptake . When paddy soils are initially flooded, the pH drops as CO2 reacts with water to form carbonic acid. However, reduction processes that occur in soils lead to a net consumption of protons and an increase in pH . To help predict the behavior of Fe and As throughout a growing season, we can identify the changes in soil parameters via a Pourbaix diagram. Figure 1.5 shows speciation of As and Fe as a function of Eh and pH. The differences between CF and HS treatments have been labeled in the diagram ,vertical farming in shipping containers and numbered for the sampling dates with number 1 being early in the season and 5 towards the end of the season.

Sampling dates and the measured soil parameters can be retrieved from Figure 1.1 and Appendix 1.1. The diagram indicates the presence of reduced forms of As and Fe in CF treatment for all dates, but later in the season, it approaches the Eh pH conditions, represented by the main diagonal orange and red lines, that indicate transformation to the oxidized forms of the elements. In contrast, HS suggests As and Fe are present in the reduced form at the beginning of the season but mid season, after the dry down event, the soil conditions suggest oxidation of both species exist for the remainder of the season. Based on the predictions of the diagram, AsV remains the dominant species after pH 7 and Eh above 100 mV, as confirmed in Marin et al., 1993 and Suthersan, 2001. Due to certain limitations, redox potential was sampled with a handheld probe which may have disturbed the surface of paddy soil when the probe was introduced for measurement. Notwithstanding, our trend indicates that Eh increased with dry down treatments, but it is understood that the anoxic levels in paddy soils could reach below 200 mV at 10 15 cm depth . Our results demonstrate that the HS treatment decreased As concentrations in white grain by 49.5% and 41.3% in brown grain and the MS treatment decreased white and brown grain As by 32.2 and 30.4%, respectively. Cd increased for the HS treatment by 393% and 264% for white and brown grain, respectively, but only by 160 185% for the MS treatment. However, it is important to recognize that these elevated Cd concentrations are very low due to low Cd in the soil and water at the experimental site, and do not pose a human health threat. It is appropriate to evaluate the differences in the water regimes of this study with field trials from Carrijo et al., 2018 and Li et al., 2019, since they were executed at the same location.

When compared to CF, their HS treatment reduced TAs concentration in grain by 54 59% for brown rice and 61 63% for white rice, while MS treatment decreased 41 52% for brown rice and 42 58% for white rice. Li et al., 2019 showed Cd concentrations in grain increased by 322% with HS treatment and 158% with MS. These authors also found that LS treatments were not very effective at reducing As grain concentrations. Evaluating water management regimes in a larger context by integrating 4 years of data can help us determine what dry down duration and frequency is ideal for developing suitable management guidelines for commercial scale rice production. II treatments in both experiments decreased As translocation to rice grains but led to elevated Cd levels. Cadmium is released from CdS as oxygen is introduced into paddy soils during dry down treatments, oxidizing sulfide and mobilizing Cd . For this reason, the MS treatment is a more suitable II recommendation for maintaining grain Cd levels low whilst reducing As accumulation; besides, longer drying times do not further decrease As uptake . Another advantage of a less severe drying treatment is to reliably maintain yield by minimizing stress in rice plants during the drying period. In this study, II treatments did not significantly impact yield or yield components, however, it is unrealistic to predict that yield will not be affected by II practices in other locations due to the differences in climate, environmental components, and rice varieties. Combined, these studies demonstrate that moderate dry downs can be effective at reducing As while minimizing Cd uptake. In developing practices for other locations,vertical grow racks these studies serve as a guide to demonstrate how II management strategies can be optimized for yield, human health, and environmental impacts over multiple growing seasons. Of course, there can be conditions where II would have low efficacy and other management strategies should be considered.

Additionally, we did not identify large differences in the reduction of As in grain between one or two dry down events in a growing season. This demonstrates that applying II treatments of a single dry down is sufficient to achieve lower As accumulation in rice. This is of great benefit as single dry down treatments are simple and easy to implement and thus suitable for implementation at the field scale. II treatments had a significant impact on grain As speciation. AsIII was identified as the dominant inorganic species translocated into the grain and organic As, as DMA, was not observed in the more severe treatments for both white and brown grain, indicating that II reduced organic As accumulation by 100%. For both years, AsIII accumulation decreased by 25 29% for the HS treatment. Our As speciation results in grain samples was in accordance with Williams et al., 2005, who reported that As in rice grain is present primarily as inorganic AsIII and AsV , with a considerable proportion of organic As, mainly as dimethylarsinic acid . AsV mainly accumulates in grain husks, therefore the only As species identified in brown and white rice are AsIII and DMA. Moreover, it has been reported As methylation does not take place within the rice plant, thus the source of methylated As species is most likely the rhizosphere where microbial methylation occurs . The transformation of AsIII to DMA is favored under anaerobic conditions in soil and rhizosphere , leading to increased DMA uptake. It is also expected that higher concentrations of As in the soil solution inhibits methylation , thus overall, and especially in areas with higher concentrations of As in the environment , AsIII is expected to be the predominant species of As in grain, as proven with our data ; this is the most toxic species to humans and strengthens the necessity to implement strategies to mitigate its uptake and accumulation in rice grain. As previously stated, the US FDA has proposed an action level for As in infant rice cereals of 0.1 mg kg 1 . At a global level, the Joint Food and Agriculture Organization and the World Health Organization Expert Committee on Food Additives suggested a maximum limit of iAs of 0.2 mg kg 1 for polished rice . Therefore, it is important to put in place paddy management strategies of high efficacy in rice growing regions worldwide, even in areas with lower background As levels as these low thresholds can be easily exceeded. In our samples, all As grain concentrations for white and brown rice were below the FDA’s proposed action level, except for 2017 brown grain which did not exceed the FAO/WHO suggested limit, and decreased to ~0.1 mg kg 1 after implementation of HS II. Higher concentrations of As in brown rice are expected because grain bran and husk accumulate higher amounts of As.

Furthermore, the soil at the experimental site had low As concentrations in comparison to other rice growing sites worldwide. As mentioned previously, soils in Southeast Asia can reach As concentrations above 100 mg kg 1 . Other studies performed in sites with higher concentrations of As in soil, in comparison to our site, show a similar decrease in As accumulation of grain with water management treatments . For our II treatments, the target volumetric water content for LS, MS, and HS was 40, 35, and 25%, respectively. These VWCs are comparable to the field trials presented by Carrijo et al., 2018 and Li et al., 2019, and aimed to shift soil redox conditions in soil to have substantial impacts on As immobilization without affecting yield . Additionally, the stage of plant growth is important when establishing II practices. It has been previously found that dry downs early , or late in the season can have a negative impact to yield and the treatment effectiveness to mitigate As accumulation in grain . Accumulation of As in rice plant tissue can also be affected by II. Our results indicate that roots accumulate the highest amounts of As, followed by straw and finally grain. Throughout the growing season, the trends of As accumulation reveal that the highest levels exist early in the season for straw and later in the season for roots. These differences in accumulation of As in biomass are explained by the detoxification mechanisms that occur when As is first absorbed into root cells , slowing its translocation to rice stems, which consequently decreases As accumulation in shoots early in the season. Regardless of the sampling time, our results show that biomass samples from the CF treatment accumulate more As in shoots and roots, indicating a higher uptake of As by rice plants with uninterrupted flooding due to higher bio availability of As in the rhizosphere.The CF treatment accumulated highest concentrations of As in the plaque extracts, while II treatments had less As within the plaque. Additionally, plaque mass is reported as the difference between root mass before and after DCB extraction, which is higher for CF treatment than compared to II . This indicates that an increase in soil moisture under anoxic conditions favors iron plaque formation and is consistent with Otte et al., 1991 and Liu et al., 2010. Other studies have found strong collocation of Fe and As on rice root surface , which explains why a larger amount of iron plaque formed under flooding conditions leads to higher As content in Fe plaque. As and Cd concentrations in rice grain have been carefully evaluated under different II treatments, however, any reduction in As grain at the expense of rice nutritional value would be unacceptable. Thus, it is crucial to understand how grain nutrient content may be influenced by II. We analyzed macro nutrient grain concentrations from II treatments of two dry downs and one dry down . The data reveal that intermittent irrigation of one or two dry downs did not negatively impact grain nutrient content. Lower concentrations of Fe present in grain of from 2017 and 2018 compared to the previous years are explained by a change in the site’s water supply to surface water with lower Fe concentrations. Moreover, we identified lower concentrations of P in CF treatments compared to all II, with higher concentrations present in HS. In paddy fields, AsV acts as a phosphate analog and is taken up by plants via phosphate transporters .