It was hypothesized that, as the water treatment pH increased, there would be greater removal of herbicide from the soil in the rinsate. As pH of the solution increased, the equilibrium of the weak acid herbicide would be pushed towards the anionic herbicide form resulting in lower sorption to the soil and greater removal in the rinsate. However, the opposite trend occurred for two of the herbicides . As water treatment pH increased, the concentration of saflufenacil and penoxsulam decreased in the aqueous solution extracted from the soil. Saflufenacil removal in the rinsate was greatest at pH 5 with about 78% and lowest at pH 8 with about 64%. At pH 5, about 35% of the penoxsulam was removal from soil and this decreased to about 22% removal at pH 8 . Indaziflam was below the detection limit of the instrumentation in all rinsate samples, regardless of pH.The results of the EC water treatments show that, as EC increased, herbicide removal decreased slightly . The effect of ionic strength on ionizable pesticide adsorption to soil has been well documented12; the common trend is that as ionic strength increases, the pesticide adsorption also increases . These data support that trend as well. The greatest amount of saflufenacil and penoxsulam was removed from soil in the 0.5 dS m-1 water solution rinse; about 70% of saflufenacil was removed from soil and about 25% of penoxsulam was removed. Meanwhile, in the 1.5 dS m-1 and 3.5 dS m-1 solutions, approximately 65% of saflufenacil and 22% of penoxsulam was removed . Indaziflam was below the detection limit of the instrumentation in all rinsate samples regardless of ionic strength of the rinse solution.Indaziflam was below the detection limit in all samples; however, it is not clear if this is due to strong sorption to soil or to degradation processes.

While indaziflam is considered moderately mobile to mobile in soil14, vertical farming racks it does have a higher Koc range than saflufenacil or penoxsulam23 meaning indaziflam would be more strongly sorbed to soil than the other herbicides in this study. Indaziflam has been reported to undergo photolysis in aqueous solutions rather quickly 14; samples were stored in the dark for much of the duration of the experiment. A brief follow-up experiment confirmed the laboratory lights did not cause photolysis of the chemical in aqueous solution under the conditions of the experiments . Saflufenacil dissipates relatively quickly in the environment23. The herbicide has biotic and abiotic degradation pathways but the most relevant pathway to this study would be hydrolysis in alkaline water13. The data set shows a significant decrease in herbicide removal from soil from pH 5 to pH 6 and 8 . The pH 7 data point was not statistically different from the other pH water treatments. There have been differing reports on penoxsulam hydrolysis. The Environmental Protection Agency states that penoxsulam is stable under hydrolysis conditions15 while Jabusch and Tjeerdema report triazolopyrimidine sulfonamide herbicides do undergo hydrolysis and the rate is dependent on pH24. There have been studies completed on two other herbicides in the TSA class which support pH dependent hydrolysis rates25-26. Given that the experimental samples were held at field capacity for seven days in this study, pH dependent hydrolysis could explain why penoxsulam concentrations were decreasing from 34% removal from soil at pH 5 to 22% removal at pH 8 .Adsorption mechanisms of pesticides are difficult to define because of the complex interactions between the soil surface, soil solution, and pesticide. Additionally, it is likely more than one adsorption mechanism occurs.

There are several mechanisms by which weak acid pesticide adsorption could be positively influenced by ionic strength – cations could displace hydrogen atoms from the soil surface resulting in a slight pH decrease that would favor a neutral pesticide form, more cations could be available to bridge the anionic form of the pesticide to the negatively charged soil surface, or cations could bond with the anionic pesticide resulting in a neutral form. A recent study on the adsorption-desorption properties of penoxsulam narrowed down the possible sorption mechanisms to H-bonding, cation bridging, and surface complexation with transition metals. The data set presented here supports the cation bridging mechanism. As ionic strength of the water treatment was increased, cation concentration increased resulting in the greater likelihood to bridge the anionic form of penoxsulam to the negatively charged soil surface. Figure 1.3 shows no statistical significance between ECw 1.5 dS m-1 and ECw 3.5 dS m-1 , this likely indicates most of the available binding sites of the soil were occupied close to ECw value 1.5 dS m-1 . Due to the similarity in size and ionizable functional group to penoxsulam, it is likely that saflufenacil is undergoing the same phenomena. The water treatments representing different irrigation water quality parameters did have a slight effect on saflufenacil and penoxsulam sorption to soil. The pH treatments indicated that both herbicides likely experience pH-dependent hydrolysis; saflufenacil and penoxsulam showed a decreasing trend in herbicide removal with increasing pH, the opposite of what the hypothesized pH effect would be. This indicates that even if irrigation water has relatively high pH, it is unlikely to substantially change the availability or movement of saflufenacil orpenoxsulam in California orchard soils. Results from the ECw treatments showed that flushing soil with a solution with moderate ionic strength could help saflufenacil and penoxsulam bind to soil versus low ionic strength. While there were statistically significant differences between water treatments, the overall effect on herbicide dissipation was minimal; the observed difference between the highest and lowest ECw treatment was only about 10% for each herbicide.

In the United States almonds are a $6 billion commodity grown solely in California making almonds the second highest grossing commodity in the state behind only dairy products . As of 2020 there were more than 500,000 bearing hectares of almond trees planted in California which produced 1.3 billion kilograms of almonds . Almonds are harvested by mechanically shaking the trees, sweeping the almonds into windrows, and picking the nuts up from the orchard floor. Preharvest herbicide programs and mowing are used to control vegetation that would otherwise reduce harvest efficiency . Glyphosate has been registered in almonds since the early 1990s and glufosinate has been registered since the early 2000s ; these are commonly used herbicides for preharvest orchard preparations because of their broad spectrum weed control and relatively short preharvest interval , three and 14 days, respectively. In 2018, over one million kilograms of glyphosate and nearly 300,000 kilograms of glufosinate-ammonium were applied in almond orchards . Because of the harvesting process, there is ample opportunity for the almond hulls, shells, and kernels to be in close contact with herbicide-treated soil. The majority of California’s almond crop, about two-thirds, is exported and generated more than $4.9 billion in 2019 . Of the exports, 22% were shipped in shell and 78% were shipped shelled . Asia is the largest aggregate market for in shell almonds while the majority of shelled almond shipments go to European markets . Exported shipments of almonds are subject to pesticide residue testing and must be at or below a maximum concentration set by the region’s food safety authority.The maximum residue limit for glyphosate and glufosinate in almonds differ by definition as well as concentration between the European Union and the US. In the United States, both glyphosate and glufosinate MRLs, which are commonly called tolerances, are defined to include the parent compound as well as its primary metabolites . For clarity these MRLs will be referred to as “total glyphosate” or “total glufosinate” if the concentrations of the metabolites are to be summed with the concentration of the parent compound. The US MRL for glyphosate in almond hulls is 25 mg kg-1 and 1 mg kg-1 for kernels. There is not a separate US MRL for in shell almonds because the residue in inshell almonds is determined by shelling the almonds and measuring the residue in only the kernels. The US MRL for total glufosinate in almond hulls and kernels is 0.5 mg kg-1 . In the European Union, the MRL for glyphosate is 0.1 mg kg-1 in almond kernels . The EU MRL for glufosinate includes its metabolites; the MRL for total glufosinate is 0.1 mg kg-1 . Glyphosate is registered in the EU until 2022 . A recent review completed by the European Food Safety Authority recommended that the MRL for glyphosate be reduced to 0.05 mg kg-1 and an optional total glyphosate MRL for the summation of glyphosate and its primary metabolites, AMPA and N-acetyl-glyphosate, set to 0.2 mg kg-1 . Hence, it is anticipated that in upcoming years glyphosate MRLs will be reduced, vertical grow rack and it is a possibility that the chemical may not be re-registered. According to statute, if at any time thesafety of a current MRL is reconsidered, the MRL can be reduced to the lowest limit of analytical detection which is 0.01 mg kg-1 .

Because of the importance of the European markets to the California almond industry and the importance of glyphosate and glufosinate to preharvest preparations, lab and field studies were conducted to evaluate the herbicide transfer from soil to almonds during harvest. The objectives were to determine if glyphosate and glufosinate residues can transfer to almonds from soil particles or directly sprayed almonds, whether increasing the PHI could substantially reduce the risk of herbicide in or on almond fractions and quantify the concentration of soil-bound herbicide in almond samples.This experiment was conducted to determine glyphosate transfer from directly-treated almonds to non-treated almonds. This was intended to mimic a situation where a small number of almonds fall to the ground very early and could conceivably be directly sprayed with preharvest treatments and then contaminate the later-harvested crop during harvest and handling steps. Two almonds were directly treated with 0.8325 MBq [14C]-glyphosate by using a microsyringe to dot the stock solution over the entire almond including the inside of the split hull and exposed shell. The two treated almonds were tumbled with nine non-treated almonds using the apparatus and methods described earlier. The almonds were tumbled using a rock tumbler for 15 minutes and let rest for 15 minutes. Before analysis the treated almonds were removed from the bottle, and the untreated almonds were dissected and analyzed for [14C]- glyphosate. This experiment was replicated four times.The whole almonds from each replicate from both soil transfer experiments and the almond-to-almond transfer experiment were separated for three different analyses: whole almond rinse, herbicide adsorption to almond fractions, and a surface swipe after a post-harvest mimicking process. All samples were analyzed using a liquid scintillation counter . The data were corrected for the background levels of radiation in the scintillation counter. The rinsate of whole almonds was used to determine how much [14C]-herbicide was loosely associated with the surface of the almonds. Three whole almonds were rinsed with water using gentle inverted shaking. The rinsate was collected into glass scintillation vials and evaporated using a vacuum evaporation system at 30°C . Once the samples were evaporated to near dryness, 10 mL of Ultima Gold™ was added to each vial. The samples were analyzed using the liquid scintillation counter. To determine how much herbicide was adsorbed to the almond fractions, three almonds were dissected into their hull, shell, and kernel components. Each component was homogenized using a mortar and pestle and liquid nitrogen. Approximately 500 mg of each homogenized almond fraction was collected into a combustion cone and combusted using a sample oxidizer . The combustion product, [14CO2], was collected in 20 mL of scintillation cocktail composed of 10 mL CarboSorb E® and 10 mL Permafluor® . Glass scintillation vials containing the [14C]-samples were analyzed using the liquid scintillation counter. The remaining three almonds went towards a post-harvest mimicking process. The almond hulls were discarded, and the shells were opened by hand cracking through a plastic barrier then discarded. The plastic was swiped using a filter paper and the swipe was added to a glass scintillation vial with 10 mL Ultima Gold™. The swipes were analyzed using the scintillation counter. The kernels were collected, homogenized and combusted, and the combustion product was mixed with scintillant and analyzed using the scintillation counter as described above.